week 3

Shabi20

15Articles.docx

Home >Nursing homework help >week 3

image4.emf

Vol.:(0123456789)

International Journal of Clinical Pharmacy (2025) 47:717–725 https://doi.org/10.1007/s11096-025-01863-w

RESEARCH ARTICLE

Pharmacist‑led deprescribing of cardiovascular and diabetes medication within a clinical medication review: the LeMON study (Less Medicines in Older Patients in the Netherlands), a cluster randomized controlled trial

Jamila Abou1 · Petra J. M. Elders2 · Danielle Huijts1 · Rob van Marum3,4 · Jacqueline Hugtenburg1

Received: 29 August 2024 / Accepted: 31 December 2024 / Published online: 23 January 2025 © The Author(s) 2025

Abstract Background Deprescribing inappropriate cardiovascular and antidiabetic medication has been shown to be feasible and safe. Healthcare providers often perceive the deprescribing of cardiovascular and antidiabetic medication as a challenge and therefore it is still not widely implemented in daily practice. Aim The aim was to assess whether training focused on conducting a deprescribing-oriented clinical medication review (CMR) results in a reduction of the inappropriate use of cardiovascular and antidiabetic medicines. Method A cluster randomized controlled trial involving 20 community pharmacists, who conducted a clinical medication review in 10 patients. The intervention group received training on deprescribing. Patients 70 years or older with polyphar- macy having a systolic blood pressure below 140 mmHg and using antihypertensive medication and/or an HbA1c level below 54 mmol/mol and using antidiabetic medication, were included. Follow-up took place within 4 weeks (T1) and after 3 months (T2). The primary outcome measure was the proportion of patients with one or more cardiovascular and antidiabetic medicine deprescribed within 3 months after the CMR (T2). Results A total of 71 patients in the intervention group and 69 patients in the control group were included. At T2, 32% of patients in the intervention group and 26% in the control group (OR 1.4, CI 0.65–2.82, p = 0.413) had one or more cardio- vascular or antidiabetic medicines discontinued. Regarding any medication, these percentages were 51% and 36%, (OR 1.8, CI 0.92–3.56, p = 0.085) respectively. Conclusion Increased awareness and ability of community pharmacists to deprescribe medication and use of general prac- titioners’ data, led community pharmacists and general practitioners to successfully conduct a more deprescribing-focused CMR in daily practice. Further research is needed to assess the necessity of additional training to optimize the deprescribing of cardiovascular and antidiabetic medication. The study was registered at The Netherlands Trial Register (registration no: NL8082).

Keywords Frail elderly · Medication reconciliation · Medication review · Pharmacists · Polypharmacy

* Jamila Abou [email protected]

1 Department of Clinical Pharmacology and Pharmacy, Amsterdam University Medical Centers, Location VUMC, Amsterdam, The Netherlands

2 Department of General Practice, Amsterdam UMC, Location VU, Amsterdam Public Health Research Institute, Amsterdam, The Netherlands

3 Department of Elderly Care Medicine, Amsterdam Public Health Research Institute, Amsterdam UMC (Location VUmc), Amsterdam, The Netherlands

4 Department of Clinical Pharmacology, Jeroen Bosch Hospital, ’s-Hertogenbosch, The Netherlands

http://crossmark.crossref.org/dialog/?doi=10.1007/s11096-025-01863-w&domain=pdf

http://orcid.org/0000-0002-6521-6203

718 International Journal of Clinical Pharmacy (2025) 47:717–725

Impact statements

• The use of medication records from community phar- macists and clinical data from general practitioners could improve the targeting of patients who may benefit from deprescribing antihypertensive and antidiabetic medication.

• The implementation of deprescribing of antihyperten- sive and antidiabetic medication within the context of clinical medication review needs further study.

• Effectively conducting deprescribing-focused clinical medication review requires both training of community pharmacists and adequate communication and collab- oration between community pharmacists and general practitioners.

Introduction

Polypharmacy is associated with medication-related hos- pital admissions [1]. Especially in older patients, poly- pharmacy can lead to an increased risk of drug-related problems. Due to age-related changes in pharmacodynam- ics and pharmacokinetics and increased frailty, potential harm can outweigh the clinical benefits of preventive medication over time. Preventive medication initiated in the past, can therefore become inappropriate, and its use may result in increased morbidity, hospitalizations and health care costs [1–5]. This indicates the need to evalu- ate prescriptions for preventive medication in older per- sons, especially frail ones. Deprescribing, defined as the process of withdrawing inappropriate medication under the supervision of a healthcare provider (HCP), aims to manage polypharmacy and improve patient outcomes [6]. It may serve as an important pharmacotherapeutic tool to address inappropriate polypharmacy in frail older patients with multimorbidity. This process can optimize medica- tion use in terms of both safety and outcomes, potentially enhancing patients' quality of life [7, 8]. Conducted at regular intervals, a clinical medication review (CMR) is performed in primary care and nursing home care to man- age polypharmacy in older patients with multimorbidity to prevent drug-related problems leading to falls, hospital admissions, a lower quality of life, increased mortality and excessive health care costs [5, 9–23].

Although evidence is not strong and clear, deprescrib- ing of inappropriate (preventive) cardiovascular and anti- diabetic medication seems feasible and safe [9, 24–32]. Therefore, incentives for deprescribing this medication have been introduced in guidelines for the treatment

of cardiovascular diseases and type 2 diabetes mellitus (T2DM) [25–28, 32–36]. HCPs, however, often experience the deprescribing of cardiovascular and antidiabetic medi- cation as a challenge. As the result, it is still not widely implemented in daily practice [9, 14, 37–41]. Barriers include a lack of evidence that discontinuation is more beneficial than continuation, HCP and patient uncertainty, fear of negative consequences, reluctance among patients or their relatives to accept changes to medication, a lack of knowledge, time, training or support to allow HCPs to effectively perform deprescribing and suboptimal HCP collaboration [9, 13, 14, 37–44]. In the Netherlands com- munity pharmacists (CPs) are responsible for offering pharmacotherapeutic care and thereby play a significant role in medication management in primary care [11–13, 21, 22]. CPs have also been tasked with conducting CMRs annually in older patients with polypharmacy in collabo- ration with general practitioners (GPs), geriatricians, specialists geriatric medicine and other HCPs [12–14, 16, 18, 44–46]. Recommendations on how to perform a patient-centered CMR, pharmacotherapeutic criteria and tools are provided by the Multidisciplinary Guide- line ‘Polypharmacy in older patients’ [44–46]. Despite its mandatory nature and remuneration, CMR implementation in practice is still not without problems, especially with regard to the deprescribing of (preventive) cardiovascular and antidiabetic medication [13–16, 37, 38, 40, 44–47]. As a remedy, CMR eligibility has been restricted to high- risk patients and recommendations on deprescribing have been added [44, 46]. Criteria and considerations have now systematically been summarized to facilitate patient-cen- tered decision-making to continue, adjust or discontinue medication. However, to achieve full implementation of the Multidisciplinary Guideline ‘Polypharmacy in older patients’ in daily practice, the CPs, GPs and other HCPs involved need training and support, especially with regard to patient selection, decision-making and HCP collabora- tion/patient communication [8, 9, 12–15, 40, 45–50].

Aim

The aim of this study was to assess whether training for CPs on how to conduct a more deprescribing-focused CMR, would result in a greater reduction in the inappropriate use of (preventive) cardiovascular and oral antidiabetic medi- cation in terms of discontinuation, as compared to CMRs conducted without prior training.

Ethics approval

The study protocol was approved by the Medical Ethics Committee of the VU University Medical Center in Amster- dam (2019.326).

719International Journal of Clinical Pharmacy (2025) 47:717–725

Method

Study design and setting

The ‘Less Medicines in Older Patients in the Netherlands’ (LeMON) study, is a cluster randomized controlled trial conducted between November 2020 and December 2021. The study evaluated the effects of training for CPs aimed to conduct a more deprescribing-focused CMR with regard to cardiovascular and oral antidiabetic medication in daily practice.

The BENU pharmacy chain is the largest pharmacy chain in the Netherlands, and the pharmacists generally follow the same procedures in patient care. The BENU pharmacy chain provided a longlist of eligible CPs work- ing in medium-sized and large BENU pharmacies across the country. In the Netherlands, a medium-sized pharmacy serves between 5000 and 10,000 patients. A large phar- macy serves between 15,000 and 20,000 patients.

Thirty-five CPs were invited to participate, 20 of these participated in the study and were block-randomized in Castor EDC, a web-based data management platform [51]. CPs randomized to the intervention group were asked to participate in a deprescribing training. To minimize per- formance bias CPs in the control group were not informed about the content of the training.

CMRs and follow-up meetings with patients were in the pharmacies or completed electronically or by telephone, depending on the CPs’ own working procedures and CP and patient preferences. Outcomes were reported within

4 weeks after each CMR (T1) and after 3 months (T2). A flowchart of the study design is presented in Fig. 1.

LeMON intervention

LeMON is part of a collaboration called DISCARDT, in which three research groups in the Netherlands have received funding from The Royal Dutch Pharmacists Asso- ciation (KNMP) to develop tools and strategies aimed at reducing the medication burden in individuals with T2DM and/or cardiovascular diseases. The training included the following key elements: 1. knowledge on the deprescribing of cardiovascular medication, 2. identification of eligible patients, 3. addressing barriers and facilitators, 4. applying shared decision-making and 5. monitoring of CMRs. More detailed information about the training can be found in the supplementary material.

Patient selection

Patients were selected by the CPs via data from the Phar- macy Information and Administration System (PIAS) and the General Practice Information and Administration Sys- tem (GIAS). This was done in the same manner for both the control and intervention groups. The inclusion criteria for patients were as follows: 70 years or older and chronic use of five or more medicines including at least one anti- hypertensive or antidiabetic medicine (Anatomical Thera- peutic Chemical [ATC] classification codes: A10: diabetes and C: cardiovascular system). In addition, patients should have had a recently measured (< 6 months before inclu- sion) systolic blood pressure < 140 mmHg and/or a HbA1c

Fig. 1 Flowchart of the outline the study design. Abbreviations: CP = community pharmacist, CMR = clinical medication review

720 International Journal of Clinical Pharmacy (2025) 47:717–725

level < 54 mmol/mol. Life expectancy should be more than 3 months as estimated by the GP.

The CPs selected eligible patients from their PIAS and contacted the GPs of these patients for collaboration and GIAS information on additional inclusion and exclusion criteria. In the case of linked systems (Pharmacom®- Medicom®) CPs had direct access to these data. CPs invited eligible patients by phone or at pharmacies. They were sub- sequently informed about the study by their CP and were asked to provide informed consent before participation.

Sample size calculation

On the basis of previous studies it was assumed that in 40% of the patients the number of cardiovascular and antidiabetic medicines could be reduced [52]. With an α of 5% and power of 90%, 80 patients were needed in each arm. Since the CPs work for the same pharmacy chain, BENU Apotheken, they are trained in the provision of pharmaceutical care and con- ducting CMRs in a similar manner. Considering 10% loss to follow up, a total of 100 patients would have to be recruited per arm. To keep the workload per CP as low as possible, in view of the estimated number of patients fulfilling the inclu- sion criteria and a recruitment period of 6 months, conduct- ing a CMR in 10 patients per CP was considered feasible.

Primary outcome measures

The primary outcome measure was the proportion of patients with one or more cardiovascular and antidiabetic medicine deprescribed within 3 months after the CMR (T2). Deprescribing was defined as stopping medication, reducing the dose of medication or substituting a medicine for a less potent alternative.

Secondary outcome measures

The secondary outcome measures were the type of medica- tion and the number of proposed deprescribing interventions at the follow-up meetings with GPs and patients 1–4 weeks after the CMR (T1).

Data analysis

All analyses were performed via the Statistical Package for the Social Sciences version 15 software (SPSS). Descrip- tive statistics included the mean and standard deviation for normally distributed variables and median and interquartile range for nonparametric variables. Study groups were com- pared via the χ2 or Fisher’s exact test for categorical vari- ables and the Student’s t-test for normally distributed con- tinuous variables. Following descriptive analysis, the data were analysed with a logistic mixed model, which included

a random effect model for CPs. Multilevel analysis was used to analyse the effect of clustering patients at the CP level. Age, systolic blood pressure and sex were analysed as poten- tial confounders. Adjustment was made for the number of medication at baseline level by including it as a fixed effect. For patients with T2DM subgroup analyses were performed. All the data were collected via Castor EDC and SPSS and were discussed within the research team and with an expert in statistics. The CONSORT statement was used to evaluate the design of this trial [53].

Results

An overview of the inclusion of CPs and intervention process is shown in Fig. 2. Between November 2020 and December 2021, 35 BENU pharmacies were invited, of which 19 indi- cated a willingness to participate and were included. The CPs were randomized to an intervention (N = 9) and control group (N = 10). One CP in the intervention withdrew after randomization (intervention) due to time constraints caused by unexpected staff shortages. The baseline characteristics of the participating CPs are shown in Table 1.

Fig. 2 Flowchart of the inclusion and intervention process

Table 1 Baseline characteristics of pharmacists

Intervention Control n = 8 n = 10

Sex, female, n (%) 8 (100) 7 (70) Experience as CP, n (%) 0–5 years 4 (50) 5 (50) 6–15 years 4 (50) 2 (20) 16–25 years 3 (30)

Scale of pharmacy Middle-large pharmacy 7 8 Large pharmacy 1 2

721International Journal of Clinical Pharmacy (2025) 47:717–725

The CPs included 140 patients: 71 in the intervention group and 69 in the control group. The baseline characteris- tics of the patients are shown in Table 2.

Table 3 shows the outcomes of the analysis for the inter- vention group compared the control group at T1 and T2, with adjustments for the covariates age, sex, blood pressure and number of medication in the regression analysis.

Among the patients who signed an informed consent form at T1, one or more proposals were made for the deprescrib- ing of cardiovascular and antidiabetic medication in 55% and 46% (p = 0.411) of the patients in the intervention and con- trol groups, respectively (Table 3). With regard to all medi- cation, these percentages were 75% and 70%, (p = 0.752) respectively.

Three months after the CMR (T2) 32% and 26% (p = 0.519) of the patients in the intervention and control groups, respectively, discontinued one or more cardiovascu- lar or antidiabetic medicines. With regard to any medication these percentages were 51% and 36%, (p = 0.221) respec- tively. The number of deprescribed medication was highest in the 'antihypertensive' medication group.

In the intervention group two thirds of the CP proposals to deprescribe resulted in the actual deprescribing of both all and cardiovascular or antidiabetic medication (68%, 59%, p = 0.221). In the control group this was the case for half of the proposals that were implemented (56%, 52%, p = 0.519) (Table 3).

The odds ratio for discontinuing cardiovascular or anti- diabetic medication at T2 was 1.4 (CI 0.65–2.28, p = 0.413)

Table 2 Baseline characteristics of patients at T0

Intervention Control p-value n = 71 n = 69

Age, mean, years (SD) 82 (5) 81 (7) 0.311 Sex, female, n (%) 39 (67%) 34 (50%) 0.098 Systolic blood pressure (mm HG) mean (SD) 129 (12) 131 (11) 0.211 Number of total medication, mean (SD) 11 (3.5) 10 (3) 0.371 Number of cardiovascular and diabetes medication before inter-

vention, mean 3.5 3.6 0.412

Table 3 Univariable and multivariable analyses of primary (T2) and secondary (T1) outcomes, with correction for the covariates age, sex, blood pressure and number of medication

Intervention n = 71 Control n = 69 Univariable regres- sion Odds ratio (CI)

P-value Multivariable regression Odds ratio (CI)

P- value

All patients n = 140 Patients at T1, proposed deprescribing Cardiovascular and/or diabetes medica-

tion, n (%) 39 (55) 32 (46) 1.41 (0.73–2.74) 0.312 1.33 (0.67–2.64) 0.411

Any medication, n (%) 53 (75) 48 (70) 1.29 (0.61–2.70) 0.503 1.14 (0.52–2.49) 0.752 Patients at T2, actual deprescribing Cardiovascular and diabetes medication,

n (%) 23 (32) 18 (26) 1.36 (0.65–2.82) 0.413 1.29 (0.59–2.79) 0.519

Any medication, n (%) 36 (51) 25 (36) 1.81 (0.92–3.56) 0.085 1.59 (0.76–3.36) 0.221 Diabetes mellitus type 2 patients n = 65 Patients at T1, proposed deprescribing Cardiovascular and/or diabetes medica-

tion, n (%) 16 (52) 14 (45) 1.03 (0.38–2.77) 0.951 1.07 (0.36–3.08) 0.908

Any medication, n (%) 22 (71) 26 (76) 0.56 (0.19–1.65) 0.292 0.55 (0.18–1.70) 0.294 Patients at T2, actual deprescribing Cardiovascular and/or diabetes medica-

tion, n (%) 12 (39) 7 (21) 1.58 (0.51–4.92) 0.432 1.69 (0.48–5.87) 0.417

Any medication, n (%) 17 (55) 12 (35) 1.72 (0.64–4.65) 0.286 1.71 (0.59–4.99) 0.322

722 International Journal of Clinical Pharmacy (2025) 47:717–725

for the intervention compared to control CPs. The differ- ences with regard to the deprescribing of cardiovascular and antidiabetic medication before and after correction for confounding variables in the regression models were not statistically significant (Table 3).

Table 4 shows the nature and frequency of medication discontinuation in the control and intervention groups. In total in the control and the intervention groups, 28 and 42 medicines were discontinued, respectively.

Discussion

The present study reports the results of a cluster-randomized controlled trial investigating the effects of a training for CPs on the deprescribing of cardiovascular and antidiabetic medication when conducting a CMR in older patients with polypharmacy. For CPs who participated in the training the number of patients in whom (cardiovascular and antidia- betic) medication was deprescribed was greater than for CPs who conducted CMRs without additional training, but the difference was not statistically significant.

In the present study, CPs proposed deprescribing any medication for 75% and 70% of the intervention and con- trol patients, respectively. Proposals for the deprescribing of cardiovascular and antidiabetic medication were made for 55% and 46% of the patients, respectively. Following discus- sions with GPs and patients, it appeared that after 3 months any medication had actually been deprescribed in 51% and 36% and cardiovascular and antidiabetic medication in 32% and 26% of these patients, respectively. Ultimately, in the intervention group about two thirds and in the control group

approximately half of the proposals were actually imple- mented. The difference was not statistically significant.

In a Canadian CP-led intervention study in older com- munity dwelling patients using medication with a high risk of inducing drug-related problems, oral antidiabetic medica- tion was deprescribed in 30% of the patients who used this medication, for the duration of a 6 months follow-up period [17]. Regardless of the medication used, in the Netherlands follow-up (3 months–1 year) deprescribing rates in CP-led CMR intervention studies in older community dwelling patients ranged from 20 to 50% with those for cardiovascular and antidiabetic medication falling within this range [11, 13, 14, 16]. In studies led by GPs but with CP assistance similar results were obtained [21, 42].

Reviews of both intervention and observational studies on the deprescribing of cardiovascular and antidiabetic medi- cation in different categories of older patients show much wider ranges of deprescribing. For patients with antidiabetic medication with or without a cardiovascular condition rates between 13 and 75% have been reported [27]. A similar wide range (27–94%) was reported in a review on the effects of various deprescribing approaches in older patients with a limited life span or living in nursing homes [32]. However, for community dwelling and nursing home patients with T2DM and low HbA1c levels oral antidiabetic medication deprescribing rates ranging from 14 to 27% were reported whereas for patients with T2DM and a low systolic blood pressure the rates for this medication were between 16 and 19% [29]. A 66% 3- month deprescribing rate of antihyper- tensive medication was reported in a UK trial in primary care patients with hypertension aged 80 years or older [31]. In this study the largest number of discontinued medicines

Table 4 Nature and frequency of medicines deprescribed in the intervention and control and groups at T2

Intervention Control

Non-CVRM medication deprescribed (ATC), n Proton pump inhibitors (A02) 5 2 Medication for benign prostatic hyperplasia (BPH) (G04) 1 1 Pain medication (N02) 1 Inhalation medication (R03) 1 3 Tricyclic antidepressants (N06) 1 1 Antihistamines (R06) 2 Bisphosphonates (M05) 2

CVRM medication group deprescribed, ATC, n Antiarrhythmic medication (C01) 1 1 Antihypertensives (C02-08) 9 7 Antiplatelet medication (B01) 6 6 Loop diuretics (C03) 4 2 Statins (C10) 6 3 Glucose lowering medication (A10) 4 1

723International Journal of Clinical Pharmacy (2025) 47:717–725

consisted of antihypertensives. The trigger for modifying or discontinuing medication is likely a clear parameter, such as blood pressure. Low blood pressure may prompt both HCPs and patients to consider discontinuing medica- tion, or perhaps these patients experienced symptoms such as hypotension. Further research is needed to gain a better understanding of the reasons for discontinuing medication.

The results of the present study are somewhat better than those of previous studies evaluating the extent to which con- ducting a CMR leads to the deprescribing of medication. The full implementation of the Multidisciplinary Guideline ‘Polypharmacy in older patients’, which makes the annual CMR in frail older patients with polypharmacy a manda- tory intervention, and the recent addition of guidance on deprescribing are likely to have increased the deprescribing rate over the years [46]. In the present study the difference between the deprescribing rates in the intervention group and the control group was not significant. Most likely this is due to the fact that in the last decade CPs have become aware of deprescribing, both as the result of experience gained in conducting increasing numbers of CMRs and the already wide attention given to deprescribing in Dutch medical and pharmaceutical journals and at local pharmacotherapeutic audit meetings. Moreover, current Dutch guidelines for the treatment of cardiovascular diseases and T2DM, which have been in use for several years, offer a wide variety of options for the tailored treatment of older patients [35, 36]. The deprescribing of medication in the majority of patients in the present study therefore appears to have been primarily due to the ability of CPs to deprescribe cardiovascular and antidiabetic medication.

Proposals for the discontinuation of medication are not always accepted by GPs and other prescribing HCPs, and patients. Proposals to discontinue or adjust medication were made for 70–75% of the patients participating in this study. Although GPs and other HCPs generally recognize the expertise of CPs, several factors may lead GPs and patients to reject CP proposals to discontinue medication [8, 9, 13, 14, 37–41].

Strengths and weaknesses

A strength of the present study is the use of both CP medica- tion records and GP clinical data in selecting older patients with polypharmacy for a CMR in both the intervention and the control groups. This allowed patients potentially eligi- ble for the deprescribing of cardiovascular and antidiabetic medication to be targeted. Another strength is that depre- scribing training and the deprescribing-focused CMR are embedded in the current practice of conducting mandatory annual CMRs by CPs and GPs.

The inclusion period coincided with the COVID-19 pan- demic, which made the inclusion process challenging. The

initial target for patient inclusion per CP at the onset of the study was 10 patients. However, during the inclusion pro- cess, it became apparent that this target was not feasible due to prevailing circumstances, and this number was not achieved by all CPs. Therefore, the non-significant results may also be due to the study being underpowered.

Selection bias may have occurred because intervention CPs could select patients who are more eligible for depre- scribing. However, we provided both intervention and control CPs with clear inclusion criteria and a protocol for selecting the patients. Moreover, CPs selected patients within the group of patients for whom a CMR with the GP was already planned. We therefore assume that selection bias is minimal. The GP collaboration required in making CMRs was also increasingly difficult due to a shortage of GPs and the resulting lack of GP time [44]. This also means that there is an obvious need for clear criteria and responsibilities, as well as (electronic) tools that allow the efficient transfer of GP clinical information, facilitate the selection of patients and the acquisition of information from patients and ensure that CMRs including the depre- scribing of medication, can be efficiently conducted [8, 9, 12, 18]. In the present study, most CPs had no direct access to blood pressure or HbA1c data. Conducting CMRs and adjusting or deprescribing cardiovascular and antidiabetic medication would be enhanced if PIAS and GIAS were (mandatory) linked so that this information is directly available to CPs. Several studies have indicated that pharmacist-led educational interventions, compared with usual care, resulted in higher rates of discontinua- tion of cardiovascular and antidiabetic medication [17, 54]. Deprescribing in patients with T2DM has been asso- ciated with a reduction in the incidence of hypoglycemia [55]. However, most of these studies were not randomized controlled trials (RCTs), and involved larger sample sizes. Despite this, the evidence regarding the impact of pharma- cist-driven interventions in deprescribing remains incon- clusive [56, 57]. In the present study the clinical effect of medication discontinuation was also not investigated. Finally, a limitation could be that the effects of deprescrib- ing were assessed only after 3 months, and that there are no longitudinal data.

Conclusion

A deprescribing training of CPs did not significantly enhance the deprescribing of cardiovascular and antidiabetic medication in older patients with polypharmacy selected on the basis of dispensing data and specific criteria for blood pressure level and HbA1c value. The increased awareness

724 International Journal of Clinical Pharmacy (2025) 47:717–725

and ability of CPs to deprescribe cardiovascular and anti- diabetic medication and use of GP data, led CPs and GPs to successfully conduct a more deprescribing-focused CMR in daily practice. The need for additional CP training and enhanced GP-CP collaboration to optimize the deprescribing of cardiovascular and antidiabetic medication in conducting CMRs requires further development and study.

Supplementary Information The online version contains supplemen- tary material available at https:// doi. org/ 10. 1007/ s11096- 025- 01863-w.

Acknowledgements The research team would like to sincerely thank all the pharmacists for giving their time and expertise in a difficult period during COVID-19 pandemic.

Funding The Royal Dutch Pharmacists Association (KNMP) provided a Grant (2018-0387). They had no role in the study or in writing and publishing the article.

Conflicts of interest The authors report no conflicts of interests. The authors alone are responsible for the content and writing of the paper.

Open Access This article is licensed under a Creative Commons Attri- bution 4.0 International License, which permits use, sharing, adapta- tion, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

References

1. Leendertse AJ, Egberts AC, Stoker LJ, et al. HARM Study Group. Frequency of and risk factors for preventable medication- related hospital admissions in the Netherlands. Arch Intern Med. 2008;168(17):1890–6.

2. Reeve E, Wiese MD, Mangoni AA. Alterations in drug dis- position in older adults. Expert Opin Drug Metab Toxicol. 2015;4:491–508.

3. Tinetti ME, Han L, Lee DS, et al. Antihypertensive medications and serious fall injuries in a nationally representative sample of older adults. JAMA Intern Med. 2014;174(4):588–95.

4. Bejan-Angoulvant T, Saadatian-Elahi M, Wright JM, et al. Treat- ment of hypertension in patients 80 years and older: the lower the better? A meta-analysis of randomized controlled trials. J Hyper- tens. 2010;28:1366–72.

5. Wallace E, Mcdowell R, Bennett K, et al. Impact of potentially inappropriate prescribing on adverse drug events, health related quality of life and emergency hospital attendance in older people attending general practice: a prospective cohort study. J Gerontol A Biol Sci Med Sci. 2017;72:271–7.

6. Reeve E, Moriarty F, Nahas R, et al. A narrative review of the safety concerns of deprescribing in older adults and strat- egies to mitigate potential harms. Expert Opin Drug Saf. 2018;17(1):39–49.

7. Reeve E, Gnjidic D, Long J, et al. A systematic review of the emerging definition of “deprescribing” with network analysis: implications for future research and clinical practice. Br J Clin Pharmacol. 2015;80(6):1254–68.

8. Radcliffe E, Servin R, Cox N, et al. What makes a multidiscipli- nary medication review and deprescribing intervention for older people work well in primary care? A realist review and synthesis. BMC Geriatr. 2023;23:591.

9. Robinson M, Mokrzecki S, Mallett AJ. Attitudes and barriers towards deprescribing in older patients experiencing polyphar- macy: a narrative review. NPJ Aging. 2024;10:6.

10. Leendertse AJ, De Koning GHP, Goudswaard AN, et al. Prevent- ing hospital admissions by reviewing medication (PHARM) in primary care: an open controlled study in an elderly population. J Clin Pharm Ther. 2013;38:379–87.

11. Kempen TGH, Van de Steeg-van Gompel CH, Hoogland P, et al. Large scale implementation of clinical medication reviews in Dutch community pharmacies: drug-related problems and inter- ventions. Int J Clin Pharm. 2014;36:630–5.

12. Mast R, Ahmad A, Hoogenboom SC, et al. Amsterdam tool for clinical medication review: development and testing of a compre- hensive tool for pharmacists and general practitioners. BMC Res Notes. 2015;8:642.

13. Chau SH, Jansen AP, van de Ven PM, et al. Clinical medication reviews in elderly patients with polypharmacy: a cross-sectional study on drug-related problems in the Netherlands. Int J Clin Pharm. 2016;38(1):46–53.

14. Geurts MM, Stewart RE, Brouwers JR, et al. Implications of a clinical medication review and a pharmaceutical care plan of polypharmacy patients with a cardiovascular disorder. Int J Clin Pharm. 2016;38:808–15.

15. Willeboordse F, Schellevis FG, Chau SH, et al. The effectiveness of optimised clinical medication reviews for geriatric patients: Opti-Med a cluster randomised controlled trial. Fam Pract. 2017;34:437–45.

16. Hurmuz MZM, Janus SIM, van Manen JG. Changes in medicine prescription following a medication review in older high-risk patients with polypharmacy. Int J Clin Pharm. 2018;40:480–7.

17. Martin P, Tamblyn R, Benedetti A, et al. Effect of a pharmacist-led educational intervention on inappropriate medication prescrip- tions in older adults: the D-PRESCRIBE randomized clinical trial. JAMA. 2018;320:1889–98.

18. Schoenmakers TW, Wensing M, De Smet PA, et al. Patient- reported common symptoms as an assessment of interventions in medication reviews: a randomised, controlled trial. Int J Clin Pharm. 2018;40:126–34.

19. Wastesson JW, Lucas Morin L, Edwin C, et al. An update on the clinical consequences of polypharmacy in older adults: a narrative review. Expert Opin Drug Saf. 2018;17(12):1185–96.

20. Alsherhi AA, Jalal Z, Cheema E, et al. Impact of the pharmacist- led intervention on the control of medical cardiovascular risk factors for the primary prevention of cardiovascular disease in general practice: a systematic review and meta-analysis of ran- domised controlled trials. Br J Clin Pharmacol. 2020;86:29–38.

21. Bosch-Lenders D, Jansen J, Stoffers HEJH, et al. The effect of a comprehensive, interdisciplinary medication review on quality of life and medication use in community dwelling older people with polypharmacy. J Clin Med. 2021;10:600.

22. Imfeld-Isenegger TL, Branco Suares I, Nabergoj Makovec U, et al. Community-pharmacist-led medication review procedures across Europe: characterization, implementation and remuneration. Res Social Adm Pharm. 2020;16:1057–66.

23. van Poelgeest E, Seppala L, Bahat G, et al. EuGMS SIG Pharma- cology. Optimizing pharmacotherapy and deprescribing strate- gies in older adults living with multimorbidity and polypharmacy:

https://doi.org/10.1007/s11096-025-01863-w

http://creativecommons.org/licenses/by/4.0/

725International Journal of Clinical Pharmacy (2025) 47:717–725

EuGMS SIG on pharmacology position paper. Eur Geriatr Med. 2023;14(6):1195–209.

24. Sussman JB, Kerr EA, Saini SD, et al. Rates of deintensification of blood pressure and glycemic medication treatment based on levels of control and life expectancy in older patients with diabetes mellitus. JAMA Intern Med. 2015;175:1942–9.

25. Farrell B, Black C, Thompson W, et al. Deprescribing anti- hyperglycemic agents in older patients. Can Fam Physician. 2017;63:832–43.

26. Boels AM, Hart HE, Rutten GE, et al. Personalised treatment targets in type 2 diabetes patients: the Dutch approach. Prim Care Diabetes. 2017;11:71–7.

27. Seidu S, Kunutsor SK, Topsever P, et al. Deintensification in older patients with type 2 diabetes: a systematic review of approaches, rates and outcomes. Diabetes Obes Metab. 2019;21:1668–79.

28. Scott IA, Hilmer SN, Le Couteur DG. Going beyond the guide- lines in individualising the use of antihypertensive drugs in older patients. Drugs Aging. 2019;36:675–85.

29. Oktora MP, Kerr KP, Hak E, et al. Rates, determinants and success of implementing deprescribing in people with type 2 diabetes: a scoping review. Diabet Med. 2021;38: e14408.

30. Reeve E, Jordan V, Thompson W et al. Withdrawal of antihy- pertensive drugs in older people. Cochrane Database Syst Rev 2020;6.

31. Sheppard JP, Burt J, Lown M, et al. Effect of antihypertensive medication reduction vs usual care on short-term blood pressure control in patients with hypertension aged 80 years and older: the OPTIMISE randomized clinical trial. JAMA. 2020;323:2039–51.

32. Hickman E, Seawoodharry CG, Khunti K, et al. Deprescribing in cardiometabolic conditions in older patients: a systematic review. GeroScience. 2023;45:3491–512.

33. Benetos A. How to obtain more evidence for the management of hypertension in frail patients over 80 years old? Eur Geriatr Med. 2018;9(2):137–40.

34. Krishnaswami A, Steinman MA, Goyal P, et al. Deprescribing in older adults with cardiovascular disease. Am Coll Cardiol. 2019;73:2584–95.

35. Guideline Cardiovascular Risk Management (CVRM). Dutch General Practitioners Association (NHG), Utrecht. 2019. https:// richt lijnen. nhg. org/ stand aarden/ cardi ovasc ulair- risic omana gement Accessed 10 July 2024.

36. Guideline Diabetes Mellitus Type 2. Dutch General Practitioners Association (NHG), Utrecht. 2018. https:// richt lijnen. nhg. org/ stand aarden/ diabe tes- melli tus- type-2 Accessed 10 July 2024.

37. van Middelaar T, Ivens SD, van Peet PG, et al. Prescribing and deprescribing antihypertensive medication in older people by Dutch general practitioners: a qualitative study. BMJ Open. 2018;8(4): e020871.

38. Crutzen S, Baas G, Abou J, et al. Barriers and enablers of older patients to deprescribing of cardiometabolic medication: a focus group study. Front Pharmacol. 2020;11:1268.

39. Goyal P, Anderson TA, Bernacki GM, et al. Physician perspectives on deprescribing cardiovascular medications for older adults. J Am Geriatr Soc. 2020;68:78–96.

40. Abou J, Crutzen S, Tromp V, et al. Barriers and enablers of health- care providers to deprescribe cardiometabolic medication in older patients: a focus group study. Drugs Aging. 2022;39(3):209–21.

41. Brunner L, Rodondi N, Aubert CE. Barriers and facilitators to deprescribing of cardiovascular medications: s systematic review. BMJ Open. 2022;12: e061686.

42. Luymes CH, Van der Kleij RMJJ, Poortvliet RKE, et al. Depre- scribing potentially inappropriate preventive cardiovascular

medication: barriers and enablers for patients and general practi- tioners. Ann Pharmacother. 2016;50:446–54.

43. Achterhof AB, Rozsnyai Z, Reeve E, et al. Potentially inappropri- ate medication and attitudes of older adults towards deprescribing. PLoS ONE. 2020;15: e0240463.

44. Hogervorst S, Adriaanse MC, Vervloet M, et al. A survey on the implementation of clinical medication reviews within a multidis- ciplinary setting. BMC Health Serv Res. 2024;24:575.

45. Bakker L, Kemper PF, Wagner C, et al. A baseline assessment by healthcare professionals of Dutch pharmacotheraeutic care for the elderly with polypharmacy. Eur J Public Health. 2017;27:679–86.

46. MDRP 2020: Nederlands Huisartsen Genootschap (NHG) [Dutch College of General Practitioners in collaboration with other pro- fessional organisations]. Multidisciplinaire Richtlijn Polyfarma- cie bij ouderen [Multidisciplinary Guideline Polypharmacy in Elderly]. Update 2020. https:// richt lijnen. nhg. org/ multi disci plina ire- richt lijnen/ polyf armac ie- bij- ouder en. Accessed 10 July 2024.

47. Baas G, Crutzen S, Smits S, et al. Process evaluation of a phar- macist-led intervention aimed at deprescribing and appropriate use of cardiometabolic medication among adult people with type 2 diabetes. Basic Clin Pharmacol Toxicol. 2023;134:83–96.

48. Sinnige J, Korevaar JC, Van Lieshout J, et al. Medication man- agement strategy for older people with polypharmacy in general practice. A qualitative study on prescribing behaviour in primary practice. Br J Gen Pract 2016;66(649):e540–51.

49. Crutzen S, Schulling J, Hugtenburg JG, et al. Development and piloting of an algorithm to select older patients for different types of medication review. Front Pharmacol. 2019;10:217.

50. Sallevelt BTGM, Huibers CJA, Knol W, et al. Evaluation of clarity of the STOPP/START criteria for clinical applicability in pre- scribing for older people: a quality appraisal study. BMJ Open. 2020;10: e33721.

51. Castor EDC. Castor Electronic Data Capture 2019 [27 Aug. 2019].: https:// casto redc. com Accessed 10 July 2024.

52. Ibrahim K, Cox NJ, Stevenson JM, et al. A systematic review of the evidence for deprescribing interventions among older people living with frailty. BMC Geriatr. 2021;21(1):258.

53. Moher D, Hopewell S, Schulz KF, et al. CONSORT 2010 expla- nation and elaboration: updated guidelines for reporting parallel group randomised trials. BMJ. 2010;340: c869.

54. McCrary M, Goldstein J. Pharmacist-led education to discontinue inappropriate prescribing. JAMA. 2019;321(13):1313–4.

55. Hui RL, Chang CC, Niu F, et al. Evaluation of a pharmacist-man- aged antidiabetic deprescribing program in an integrated health care system. J Manag Care Spec Pharm. 2019;25(8):927–34.

56. Bužančić I, Kummer I, Držaić M, et al. Community-based phar- macists’ role in deprescribing: a systematic review. Br J Clin Phar- macol. 2022;88(2):452–63.

57. Christopher CM, Blebil AQ, Bhuvan KC, et al. Assessing feasibil- ity of conducting medication review with follow-up among older adults at community pharmacy: a pilot randomised controlled trial. Int J Clin Pharm. 2024;46(4):843–53. https:// doi. org/ 10. 1007/ s11096- 024- 01711-3.

Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

https://richtlijnen.nhg.org/standaarden/cardiovasculair-risicomanagement

https://richtlijnen.nhg.org/standaarden/diabetes-mellitus-type-2

https://richtlijnen.nhg.org/multidisciplinaire-richtlijnen/polyfarmacie-bij-ouderen

https://castoredc.com

https://doi.org/10.1007/s11096-024-01711-3

Pharmacist-led deprescribing of cardiovascular and diabetes medication within a clinical medication review: the LeMON study (Less Medicines in Older Patients in the Netherlands), a cluster randomized controlled trial

Abstract

Background
Aim
Method
Results
Conclusion

Impact statements
Introduction

Aim
Ethics approval

Method

Study design and setting
LeMON intervention
Patient selection
Sample size calculation
Primary outcome measures
Secondary outcome measures
Data analysis

Results
Discussion

Strengths and weaknesses

Conclusion
Acknowledgements
References

image5.emf

https://doi.org/10.1177/15394492261442163

OTJR: Occupational Therapy Journal of Research

1 –11 © The Author(s) 2026

Article reuse guidelines: sagepub.com/journals-permissions

DOI: 10.1177/15394492261442163 journals.sagepub.com/home/otj

1442163OTJXXXXXX10.1177/1539449226144216310.1177/15394492261442163OTJR: Occupational Therapy Journal of ResearchSomerville et al. research-article20262026

Feasibility, Acceptability, and Preliminary Efficacy of a Medication Intervention for Older Adults

Emily Somerville1 , Rebecca M. Bollinger1 , Meghan Haxton1, Brianna Holden1, Szu-Wei Chen1, Audrey Keleman1, Behnaz Sarrami2, Yan Yan1, and Susan Stark1

Abstract Interventions to address polypharmacy and improve adherence are insufficient for community-dwelling older adults. To examine the feasibility, acceptability, and preliminary efficacy of a tailored medication management intervention (TIMM) to decrease inappropriate polypharmacy and improve medication management in community-dwelling older adults. We completed a single-blind, parallel-group randomized controlled trial. The intervention group received TIMM, including a comprehensive medication review by a pharmacist and an in-home intervention by an occupational therapist. About 78.8% of participants completed the study (n = 26). TIMM participants adhered to treatment recommendations from the occupational therapist (94%) and pharmacist (58%). At 6 months, TIMM participants experienced a significant reduction in medication management barriers (p < .01) and improvement in performance (p < .01). No significant differences were found between groups in reducing inappropriate medications (p = .302) or adherence (p = .458). TIMM is a feasible and acceptable intervention for community-dwelling older adults. It reduces barriers to medication management and improves medication management.

Plain Language Summary Is a personalized medication management program practical, well-liked, and helpful for older adults who live at home? Why was the study done? Many older adults take several medications at the same time that can cause harmful side effects, falls, and confusion. Many older adults are also not able to take their medications correctly, which leads to more problems. Most programs to help older adults with their medications only focus on one medication or use the same solution for everyone, like giving out pillboxes. These programs usually occur in clinics and don’t always consider the person’s needs or home setup, where they usually take medication. What did the researchers do? The researchers developed a new program called TIMM (tailored intervention for medication management). During TIMM, a pharmacist and an occupational therapist work together with the older adult to reduce unnecessary medications and to help the older adult manage their medications more easily at home. What did the researchers find? Thirty-three older adults participated, and most completed the study. Participants used most of the occupational therapy recommendations, such as using pill organizers and alarm clocks, but followed fewer of the medication changes suggested by the pharmacist. TIMM helped older adults fix problems in their homes that made it hard to take medications and helped them perform better on medication tasks, such as reading labels and opening pill bottles. There were no significant differences between the TIMM group and the control group in reducing the number of unnecessary medications or taking medication correctly. One reason for this was that some doctors did not respond to the pharmacist about stopping or simplifying medications. What do the findings mean? TIMM is a feasible and well-accepted program that improves how older adults manage medications at home. More work was needed to improve collaboration with doctors and to test whether TIMM can reduce unnecessary medications and improve the ability to take medication correctly for more older adults.

Keywords polypharmacy, medication management, older adults

Article

http://crossmark.crossref.org/dialog/?doi=10.1177%2F15394492261442163&domain=pdf&date_stamp=2026-05-14

2 OTJR: Occupational Therapy Journal of Research 00(0)

Introduction

Polypharmacy, or concurrent use of multiple medications, is highly prevalent among community-dwelling older adults (Charlesworth et al., 2015). It is estimated that 50% of older adults take five or more medications to treat multiple chronic health conditions (Brown & Bussell, 2011; Gould & Mitty, 2010). Polypharmacy can be dangerous in older adults. It increases the risk of inappropriate medication, duplicate medication therapies, drug–drug interactions, and geriatric syndromes including falls, delirium, cognitive impairment, and functional decline (Charlesworth et al., 2015).

In addition to the risks of polypharmacy, an estimated 40% to 75% of older adults are not able to correctly manage their medications (Brown & Bussell, 2011). Decreased medi- cation adherence is responsible for approximately $290 bil- lion in added health care costs each year and is responsible for one-third to one-half of all medication-related hospital admissions (Iuga & McGuire, 2014; Wolf et al., 2011). Older adults who are not able to take their medications as pre- scribed experience negative health outcomes, including additional functional decline, increased medication side effects, institutionalization, and even death (Chisholm-Burns & Spivey, 2012; Haynes et al., 2008; Nieuwlaat et al., 2014; Osterberg & Blaschke, 2005).

Deprescribing interventions are effective in reducing polypharmacy, but alone they are insufficient to impact med- ication adherence in community-dwelling older adults (Cooper et al., 2015; Marcum et al., 2021; Nieuwlaat et al., 2014; Ulley et al., 2019). Most medication management interventions usually only target one type or class of medica- tion or provide only one intervention strategy (e.g., a pillbox) to all participants (Clifford-Middel, 2004; Fritz, 2014; Haltiwanger, 2012; Kini & Ho, 2018; Sabaté, 2003; Webb et al., 2001). These interventions neglect the individual abili- ties, circumstances, and complexity of the medication rou- tines of each older adult (Gellad et al., 2011; George et al., 2008; Holt et al., 2014; Osterberg & Blaschke, 2005). Furthermore, most interventions designed for older adults occur in a clinical setting (Bakar et al., 2016; Nazareth et al., 2001), which does not consider the complexity of the home environment where medication management typically occurs (Farris & Phillips, 2008; Sorensen et al., 2006).

Occupational therapy plays a vital role in improving med- ication management for older adults with polypharmacy. As outlined in the Occupational Therapy Practice Framework: Domain and Process Fourth Edition (Boop et al., 2020) health management domain and the American Occupational

Therapy Association (AOTA, 2017) position paper on occu- pational therapy and medication management, occupational therapists (OTs) analyze, assess, and address older adults’ actual performance of specific medication management activities as part of their daily routines. However, evidence supporting occupational therapy interventions for medica- tion management in the home are limited to preliminary find- ings (Schwartz et al., 2017) or no intervention (Sanders & Van Oss, 2013; Schwartz & Smith, 2017; Yau, 1995).

To fill this gap, we developed the tailored intervention for medication management (TIMM). TIMM uses an interpro- fessional team including an OT and a pharmacist in collabo- ration with the older adult’s prescribing physicians to accomplish two important goals: (a) reduce inappropriate polypharmacy and (b) reduce barriers to successful medica- tion management, or taking medications as prescribed, for community-dwelling older adults. The purpose of this study was to determine the feasibility, acceptability, and prelimi- nary efficacy of TIMM.

Method

Study Design and Procedures

This study was a single-blind, parallel-group randomized controlled trial that adhered to the Consolidated Standards of Reporting Trials (CONSORT) guidelines. All participants received a baseline evaluation (T1) via telehealth with an OT interventionist. The T1 evaluation consisted of (a) acquiring a medication list from participants including medication name, dosage, frequency, and prescribing physician; (b) a performance-based assessment to measure medication man- agement ability; (c) evaluation of intrinsic and extrinsic risk factors of medication nonadherence; and (d) evaluation of environmental barriers impacting the older adult’s ability to successfully manage medication or take medications as pre- scribed (AOTA, 2017). Assessments delivered at all time- points, including T1, are listed in Table 1. The Washington University in St. Louis Human Research Protection Office reviewed and approved the study procedures (#202008139). The protocol, including a description of the sample size, has been published elsewhere (Somerville et al., 2023).

After the T1 evaluation was completed, participants were randomized to receive TIMM or control using Research Electronic Data Capture (REDCap; Harris et al., 2009). Randomization was stratified by number of medications (9 or fewer medications vs. 10 or more medications). Within each stratum, participants were allocated using a 1:1 ratio via

1Washington University in St. Louis, USA 2Missouri Pharmacogenomics Consulting, St. Louis, MO, USA

Corresponding Author: Emily Somerville, Program in Occupational Therapy, School of Medicine, Washington University in St. Louis, 5232 Oakland Ave., St. Louis, MO 63110, USA. Email: [email protected]

Table 1. Primary and Secondary Endpoints for the Study and Rationale for Inclusion.

Endpoints Outcome Measure Explanation Collection point

Primary endpoints: feasibility and acceptability

Recruitment and retention

Screening log Recruitment; retention rate T1, T2, T3

Dose Treatment notes Minutes and number of treatment sessions

Fidelity Visit checklist Session checklist of components delivered by pharmacist and OT

Safety Fall questionnaire ED visits, unplanned hospitalizations, doctor visits, falls

T1, T2, T3

Cost Cost log Pharmacist time, OT time, cost of modifications

Secondary endpoint: Efficacy

Medication management ability

In-Home Medication Management Assessment (HOME-Rx)

Performance-based assessment used to measure medication management ability (Murphy et al., 2017)

T1, T2, T3

Medication appropriateness

Beers Criteria for Potentially Inappropriate Medication (Panel et al., 2019)

Identifies inappropriate medications T1, T2, T3

Note. T1 = baseline; T2 = immediate post-treatment; T3 = 6-month postevaluation; Tx = treatment. ED = emergency department; OT = occupational therapist.

Somerville et al. 3

randomization sequences generated a priori by the study stat- istician. Follow-up assessments occurred immediately after the intervention or control visits (T2) and again 6-month postbaseline evaluation (T3) with a rater blind to interven- tion and group allocation.

Participants

The study took place in the homes of older adults living in the greater St. Louis metropolitan area in the United States. Potential participants were recruited from a local aging ser- vices organization, a research participant registry, and a list of previous research participants who agreed to be contacted for future studies. Participants were recruited if they met the following inclusion criteria: (a) age 65 years or older, (b) pre- scribed five or more medications, and (c) reported decreased medication adherence as indicated by one or more “yes” responses on the Medication Adherence Rating Scale (MARS; Fialko et al., 2008). Older adults were excluded if they resided in an institutional setting or had a significant cognitive impairment as indicated by a score of 10 or more on the Short Blessed Test (Katzman et al., 1983).

Intervention Development

TIMM was developed based on a competence-press model (Lawton & Nahemow, 1973) that posits changes to the physi- cal environment (e.g., easy-open pill bottles), matched with the older adult’s pattern of functional loss (e.g., decreased grip strength), will improve performance. The two essential elements (Whyte & Hart, 2003) of TIMM are as follows: (a) medication deprescribing and (b) tailored compensatory strat- egies and supports provided in the home to reduce barriers to

medication management. Individualized, home-based occu- pational therapy interventions for community-dwelling older adults using environmental support are effective in improving daily activity performance (Keglovits & Stark, 2020; Stark et al., 2017; Stark et al., 2009), so the same approach was implemented in TIMM.

Treatment Session Overview. TIMM includes a total of three visits (one evaluation/baseline visit and two treatment visits) that last 75 min each, over 4 weeks. Sessions/visits are spaced 1 week apart. Figure 1 outlines the study sessions.

After the baseline visit, the results were sent to the study pharmacist for review. The pharmacist used a standardized protocol (Garfinkel & Mangin, 2010) to review prescription and over-the-counter medications, vitamins, and supple- ments to identify any potentially inappropriate medications. This is an essential first step in a medication management intervention because improving adherence to inappropriate medications would only increase the negative impact of polypharmacy (Ulley et al., 2019). The pharmacist made additional recommendations as indicated, including sugges- tions for routine simplification (e.g., recommended taking a medication at the same time as other medications instead of the middle of the day by itself) or decreases in pill burden (e.g., recommended one 40 mg pill instead of two 20 mg pills of the same medication). All recommended changes were presented to the physician, and upon approval, changes were communicated to the older adult and the OT and were updated on the older adult’s medication list.

The next step of the intervention included two visits from an OT. The OT and the older adult worked together to identify tailored compensatory strategies and supports, including adap- tive equipment, to remediate extrinsic risk factors and modify

Method

Study Design and Procedures

Table 1. Primary and Secondary Endpoints for the Study and Rationale for Inclusion.

Endpoints Outcome Measure Explanation Collection point

Primary endpoints: feasibility and acceptability

Recruitment and retention

Screening log Recruitment; retention rate T1, T2, T3

Dose Treatment notes Minutes and number of treatment sessions

Fidelity Visit checklist Session checklist of components delivered by pharmacist and OT

Safety Fall questionnaire ED visits, unplanned hospitalizations, doctor visits, falls

T1, T2, T3

Cost Cost log Pharmacist time, OT time, cost of modifications

Secondary endpoint: Efficacy

Medication management ability

In-Home Medication Management Assessment (HOME-Rx)

Performance-based assessment used to measure medication management ability (Murphy et al., 2017)

T1, T2, T3

Medication appropriateness

Beers Criteria for Potentially Inappropriate Medication (Panel et al., 2019)

Identifies inappropriate medications T1, T2, T3

Note. T1 = baseline; T2 = immediate post-treatment; T3 = 6-month postevaluation; Tx = treatment. ED = emergency department; OT = occupational therapist.

4 OTJR: Occupational Therapy Journal of Research 00(0)

the home environment to improve medication adherence. For example, if an older adult had difficulty remembering to take their medication on time, the OT and the older adult worked together to identify feasible strategies to remind the client (e.g., a medication alarm clock, setting an alarm on their phone, or a reminder note placed by an already occurring activity). The older adult chose which solution they preferred, and the OT interventionist provided education and training on its use. The older adult engaged in active practice using the compensatory strategies and supports in the home in an effec- tive and safe manner. The OT ensured that the new strategies and supports met the older adult’s needs so they could com- plete medication management as independently as possible.

Telehealth Adaptation

Prior to the start of the study, TIMM was adapted from in- person to remote delivery due to the COVID-19 pandemic. After pilot testing multiple devices and HIPAA-compliant telehealth platforms, a study-provided cell phone mounted on a chest harness was used for a hands-free, outward-facing iPhone with a camera, using Zoom or Doxy.me with an OT present virtually. Prior to each study visit, the technology was delivered to the participant’s home by a study team member who was also available to provide assistance setting up the technology if necessary. Once the chest harness and cell phone were set up, the OT verified that the participant could hear adequately prior to beginning the T1 evaluation or TIMM portion of the study.

Attention Control

Participants in the control group experienced the same effects of time and attention in the home but no effect on the out- come of interest. Because this was a study of a new

intervention, there was no opportunity for a usual care arm with dose equivalency. The attention control group received two 75-min attention phone calls during which a trained research assistant completed a semistructured interview using a discussion guide about leisure interests and experi- ences. All control group participants were offered the inter- vention upon study completion.

Therapist Training

OT interventionists and raters were trained by the study prin- cipal investigator (E.S.) using didactic lecture, group discus- sion, and case studies to complete T1 and TIMM. All OTs completed a written examination and checkout and were required to achieve a score of 95% using an intervention fidelity checklist. Raters were trained to conduct semistruc- tured interviews for the control group. Additional raters were trained to administer assessments during the in-home T2 and T3 visits. Inter-rater reliability of 0.90 was achieved through a simulated case with an older adult before raters were certi- fied to collect data.

Due to the COVID-19 pandemic, the OTs participated in additional training to deliver TIMM remotely. Training included review of additional clarifying questions for tasks that might be harder for the OT to see on camera or prompts to tilt the camera up or down. For example, for medication storage, the OT may need to ask questions such as, “Show me where you keep your medication,” and “Are there any additional pill bottles on your bedside table or dresser? How about in any drawers?” When preparing to sort medications, after the participant reads the medication label out loud, the OT may need to say, “Bring the pill bottle closer to you so I can see the label,” or “Was it hard for you to open that bottle or read the label?” When opening the pill bottles or sorting the medication, the OT may say, “Please check the floor to

Figure 1. Overview of the tailored intervention for medication management (TIMM).

Somerville et al. 5

see if you dropped any pills,” or “Show me your pill orga- nizer” to verify that the correct pills and the correct number of pills are located in the correct location.

Measures

Demographic characteristics of the study population, includ- ing age, gender, race, years of education, and living situation, were documented. All outcome measures were administered at the baseline evaluation (T1), immediately after the inter- vention or control visits (T2), and 6-month postbaseline evaluation (T3).

Primary Outcome Measures: Feasibility and Acceptability. To determine the feasibility and acceptability of TIMM, we examined recruitment (feasibility) and retention (acceptabil- ity) rates of study participants using the study screening log and study dropout form. Feasibility and acceptability rates of 70% to 80% or higher are considered good in behavioral research studies (Abshire et al., 2017; Sekhon et al., 2017). The dose of the intervention was tracked using treatment notes, including the treatment minutes per visit as well as the number of visits. Fidelity to the intervention was measured using a visit-by-visit checklist. Both the OTs and pharmacist used a checklist (included in study materials) to document completion of the required components of each visit. Safety was measured by the number of falls, emergency department visits, unplanned hospitalizations, and doctor visits experi- enced by older adults in each group, which were collected via self-report over the phone.

Secondary Outcome Measures. The In-Home Medication Management Performance Evaluation (HOME-Rx; Murphy et al., 2017) is a three-part performance-based medication management assessment that was administered by the OT. It assesses an individual’s ability to perform 11 tasks related to medication management (e.g., reading labels and opening pill bottles). The older adult’s ability to perform each task is scored on a scale of 0 to 4, with 4 indicating independence in the task and 0 indicating dependence or inability to complete the task, for a total possible score of 44 for the performance subscale. Higher total scores indicate greater independence in tasks related to medication management. Environmental barriers that impact the older adult’s ability to perform each task are also identified and scored on a scale from 0 to 5, with 5 indicating that the barrier completely impedes performance and 0 indicating that the environment does not impact the performance of the task. Although the HOME-Rx had not previously been delivered via telehealth, it was adapted for remote administration and was pilot tested iteratively with eight older adults prior to use in the study to ensure that the assessment could be conducted accurately.

Medication adherence was measured by the pharmacist using the MARS (Fialko et al., 2008). The MARS is a 10-item self-report measure with scores ranging from 0 to

10. Lower scores indicate greater adherence to medication. Potentially inappropriate medications were identified accord- ing to the Beers Criteria for Potentially Inappropriate Medication (Fick et al., 2003), which identifies medications for which potential harm outweighs the expected benefit in older adults. The number of inappropriate medications, iden- tified by pharmacist and confirmed by physician, was docu- mented for each older adult.

Statistical Analysis

Descriptive statistics were used to explore baseline charac- teristics. Independent t-tests and chi-square tests, or similar nonparametric tests, were used to compare the differences between the two groups at baseline to confirm that random- ization was successful.

Primary Outcomes: Feasibility and Acceptability. Recruitment and retention of TIMM were measured using recruitment metrics, including participation rate, reasons for decline (e.g., not interested in the intervention), and demographic characteristics of the study population. Dose, fidelity, and safety were compared between groups using independent t-tests or chi-square tests. Characteristics of older adults who completed the intervention were compared with those who did not.

Secondary Outcome: Efficacy. Because this was a feasibility study, the analyses were not focused on hypothesis testing, as the study was not powered to detect significant differences (Orsmond & Cohn, 2015). However, we used nonparametric tests to examine the preliminary efficacy of TIMM to improve performance and decrease environmental barriers related to medication management, reduce potentially inap- propriate polypharmacy, and improve medication adherence compared with the control group.

Results

Primary Outcomes: Feasibility and Acceptability

Thirty-four participants were enrolled in a 12-month recruit- ment period. Seventeen participants were allocated to the treatment group and 16 to the control. One participant from the control group was withdrawn after randomization because he was no longer eligible. Fifteen participants from the TIMM intervention group and 12 from the control group completed T2 (immediate postintervention or control visits). Thirteen participants from the treatment group and 13 from the control completed T3 (6-month postevaluation follow-up visit; Figure 2). There were no significant differences in age, education, gender, or living situation between groups at baseline (Table 2).

TIMM had a high retention rate of 94.1% at treatment completion and 76.5% at study completion. Treatment group

6 OTJR: Occupational Therapy Journal of Research 00(0)

participants received an average of 52 min of treatment, with seven participants requiring only one visit, and 10 requiring two visits. Approximately 65.2% of older adults needed some level of assistance from study personnel to start the call for T1, such as answering the Zoom call, adjusting the call vol- ume, or adjusting the placement and positioning of the cam- era (Table 3). There were no significant differences in falls or

unplanned hospitalizations, emergency department visits, or doctor visits between groups at T3 (Table 3). The OT and the pharmacist were able to implement the intervention with high fidelity (84.7% and 89.0%, respectively). The majority of modifications included individualized pill sorters, large-print medication lists, and medication alarm clocks. The most com- mon types of OT recommendations were adaptive equipment

(n = 67, 75.3%), including pill organizers (daily, weekly, travel; n = 28), updated or laminated medication lists (n = 9), medication alarm clocks/setting alarms (n = 16), magnifying glasses (n = 7), and other (e.g., pill cutters and pill bags from pharmacy; n = 7). The average cost of adaptive equipment was $51.10 USD (SD $18.20). Adherence at 6 months to the implemented adaptive equipment recommendations was excellent (94.7%), and adherence to the medication regimen corrections was 58% (Table 3).

Secondary Outcome: Efficacy

The secondary outcome of this study was to evaluate the pre- liminary efficacy of TIMM to improve performance in

Table 2. Baseline Demographic Characteristics.

Characteristics TIMM intervention (n = 17) Control (n = 15)

Age, M (SD) 72.9 (5.5) 73.5 (4.5) Gender, female, n (%) 8 (47.1) 9 (60.0) Race, Black, n (%) 9 (52.9) 7 (46.7) Years of education, M (SD) 14.4 (3.1) 13.9 (1.7) Living situation, lives alone, n (%) 12 (70.6) 12 (80.0) Number of prescriptions, M (SD) 13.7 (6.6) 13.5 (8.7) Number of vitamins, M (SD) 2.6 (3.4) 1.9 (2.4) Number of OTC, M (SD) 5.7 (4.9) 3.9 (3.5) Number of HS, M (SD) 1.53 (3.3) 0.3 (.6) Total number of meds, M (SD) 23.5 (9.6) 19.6 (10.7) HOME-Rx Performance subscale 35.6 (5.4) 33.8 (6.3) MARS 4.5 (1.1) 3.7 (0.8) Fall in home in past year, n (%) 8 (47.1) 6 (40.0)

Note. M = mean; SD = standard deviation; OTC = over-the-counter medications; HS = herbal supplements; HOME-RX = In-Home Medication Management Performance Evaluation; MARS = Medication Adherence Rating Scale.

Table 3. Process Outcomes at 6 Months.

Characteristics TIMM intervention (n = 13) Control (n = 13)

Recruitment rate (%) 65.4 65.4 Retention rate (%) 82.4 81.3 Assistance required to start call, n (%) 15 (65.2) – Owns technology to complete telehealth visits, n (%) 9 (52.9) – Dose Number of treatment sessions, M (SD) 1.6 (0.5) – Minutes of sessions, M (SD) 52.1 (26.7) – Tx fidelity, OT (%) 84.7 – Tx fidelity, pharmacist (%) 89.0 – Safety, M (SD) ED visits 0.17 (0.0) 0.50 (0.1) Unplanned hospitalizations 0.25 (0.1) 0.36 (0.5) Doctor visits 2.0 (1.6) 2.3 (1.4) Falls 0.61 (0.8) 0.92 (1.7) Cost of modifications, M (SD) 51.1 (18.2) – Adherence to adaptive equipment (%) 94.7 – Adherence medication regimen corrections (%) 58 –

Note. M = mean, SD = standard deviation; Tx = treatment; OT = occupational therapist; ED = emergency department.

Figure 2. Study flow through.

Somerville et al. 7

Secondary Outcome: Efficacy

The secondary outcome of this study was to evaluate the pre- liminary efficacy of TIMM to improve performance in

medication management tasks, decrease environmental barriers to medication management, reduce polypharmacy, and improve adherence to the medication routine. The HOME-Rx was used to measure the performance of medication management tasks and environmental barriers impacting the older adult’s ability to complete those tasks. There were no significant differences in the performance or environmental barriers subscales of the HOME-Rx between groups at baseline. For the TIMM inter- vention group, there was a significant decrease in the impact of the barriers on performance between baseline (T1; Mdn = 19.0) and 6-month follow-up (T3; Mdn = 2.5), T = 0.00, p = .001, r = −.062. There was also a significant difference for the control group between T1 (Mdn = 21.0) and T3 (Mdn = 11.0), T = 5.5, p = .005, r = −.53. Barriers were reduced significantly for the

Table 2. Baseline Demographic Characteristics.

Characteristics TIMM intervention (n = 17) Control (n = 15)

Table 3. Process Outcomes at 6 Months.

Characteristics TIMM intervention (n = 13) Control (n = 13)

Note. M = mean, SD = standard deviation; Tx = treatment; OT = occupational therapist; ED = emergency department.

Figure 2. Study flow through.

8 OTJR: Occupational Therapy Journal of Research 00(0)

TIMM intervention group (mean rank = 10.0) compared with the control group (mean rank = 18.3) at T3 (U = 146.5, z = 2.70, p = .006), with a medium effect size (r = .51). For the TIMM intervention group, there was a significant improvement in performance of medication management tasks between T1 (Mdn = 36.5) and T3 (Mdn = 42.0), T = 66.0, p = .003, r = .55. There was not a significant difference in the performance scores for the control group between T1 (Mdn = 36.0) and T3 (Mdn = 36.0), T = 60.5, p = .382, r = .20. The TIMM inter- vention group had significant improvements in performance of medication management tasks (e.g., reading labels, opening pill bottles, and remembering to take medication on time; mean rank = 17.9) than the control group (mean rank = 9.8) at T3 (U = 36.0, z = −2.69, p = .007), with a large effect size (r = −.52). Potentially inappropriate medications in the TIMM intervention group (mean rank = 12.5) did not differ significantly from those in the control group (mean rank = 15.7) at T3 (U = 112.5, z = 1.26, p = .302, r = .24). Medication adherence rates in the TIMM intervention group (mean rank = 13.0) did not differ significantly from those in the control group (mean rank = 15.0) at T3 (U = 75.0, z = −0.792, p = .458, r = −.15).

Discussion

This study demonstrated that TIMM is feasible and accept- able for community-dwelling older adults. Participants were willing to participate in this randomized tailored interven- tion. Of the 33 eligible participants who enrolled in this study, most participants completed the study (78.8%). The results also indicated that TIMM can be delivered in a stan- dardized manner with high fidelity. The OT and the pharma- cist were able to achieve 84.7% and 89.0% adherence to the treatment protocol, respectively. The results also show that TIMM is acceptable. Study retention, a proxy for acceptabil- ity, was good, at 76.5% (Abshire et al., 2017; Sekhon et al., 2017). There was also high adherence to OT recommenda- tions (94%; Cumming et al., 2001; Stark et al., 2009), another indication that the intervention was acceptable to the older adults. Implementation and adherence to pharmacist recom- mendations for medication regimen changes were 58%, which is higher than some previous studies (20% in Clark et al., 2020 and 43% in Martin et al., 2018) but lower than other studies (91% in Kimura et al., 2022). The primary rea- son for lack of implementation of pharmacist recommenda- tions was lack of response from the physician, highlighting the need for additional education and collaboration between pharmacists and physicians.

TIMM can also improve performance of medication man- agement tasks over time and decrease barriers negatively impacting performance. Adherence to the medication routine improved significantly for the TIMM intervention group over time, but the difference between groups was not signifi- cant. It is possible that a longer follow-up period would have revealed sustained changes in adherence for the TIMM inter- vention group compared with the control group. Follow-up

periods for medication adherence studies are often 1 to 2 years (Martinez et al., 2020; Nieuwlaat et al., 2014); how- ever, our study was not designed to track participants for that length of time. Future studies should consider this.

We did not detect a difference in the number of potentially inappropriate medications at 6 months for those in the TIMM intervention group compared with the control group, likely because we had a low response rate from physicians regard- ing suggested medication regimen changes from the pharma- cist. This could be because both the study pharmacist and OT did not have any previous relationship with the physicians. Future studies could incorporate the use of the older adult’s current pharmacist to conduct the medication review and facilitate any recommendations for routine simplification or deprescribing with the physician. The OT should also feel empowered to take a more active role in the medication man- agement process, particularly within the interprofessional team. The AOTA specifically recommends that OTs “work with prescribers and other team members to resolve discrep- ancies and promote a shared understanding of the medication regimen” (AOTA, 2017).

In addition, future studies can incorporate evidence-based strategies to improve the relationship between pharmacist and physician, including face-to-face visits between provid- ers, role specification, and relationship intention (Amin & McKeirnan, 2022; Cromer et al., 2009), with hope that improved collaboration between physician and pharmacist would lead to deprescribing of potentially inappropriate medications.

While this intervention was initially designed for in-per- son delivery (Somerville et al., 2023), it was still feasible to deliver remotely if participants had or were provided with the appropriate technology (i.e., iPhone with camera) and in- person support to complete the visits remotely. All partici- pants were willing to wear the chest harness with a camera. They also benefited from having in-person support from a rater to initiate the call and provide any needed support dur- ing the call. Further analysis is needed to determine the type and amount of support provided from the rater (e.g., one- time minimal assistance to start the telehealth session vs. continuous support throughout the entire visit). Information from this analysis can help determine who could complete a remote visit successfully and can also help identify those for whom an in-person visit is recommended.

TIMM can be implemented within the U.S. health care system. In-home occupational therapy services addressing medication management, billed as either home health and/or mobile outpatient, are covered by Medicare and commercial insurance plans (Centers for Medicare & Medicaid Services, 2025b). Up until very recently (October 1, 2025), in-home telehealth services were also covered by Medicare (Centers for Medicare & Medicaid Services, 2025a). While some pri- vate insurers continue to cover telehealth services, it is likely that they will follow Medicare’s guidance and discontinue reimbursement. Should coverage for telehealth services

Somerville et al. 9

resume, TIMM can be delivered remotely. In addition, medi- cation therapy management, a service provided by pharma- cists that includes medication reviews, checking for potential drug interactions, duplicative therapies, and communication with a physician if any changes are needed, can be billed through Medicare Part D (Centers for Medicare & Medicaid Services, 2025c) and other commercial insurance plans. The adaptive equipment provided in this intervention is not cov- ered by insurance; however, the cost was minimal ($51 USD), and equipment can often be found at no cost through adaptive equipment reuse programs or disease-specific organizations.

Our results support previous research indicating that tai- lored, in-home interventions can improve older adults’ per- formance of daily activities (Keglovits & Stark, 2020; Stark et al., 2017), older adults are willing to participate in a remote intervention (Greenwald et al., 2018), and telehealth can be used to deliver occupational therapy services (Cason, 2014). Due to the timing of the study during the COVID-19 pandemic, technology use was required to deliver the inter- vention remotely. For older adults without access to tech- nology that can be used to conduct telehealth sessions, this is a limiting factor. However, TIMM was originally designed to be implemented in person (Somerville et al., 2023), and any older adult who has Medicare, even if they do not have access to technology, can receive TIMM through traditional home health or mobile outpatient occu- pational therapy services.

Limitations

One important limitation of the study was that some physi- cians did not reply to the pharmacist’s request to deprescribe inappropriate medications. Because of this, some of the pharmacist’s recommendations for deprescribing medica- tions were not implemented. Therefore, it is not surprising that there was not a significant reduction in potentially inap- propriate medications in the treatment group. An electronic medical record (EMR) system was not used during this study. The use of EMR features such as inbox messages could help keep the physician up to date or alert them of communication coming from the pharmacist. Future studies should evaluate whether the use of the EMR system, along with education, would improve communication between the pharmacist and prescribing physicians to reduce polyphar- macy for older adults.

Another limitation of the study was the ability of the MARS to measure medication adherence. Self-report meth- ods are generally the most time- and cost-effective way of obtaining adherence. However, self-report, especially as it relates to medication adherence, can be biased because older adults are more likely to report themselves as more adherent than they actually are. In addition, the MARS measure used in this study was originally designed to be used with adults with psychiatric illness. While it has been used in a more general population (Fialko et al., 2008), a more sensitive and

objective measure of adherence might be more appropriate for this population of older adults.

Conclusion

TIMM is a feasible and acceptable intervention for commu- nity-dwelling older adults. It demonstrates preliminary efficacy in decreasing barriers to medication management and improves the performance of tasks related to medication management. Developing an effective intervention to support older adults’ independence in medication management is of significant clin- ical importance as it has the potential to reduce medication- related hospital admissions, improve health outcomes, reduce falls and functional decline, and enable older adults to remain in their homes. Additional research is needed to (a) determine the effectiveness of TIMM to reduce potentially inappropriate medications and improve medication management for older adults and (b) examine the barriers and facilitators to imple- menting TIMM as part of occupational therapy practice.

Acknowledgments

The authors would like to thank all of the participants for making this study possible.

ORCID iDs

Emily Somerville https://orcid.org/0000-0003-2843-9640 Rebecca M. Bollinger https://orcid.org/0000-0002-1931-8372

Susan Stark https://orcid.org/0000-0002-2816-7158

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: American Occupational Therapy Foundation: AOTFIRG20SOMERVILLE.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Trial Registration

ClinicalTrials.gov identifier NCT04717297 (https://clinicaltrials. gov/study/NCT04717297?cond=NCT04717297&rank=1)

Supplemental Material

Supplemental material for this article is available online.

References

Abshire, M., Dinglas, V. D., Cajita, M. I., Eakin, M. N., Needham, D. M., & Himmelfarb, C. D. (2017). Participant retention prac- tices in longitudinal clinical research studies with high retention rates. BMC Medical Research Methodology, 17(1), Article 30. https://doi.org/10.1186/s12874-017-0310-z

American Occupational Therapy Association. (2017). Occupational therapy’s role in medication management. American Journal of Occupational Therapy, 71, 1–20.

10 OTJR: Occupational Therapy Journal of Research 00(0)

Amin, S., & McKeirnan, K. C. (2022). The physician perspective on pharmacist-physician collaboration and trust. Journal of the American Pharmacists Association, 62(4), 1304–1312.

Bakar, Z. A., Fahrni, M. L., & Khan, T. M. (2016). Patient satisfaction and medication adherence assessment amongst patients at the dia- betes medication therapy adherence clinic. Diabetes & Metabolic Syndrome: Clinical Research & Reviews, 10(2), S139–S143.

Boop, C., Cahill, S. M., Davis, C., Dorsey, J., Gibbs, V., Herr, B., Kearney, K., Lannigan, E. Q. L. Q. G., Metzger, L., Miller, J., Owens, A., Rives, K., Synovec, C., Winistorfer, W. L., & Li- eberman, D. (2020). Occupational therapy practice framework: Domain and process fourth edition. AJOT: American Journal of Occupational Therapy, 74(S2), Article 74S2001.

Brown, M. T., & Bussell, J. K. (2011). Medication adherence: WHO Cares? Mayo Clinic Proceedings, 86(4), 304–314. https:// doi.org/10.4065/mcp.2010.0575

By the 2019 American Geriatrics Society Beers Criteria Update Ex- pert Panel. (2019). American Geriatrics Society 2019 updated AGS beers criteria for potentially inappropriate medication use in older adults. Journal of the American Geriatric Society, 67, 674–694. https://doi.org/10.1111/jgs.15767

Cason, J. (2014). Telehealth: A rapidly developing service delivery model for occupational therapy. International Journal of Telere- habilitation, 6(1), Article 29.

Centers for Medicare & Medicaid Services. (2025a). 10/1 Updated: Telehealth FAQ. Calenda Year, 2025. https://www.cms.gov/ files/document/telehealth-faq-updated-10-15-2025.pdf

Centers for Medicare & Medicaid Services. (2025b). Billing and Coding: Outpatient Physical and Occupational Therapy Ser- vices. U.S. Centers for Medicare & Medicaid Services. https:// www.cms.gov/medicare-coverage-database/view/article.aspx?a rticleid=57067&ver=29&=

Centers for Medicare & Medicaid Services. (2025c). Medica- tion Therapy Management. U.S. Centers for Medicare & Medicaid Services. https://www.cms.gov/medicare/coverage/ prescription-drug-coverage-contracting/medication-therapy- management

Charlesworth, C. J., Smit, E., Lee, D. S., Alramadhan, F., & Odden, M. C. (2015). Polypharmacy among adults aged 65 years and older in the United States: 1988–2010. Journals of Gerontol- ogy Series A: Biomedical Sciences and Medical Sciences, 70(8), 989–995.

Chisholm-Burns, M. A., & Spivey, C. A. (2012). The “cost” of med- ication nonadherence: Consequences we cannot afford to ac- cept. Journal of the American Pharmacists Association, 52(6), 823–826. https://doi.org/10.1331/JAPhA.2012.11088

Clark, C. M., LaValley, S. A., Singh, R., Mustafa, E., Monte, S. V., & Wahler, R. G., Jr. (2020). A pharmacist-led pilot program to facilitate deprescribing in a primary care clinic. Journal of the American Pharmacists Association, 60(1), 105–111.

Clifford-Middel, M. (2004). Review: Simplifying dosing regimens appears to improve treatment adherence in patients with high blood pressure in ambulatory settings. Evidence Based Nursing, 7(4), Article 110.

Cromer, J., Hojjat, R., Peker, S., & Aprile, J. (2009). Fostering the pharmacist-physician relationship. American Journal of Health- Stystem Pharmacy, 66(2), 118–119.

Cooper, J. A., Cadogan, C. A., Patterson, S. M., Kerse, N., Brad- ley, M. C., Ryan, C., & Hughes, C. M. (2015). Interventions to improve the appropriate use of polypharmacy in older people:

A Cochrane systematic review. British Medical Journal Open, 5(12), Article e009235.

Cumming, R. G., Thomas, M., Szonyi, G., Frampton, G., Salkeld, G., & Clemson, L. (2001). Adherence to occupational therapist recommendations for home modifications for falls prevention. The American Journal of Occupational Therapy, 55(6), 641– 648.

Farris, K. B., & Phillips, B. B. (2008). Instruments assessing capac- ity to manage medications. Annals of Pharmacotherapy, 42(7), 1026–1036. https://doi.org/10.1345/aph.1G502

Fialko, L., Garety, P. A., Kuipers, E., Dunn, G., Bebbington, P. E., Fowler, D., & Freeman, D. (2008). A large-scale valida- tion study of the Medication Adherence Rating Scale (MARS). Schizophrenia Research, 100(1–3), 53–59.

Fick, D. M., Cooper, J. W., Wade, W. E., Waller, J. L., Maclean, J. R., & Beers, M. H. (2003). Updating the Beers criteria for poten- tially inappropriate medication use in older adults: Results of a U.S. consensus panel of experts. Archives of Internal Medicine, 163(22), 2716–2724.

Fritz, H. (2014). The influence of daily routines on engaging in dia- betes self-management. Scandinavian Journal of Occupational Therapy, 21(3), 232–240. https://doi.org/10.3109/11038128.20 13.868033

Garfinkel, D., & Mangin, D. (2010). Feasibility study of a sys- tematic approach for discontinuation of multiple medications in older adults: Addressing polypharmacy. Archives of Internal Medicine, 170(18), 1648–1654. https://doi.org/10.1001/archin- ternmed.2010.355

Gellad, W. F., Grenard, J. L., & Marcum, Z. A. (2011). A system- atic review of barriers to medication adherence in the elderly: Looking beyond cost and regimen complexity. The American Journal of Geriatric Pharmacotherapy, 9(1), 11–23. https://doi. org/10.1016/j.amjopharm.2011.02.004

George, J., Elliott, R. A., & Stewart, D. C. (2008). A systematic review of interventions to improve medication taking in elderly patients prescribed multiple medications. Drugs & Aging, 25(4), 307–324. https://doi.org/2544

Gould, E., & Mitty, E. (2010). Medication adherence is a partner- ship, medication compliance is not. Geriatic Nursing, 31(4), 290–298. https://doi.org/10.1016/j.gerinurse.2010.05.004

Greenwald, P., Stern, M. E., Clark, S., & Sharma, R. (2018). Older adults and technology: In telehealth, they may not be who you think they are. International Journal of Emergency Medicine, 11(1), 1–4.

Haltiwanger, E. P. (2012). Effect of a group adherence intervention for Mexican-American older adults with type 2 diabetes. Ameri- can Journal of Occupational Therapy, 66(4), 447–454. https:// doi.org/10.5014/ajot.2012.004457

Harris, P. A., Taylor, R., Thielke, R., Payne, J., Gonzalez, N., & Conde, J. G. (2009). Research Electronic Data Capture (REDCap)-A metadata-driven methodology and workflow pro- cess for providing translational research informatics support. Journal of Biomedical Informatics, 42(2), 377–381. https://doi. org/10.1016/j.jbi.2008.08.010

Haynes, R. B., Ackloo, E., Sahota, N., McDonald, H. P., & Yao, X. (2008). Interventions for enhancing medication adherence. Cochrane Database of Systematic Reviews, 2008, Article CD000011. https://doi.org/10.1002/14651858.CD000011. pub3

Holt, E. W., Rung, A. L., Leon, K. A., Firestein, C., & Krousel- Wood, M. (2014). Medication adherence in older adults: A

Somerville et al. 11

qualitative study. Educational Gerontology, 40(3), 198–211. Iuga, A. O., & McGuire, M. J. (2014). Adherence and health care

costs. Risk Management and Healthcare Policy, 7, Article 35. Katzman, R., Brown, T., Fuld, P., Peck, A., Schechter, R., & Schim-

mel, H. (1983). Validation of a short orientation-memory-con- centration test of cognitive impairment. American Journal of Psychiatry, 140(6), 734–739.

Keglovits, M., & Stark, S. (2020). Home modifications to improve function and safety in the United States. Journal of Aging and Environment, 34(2), 110–125.

Kimura, T., Fujita, M., Shimizu, M., Sumiyoshi, K., Bansho, S., Yamamoto, K., . . .Yano, I. (2022). Effectiveness of pharmacist intervention for deprescribing potentially inappropriate medica- tions: A prospective observational study. Journal of Pharmaceu- tical Health Care and Sciences, 8(1), Article 12.

Kini, V., & Ho, P. M. (2018). Interventions to improve medication adherence: A review. Journal of the American Medical Associa- tion, 320(23), 2461–2473.

Lawton, M., & Nahemow, L. (1973). Ecology and the aging pro- cess. In C. Eisdorfer & M. Lawton (Eds.), The psychology of adult development and aging (pp. 619–674). American Psycho- logical Association. https://doi.org/10.1037/10044-020

Marcum, Z. A., Jiang, S., Bacci, J. L., & Ruppar, T. M. (2021). Pharmacist-led interventions to improve medication adherence in older adults: A meta-analysis. Journal of the American Geri- atrics Society, 69(11), 3301–3311.

Martin, P., Tamblyn, R., Benedetti, A., Ahmed, S., & Tannenbaum, C. (2018). Effect of a pharmacist-led educational intervention on in- appropriate medication prescriptions in older adults: The D-PRE- SCRIBE randomized clinical trial. JAMA, 320(18), 1889–1898.

Martinez, A. I., Abner, E. L., Jicha, G. A., Rigsby, D. N., Eck- mann, L. C., Huffmyer, M. J., & Moga, D. C. (2020). One-year evaluation of a targeted medication therapy management in- tervention for older adults. Journal of Managed Care & Spe- cialty Pharmacy, 26(4), 520–528. https://doi.org/10.18553/ jmcp.2020.26.4.520

Murphy, M. C., Somerville, E., Keglovits, M., Hu, Y. L., & Stark, S. (2017). The in-home medication management performance evaluation (HOME-Rx): A validity study. American Journal of Occupational Therapy, 71(4), Article 7104190020. https://doi. org/10.5014/ajot.2017.022756

Nazareth, I., Burton, A., Shulman, S., Smith, P., Haines, A., & Tim- berall, H. (2001). A pharmacy discharge plan for hospitalized elderly patients: A randomized controlled trial. Age and Ageing, 30(1), 33–40.

Nieuwlaat, R., Wilczynski, N., Navarro, T., Hobson, N., Jef- fery, R., Keepanasseril, A., Agoritsas, T., Mistry, N., Iorio, A., Jack, S., Sivaramalingam, B., Iserman, E., Mustafa, R. A., Je- draszewski, D., Cotoi, C., & Haynes, R. B. (2014). Interven- tions for enhancing medication adherence. Cochrane Database of Systematic Reviews, 2014, Article CD000011. https://doi. org/10.1002/14651858.CD000011.pub4

Orsmond, G. I., & Cohn, E. S. (2015). The distinctive features of a feasibility study: Objectives and guiding questions. OTJR: Oc- cupation, Participation and Health, 35(3), 169–177.

Osterberg, L., & Blaschke, T. (2005). Adherence to medication. New England Journal of Medicine, 353(5), 487–497. https://doi. org/10.1056/NEJMra050100

Sabaté, E. (2003). Adherence to long-term therapies: Evidence for action. World Health Organization.

Sanders, M. J., & Van Oss, T. (2013). Using daily routines to pro- mote medication adherence in older adults. American Journal of Occupational Therapy, 67(1), 91–99.

Schwartz, J. K., Grogan, K. A., Mutch, M. J., Nowicki, E. B., Se- idel, E. A., Woelfel, S. A., & Smith, R. O. (2017). Intervention to improve medication management: Qualitative outcomes from a phase I randomized controlled trial. American Journal of Oc- cupational Therapy, 71(6), Article 7106240010.

Schwartz, J. K., & Smith, R. O. (2017). Integration of medication management into occupational therapy practice. American Jour- nal of Occupational Therapy, 71(4), Article 7104360010.

Sekhon, M., Cartwright, M., & Francis, J. J. (2017). Acceptability of healthcare interventions: An overview of reviews and develop- ment of a theoretical framework. BMC Health Services Research, 17, Article 88. https://doi.org/10.1186/s12913-017-2031-8

Somerville, E., Bollinger, R., Keleman, A., Haxton, M., Sarrami, B., Chen, S. W., Holden, B., Yan, Y., & Stark, S. (2023). Tai- lored medication management intervention delivered by oc- cupational therapists for older adults: A study protocol. British Journal of Occupational Therapy, 86(4), 257–264. https://doi. org/10.1177/03080226221135366

Sorensen, L., Stokes, J., Purdie, D., Woodward, M., & Roberts, M. (2006). Medication management at home: Medication risk fac- tor prevalence and inter-relationships. Journal of Clinical Phar- macy and Therapeutics, 31(5), 485–491.

Stark, S., Keglovits, M., Arbesman, M., & Lieberman, D. (2017). Effect of home modification interventions on the participation of community-dwelling adults with health conditions: A system- atic review. American Journal of Occupational Therapy, 71(2), Article 7102290010.

Stark, S., Landsbaum, A., Palmer, J. L., Somerville, E. K., Mor- ris, J. C., Harvey, A., & Friedman, D. H. (2009). Client-centred home modifications improve daily activity performance of older adults. Canadian Journal of Occupational Therapy, 76(1 Sup- pl.), 235–245.

Ulley, J., Harrop, D., Ali, A., Alton, S., & Fowler Davis, S. (2019). Deprescribing interventions and their impact on medication ad- herence in community-dwelling older adults with polypharma- cy: A systematic review. BMC Geriatrics, 19(1), 1–13.

Webb, D. G., Horne, R., & Pinching, A. J. (2001). Treatment-relat- ed empowerment: Preliminary evaluation of a new measure in patients with advanced HIV disease. International Journal of STD & AIDs, 12(2), 103–107.

Whyte, J., & Hart, T. (2003). It’s more than a black box; it’s a Rus- sian doll: Defining rehabilitation treatments. American Journal of Physical Medicine & Rehabilitation, 82(8), 639–652. https:// doi.org/10.1097/01.PHM.0000078200.61840.2D

Wolf, M. S., Curtis, L., Waite, K., Cooper Bailey, S., Hedlund, L. A., Davis, T. C., Shrank, W. H., Parker, R. M., & Wood, A. J. (2011). Helping patients simplify and safely use complex prescription regimens. Archives of Internal Medicine, 171(4), 300–305.

Yau, M. (1995). Occupational therapy in community mental health: Do we have a unique role in the interdisciplinary en- vironment? Australian Occupational Therapy Journal, 42(3), 129–113.

image6.emf

Archives of Gerontology and Geriatrics 107 (2023) 104910

Deprescribing for older people living in residential aged care facilities: Pharmacist recommendations, doctor acceptance and implementation

Hui Wen Quek a,b,*, Christopher Etherton-Beer c, Amy Page b,c, Andrew J McLachlan a, Sarita Y Lo d,e,f, Vasi Naganathan d,e, Leanne Kearney d, Sarah N Hilmer f, Tracy Comans g, Derelie Mangin h, Richard I Lindley i,j, Kathleen Potter c,k

a Sydney Pharmacy School, Faculty of Medicine and Health, University of Sydney, Sydney, New South Wales, Australia b School of Allied Health, University of Western Australia, Perth, Western Australia, Australia c Western Australia Centre for Health and Aging, School of Medicine, University of Western Australia, Perth, Western Australia, Australia d Centre for Education and Research on Ageing, Department of Geriatric Medicine, Concord Repatriation General Hospital, Concord, New South Wales, Australia e Concord Clinical School, Faculty of Medicine and Health, University of Sydney, Sydney, New South Wales, Australia f Kolling Institute, Faculty of Medicine and Health, Northern Clinical School, The University of Sydney and Royal North Shore Hospital, St Leonards, New South Wales, Australia g Menzies Health Institute Queensland, Griffith University, University Drive Meadowbrook, Brisbane, Queensland Australia h Primary Care Research Unit, Department of Public Health and General Practice, Christchurch School of Medicine & Health Sciences, University of Otago, Christchurch, New Zealand i Sydney Medical School, University of Sydney, Sydney, New South Wales j Australia and The George Institute for Global Health, Sydney, New South Wales, Australia k Ryman Healthcare, Christchurch, New Zealand

A R T I C L E I N F O

Keywords: Deprescribing PIMs Polypharmacy Ageing Elderly Nursing home Nursing home residents Acceptability Interprofessional relations Medication therapy management Medication review

A B S T R A C T

Background: Deprescribing is an intervention to address the high prevalence of inappropriate polypharmacy in older people living in residential aged care facilities (RACFs). Many deprescribing interventions are complex and involve several stages including initial pharmacist recommendation, subsequent acceptance of the recommen- dations by a prescriber and the patient, and then actual implementation. Objectives: This study aimed to investigate pharmacist deprescribing recommendations for residents within RACFs, general practitioner (GP) acceptance, and the actual implementation of the accepted recommendations at 12-month. Methods: The intervention occurred as part of a randomised controlled trial and comprised a pharmacist-led medication review using an evidence-based algorithm, with the focus on identifying medications to poten- tially deprescribe. Consent to participate was obtained from residents (or surrogate decision-makers), RACF nursing staff and the resident’s GP. Deprescribing recommendations were reviewed by GPs before imple- mentation as part of the intervention and control arms of the trial, although control group participants continued to receive their usual medications in a blinded manner. Results: There were 303 participants enrolled in the study, and 77% (941/1222) of deprescribing recommen- dations suggested by the pharmacists were accepted by GPs. Of the recommendations accepted by GPs, 74% (692/ 941) were successfully implemented at the end of the follow-up visit at 12 months. The most common reason for deprescribing was because medications were no longer needed (42%, 513/ 1231). Conclusion: Pharmacist-led deprescribing recommendations arising from an algorithm-based medication review are acceptable to doctors and can have a significant impact on reducing the number of inappropriate medications consumed by older people in RACFs. Trial registration: Australian New Zealand Clinical Trials Registry ACTRN12613001204730

Abbreviations: ADWE, Adverse Drug Withdawal Event; ATC, Anatomical Therapeutic Classification; CI, Confidence Interval; DBI, Drug Burden Index; GP, General Practitioner; MBI, Modified Barthel Index; MMSE, Mini Mental State Examination; MWP, Medication Withdrawal Plan; NOK, Next of kin; NPI-NH, Neuro-Psychiatric Index Nursing Home Edition; NSW, New South Wales, Australia; PIM, Potentially Inappropriate Medicines; PRN, Pro Ne Rata; RACF, Residential Aged Care Facility; WA, Western Australia, Australia.

* Corresponding author. E-mail address: [email protected] (H.W. Quek).

Contents lists available at ScienceDirect

Archives of Gerontology and Geriatrics

journal homepage: www.elsevier.com/locate/archger

https://doi.org/10.1016/j.archger.2022.104910 Received 21 October 2022; Received in revised form 7 December 2022; Accepted 17 December 2022

mailto:[email protected]

www.sciencedirect.com/science/journal/01674943

https://www.elsevier.com/locate/archger

https://doi.org/10.1016/j.archger.2022.104910

http://crossmark.crossref.org/dialog/?doi=10.1016/j.archger.2022.104910&domain=pdf

Archives of Gerontology and Geriatrics 107 (2023) 104910

1. Introduction

Polypharmacy, defined as the concomitant use of five or more reg- ular medications, is common in older people, especially those who live in residential aged care facilities (RACFs) (Masnoon et al., 2017). A systematic review found that up to 74% of the residents in RACFs were taking more than ten medications (Jokanovic et al., 2015). The high prevalence is due to multimorbidity with advancing age, and the involvement of multiple prescribers for their medications before admission to the facility (Jokanovic et al., 2015; Jokanovic et al., 2016). Polypharmacy increases the risk of adverse health outcomes such as geriatric syndromes and reduced functional capacity (Maher et al., 2014; Gnjidic et al., 2012; Turner et al., 2016). In addition to poly- pharmacy per se, the use of potentially inappropriate medications (PIMs) for which the possible harms outweigh the benefits is of concern (Turner et al., 2016). Explicit criterion-based tools are widely used to evaluate PIMs in older people (Beers et al., 1991; Lee et al., 2022; Lee, Etherton-Beer, Hosking, Pasco, & Page, 2022). It has been estimated in an international systematic review that 43% of residents in RACFs were exposed to at least one PIM (Morin et al., 2016). In a more recent Australian study, the prevalence of PIM in RACF residents was higher (83%), and they were found to have a greater risk of being prescribed more than one PIM than community dwellers (Bony et al., 2020). Medications with anticholinergic or sedative effects, in particular, have been associated with a decline in both physical and cognitive function (Harrison et al., 2018; 11. Kouladjian et al., 2014). For this reason, risk assessment tools such as the Drug Burden Index (DBI) was developed to quantify individual exposure to anticholinergic and sedative medica- tions (Hilmer et al., 2007).

There is a growing body of research on medication management in older people living in RACFs, and suboptimal prescribing is often re- ported (Royal Commission into Aged Care Quality and Safety, 2019). Deprescribing, defined as the supervised withdrawal of an inappropriate medication with the aim of improving patient well-being, is an effective way to address polypharmacy (Reeve et al., 2015; Scott et al., 2015). In randomised controlled studies analysed in a systematic review, depres- cribing had no clear impact on mortality (odds ratio 0.82, 95% confi- dence interval 0.61-1.11) (Page et al., 2016). In another systematic review, deprescribing interventions among older people living with frailty were found to be safe, feasible, well tolerated and can lead to improved outcomes on mental health and frailty (Ibrahim et al., 2021).

Although the outcomes of deprescribing are assumed to be positive, barriers exist when implementing the intervention into routine clinical practice, including time constraints on general practitioners (GPs) who prescribe medications in RACFs and the complex clinical profile of res- idents in RACFs, especially when multiple prescribers are involved (Garfinkel, 2017). Deprescribing also requires the skills and knowledge from nursing staff to identify potential medications to deprescribe and monitor for adverse drug withdrawal effects (Palagyi et al., 2016). Deprescribing interventions benefit from a multidisciplinary approach by combining the expertise of all team members. Pharmacist input, in addition to input from the resident’s usual practitioner, is particularly important. As experts in pharmacotherapy, pharmacists can identify medication-related problems through formal medication reviews. Several studies have shown that pharmacist-led medication reviews achieve a significant reduction in the number of PIM, DBI, and falls in older people (Chen, 2016; Nishtala et al., 2009; Ailabouni et al., 2019; Gheewala et al., 2014).

While studies support involving pharmacists in medication man- agement, the full potential for pharmacists to guide and support for deprescribing is rarely utilised in practice. It is unknown how receptive prescribers are to deprescribing recommendations made by pharmacists. A qualitative study on GPs’ views towards the HMRs conducted by pharmacists revealed varied attitudes, ranging from “useful” or “ambivalent” to “sceptical” (Weir et al., 2019). Pharmacists, interviewed by the same research team, reported a lack of follow-up from the

referring GP after the HMR (Weir et al., 2020). A close working rela- tionship between the two professions facilitates implementation of pharmacist recommendations, but evidence of collaborative depres- cribing in the RACF setting is scarce (Kwint, 2013).

This study aimed to report the deprescribing recommendations of pharmacists for residents within RACFs, evaluate the acceptance by doctors of these recommendations and review implementation of accepted recommendations.

2. Methods

2.1. Study design

This study is part of a larger deprescribing trial (Opti-Med) pro- spectively registered with the Australian New Zealand Clinical Trials Registry (registration number: ACTRN12613001204730). Opti-Med is a randomised, double-blind, controlled trial with an additional open intervention arm. The study was approved by the Human Research Ethics Committees from The University of Western Australia in Western Australia (RA/4/1/5930) and Concord Repatriation General Hospital in New South Wales (HREC13/CRGH/77). This process evaluation in- vestigates pharmacists’ medication reviews and recommendations and GP acceptance and implementation of these recommendations through analyses of medicines data from participants enrolled in the Opti-Med study.

2.2. Setting

Opti-Med was a multicentre trial conducted in 17 residential aged care facilities (RACFs) in two Australian states (New South Wales and Western Australia).

2.3. Recruitment, participants, inclusion and exclusion criteria

RACF managers or owners were provided with the Opti Med study information and asked for written permission to recruit participants from their facilities. All residents living in participating RACFs were screened for eligibility to participate in the trial. We recruited residents aged over 65 years who were taking at least one regular medication. Residents were excluded if they were moribund or had a short life ex- pectancy. People who met the inclusion criteria were invited to partic- ipate and were provided with a participant information sheet and consent form. If the residents had impaired capacity to make decisions, then consent was sought from their next-of-kin or surrogate decision- makers. In addition, the resident’s usual GP and facility manager were required to agree to the resident’s participation.

2.4. Description of intervention

Step 1: Study Cohort Characteristics A baseline assessment and initial consultation was completed by

trained research staff. We used the Modified Barthel Index to measure activities of daily

living (lower scores indicating increased dependency), Charlson Co- morbidity Index to assess severity of medical comorbidities (score of >2 considered “high”), EuroQol5-Dimensions (EQ-5D) to assess health- related quality of life (maximal score = 100), Neuro-Psychiatric In- ventory Nursing Home Version (NPI-NH) to assess the behavioural and psychological symptoms of dementia, NPI-NH occupational disruptive- ness scale to assess the impact of behavioural disturbances on the pro- fessional caregivers, Rockwood’s 40-item frailty index to measure the health status (score of ≥0.2 indicating frailty), and Mini-Mental State Examination (MMSE) to assess cognitive functions (score of <25 indi- cating impaired cognition (Shah et al., 1989; Charlson et al., 1987; Herdman et al., 2011; Pickard et al., 2007; Westaway et al., 2020; Cummings et al., 1994; Rockwood et al., 2004; Folstein et al., 1975).

H.W. Quek et al.

Archives of Gerontology and Geriatrics 107 (2023) 104910

Participants with intact cognition (MMSE >23) were additionally administered the Beliefs About Medicines questionnaire, with higher scores indicating stronger beliefs in the concepts of the scale (Horne et al., 1999).

Step 2: Medication Review and Medication Withdrawal Plan A comprehensive clinical history for each participant was con-

structed from the RACF medical and pharmacy records and interviews with the resident or their NOK, their GPs, and specialists. The research pharmacists compiled a comprehensive list of all medications consumed by each participant from their RACF medication chart and most recent administration record. The list included all prescription medications, supplements, over-the-counter medications, and herbal medicine prod- ucts taken regularly and as needed (pro re nata; “PRN”). The active ingredient name of the medication, dosage, frequency, and adminis- tration times were recorded.

An individualised patient-centred medication review was conducted independently by the study pharmacists AP and MS for all pS for all participants and a third pharmacist (SL) for NSW participants)artici- pants to identify a list of suitable deprescribing medication targets and to determine the order in which targeted medicines would be depres- cribed (medication withdrawal plan; “MWP”). The medication review was guided by an evidence-based deprescribing algorithm (Fig. 1), previously adapted from the Good Palliative-Good Practice tool (Gar- finkel et al., 2007). The algorithm was validated for inter-rater reli- ability between two pharmacists and two physicians prior to implementation, and remains, to our knowledge, the only explicit deprescribing tool that has been validated (Page et al., 2016). This tool was used in the first Australian randomised controlled trial of depres- cribing (Potter et al., 2016).

All regular solid-dose oral medications were considered as potential deprescribing targets when formulating the MWP. PRN or self- administered medications used more frequently than once a week on average, over the previous four weeks were classified as “regular med- ications”. Non-solid oral medications, such as topical medications, transdermal patches, injections and eye/ear drops, were excluded from consideration for deprescribing due to the inability to blind the with- drawal process. Each pharmacist independently recorded their decision to continue or cease each regular medication. The same two pharmacists developed all MWPs, with a third pharmacist additionally participating providing peer review of the MWP for participants based in the New South Wales facilities.

The research pharmacists resolved any differences in the target medications or withdrawal order by consensus to generate a single consensus MWP which was then sent to the GPs. The plan included all medications targeted for deprescribing, the rationale for targeting, dose tapering schedule (if needed), order of withdrawal, and monitoring plan. The sequence of withdrawal was determined by the characteristics of the medication and the priority for the participant. Medications that were least likely to be beneficial to the participant and least likely to cause an adverse drug withdrawal event (ADWE) were ceased first. Medications were tapered slowly and withdrawn last if they were used for symptomatic relief or more likely to cause an ADWE. For many medications, the general approach for tapering is to halve the dose at 2-4 weeks intervals (Steinman & Reeve, 2021). Hence in this study, tapering was generally carried out by halving the doses at a fortnightly interval until half of the lowest dose was reached, following which the medica- tion was ceased. Some medications with longer half-lives were gradually tapered to every second- or third-day dosing until ceased. All partici- pants were closely monitored for potential ADWEs after each medication adjustment.

Step 3: General Practitioners’ Agreement to the Medication Withdrawal Plan

GPs were asked to either approve or decline each recommendation based on their clinical knowledge of the patient. The MWPs were then finalised for implementation in the study (S1 Figure for a sample).

Step 4: Randomisation

Randomisation was undertaken after the MWPs were finalised. Par- ticipants were randomised using block sizes of 6, 9 or 12 in a 1:1:1 and stratified by gender, number of medications, and frailty index to one of three groups: blinded control, blinded intervention (deprescribing) or open intervention (deprescribing). The randomisation process was per- formed by a biostatistician using computer-generated randomisation tables for allocation.

Step 5: Implementation of the Medication Withdrawal Plan The participants who were randomised to the blinded intervention

group had their medications deprescribed according to the final approved MWP. Blinding was maintained by over-encapsulating the study medication or by replacing the study medication with a capsule filled with inert filler. Participants who were randomised to the blinded control group continued to receive their targeted medications with over- encapsulation to maintain blinding. All personnel involved in the trial (including participants, their relatives, research assistants, RACF staff, GPs, and investigators) were blinded to the treatment allocation, except for the biostatistician who undertook the randomisation (KM) and pharmacy staff (one research pharmacist, MS and SL, in each state) who prepared the study medications. No blinding occurred for participants in the open intervention arm, medications were simply withdrawn ac- cording to the MWP.

Step 6: Safety Monitoring The participants in all three groups were monitored for any potential

ADWEs or clinical incidents (such as an intercurrent illness) by the research staff one week after each adjustment in their medicine regimen. The MWPs specified any specific additional monitoring that was required and the time frame. An example of monitoring may have been scheduled monitoring was symptoms, heart rate and blood pressure at two weekly intervals for a defined time frame. The study logistics were monitored by unblinded, accredited consultant pharmacists who over- saw all aspects of the study conduct and provided safety oversight. Unblinding was also a possible option in the event of an emergency.

Step 7: Follow-up Assessments Follow-up assessments, including documentation of current medi-

cations, were completed at the intervals of 3, 6, and 12 months. All participants were followed for up to 12 months or until death, whichever occurred sooner.

2.5. Data analysis

The primary outcome of this process evaluation was a description of pharmacists’ recommendations, the acceptance rate by GPs, and the rate of recommended medication withdrawals achieved at the 12-month follow-up. The number of regular and PRN medications were assessed at baseline and 12-month follow-up. Additionally, we also assessed the number of medications based on the assumptions that all pharmacist recommendations had been accepted and successfully implemented, and by assuming all GP accepted recommendations had been successfully implemented.

All medications chosen for deprescribing were classified according to the World Health Organisation’s Anatomical Therapeutic Chemical (ATC) classification system (World Health Organisation, 2011). Com- bination products were recorded as each distinct ingredient in the medication review unless not available in that dose or form individually. For example, decarboxylase inhibitor is commonly combined with levodopa but is not available as a single active ingredient. Medications were grouped by the anatomical main group (ATC first level), thera- peutic subgroup (second level), pharmacological subgroup (third level), chemical subgroup (fourth level), and chemical substance (fifth level). GP acceptance rates in specific medication classes were assessed on the level of therapeutic subgroup (ATC second level). The reasons for rec- ommendations made by pharmacists were coded into at least one of the four relevant categories as described in the deprescribing algorithm (Fig. 1).

H.W. Quek et al.

Archives of Gerontology and Geriatrics 107 (2023) 104910

Fig. 1. Deprescribing algorithm.

H.W. Quek et al.

Archives of Gerontology and Geriatrics 107 (2023) 104910

2.5.1. High risk prescribing The Drug Burden Index (DBI) was calculated using validated meth-

odology, excluding any PRN, topical dermatological, inhaled, and ophthalmological medications (Hilmer et al., 2007). The cumulative DBI for each participant was calculated by summing the DBI for each medication with anticholinergic or sedative properties. We additionally considered high risk prescribing as the use of medicines included in the Beers Criteria, and the Anticholinergic Drug Scale in addition to po- tential prescribing cascades identified in the literature (Fick et al., 2019; Somers et al., 2010; Carnahan et al., 2006).

2.5.2. Statistical methods Data were analysed using Stata Statistical Software: Release 17

(College Station, Texas: StataCorp LLC) on an intention-to-treat basis. Descriptive statistics, including mean, standard deviation, and 95% confidence interval (95% CI) were calculated to describe the number of medications, and DBI as appropriate using two-sided t tests. Categorical data for the identification of anticholinergic and sedative medicines contributing to DBI, anticholinergic medicines, Beers medicines, and potential prescribing cascades were compared using chi-square tests. P- values of less than 0.05 were considered significant.

3. Results

3.1. Study cohort characteristics

A total of 303 residents were eligible, consented to participate, and subsequently randomised: 102 in the blinded intervention group, 101 in the open intervention, and 100 in the blinded control group. A summary of the medications taken by the participants at baseline is presented in Table 1. A detailed description of the participants characteristics and their medicine use at baseline are presented elsewhere in this journal (Page et al., 2023) The average age of all participants was 85.0 (± 7.5). Most participants had cognitive impairment with a mean MMSE score of 15.2 (± 10.3). By 12-month follow-up, 85 participants discontinued intervention (due to death, trial withdrawal, or not living at RACF), 1 was lost to follow up, and 3 did not receive allocated intervention after randomisation. A detailed report of the study outcomes and the baseline characteristics of the participants as well as their medication regimens are published elsewhere.

3.2. Commonly targeted medication classes

Following medication review, the pharmacists identified 1222 reg- ular medications that were in solid oral dosage forms as potential targets for deprescribing (Fig. 2, S2 Table). The top five most commonly tar- geted medication classes were vitamins (167, 13.7%), psycholeptics including antipsychotics, anxiolytics, hypnotics and sedatives (97, 7.9%), lipid-modifying agents (90, 7.4%), drugs for acid-related disor- ders (88, 7.2%), and psychoanaleptics including antidepressants and anti-dementia drugs (82, 6.7%). Among these top five medication clas- ses, the most frequently targeted medications were cholecalciferol (vitamin D), temazepam, atorvastatin, pantoprazole, and mirtazapine.

3.3. Reasons for deprescribing

As depicted in Table 2, the reasons for deprescribing (predefined in the study algorithm) were provided to the GPs for each targeted medi- cation. The most common reason for a pharmacist’s recommendation to deprescribe was because the medications were “no longer needed” (42%, 513/ 1231). “Inappropriate prescriptions” were the second most common (28%, 348/ 1231). For example, lack of documentation for indication, obvious contraindication, or prescribing cascade. This is followed by “possible adverse effects or interactions outweigh potential benefits” which made up 17% (207/1231) of the recommendations, and “potential benefit uncertain or unlikely to be realised” which made up

13% (163/1231) of the recommendations.

3.4. GPs’ acceptance rate

Overall, three out of every four recommendations made by the pharmacists were accepted by the GP (77%, 941/1222, Fig. 2, S2 Table). Among the top five commonly targeted medication classes, the accep- tance rates for vitamins, psycholeptics, lipid-modifying agents, drugs for acid-related disorder, and psychoanaleptics were 80% (133/167), 76% (74/97), 79% (71/90), 88% (77/88), and 63% (52/82) respectively. Out of each of the four rationales for targeting, “inappropriate prescriptions” had the highest acceptance rate of 82% (287/ 348).

3.5. Number of medications

If all the pharmacists’ recommendations had been accepted and successfully implemented, the average number of regular medications would have been reduced from 10.1 (± 4.4) per participant at baseline to 6.0 (± 3.7) [Table 3]. If all the recommendations accepted by GPs were successfully implemented, the number of regular medications would have been reduced to 7.0 (± 3.9).

At 12-month follow up, approximately every three out of four rec- ommendations accepted by GPs (74%, 938/1220) were successfully implemented. Overall study outcomes are reported in detail elsewhere. In brief, the average number of regular medications was significantly reduced by a mean of 2.4 (95% CI -3.2, -1.6) for the blinded intervention group and 1.9 (95% CI -2.7, 1.0) for the open intervention group at 12 months when compared to baseline (p < 0.0001). On the other hand, the number remained relatively constant across all periods for participants in the control group.

3.6. High risk prescribing

Medications identified as potentially suboptimal prescribing using standard measures, namely the Drug Burden Index, Anticholinergic

Table 1 Medication classes at baseline for all participants (n=303).

Regular Medications, n (%)

Anti-infectives for systemic use 43 (1.4) Anti-neoplastic and immune-modulating agents 15 (0.5) Anti-parasitic products, insecticides and repellents 1 (0) Alimentary tract and metabolism 928 (29.8) Blood and blood forming organs 212 (6.8) Cardiovascular system 543 (17.4) Dermatologicals 110 (3.5) Genito urinary system and sex hormones 58 (1.9) Systemic hormonal preparations, excluding sex hormones and insulins 72 (2.3) Musculo-skeletal system 83 (2.7) Nervous system 747 (23.9) Respiratory system 157 (5.0) Sensory organs 131 (4.2) Various 21 (0.7) PRN (pro-re-nata) Medications, n (%) Anti-infectives for systemic use 5 (0.5) Alimentary tract and metabolism 423 (41.1) Blood and blood forming organs 1 (0.1) Cardiovascular system 48 (4.7) Dermatologicals 100 (9.7) Genito urinary system and sex hormones 3 (0.3) Systemic hormonal preparations, excluding sex hormones and insulins 16 (1.6) Musculo-skeletal system 18 (1.8) Nervous system 263 (25.6) Respiratory system 109 (10.6) Sensory organs 40 (3.9) Various 2 (0.2)

Note. Medications were classified by anatomical main group according to the World Health Organisation Anatomical Therapeutic Chemical (ATC) classifica- tion system. “Various” includes nutritional supplements.

H.W. Quek et al.

Archives of Gerontology and Geriatrics 107 (2023) 104910

Fig. 2. Stacked column graph of medication classes frequently selected for deprescribing. Note. Horizontal axis is the frequency of medications being targeted for deprescribing, whereas vertical axis is the therapeutic subgroup (second level) by the World Health Organisation Anatomical Therapeutic Chemical (ATC) classification system. Bars filled with lines represent the frequency of recommended medications declined by GPs. Grey-shaded bars represent the frequency of medications accepted by GPs and implemented at 12-month, whereas bars filled with dots represent the frequency of which accepted by GPs but not implemented at 12-month. * Medication classes that include medications that the participants were exposed to, of which were used in the Drug Burden Index (DBI) calculation. † “Other nervous system drugs” include betahistine. Individual medications in each medication class are presented in S2 Table.

H.W. Quek et al.

Archives of Gerontology and Geriatrics 107 (2023) 104910

Drug Scale, and Beers Criteria, were significantly more likely to be identified as potential deprescribing targets (Table 4, S3 Table). On the other hand, medications that could indicate a prescribing cascade were not more likely to be identified as potential deprescribing targets. Table 5 summarises the change in DBI following pharmacist recom- mendations, GP acceptance, and successful implementation at 12 months. There was no significant difference in DBI between baseline and 12 months for any groups.

4. Discussion

This process evaluation study highlights that there is substantial opportunity to deprescribe potentially suboptimal medications for older people living in RACFs. GPs accepted more than three quarters (77%) of the pharmacists’ recommendations and the collaboration between GPs

and pharmacists significantly reduced the number of medications taken by RACF residents.

Instead of medication withdrawals, previous studies have focused on other aspects of pharmacist recommendations such as changes in ther- apeutic agent and dosage. When compared to other studies investigating pharmacist-led deprescribing intervention in RACFs, our prescriber acceptance rate is similar to the 72% and 86% in New Zealand and Canadian studies, and much higher than 26% in a Pennsylvanian study (Ailabouni et al., 2019; Balsom et al., 2020; Pruskowski & Handler, 2017). Some previous studies focused on deprescribing a single specific medication class (Ailabouni et al., 2019; Pruskowski & Handler, 2017), but our study was designed to target all medication classes, to better reflect actual clinical practice. Our findings may underestimate the po- tential for deprescribing in this population, however, as we limited medications targeted for deprescribing to solid oral dosage form to allow “blinding” by encapsulation. As such, some medications suitable for deprescribing were excluded from our medication withdrawal plans. Regardless of the differences in acceptance rate, deprescribing has led to a significant reduction in the number of medications, in agreement with the findings from previous studies (Ailabouni et al., 2019; Balsom et al., 2020).

In this study, deprescribing vitamins was a common pharmacist recommendation. While vitamin D supplementation is routinely rec- ommended for all RACF residents to reduce the frequency of falls, with sufficient monitoring in place, GPs in this study were comfortable with ceasing every four out of five vitamins targeted (the majority of which was vitamin D) (Duque et al., 2016). This finding is consistent with another study which speculated that pill burden and minimal negative consequences from deprescribing of vitamins could be a key motivator (Cheong et al., 2018). Prescribing of vitamin D is recommended when there is vitamin D deficiency or insufficient sun exposure. The current best practice is to balance the decision to prescribe with the intention to minimise polypharmacy. Primary or secondary prevention may be un- necessary when taking into account an individual’s prognosis, frailty, and personal goals (The Royal Australian College of General Practi- tioners, 2019).

In this study, medications included in the DBI, Anticholinergic Drug Scale, and Beers Criteria were commonly identified as deprescribing targets, of which anticholinergic and sedating agents appear in all referenced measures. We followed the changes in DBI specifically, as high DBI scores have been associated with adverse outcomes in older people living in RACFs including falls (Wilson et al., 2011). However, we did not observe a significant reduction in DBI in this study. Previous studies investigating the impact of a pharmacist-led intervention on the exposure to anticholinergic or sedative medications have also shown

Table 2 Rationales for deprescribing and acceptance rate by GPs (%).

Nature of deprescribing recommendation

Frequency, n (%)

Example of evidence- based rationale

Intervention accepted, n (%)

1. Inappropriate prescriptions (i.e. prescribing cascade, no clear indication, obvious contraindication)

348 (28) Atorvastatin: No documented cardiac history and no recent lipid test was carried out.

287 (82)

2. Adverse effects or interactions outweigh potential benefits

207 (17) Furosemide (frusemide): Symptoms have been stable. May exacerbate urinary incontinence.

158 (76)

3. No longer needed (i.e. symptoms resolved and unlikely to recur OR symptoms stable)

513 (42) Pantoprazole: No recent symptoms reported, suggested potential tapering, and eventually cease.

397 (77)

4. Potential benefit uncertain or unlikely to be realised (i.e. preventive medication)

163 (13) Aspirin: Lack of long- term benefit for primary prevention of cardiovascular events (and risk of gastrointestinal harm) in older people.

127 (78)

Note. The reasons for recommendations made by pharmacists were coded into the four categories (1-4) as described in the deprescribing algorithm (Fig. 1). Recommendations were coded into each category that was relevant.

Table 3 Number of medications at baseline, after pharmacist recommendations, after GP acceptance, and at 12-month follow-up.

Group All regular medicationsa PRN medications Baseline Pharmacist

recommen- dationsb

After GP inputc

12 monthsd

Difference between baseline and 12 months (95% CI)

P-value Baseline 12 months

Difference between baseline and 12 months (95% CI)

P- value

Blinded intervention (n =102)

10.1 ± 4.4

6.0 ± 3.4 6.8 ± 3.5

7.6 ± 3.9

-2.4 (-3.2, -1.6) <

0.0001 3.9 ± 2.7

4.5 ± 3.1

0.7 (0.3, 1.2)

0.0017

Open intervention (n = 101)

10.7 ± 4.4

6.4 ± 3.6 7.4 ± 3.6

9.4 ± 4.5

-1.9 (-2.7, 1.0) <

0.0001 3.5 ± 2.4

3.6 ± 2.6

0.5 (0.1, 0.9)

0.0084

Control (n = 100) 10.1 ± 4.8

5.7 ± 4.0 6.8 ± 4.5

10.4 ± 5.0

0.0 (-0.7, 0.8) 0.9116 3.6 ± 2.6

4.1 ± 2.7

0.8 (0.4, 1.3)

0.0010

Numbers are mean ± SD unless otherwise stated. Data presented are from participants who completed the 12 months follow-up. P-values are from a two-sided t-test. PRN pro re nata.

a Regular medications include all systemically active doses (all oral medicines, topical patches, injections, all inhaled and eye drops, nutritional supplements, mouthwashes, gargles, artificial tears and oxygen). However, only medications in solid oral dosage form were eligible targets for deprescribing.

b Assuming all pharmacist recommendations were accepted and successfully implemented. c Assuming all GP accepted recommendations were successfully implemented. d Based on actual implementation at 12-month follow-up.

H.W. Quek et al.

Archives of Gerontology and Geriatrics 107 (2023) 104910

mixed results (Nishtala et al., 2009; Ailabouni et al., 2019; McDerby et al., 2020; Castelino et al., 2010; van der Meer et al., 2018; Gnjidic et al., 2010). Although a significant reduction in DBI was reported in some studies, these findings were based on pharmacist recommenda- tions or GP acceptance and not the actual implementation (Nishtala et al., 2009; Castelino et al., 2010). Among the two studies that imple- mented a pharmacist-led medication review to explicitly target anti- cholinergic and sedative medications, only one reported a significant decrease in DBI (Ailabouni et al., 2019; van der Meer et al., 2018). Studies suggested that preventative strategies aiming at reducing the initiation of DBI-contributing medications may be more effective. One example is by introducing educational interventions (e.g. academic de- tailing) to influence GPs’ prescribing practices (Gnjidic et al., 2010).

The major strength of this study is a randomised controlled design that allows robust comparison of differences between intervention and control groups, and the use of a deprescribing algorithm modified from explicit and implicit evidence-based tools to standardise the process and minimise subjective bias. However, our study also has several limita- tions. Only one quarter of eligible residents consented to participate in the study, and due to volunteer bias, the participants in this randomised controlled trial are unlikely to be representative of the general RACF population, although baseline data reported elsewhere show the age and gender of the study participants are similar to those who did not consent. In addition, our intervention may not be directly translatable to clinical practice as this rigorous collaborative model of care (two pharmacists and discussions with GPs) is resource- and time-intensive. We also did not document reasons for non-acceptance by GPs.

5. Conclusion

The role of a pharmacist in today’s world is constantly expanding. The effective collaboration between physicians and pharmacists has consistently been shown in the literature as a key driver for positive patient outcomes. Our findings demonstrated that pharmacists, as part of a multidisciplinary team, can make recommendations on depres- cribing that are acceptable to doctors. With a robust and structured

algorithm-based medication review, these recommendations can have a significant impact on reducing the number of inappropriate medications consumed by older people in RACFs. The present study provides a deeper understanding of the prevalence of GP acceptance or non- acceptance by specific medication classes and reasons to deprescribe. The findings in this study point toward a predictor for deprescribing and the potential for enhancing future strategies by focusing the resources on certain medication classes or groups of medications commonly identi- fied as potentially inappropriate. As the evidence for collaborative deprescribing intervention grows, these findings are crucial for future studies implementing and expanding the deprescribing intervention in another setting (community or hospital).

Description of authors’ roles

HWQ analysed the data for NSW and drafted the paper. AP was one of the research pharmacists conducted the medication reviews for all participants. AP analysed the overall data set for Western Australia and NSW as well as reviewed the drafts and assisted with revisions. SL was the research pharmacist for participants in NSW. AM, SL, VN, LK, SH, CEB supervised the data acquisition and interpretation for NSW as well as commented on the draft manuscript for NSW. CEB, KP, SH, VN, AM, TC, DM, RL, AP were the investigators directly involved in the design, development and implementation of the study. All authors read, revised and accepted the final draft.

Compliance with ethical standards

Funding

The study was funded by the National Health and Medical Research Council (GNT1045662). Sponsors played no role in the design, execu- tion, analysis and interpretation of data, or the writing of the study. AP was supported by a University Postgraduate Award during the study and now AP is supported by an NHMRC Early Career Fellowship (1156892).

Table 4 Solid oral medicines commonly identified as potentially suboptimal.

Number of solid oral medicines across all participants

Not identified by pharmacists as potential deprescribing targets

Pharmacists identified medicines as potential deprescribing targets

P-value

GP agreed

GP disagreed

Anticholinergic and sedative medicines contributing to Drug Burden Index

488 208 (42%) 203 (42%)

77 (16%) 0.018

Anticholinergic medicines 328 129 (39%) 147 (45%)

52 (16%) 0.018

Beers medicines 359 116 (32%) 194 (54%)

49 (14%) <0.0001

Potential prescribing cascades 119 47 (40%) 56 (47%)

16 (13%) 0.40

Numbers are presented as count values. P-values are from a chi-square test.

Table 5 Drug Burden Index at baseline, after pharmacist recommendations and after GP acceptance, and at 12-month follow-up.

Group Drug Burden Index Baseline Pharmacist

recommendationsa After GP inputb 12 monthsc Difference between baseline and 12 months (95% CI) P-

value

Blinded intervention (n =102) 1.0 ± 0.9 0.6 ± 0.7 0.7 ± 0.8 1.0 ± 0.8 0.0 (-0.1, 0.1) 0.9377 Open intervention (n = 101) 1.1 ± 0.9 0.6 ± 0.7 0.7 ± 0.7 1.0 ± 0.8 -0.1 (-0.2, 0.0) 0.1596 Control (n = 100) 1.1 ± 0.9 0.6 ± 0.7 0.7 ± 0.8 1.1 ± 0.9 0.1 (-0.0, 0.2) 0.1396

Numbers are mean ± SD unless otherwise stated. P-values are from a two-sided t-test. Data presented are from participants who completed the 12 months follow-up. a Assuming all pharmacist recommendations were accepted and successfully implemented. b Assuming all GP accepted recommendations were successfully implemented. c Based on actual implementation at 12-month follow-up.

H.W. Quek et al.

Archives of Gerontology and Geriatrics 107 (2023) 104910

Ethics approval

This study was approved by the University of Western Australia (Western Australia) and Concord Repatriation General Hospital (New South Wales) human research ethics committees (RA/4/1/5930 and HREC13/CRGH/77). The study was registered with the Australian New Zealand Clinical Trial Registry number (ACTRN12613001204730; World Health Organisation Universal Trial Number U1111-1148-6094).

Consent to participate

Informed consent was obtained from all individual participants included in the study, or their surrogate decision-makers.

Availability of data and materials

The data that support the findings of this study are available from the corresponding author on request.

Code availability

Not applicable.

Acknowledgments

We thank the participating residents, their families and carers, GPs, nursing staff, and community pharmacists involved in the Opti-Med trial. We thank the study personnel involved (David Le Couteur, Jenny Tasker, Marnee Eames, Terry Jin, Rachael Kelly, Kerri Schoenauer, Ash Osborne, Melissa Casey and Georgie Lee). We thank and acknowledge the study biostatistician who undertook the randomisation, the late Kieran McCaul.

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.archger.2022.104910.

References

Ailabouni, N., Mangin, D., & Nishtala, P. S. (2019). DEFEAT-polypharmacy: Deprescribing anticholinergic and sedative medicines feasibility trial in residential aged care facilities. International Journal of Clinical Pharmacy, 41, 167–178.

Balsom, C., Pittman, N., King, R., & Kelly, D. (2020). Impact of a pharmacist- administered deprescribing intervention on nursing home residents: a randomized controlled trial. International Journal of Clinical Pharmacy, 42(4), 1153–1167.

Beers, M. H., Ouslander, J. G., Rollingher, I., Reuben, D. B., Brooks, J., & Beck, J. C. (1991). Explicit criteria for determining inappropriate medication use in nursing home residents. Archives of internal medicine, 151(9), 1825–1832.

Bony, H., Lloyd, R. A., Hotham, E. D., Corre, L. J., Corlis, M. E., Loffler, H. A., …, & Suppiah, V. (2020). Differences in the prescribing of potentially inappropriate medicines in older Australians: comparison of community dwelling and residential aged care residents. Scientific Reports, 10(1), 1–8.

Carnahan, R. M., Lund, B. C., Perry, P. J., Pollock, B. G., & Culp, K. R. (2006). The anticholinergic drug scale as a measure of drug-related anticholinergic burden: Associations with serum anticholinergic activity. The Journal of Clinical Pharmacology, 46(12), 1481–1486.

Castelino, R. L., Hilmer, S. N., Bajorek, B. V., Nishtala, P., & Chen, T. F. (2010). Drug burden index and potentially inappropriate medications in community-dwelling older people. Drugs & Aging, 27(2), 135–148.

Charlson, M. E., Pompei, P., Ales, K. L., & MacKenzie, C. R. (1987). A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. Journal of Chronic Diseases, 40(5), 373–383.

Chen, T. F. (2016). Pharmacist-led home medicines review and residential medication management review: The Australian model. Drugs & Aging, 33(3), 199–204.

Cheong, S. T., Ng, T. M., & Tan, K. T. (2018). Pharmacist-initiated deprescribing in hospitalised elderly: Prevalence and acceptance by physicians. European Journal of Hospital Pharmacy, 25(e1), e35–e39.

Cummings, J. L., Mega, M., Gray, K., Rosenberg-Thompson, S., Carusi, D. A., & Gornbein, J. (1994). The Neuropsychiatric Inventory: Comprehensive assessment of psychopathology in dementia. Neurology, 44(12), 2308-2308.

Duque, G., Lord, S. R., Mak, J., Ganda, K., Close, J. J., Ebeling, P., …, & Inderjeeth, C. A. (2016). Treatment of osteoporosis in Australian residential aged care facilities:

Update on consensus recommendations for fracture prevention. Journal of the American Medical Directors Association, 17(9), 852–859.

2019 American Geriatrics Society Beers Criteria® Update Expert Panel, Fick, D. M., Semla, T. P., Steinman, M., Beizer, J., Brandt, N., …, & Sandhu, S. (2019). American Geriatrics Society 2019 updated AGS Beers Criteria® for potentially inappropriate medication use in older adults. Journal of the American Geriatrics Society, 67(4), 674–694.

Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). Mini-mental state”: A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12(3), 189–198.

Garfinkel, D., Zur-Gil, S., & Ben-Israel, H. (2007). The war against polypharmacy: A new cost-effective geriatric-palliative approach for improving drug therapy in disabled elderly people. Israel Medical Association Journal, 9(6), 430–434.

Garfinkel, D. (2017). Overview of current and future research and clinical directions for drug discontinuation: Psychological, traditional and professional obstacles to deprescribing. European Journal of Hospital Pharmacy, 24(1), 16–20.

Gheewala, P. A., Peterson, G. M., Curtain, C. M., Nishtala, P. S., Hannan, P. J., & Castelino, R. L. (2014). Impact of the pharmacist medication review services on drug-related problems and potentially inappropriate prescribing of renally cleared medications in residents of aged care facilities. Drugs & Aging, 31(11), 825–835.

Gnjidic, D., Couteur, D. G. L., Abernethy, D. R., & Hilmer, S. N. (2010). A pilot randomized clinical trial utilizing the drug burden index to reduce exposure to anticholinergic and sedative medications in older people. Annals of Pharmacotherapy, 44(11), 1725–1732.

Gnjidic, D., Hilmer, S. N., Blyth, F. M., Naganathan, V., Waite, L., Seibel, M. J., …, & Le Couteur, D. G. (2012). Polypharmacy cutoff and outcomes: Five or more medicines were used to identify community-dwelling older men at risk of different adverse outcomes. Journal of Clinical Epidemiology, 65(9), 989–995.

Harrison, S. L., Kouladjian O’Donnell, L., Bradley, C. E., Milte, R., Dyer, S. M., Gnanamanickam, E. S., …, & Crotty, M. (2018). Associations between the drug burden index, potentially inappropriate medications and quality of life in residential aged care. Drugs & Aging, 35(1), 83–91.

Herdman, M., Gudex, C., Lloyd, A., Janssen, M. F., Kind, P., Parkin, D., …, & Badia, X. (2011). Development and preliminary testing of the new five-level version of EQ-5D (EQ-5D-5L). Quality of Life Research, 20(10), 1727–1736.

Hilmer, S. N., Mager, D. E., Simonsick, E. M., Cao, Y., Ling, S. M., Windham, B. G., …, & Abernethy, D. R. (2007). A drug burden index to define the functional burden of medications in older people. Archives of Internal Medicine, 167(8), 781–787.

Horne, R., Weinman, J., & Hankins, M. (1999). The beliefs about medicines questionnaire: The development and evaluation of a new method for assessing the cognitive representation of medication. Psychology and Health, 14(1), 1–24.

Ibrahim, K., Cox, N. J., Stevenson, J. M., Lim, S., Fraser, S. D. S., & Roberts, H. C. (2021). A systematic review of the evidence for deprescribing interventions among older people living with frailty. BMC Geriatrics, 21(1), 258.

Jokanovic, N., Tan, E. C., Dooley, M. J., Kirkpatrick, C. M., & Bell, J. S. (2015). Prevalence and factors associated with polypharmacy in long-term care facilities: A systematic review. Journal of the American Medical Directors Association, 16(6), 535. -e1.

Jokanovic, N., Tan, E. C., Dooley, M. J., Kirkpatrick, C. M., Elliott, R. A., & Bell, J. S. (2016). Why is polypharmacy increasing in aged care facilities? The views of Australian health care professionals. Journal of Evaluation in Clinical Practice, 22(5), 677–682.

Kouladjian, L., Gnjidic, D., Chen, T. F., Mangoni, A. A., & Hilmer, S. N. (2014). Drug Burden Index in older adults: Theoretical and practical issues. Clinical Interventions in Aging, 9, 1503.

Kwint, H. F., Bermingham, L., Faber, A., Gussekloo, J., & Bouvy, M. L. (2013). The relationship between the extent of collaboration of general practitioners and pharmacists and the implementation of recommendations arising from medication review. Drugs & Aging, 30(2), 91–102.

Lee, G., Lim, J. F., Page, A. T., Etherton-Beer, C., Clifford, R., & Wang, K. (2022). Applicability of explicit potentially inappropriate medication lists to the Australian context: A systematic review. Australasian Journal on Ageing.

Lee, G. B., Etherton-Beer, C., Hosking, S. M., Pasco, J. A., & Page, A. T. (2022). The patterns and implications of potentially suboptimal medicine regimens among older adults: a narrative review. Therapeutic Advances in Drug Safety, 13, Article 20420986221100117.

Maher, R. L., Hanlon, J., & Hajjar, E. R. (2014). Clinical consequences of polypharmacy in elderly. Expert Opinion on Drug Safety, 13(1), 57–65.

Masnoon, N., Shakib, S., Kalisch-Ellett, L., & Caughey, G. E. (2017). What is polypharmacy? A systematic review of definitions. BMC Geriatrics, 17(1), 1–10.

McDerby, N., Kosari, S., Bail, K., Shield, A., Peterson, G., & Naunton, M. (2020). Residential aged care pharmacist: An Australian pilot trial exploring the impact on quality use of medicines indicators. Medicines, 7(4), 20.

Morin, L., Laroche, M. L., Texier, G., & Johnell, K. (2016). Prevalence of potentially inappropriate medication use in older adults living in nursing homes: A systematic review. Journal of the American Medical Directors Association, 17(9), 862. -e1.

Nishtala, P. S., Hilmer, S. N., McLachlan, A. J., Hannan, P. J., & Chen, T. F. (2009). Impact of residential medication management reviews on Drug Burden Index in aged-care homes. Drugs & Aging, 26(8), 677–686.

Page, A. T., Etherton-Beer, C. D., Clifford, R. M., Burrows, S., Eames, M., & Potter, K. (2016). Deprescribing in frail older people–do doctors and pharmacists agree? Research in Social and Administrative Pharmacy, 12(3), 438–449.

Page, A. T., Clifford, R. M., Potter, K., Schwartz, D., & Etherton-Beer, C. D. (2016). The feasibility and effect of deprescribing in older adults on mortality and health: A systematic review and meta-analysis. British journal of Clinical Pharmacology, 82(3), 583–623.

H.W. Quek et al.

https://doi.org/10.1016/j.archger.2022.104910

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0002

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0003

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0004

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0005

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0006

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0007

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0008

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0009

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0010

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0011

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0012

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0001

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0013

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0014

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0015

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0016

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0017

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0018

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0019

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0020

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0021

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0022

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0023

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0024

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0025

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0026

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0027

http://refhub.elsevier.com/S0167-4943(22)00297-7/optV4KWsbJ35R

http://refhub.elsevier.com/S0167-4943(22)00297-7/optvveZExeIFl

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0028

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0029

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0030

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0031

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0032

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0033

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0034

Archives of Gerontology and Geriatrics 107 (2023) 104910

Page, A. T., Potter, K., Naganathan, V., Hilmer, S., McLachlan, A. J., Lindley, R. I., Coman, T., Mangin, D., & Etherton-Beer, C. (2023). Polypharmacy and medicine regimens in older adults in residential aged care. Archives of Gerontology and Geriatrics, 105, Article 104849.

Palagyi, A., Keay, L., Harper, J., Potter, J., & Lindley, R. I. (2016). Barricades and brickwalls–A qualitative study exploring perceptions of medication use and deprescribing in long-term care. BMC Geriatrics, 16(1), 1–11.

Pickard, A. S., De Leon, M. C., Kohlmann, T., Cella, D., & Rosenbloom, S. (2007). Psychometric comparison of the standard EQ-5D to a 5 level version in cancer patients. Medical care, 259–263.

Potter, K., Flicker, L., Page, A., & Etherton-Beer, C. (2016). Deprescribing in frail older people: A randomised controlled trial. PLoS One, 11(3), Article e0149984.

Pruskowski, J., & Handler, S. M. (2017). The DE-PHARM project: A pharmacist-driven deprescribing initiative in a nursing facility. The Consultant Pharmacist®, 32(8), 468–478.

Reeve, E., Gnjidic, D., Long, J., & Hilmer, S. (2015). A systematic review of the emerging definition of ‘deprescribing’with network analysis: Implications for future research and clinical practice. British Journal of Clinical Pharmacology, 80(6), 1254–1268.

Rockwood, K., Mogilner, A., & Mitnitski, A. (2004). Changes with age in the distribution of a frailty index. Mechanisms of Ageing and Development, 125(7), 517–519.

Royal Commission into Aged Care Quality and Safety. Interim report: neglect. Canberra, ACT: Commonwealth of Australia. 2019 September. Retrieved from https://agedcar e.royalcommission.gov.au/publications/interim-report. Accessed 2020 October 23.

Scott, I. A., Hilmer, S. N., Reeve, E., Potter, K., Le Couteur, D., Rigby, D., …, & Martin, J. H. (2015). Reducing inappropriate polypharmacy: The process of deprescribing. JAMA Internal Medicine, 175(5), 827–834.

Shah, S., Vanclay, F., & Cooper, B. (1989). Improving the sensitivity of the Barthel Index for stroke rehabilitation. Journal of Clinical Epidemiology, 42(8), 703–709.

Somers, M., Rose, E., Simmonds, D., Whitelaw, C., Calver, J., & Beer, C. (2010). Quality use of medicines in residential aged care. Australian Family Physician, 39(6), 413–416.

Steinman, M., & Reeve, E. (2021). Deprescribing. UptoDate. Retrieved from https://www. uptodate.com/contents/deprescribing/print. Accessed 2022 May 15.

The Royal Australian College of General Practitioners. (2019). RACGP aged care clinical guide (Silver Book) (5th ed.). East Melbourne, VIC: RACGP. Retrieved from https: //www.racgp.org.au/silverbook.

Turner, J. P., Jamsen, K. M., Shakib, S., Singhal, N., Prowse, R., & Bell, J. S. (2016). Polypharmacy cut-points in older people with cancer: How many medications are too many? Supportive Care in Cancer, 24(4), 1831–1840.

van der Meer, H. G., Wouters, H., Pont, L. G., & Taxis, K. (2018). Reducing the anticholinergic and sedative load in older patients on polypharmacy by pharmacist- led medication review: a randomised controlled trial. BMJ open, 8(7), Article e019042.

Weir, K. R., Naganathan, V., Bonner, C., McCaffery, K., Rigby, D., McLachlan, A. J., & Jansen, J. (2020). Pharmacists’ and older adults’ perspectives on the benefits and barriers of Home Medicines Reviews–A qualitative study. Journal of Health Services Research & Policy, 25(2), 77–85.

Weir, K. R., Naganathan, V., Rigby, D., McCaffery, K., Bonner, C., Trevena, L., … Jansen. (2019). Home medicines reviews: A qualitative study of GPs’ experiences. Australian Journal of Primary Health, 26(1), 24–30.

Westaway, K., Sluggett, J., Alderman, C., Moffat, A., Procter, N., & Roughead, E. (2020). The extent of antipsychotic use in Australian residential aged care facilities and interventions shown to be effective in reducing antipsychotic use: A literature review. Dementia, 19(4), 1189–1202.

Wilson, N. M., Hilmer, S. N., March, L. M., Cameron, I. D., Lord, S. R., Seibel, M. J., …, & Sambrook, P. N. (2011). Associations between drug burden index and falls in older people in residential aged care. Journal of the American Geriatrics Society, 59(5), 875–880.

World Health Organisation. (2011). WHO collaborating centre for drug statistics methodology. ATC/DDD index 2011. World Health Organization2011WHO Collaborating Centre for Drug Statistics Methodology. ATC/DDD index.

H.W. Quek et al.

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0035

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0036

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0037

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0038

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0039

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0040

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0041

https://agedcare.royalcommission.gov.au/publications/interim-report

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0043

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0044

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0045

https://www.uptodate.com/contents/deprescribing/print

https://www.racgp.org.au/silverbook

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0048

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0049

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0050

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0051

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0052

http://refhub.elsevier.com/S0167-4943(22)00297-7/sbref0053

Deprescribing for older people living in residential aged care facilities: Pharmacist recommendations, doctor acceptance an ...

1 Introduction
2 Methods

2.1 Study design
2.2 Setting
2.3 Recruitment, participants, inclusion and exclusion criteria
2.4 Description of intervention
2.5 Data analysis

2.5.1 High risk prescribing
2.5.2 Statistical methods

3 Results

3.1 Study cohort characteristics
3.2 Commonly targeted medication classes
3.3 Reasons for deprescribing
3.4 GPs’ acceptance rate
3.5 Number of medications
3.6 High risk prescribing

4 Discussion
5 Conclusion
Description of authors’ roles
Compliance with ethical standards

Funding
Ethics approval
Consent to participate
Availability of data and materials
Code availability

Acknowledgments
Supplementary materials
References

image7.emf

Vol.:(0123456789)

The European Journal of Health Economics (2025) 26:427–454 https://doi.org/10.1007/s10198-024-01718-7

ORIGINAL PAPER

Cost effectiveness of a GP delivered medication review to reduce polypharmacy and potentially inappropriate prescribing in older patients with multimorbidity in Irish primary care: the SPPiRE cluster randomised controlled trial

Paddy Gillespie1 · Frank Moriarty2 · Susan M. Smith3 · Anna Hobbins1 · Sharon Walsh1 · Barbara Clyne4 · Fiona Boland4 · Tara McEnteggart2 · Michelle Flood4 · Emma Wallace5 · Caroline McCarthy6 · for the SPPiRE Study team

Received: 16 November 2023 / Accepted: 7 August 2024 / Published online: 27 August 2024 © The Author(s) 2024

Abstract Background Evidence on the cost effectiveness of deprescribing in multimorbidity is limited. Objective To investigate the cost effectiveness of a general practitioner (GP) delivered, individualised medication review to reduce polypharmacy and potentially inappropriate prescribing in older patients with multimorbidity in Irish primary care. Methods Within trial economic evaluation, from a healthcare perspective and based on a cluster randomised controlled trial with a 6 month follow up and 403 patients (208 Intervention and 195 Control) recruited between April 2017 and December 2019. Intervention GPs used the SPPiRE website which contained educational materials and a template to support a web- based individualised medication review. Control GPs delivered usual care. Incremental costs, quality adjusted life years (QALYs) generated using the EQ-5D-5L instrument, and expected cost effectiveness were estimated using multilevel model- ling and multiple imputation techniques. Uncertainty was explored using parametric, deterministic and probabilistic methods. Results On average, the SPPiRE intervention was dominant over usual care, with non-statistically significant mean cost sav- ings of €410 (95% confidence interval (CI): − 2211, 1409) and mean health gains of 0.014 QALYs (95% CI − 0.011, 0.039). At cost effectiveness threshold values of €20,000 and €45,000 per QALY, the probability of SPPiRE being cost effective was 0.993 and 0.988. Results were sensitive to missing data and data collection period. Conclusions The study observed a pattern towards dominance for the SPPiRE intervention, with high expected cost effec- tiveness. Notably, observed differences in costs and outcomes were consistent with chance, and missing data and related uncertainty was non trivial. The cost effectiveness evidence may be considered promising but equivocal. Trial registration ISRCTN: 12752680, 20th October 2016.

Keywords Multimorbidity · Deprescribing · Cost effectiveness

JEL Classification I10 · I11

* Paddy Gillespie [email protected]

1 Health Economics and Policy Analysis Centre, Institute for Lifecourse and Society, CURAM, SFI Research Centre for Medical Devices, School of Business and Economics, University of Galway, University Road, Galway, Ireland

2 School of Pharmacy and Biomolecular Sciences, Royal College of Surgeons in Ireland, University of Medicine and Health Sciences, Dublin, Ireland

3 Discipline of Public Health and Primary Care, Institute of Population Health, Trinity College Dublin, Dublin, Ireland

4 School of Population Health, Royal College of Surgeons in Ireland, University of Medicine and Health Sciences, Dublin, Ireland

5 Department of General Practice, University College Cork, Cork, Ireland

6 Department of General Practice, Royal College of Surgeons in Ireland, University of Medicine and Health Sciences, Dublin, Ireland

http://crossmark.crossref.org/dialog/?doi=10.1007/s10198-024-01718-7&domain=pdf

http://orcid.org/0000-0001-6297-9478

428 P. Gillespie et al.

Introduction

Multimorbidity is associated with adverse outcomes including increased mortality and reduced quality of life, and increased healthcare utilisation and costs [1–3]. Unplanned hospital admissions are a key driver of excess healthcare costs related to multimorbidity [4], and these in turn, are frequently the result of adverse drug reactions [5]. Indeed, prescribing for patients with multimorbidity is particularly complex due to related polypharmacy, which is associated with potentially inappropriate prescribing and adverse drug related events [6–8]. This polyphar- macy is often necessary and appropriate in the context of managing multiple chronic conditions and complex needs. However, higher levels of polypharmacy in multimorbid- ity have been shown to be associated with higher levels of adverse outcomes, hospital admissions, and related health- care costs [4, 5, 9]. In this context, healthcare providers caring for patients with multimorbidity are increasingly engaged in medications management and deprescribing practices, which involve the ongoing assessment of both the effectiveness and risks of treatments, and incorpora- tion of their patients’ preferences [10]. To this end, an individualised approach to deprescribing in multimorbid- ity has been proposed in published multimorbidity and polypharmacy guidelines [11–14], which highlight that a single disease focus may not be optimal for patients with multimorbidity [15].

Evidence from randomised controlled trials is limited generally on the clinical and cost effectiveness of inter- ventions targeting patients with multimorbidity [16], and interventions targeting patients with polypharmacy [17], with respective 2021 and 2018 Cochrane reviews reporting sparse and mixed results for health benefits and value for money. Evidence of cost effectiveness of medications man- agement and deprescribing interventions for multimorbid patients is even more limited, and reviews have highlighted the need for further research in this area [16–19]. In the Irish context, the Supporting Prescribing in Older Adults with Multimorbidity in Irish Primary Care (SPPiRE) study reported the clinical effectiveness of a general practitioner (GP) delivered, individualised medication review inter- vention, that was developed incorporating the concepts of treatment burden and deprescribing and which focused on higher levels of polypharmacy for patients with mul- timorbidity [20]. The intervention resulted in a small but statistically significant effect in reducing the number of medicines (IRR 0.95, 95% CI 0.899–0.999, p = 0.045) and a weakly significant effect on potentially inappropri- ate prescriptions (PIP) (OR 0.39, 95% CI 0.140–1.064, p = 0.066). In addition to clinical effectiveness, any

decision regarding the adoption of an intervention in clini- cal practice will depend upon its expected cost effective- ness [21]. The technique of health economic evaluation is concerned with the exploration of cost effectiveness by relating the mean difference in cost between alternative treatment options to their mean difference in effectiveness, and by quantifying the related uncertainty [21]. This paper reports the cost effectiveness results from the health eco- nomic evaluation conducted alongside the SPPiRE cluster RCT to assess an intervention targeting reductions in poly- pharmacy and potentially inappropriate prescribing among older patients with multimorbidity in Irish primary care.

This study adds to the limited evidence base on the cost effectiveness of interventions targeting medications man- agement and desprescribing in patients with multimorbid- ity. Laberge et al. [18] conducted a systematic review of the economic impact of interventions intended at optimiz- ing medication use in older adults with multimorbidity and polypharmacy. The review included 11 studies and reported that interventions to optimize medication use may provide benefits that outweigh their implementation costs, but the evidence remains limited [18]. In terms of the related and relevant cost effectiveness studies, two recent papers based on randomised controlled trials examined the cost effective- ness of interventions targeting medications management and desprescribing in patients with multimorbidity, one in pri- mary care [22] and one in hospital care [23]. Thorn et al. [22] conducted a health economic evaluation of the 3D ran- domised control trial study, and found that the evidence for the cost effectiveness of the 3D intervention was equivo- cal; the results suggesting that there was just over a 50% chance of cost effectiveness at the established UK threshold of £20,000 per QALY from the healthcare perspective [22]. More recently, Salari et al. [23] reported the cost effective- ness findings alongside the OPERAM cluster-randomized trial aimed at testing the effect of a structured pharmacother- apy optimization intervention on preventable drug-related hospital admissions in adults with multimorbidity and polypharmacy aged 70 years or older. The authors reported that the results were not definitive, but were indicative of a pattern towards dominance, potentially resulting from an accumulation of multiple, small, positive intervention effects [23]. This study also adds to the evidence base for the cost effectiveness of interventions targeting multimorbidity and polypharmacy more generally [16–19]. While comparison of international studies is complicated by the variety of defini- tions used for multimorbidity, and by the heterogeneity in study designs, further studies are required to explore the health and economic implications of interventions targeting the multimorbid patient population.

429Cost effectiveness of a GP delivered medication review to reduce polypharmacy and potentially…

Methods

Overview

The economic evaluation was conducted insofar as pos- sible, in accordance to the guidelines for the conduct of health economic evaluation in Ireland [24], and the find- ings are reported in line with the best-practice CHEERS checklist [25]. The perspective of the healthcare system (that is, the Irish health service executive (HSE)) was adopted with respect to costing and health outcomes were expressed in terms of quality adjusted life years (QALYs) gained. The mode of analysis consisted of a trial-based evaluation with a time horizon of 6 months, the trial fol- low-up period. Given the length of follow up, neither costs nor outcomes were discounted. Data on resource use were collected directly from general practice records, while health outcome data was collected via patient question- naires at baseline and follow up. The statistical analysis was conducted on an intention to treat basis, and in accord- ance with guidelines for cluster RCTs [26–30]. Results are presented from complete case and multiple imputation [31] analyses, which was conducted following guidance for hierarchical datasets [30]. Uncertainty was addressed using statistical inference methods, in the form of 95% confidence intervals and hypothesis tests, probabilistic sensitivity analysis, reported in the form of estimated probabilities of the intervention being cost effective at a range of potential threshold values (λ) that the health sys- tem may be willing to pay per additional QALY gained [21], and deterministic sensitivity analysis. All analyses were undertaken using Stata 15 and Microsoft Excel sta- tistical software packages.

Randomised controlled trial (RCT)

The methods for the SPPiRE RCT (ISRCTN: 12752680) have been described elsewhere [32]. In brief, SPPiRE was a pragmatic two arm cluster RCT, which was conducted in line with the CONSORT guidelines for cluster RCTs [26]. Ethical approval was granted by the Irish College of General Practitioners Research Ethics Committee. Infor- mation about the trial was publicised through a variety of research, teaching and training networks throughout Ire- land. Eligible practices that expressed an interest in tak- ing part were formally invited. Practices were eligible to participate if they had at least 300 patients aged ≥ 65 years on their patient panel and they used either one of the two Irish practice management systems in use in over 80% of practices nationally. Practices were excluded if they were currently involved in a medication management or

prescribing trial or if they were unable to recruit at least five participants. Eligible patients were aged ≥ 65 years and prescribed ≥ 15 repeat medicines. Patients were excluded if they had been recruited into a practice that was unable to recruit at least four other participants, they were unable to give informed consent, as judged by their GP, or they were unable to attend the practice for a face to face medication review. Recruited GPs ran a patient finder tool embedded in their practice management systems and screened the generated list to ensure only eligible patients were invited. All recruited practice and patients gave full informed consent and baseline data was collected prior to treatment arm allocation.

Between April 2017 and December 2019, 139 practices and 1626 patients were invited to take part. A total of 51 practices were recruited giving an overall practice enrolment rate of 36.7%.

Of the patients invited, 403 were recruited into the trial giving an enrolment rate of 24.8%. Recruited practices were randomly allocated using minimisation to the SPPiRE inter- vention (n = 26) or the usual care control (n = 25) by the independent trial statistician, resulting in 208 patients being randomised to the SPPiRE intervention arm and 195 to the usual care arm. Considering the nature of the intervention, it was not possible to blind GPs or patients to the intervention. Descriptive statistics for the baseline characteristics of the practices and patients for each treatment arm are presented in Table 1. Recruited patients had a mean age of 76.5 years (SD 6.83), a mean number of medicines of 17.37 (SD 3.50), and a mean number of PIPs per person of 2.52 (SD 1.48), identified from a list of 34 pre-specified indicators (see Appendix Table 5). Practices and patients in each group had similar characteristics at baseline.

The SPPiRE intervention, in terms of its implementa- tion strategies and intervention components, are described in detail elsewhere [20] and in the accompanying appendix. In brief, intervention GPs received unique login details to the SPPiRE website where they had access to five training videos and a template for performing the SPPiRE medica- tion review (Appendix Fig. 1). The training videos provided background information on multimorbidity and polyphar- macy, PIP, eliciting patient treatment priorities and conduct- ing a brown bag medication review, in which the GP and patient jointly reviewed each medication. In terms of the key intervention components, GPs were instructed to book a double appointment and to ask their patients to bring all their medicines in to the medication review visit with them. The SPPiRE medication review process had two elements; gather and record information and then to discuss and agree changes with their patient based on the recorded informa- tion, with a focus on deprescribing medicines that were inap- propriate. GPs initially screened the prescription for PIP and

430 P. Gillespie et al.

Table 1 Practice and patient characteristics by treatment arm at baseline

Practice variable Intervention (N = 26) Control (N = 25)

No. of GP sessions per week Mean (SD) 30.42 (17.35) 27.54 (13.78) Median (IQR) 29.50 (18–37) 26 (16–35.5) No. of PN sessions per week Mean (SD) 12.40 (7.09) 10.79 (5.68) Median (IQR) 10 (9–15) 10 (7.5–12.5) Practice manager None (%) 3 (11.5) 6 (24.0) Part-time N (%) 9 (34.6) 9 (36.0) Full-time N (%) 14 (53.8) 10 (40.0) No. of patients Mean (SD) 6877.72 (3354.24) 6512.56 (3942.18) Median (IQR) 6850 (5484–7994) 5948 (3265–8519) No. of Patients aged ≥ 65 years Mean (SD) 1192.78 (916.78) 1192.78 (650.66) Median (IQR) 974.5 (625–1248) 714 (591–1422) Location Urban N (%) 14 (53.8) 16 (64) Rural N (%) 4 (15.4) 2 (8) Mixed N (%) 8 (30.8) 7 (28) Written repeat prescribing policy N (%) 14 (53.8) 11 (44) Patient variables Intervention (N = 208) Control (N = 195) Age Mean (SD) 76.67 (6.80) 76.32 (6.90) Median (IQR) 76 (71–82) 76 (70–82) Sex

N % N % Male 89 42.79 83 42.56 Female 119 57.21 112 57.44 General Medical Scheme (GMS) medical card holder Full medical card 164 82.00 166 89.73 Doctor visit card 32 16.00 16 8.65 Private Health Insurance (PHI) holder PHI 69 34.50 65 35.14 Language Language other than English 3 1.52 2 1.10 English 195 98.48 179 98.90 Social class Professional worker 13 6.25 11 5.61 Managerial and technical 38 18.27 25 12.76 Non-manual 30 14.42 26 13.27 Skilled manual 26 12.50 26 13.27 Semi-skilled 11 5.29 18 9.18 Unskilled 9 4.33 7 3.57 Farmer, size of farm unspecified 5 2.40 10 5.10 Unknown 52 25.00 51 26.02 Homemaker 24 11.54 22 11.22 Education No schooling 0 0.00 3 1.63 Primary school education only 69 34.85 85 46.20

431Cost effectiveness of a GP delivered medication review to reduce polypharmacy and potentially…

then discussed patient treatment priorities and performed a brown bag review where each medicine was discussed in turn with the patient, and issues such as effectiveness, adverse effects and actual drug utilisation were discussed. The website provided suggested treatment alternatives for identified PIP but all treatment decisions were ultimately at the discretion of the individual GP, based on their clinical judgement and their patients’ individual priorities.

Control GPs delivered usual care in Irish general practice. At the time of intervention delivery there was no structured chronic disease management programme in Irish primary care and many patients with multimorbidity attended mul- tiple hospital specialists. All people aged ≥ 70 years of age have access to a medical card which grants free primary care, whereas access in the 65–69 year old age category is means tested on the basis of income. Access to special- ists and diagnostics in secondary care is free for the entire population, but shorter waiting lists exist in the private care pathway and for those with private health insurance.

The two primary outcomes in the clinical effectiveness analysis were the number of repeat medicines and the pro- portion of patients with any PIPs. Data on demographics, socioeconomics, medical history, healthcare utilisation and health outcomes were collected at baseline and at 6 months after intervention delivery. Data for prescribed medicines and healthcare utilisation were collected by participating GPs and submitted to the study manager. Data for patient reported health outcome measures were collected by postal questionnaires. Overall, 35 patients (8.66%) were lost to

follow-up, 21 of whom died (12 in intervention and 9 in control) during the study period (Appendix Table 6). In addition, 174 or 43% of participants (90 in intervention and 84 in control) did not complete the patient questionnaire at follow up, and had no available data for patient reported health outcomes. There did not appear to be systematic dif- ferences between intervention and control in respect of par- ticipants lost to follow up or missing data and as a result, such data was assumed to be missing at random (Appendix Table 7). Finally, the study was ongoing during the onset of the COVID-19 pandemic, and follow up of 106 participants (57 in intervention and 49 in control) was after March 2020 and in the midst of the pandemic.

Cost analysis

Three cost components were included in the analysis, all of which were expressed in Euros (€) in 2022 prices (see Appendix Table 8). Unit cost estimates for each activity were based on national data sources and, where necessary, were transformed to Euros (€) in 2022 prices using appropri- ate indices [33].

The first cost component related to the resources required to implement and operate the SPPiRE interven- tion in clinical practice. This included a number of fixed, once-off resource outlays including the establishment of the SPPiRE website, and the related educator and admin- istrator time input. In addition, it included a range of vari- able operation items relating to healthcare professional

SD standard deviation, IQR interquartile range, GP general practitioner, PN practice nurse, PIP potentially inappropriate prescriptions

Table 1 (continued) Practice variable Intervention (N = 26) Control (N = 25)

Some secondary education 53 26.77 44 23.91 Complete secondary education 40 20.20 20 10.87 Some third level education 20 10.10 19 10.33 Complete third level education 16 8.08 13 7.07 Employment Employed 1 0.51 0 0.00 Self employed 7 3.54 4 2.17 Retired 156 78.79 145 78.80 Homemaker 32 16.16 31 16.85 Other 2 1.01 4 2.17 Number of prescribed medicines Mean (SD) 16.96 (3.25) 17.83 (3.71) Median (IQR) 16 (15–19) 17 (15–20) Proportion of patients with at least 1 Potentially Inappropriate Prescription (PIP) N (%) 193 (93.24) 180 (92.78) Number of PIP Mean (SD) 2.49 (1.52) 2.56 (1.45) Median (IQR) 2 (1–3) 3 (2–3)

432 P. Gillespie et al.

time input, educational materials and consumables, post, packaging, telephone and travel expenses. This data was recorded prospectively by the study research team. This cost was allocated to all patients in the intervention arm. The impact of halving and doubling the intervention cost estimate were tested in sensitivity analysis.

Second, the costs of medication prescriptions over the trial follow up period were estimated for both treatment arms. A marginal analysis approach was adopted given the volume of prescriptions (approximately 13,800 in total) involved, in that only the costs of medications that were stopped (that is, present at baseline but not follow up: n = 1398), and the costs

Table 2 Summary statistics for resource use, healthcare costs, EQ-5D-5L scores and QALY estimates at baseline and follow up

SD standard deviation, % = percentage, GP general practitioner, QALYs quality adjusted life years Missing Data: Intervention: Baseline—7% missing data for GP consultations, PN visits, emergency department visits, inpatient nights, outpatient visits, 9% for EQ-5D-5L index scores, and 7% for Total Cost. Follow up—9% missing data for GP consultations, PN visits, emergency department visits, inpa- tient nights, outpatient visits, 50% for EQ-5D-5L, 2% for medications started and medications stopped, 11% for Total Cost, and 53% for QALYs gained Control: Baseline—10% missing data for GP consultations, PN visits, emergency department visits, 14% inpatient nights, 13% outpatient visits, 14%, 6% for EQ-5D-5L index score, and 15% for Total Cost. Follow up—12% missing data for GP consultations, PN visits, emergency depart- ment visits, 14% inpatient nights, 13% outpatient visits, 46% for EQ-5D-5L index score, 5% for medications started and medications stopped, 19% for Total Cost, and 49% for QALYs gained

Healthcare Resource Items

Intervention Control Intervention Control

Baseline—6 Months—Mean (SD) / % Follow Up—6 Months—Mean (SD) / %

Usage Cost € Usage Cost € Usage Cost € Usage Cost €

Intervention SPPiRE 100% 257 0% 0 Medication

Prescriptions Number of

medicines stopped

3.97 (3.15) 598.80 (1063.18)

2.92 (3.17) 448.29 (1026.34)

Number of medicines started

3.02 (3.03) 487.01 (763.99) 2.67 (2.91) 519.98 (1443.07)

Other Health- care Services

GP Consulta- tions

4.81 (3.67) 254.89 (194.30) 4.45 (3.06) 238.79 (162.32) 4.42 (3.51) 234.32 (186.10) 3.84 (3.27) 203.32 (173.39)

GP Phone Con- sultations

0.94 (1.23) 49.72 (65.11) 0.99 (1.90) 52.69 (105.16) 1.55 (2.12) 82.29 (112.33) 1.42 (2.21) 75.14 (117.27)

GP House Call Consultations

0.17 (0.59) 8.74 (31.16) 0.15 (0.63) 8.13 (33.25) 0.20 (0.83) 10.60 (44.02) 0.13 (0.68) 6.78 (36.07)

GP Out of Hours Consultations

0.31 (1.09) 16.67 (57.84) 0.11 (0.39) 5.72 (20.80) 0.33 (1.15) 17.29 (60.68) 0.16 (0.44) 8.68 (23.50)

GP Prescription Consultations

2.69 (2.93) 142.34 (155.15) 2.90 (3.57) 153.58 (188.85) 2.51 (2.55) 132.78 (134.96) 2.40 (2.58) 127.08 (136.46)

Practice Nurse Consultations

1.93 (2.23) 81.19 (93.58) 2.25 (3.37) 94.50 (141.66) 1.76(1.98) 74.00 (83.12) 1.90 (2.96) 79.83 (124.40)

Outpatient Clinic Visits

2.38 (2.07) 416.97 (362.90) 2.73 (2.26) 467.23 (385.72) 2.62(5.54) 458.99 (971.62) 2.39 (2.29) 419.06 (400.95)

Hospital Inpa- tient Nights

2.42 (6.17) 2388.52 (6098.40)

2.71 (9.95) 2691.86 (9832.07)

2.43(6.19) 2399.43 (6110.99)

3.09 (9.81) 3052.21 (9695.97)

Emergency Department Visits

0.38 (0.77) 116.93 (237.43) 0.31(0.62) 94.06 (190.29) 0.46 (1.01) 140.37 (309.91) 0.33 (0.85) 102.18 (259.69)

Total Health- care Cost

3475.96 (6318.62)

3814.78 (9947.10)

3744.80 (6440.69)

4137.97 (10347.80)

Health Outcomes EQ-5D-5L Index

Score 0.496 (0.362) 0.471 (0.383) 0.517 (0.382) 0.452 (0.357)

QALYs Gained 0.261 (0.171) 0.234 (0.167)

433Cost effectiveness of a GP delivered medication review to reduce polypharmacy and potentially…

Ta bl

e 3

In cr

em en

ta l c

os t e

ffe ct

iv en

es s a

na ly

si s r

es ul

ts a

t 6 m

on th

s f ol

lo w

u p

Va ria

bl e/

an al

y- si

s C

om pl

et e

ca se

a na

ly si

s M

ul tip

le im

pu ta

tio n

an al

ys is

In cr

em en

ta l

C os

t A na

ly -

sis To

ta l H

ea lth

- ca

re C

os t

In te

rv en

tio n

C on

tr ol

In te

rv en

tio n

C on

tr ol

M ea

n Es

tim at

e (9

5% C

I) 38

15 .3

3 (3

64 0.

25 ,

39 90

.4 2)

44 30

.4 8

(4 25

5. 40

, 46

05 .5

37 31

.2 3

(3 58

4. 84

, 3 87

7. 63

) 41

32 .5

0 (3

98 6.

10 , 4

27 8.

90 )

M ea

n D

iff er

- en

ce E

sti m

at e

(S E)

(p -v

al ue

) (9

5% C

I) (N

) ( Q

IC )

− 61

5. 15

(9 95

.2 5)

(0 .5

37 ) (

− 25

65 .8

13 35

.5 1)

(3 31

) ( 61

64 .9

27 )

− 40

1. 27

(9 22

.2 3)

(0 .6

64 ) (

− 22

11 .0

0, 1

40 8.

47 )

In cr

em en

ta l

Eff ec

tiv en

es s

A na

ly sis

Q A

LY s G

ai ne

d In

te rv

en tio

n C

on tr

ol In

te rv

en tio

n C

on tr

ol M

ea n

Es tim

at e

(9 5%

C I)

0. 24

0 (0

.2 23

, 0.

25 7)

0. 23

5 (0

.2 18

, 0.

25 2)

0. 24

7 (0

.2 35

, 0 .2

58 )

0. 23

3 (0

.2 21

, 0 .2

44 )

M ea

n D

iff er

- en

ce E

sti m

at e

(S E)

(p -v

al ue

) (9

5% C

I) (N

) ( Q

IC )

0. 00

5 (0

.0 08

) ( 0.

53 8)

(− 0.

01 1,

0 .0

20 )

(1 96

) ( 6.

28 3)

0. 01

4 (0

.0 12

) ( 0.

26 9)

(− 0.

01 1,

0 .0

39 )

In cr

em en

- ta

l C os

t Eff

ec tiv

en es

s A

na ly

sis N

et B

en efi

t St

at ist

ic (λ

) In

te rv

en tio

n C

on tr

N et

B en

efi t (

λ)

M ea

n Es

ti- m

at e

λ = €2

0, 00

0 (9

5% C

I) λ =

€4 5,

00 0

(9 5%

C I)

13 28

.1 8

(9 69

.3 7,

1 68

6. 99

) 76

23 .3

0 (6

88 4.

53 , 8

36 2.

06 )

37 9.

18 (−

10 2.

89 , 8

61 .2

4) 60

49 .8

5 (5

15 4.

72 , 6

94 4.

99 )

434 P. Gillespie et al.

Ta bl

e 3

(c on

tin ue

Va ria

bl e/

an al

y- si

s C

om pl

et e

ca se

a na

ly si

s M

ul tip

le im

pu ta

tio n

an al

ys is

In cr

em en

ta l

N et

B en

efi t

(I N

B ) (

λ) IN

λ = €2

0, 00

0 (S

E) (p

-v al

ue )

(9 5%

C I)

IN B

λ =

€4 5,

00 0

(S E)

(p -v

al ue

) (9

5% C

SP Pi

R E

In te

rv en

tio n

D om

in an

t 98

4. 83

(3 91

.7 1)

(0 .0

12 ) (

21 7.

09 , 1

75 2.

57 )

16 46

.6 0

(7 45

.9 4)

(0 .0

27 ) (

18 4.

59 , 3

10 8.

62 )

Pr ob

ab ili

st ic

na ly

sis Pr

ob ab

ili ty

th at

S PP

iR E

In te

rv en

tio n

is C

os t E

ffe ct

iv e

fo r

Th re

sh ol

d Va

lu e

(λ )

Pr ob

ab ili

ty th

th e

SP Pi

R E

In te

rv en

- tio

n is

C os

t Eff

ec tiv

e fo

r a

ra ng

e of

re sh

ol d

Va lu

es (λ

) es

tim at

us in

g a

tw o-

st ag

e, n

on -

pa ra

m et

ric

bo ot

str ap

pi ng

ch ni

qu e

λ = €0

λ = €1

0, 00

0 λ =

€2 0,

00 0

λ = €3

0, 00

0 λ =

€4 00

00 λ =

€4 50

00 0.

99 5

0. 99

8 0.

99 3

0. 99

3 0.

98 8

0. 98

8 λ =

€5 00

00 λ =

€6 00

00 λ =

€7 00

00 λ =

€8 00

00 λ =

€9 00

00 λ =

€1 00

00 0

0. 98

0 0.

97 9

0. 97

7 0.

97 4

0. 97

5 0.

97 7

C I c

on fid

en ce

in te

rv al

s, Q

IC q

ua si

-li ke

lih oo

d cr

ite rio

n, S

E st

an da

rd e

rr or

, G P

ge ne

ra l p

ra ct

iti on

er , Q

AL Ys

q ua

lit y

ad ju

ste d

lif e

ye ar

s Q

A LY

s A na

ly sis

: G EE

re gr

es si

on m

od el

, w ith

id en

li nk

fu nc

tio n,

G au

ss ia

n va

ria nc

e fu

nc tio

n, e

sti m

at ed

c on

tro lli

ng fo

r t re

at m

en t g

ro up

a nd

b as

el in

e EQ

-5 D

-5 L

C os

t A na

ly sis

: G EE

re gr

es si

on m

od el

, w ith

id en

li nk

fu nc

tio n,

G am

m a

va ria

nc e

fu nc

tio n,

e sti

m at

ed c

on tro

lli ng

fo r t

re at

m en

t g ro

up a

nd b

as el

in e

to ta

l c os

t Pr

ob ab

ili st

ic A

na ly

sis : B

as ed

o n

20 00

si m

ul at

io ns

g en

er at

ed b

y a

tw o-

st ag

e no

n- pa

ra m

et ric

b oo

tst ra

pp in

g to

ol (N

g et

a l.)

a nd

u si

ng th

e in

cr em

en ta

l c os

t a nd

in cr

em en

ta l Q

A LY

e sti

m at

es fr

th e

G EE

re gr

es si

on s

M ul

tip le

I m

pu ta

tio n:

M =

10 . P

re di

ct iv

e m

ea n

m at

ch in

g fo

r EQ

-5 D

-5 L

in de

x sc

or es

, a nd

in di

vi du

al r

es ou

rc e

co st

ite m

s (K

N N

= 5;

V ar

ia bl

es : a

ge , g

en de

r, pr

iv at

e he

al th

in su

ra nc

e, g

en er

m ed

ic al

sc he

m e,

n um

be r o

f m ed

ic at

io ns

a t b

as el

in e,

g en

er al

p ra

ct ic

e ID

)

435Cost effectiveness of a GP delivered medication review to reduce polypharmacy and potentially…

Table 4 Sensitivity and Subgroup Incremental Cost Effectiveness Analysis Results at 6 Months Follow Up

Description Δ Cost Mean Difference Estimate (SE) (p-value) (95% CI)

Δ QALY Mean difference estimate (SE) (p-value) (95% CI)

Probabilistic method Probability that intervention is cost effective: λ = €20,000

Probability that intervention is cost effective: λ = €45,000

Base Case Analy- sis—Complete Case Analysis

− 615.15 (995.25) (0.537) (− 2565.80, 1335.51)

0.005 (0.008) (0.538) (− 0.011, 0.020)

Monte Carlo Simula- tion

0.749 0.777

Base Case Analysis – Multiple Imputation Analysis

− 401.27 (922.23) (0.664) (− 2211.00, 1408.47)

0.014 (0.012) (0.269) (− 0.011, 0.039)

Monte Carlo Simula- tion

Parametric Formula Nonparametric Boot-

strap

0.770 0.994 0.993

0.841 0.986 0.988

Sensitivity Analy- sis—Complete Case Analysis

Unadjusted (i.e. control for treatment arm only)

− 348.08 (996.50) (0.727) (− 2301.18, 1605.02)

0.029 (0.027) (0.272) (− 0.023, 0.082)

Monte Carlo Simula- tion

0.803 0.864

Sensitivity Analysis— Multiple Imputation Analysis

Unadjusted (i.e. control for treatment arm only)

− 359.11 (929.21) (0.699) (− 2181.86, 1463.65)

0.027 (0.020) (0.195) (− 0.014, 0.067)

Monte Carlo Simula- tion

0.821 0.896

Sensitivity Analysis— Multiple Imputation Analysis

Variables included: Age Gender No of Medications, Arm, GP ID

− 529.66 (901.67) (0.557) (− 2298.04, 1238.72)

0.013 (0.011) (0.218) (− 0.008, 0.034)

Parametric Formula Nonparametric Boot-

strap

0.998 0.999

0.991 0.988

Sensitivity Analysis— Multiple Imputation Analysis

Variables included: Age Gender, Arm, GP ID

− 495.70 (896.37) (0.580) (− 2253.61, 1262.22)

0.015 (0.011) (0.156) (− 0.006, 0.037)

Parametric Formula Nonparametric Boot-

strap

0.998 0.998

0.991 0.987

Sensitivity Analysis— Multiple Imputation Analysis

Employ M = 5 imputed datasets

− 278.34 (882.51) (0.752) (− 2008.87, 1452.19)

0.013 (0.015) (0.409) (− 0.021, 0.047)

Parametric Formula Nonparametric Boot-

strap

0.978 0.976

0.970 0.963

Sensitivity Analysis— Multiple Imputation Analysis

Predictive mean match- ing, with nearest neighbour setting: knn = 10

− 183.28 (993.46) (0.854) (− 2136.85, 1770.28)

0.012 (0.012) (0.290) (− 0.011, 0.036)

Parametric Formula Nonparametric Boot-

strap

0.974 0.972

0.968 0.960

Sensitivity Analysis— Multiple Imputation Analysis

Exclude Medication Stopped from Total Healthcare Cost estimate

− 213.35 (926.87) (0.818) (− 2031.55, 1604.84)

0.014 (0.012) (0.269) (− 0.011, 0.039)

Parametric Formula Nonparametric Boot-

strap

0.961 0.961

0.957 0.951

436 P. Gillespie et al.

Table 4 (continued)

Description Δ Cost Mean Difference Estimate (SE) (p-value) (95% CI)

Δ QALY Mean difference estimate (SE) (p-value) (95% CI)

Probabilistic method Probability that intervention is cost effective: λ = €20,000

Probability that intervention is cost effective: λ = €45,000

Sensitivity Analysis— Multiple Imputation Analysis

Exclude Hospitalisa- tion Costs from Total Healthcare Cost estimate

321.91 (159.45) (0.044) (9.26, 634.55)

0.014 (0.012) (0.269) (− 0.011, 0.039)

Parametric Formula Nonparametric Boot-

strap

0.754 0.718

0.900 0.880

Sensitivity Analysis— Multiple Imputation Analysis

Assume 3 Month Time Horizon for Medication Started Costs and Medication Stopped Costs

− 375.57 (897.97) (0.676) (− 2137.95, 1386.81)

0.014 (0.012) (0.269) (− 0.011, 0.039)

Parametric Formula Nonparametric Boot-

strap

0.993 0.993

0.985 0.981

Sensitivity Analy- sis—Complete Case Analysis

Exclude negative values (n = 17) from Total Healthcare Cost estimate

− 598.33 (1024.70) (0.559) (− 2606.71, 1410.05)

0.005 (0.008) (0.538) (− 0.011, 0.020)

Sensitivity Analysis— Multiple Imputation Analysis

Halve SPPiRE Inter- vention Cost

− 531.92 (917.00) (0.562) (− 2331.42, 1267.58)

0.014 (0.012) (0.269) (− 0.011, 0.039)

Parametric Formula Nonparametric Boot-

strap

0.998 0.999

0.991 0.989

Sensitivity Analysis— Multiple Imputation Analysis

Double SPPiRE Inter- vention Cost

− 140.30 (935.26) (0.881) (− 1975.50, 1694.90)

0.014 (0.012) (0.269) (− 0.011, 0.039)

Parametric Formula Nonparametric Boot-

strap

0.968 0.966

0.969 0.962

Sensitivity Analysis— Multiple Imputation Analysis

Unit Costs for Primary and Secondary Care – Deflate by 15%

− 310.07 (793.78) (0.696) (− 1867.65, 1247.51)

0.014 (0.012) (0.269) (− 0.011, 0.039)

Parametric Formula Nonparametric Boot-

strap

0.991 0.992

0.983 0.977

Sensitivity Analysis— Multiple Imputation Analysis

Unit Costs for Primary and Secondary Care – Inflate by 15%

− 492.69 (1051.18) (0.639) (− 2555.56, 1570.18)

0.014 (0.012) (0.269) (− 0.011, 0.039)

Parametric Formula Nonparametric Boot-

strap

0.996 0.999

0.989 0.985

Sensitivity Analysis— Multiple Imputation Analysis

Total Healthcare Cost – GEE Regression— Family/Link = Gauss- ian/Identity

− 171.26 (1005.85) (0.865) (− 2143.93, 1801.41)

0.014 (0.012) (0.269) (− 0.011, 0.039)

Parametric Formula Nonparametric Boot-

strap

0.983 0.982

0.975 0.966

437Cost effectiveness of a GP delivered medication review to reduce polypharmacy and potentially…

of medications started (that is, present at follow up but not baseline: n = 1153), were estimated. This process involved applying medication unit cost data to medication utilisation data, and directly informed by recorded information on the prescribed medication name, strength, dosage and quantity.

Data on medication utilisation were captured directly from general practice records and categorised by the study team using the World Health Organisation (WHO) Anatomical Therapeutic Chemical (ATC) classification system. Medi- cations covered by Ireland's state drug schemes, unlicensed

Table 4 (continued)

Description Δ Cost Mean Difference Estimate (SE) (p-value) (95% CI)

Δ QALY Mean difference estimate (SE) (p-value) (95% CI)

Probabilistic method Probability that intervention is cost effective: λ = €20,000

Probability that intervention is cost effective: λ = €45,000

Sensitivity Analysis— Multiple Imputation Analysis

Total Healthcare Cost – GEE Regres- sion—Family/ Link = Gamma/Log

− 501.84 (1248.16) (0.688) (− 2948.19, 1944.51)

0.014 (0.012) (0.269) (− 0.011, 0.039)

Parametric Formula Nonparametric Boot-

strap

0.961 0.949

0.961 0.948

Sensitivity Analysis— Multiple Imputation Analysis

Total Healthcare Cost – GEE Regres- sion—Family/ Link = Gamma/ Power 1.5

− 565.46 (1228.34) (0.645) (− 2972.96) (1842.04)

0.014 (0.012) (0.269) (− 0.011, 0.039)

Parametric Formula Nonparametric Boot-

strap

0.923 0.917

0.926 0.927

Sensitivity Analy- sis—Complete Case Analysis

Covariate for post COVID 19 Follow Up data collection period included in regression analysis for QALYs and Total Healthcare Cost

− 110.57 (924.79) (0.905) (− 1923.13, 1701.98)

0.005 (0.008) (0.554) ( − 0.011, 0.020)

Sensitivity Analysis— Multiple Imputation Analysis

Covariate for post COVID 19 Follow Up data collection period included in regression analysis for QALYs and Total Healthcare Cost

89.34 (975.31) (0.927) (− 1830.64, 2009.32)

0.014 (0.012) (0.272) (− 0.011, 0.039)

Parametric Formula Nonparametric Boot-

strap

0.861 0.847

0.928 0.925

QALYs Analysis: GEE regression model, with identity link function, Gaussian variance function, estimated controlling for treatment group and baseline EQ-5D-5L Cost Analysis: GEE regression model, with identity link function, Gamma variance function, estimated controlling for treatment group and baseline total cost Probabilistic Analysis: Monte Carlo Simulation—Based on 2000 Monte Carlo simulations assuming normal distribution for incremental cost and QALY estimates from the GEE regression Probabilistic Analysis: Parametric—Based on net benefit method using the incremental net benefit estimates from the independent GEE regression Probabilistic Analysis: Nonparametric Bootstrap: Based on 2000 simulations generated by a two-stage non-parametric bootstrapping tool and using the incremental cost and incremental QALY estimates from the GEE regressions Multiple Imputation: M = 10. Predictive mean matching for EQ-5D-5L index scores, and individual resource cost items (KNN = 5; Variables: age, gender, private health insurance, general medical scheme, number of medications at baseline, general practice ID)

438 P. Gillespie et al.

prescription-only medicines, and non-prescription medi- cines which are covered or are therapeutic alternatives to covered medications, were included in the analysis. Other non-prescription medicines, non-drug products, and high- tech drugs were excluded. High-tech drugs are prescribed by hospital consultants and their recording in GP records are inconsistent. Unit costs for each medication (based on the specified brand or the most common brand of the medication in Ireland) were obtained from the Irish Pharmacy Union Product File. Unit cost data were applied manually to the medication utilisation data by one member of the study team and checked by a second member. As per guidance from Irish National Centre for Pharmacoeconomics a pharmacy dispensing fee was added to each medication ingredient cost [34]. For the purposes of the incremental cost analysis, a fol- low up period of 6 months was assumed for the medication costs, but in a manner that reflected the recorded medication dispensing interval. Alternative follow up period assump- tions were tested in sensitivity analysis.

Third, costs relating to the use of primary and secondary healthcare services over the course of the trial follow up period were estimated for both treatment arms. This included the costs of GP consultations, outpatient clinic visits, acci- dent and emergency department visits, and hospital inpatient admissions Resource use was captured directly from general practice records at baseline and follow up and for a period of 6 months. A vector of unit costs was applied to calculate the cost associated with each resource activity. In sensitivity analysis, the impact of inflating and deflating the unit costs by an arbitrary figure of 15% for the primary and secondary care services was examined.

For the purposes of the incremental analysis, a ‘total healthcare cost at 6 months follow up’ variable was con- structed by aggregating individual resource costs across the follow up period. This comprised of the sum of the costs of the intervention, medications stopped, medications started, and primary and secondary care. A number of alterative total healthcare cost estimates, based on variations in the estima- tion approach for the intervention and medication cost inputs as detailed above, were tested in sensitivity analysis.

For the complete case analysis, estimation of incremental cost was undertaken using a generalized estimating equa- tions (GEE) [35] regression model, controlling for treatment arm, baseline total cost and clustering. To account for the hierarchical and distributional nature of the cost data, an exchangeable correlation structure, a Gamma variance func- tion and identity link function, was assumed [36, 37]. In addition to the complete case analysis, a multiple imputa- tion analysis was undertaken using the MI package in Stata 15 to generate missing values for individual resource use costs at each time point, which were then summed to gener- ate the imputed total healthcare cost variable. The imputa- tion models employed predictive mean matching drawing

from KNN = 5 closest observations, and were estimated using available data on age, gender, treatment arm, num- ber of baseline medications, private health insurance sta- tus, medical card status, and general practice setting [30]. For the multiple imputation analysis, estimation was based on M = 10 completed datasets and Rubin’s rules [31] were employed to combine values and produce the coefficients of interest. Alternative multiple imputation assumptions were tested in sensitivity analysis. The analysis was undertaken using the XTGEE, MI estimate and MI predict commands in Stata 15. The mean cost estimates of interest were obtained from the linear predictions from the regression analysis, using the method of recycled predictions [21].

Effectiveness analysis

Health outcomes were expressed in terms of QALYs gained at 6 months, calculated based on participant responses to the EuroQol [36] EQ-5D-5L instrument, collected via question- naire at baseline and follow up. The EQ-5D-5L consists of five dimensions: mobility, self-care, usual activities, pain or discomfort and anxiety or depression; and each dimension has five levels of severity: no problems, slight problems, moderate problems, severe problems, or unable/extreme problems. In completing the EQ-5D-5L, an individual is located in one of 3125 health state, each of which may be transformed into a health state index score or ‘utility’ using values elicited from the relevant general population. The index score ranges from 1, which is equivalent to perfect health, to 0, which is equivalent to death, and below 0, with negative scores for those health states that are valued as worse than death. The EQ-5D-5L value set scoring algo- rithm for Ireland, which was generated using a hybrid time trade-off and discrete choice experiment approach, produces health utility index scores ranging from − 0.974 to 1 [39]. For economic evaluation, QALYs gained over a period of time are calculated using the area under the curve method, which involves weighting the time spent in EQ-5D-5L health

Fig. 1 SPPiRE Intervention

439Cost effectiveness of a GP delivered medication review to reduce polypharmacy and potentially…

states by their relevant index scores [40]. For the purposes of the incremental analysis, a ‘QALYs gained at 6 months variable’ was constructed using the EQ-5D-5L index scores at baseline and 6 months follow up. The statistical and mul- tiple imputation analysis techniques adopted were similar to those for the cost analysis described above; adopting a GEE regression model estimated controlling for treatment arm, baseline EQ-5D-5L score and clustering, assuming an exchangeable correlation structure, a Gaussian variance function and an identity link function [36].

Cost effectiveness analysis

The net benefit framework, which allows for costs and effects, and their correlation, to be combined into a single variable, enables the identification of the cost effectiveness of a treatment relative to a comparator [21, 41]. In this case, net benefit (nb) is defined as follows:

where eijk is the health outcome for the ith person in the jth cluster in treatment arm k, λ is the cost effectiveness thresh- old value, and cijk is the cost. Applying this framework, a treatment is defined to be cost effective, at a given threshold value, λ, if its corresponding net benefit is greater than that of its comparator: that is, if the incremental net benefit is greater than zero. Point estimates for the mean differences in cost and effectiveness between the alternatives must be estimated, and an explicit examination of the uncertainty surrounding these point estimates conducted. The probabil- istic analysis of uncertainty incorporates both the sampling uncertainty around the mean cost effectiveness estimates and the uncertainty around the true cost effectiveness threshold value [42]. In the Irish context, cost effectiveness thresholds in the range of €20,000 to €45,000 per QALY are gener- ally recommended for decision-making [24], although these are not universally employed. Further, when undertaking analysis using data from cluster RCTs, techniques which recognise the correlation and clustering in the cost and effect data are recommended [27–29].

Given the divergence in missing data patterns between cost and QALY variables, and the need to account for the correlation between both variables, the incremental cost effectiveness findings, both in the form of the net benefits point estimates and the expected cost effectiveness prob- abilities, are presented solely for the multiple imputation analysis. Net benefit statistics at thresholds of €20,000 and €45,000 were generated, and incremental net benefits were estimated using a GEE regression model, controlling for treatment arm and clustering, assuming an exchange- able correlation structure, a Gaussian variance function and an identity link function. The probabilistic analysis was

nbijk = eijk�− cijk,

conducted using a two-stage, non-parametric bootstrapping technique [43], which was based on 2000 bootstrap replica- tions of the linear predictions for total healthcare cost and QALYs. The analysis was undertaken using the XTGEE, MI estimate, MI predict and TSB commands in Stata 15.

Deterministic sensitivity analysis

An extensive range of deterministic sensitivity analysis was conducted to test the robustness of the base-case results to variations in the methods and assumptions employed. The results are presented for comparison purposes in terms of incremental costs, QALYs and cost effectiveness prob- abilities at the €20,000 and €45,000 per QALY thresholds. First, alternative multiple imputation approaches, employ- ing different variables and assumptions were employed. Second, alternative regression model specifications were tested. Third, a number of assumptions relating to the cost- ing methods were varied. Fourth, an analysis was conducted which explicitly accounted for the impact of the COVID-19 pandemic on the study and for the 106 individuals (inter- vention = 57 and control = 49) whose data were followed up after March 2020. Fifth, alternative probabilistic methods for the generation of the cost effectiveness probabilities were tested.

Results

Descriptive statistics for health outcomes, resource use and costs at baseline and follow up are summarised in Table 2. The total cost of implementing the intervention was €53,492, resulting in a mean cost per participant estimate of €257. In terms of total healthcare cost over the 6 month follow up period, the mean cost per patient in the SPPiRE interven- tion arm was €3745 (standard deviation (SD): 6441) and €4115 (SD: 10,319) in the usual care control arm. In terms of health outcomes, mean QALYs gained per patient at 6 months was 0.261 (SD: 0.171) in the SPPiRE intervention arm and 0.235 (SD: 0.167) in the usual care arm. Descriptive statistics for EQ-5D-5L responses are presented in Appendix Table 9. Missing data for the intervention and control arms at baseline and follow up are presented alongside Table 2.

The results from the incremental cost, QALY and cost effectiveness analyses are presented in Table 3 for the com- plete case analysis and for the multiple imputation analy- sis. On average, the SPPiRE intervention was the dominant strategy over usual care. In the complete case analysis, the intervention was associated with a non-statistically signifi- cant mean cost saving of €615 (95% CI − 2566, 1336) and a non-statistically significant mean gain of 0.005 QALYs (95% CI − 0.011, 0.020) relative to the control. In the mul- tiple imputation analysis, the intervention was associated

440 P. Gillespie et al.

with a mean cost saving of €401 (95% CI (− 2211, 1409) and a mean gain of 0.014 QALYs (95% CI − 0.011, 0.039). Univariate analysis comparing the means for individual resource costs and total costs, and EQ-5D-5L and QALYs gained scores are presented in Appendix Table 10. Moving to the incremental cost effectiveness results, the incremental net monetary benefit statistics at the €20,000 and €45,000 thresholds were at estimated €985 (95% CI 217.09, 1752.57) and €1647 (95% CI 184.59, 3108.62) respectively. In terms of expected cost effectiveness based on the available and imputed data, the probability that the SPPiRE intervention is cost effective was 0.993 and 0.988 at threshold values of €20,000 and €45,000 per QALY respectively.

The results from the sensitivity analysis are presented in Table 4 and in Appendix Tables 13 and 14. The results generally confirmed the robustness of the findings from the base case analyses. The results from the analysis for the post COVID-19 data collection cohort revealed stronger evidence in favour of the SPPiRE intervention. In the complete case analysis, the intervention was associated with a statistically significant cost saving of €6084 (95% confidence interval (CI): − 11268, − 901) and a non-statistically significant gain of 0.019 QALYs (95% CI: − 0.017; 0.055) per patient relative to the control (see Appendix Table 13). The cost savings were driven by statistically significant reductions in hospitalisation and emergency department costs in the inter- vention arm within this subgroup (see Appendix Table 14). Finally, employing the Monte Carlo simulation method for the probabilistic analysis resulted in lower probability esti- mates of 0.770 and 0.841 at threshold values of €20,000 and €45,000 respectively.

Discussion

This paper reports the findings from a within trial economic evaluation which observed a pattern towards dominance for the SPPiRE intervention over usual care for patients with multimorbidity in Irish general practice. This potentially resulted from an accumulation of multiple, small, positive, albeit statistically insignificant intervention impacts. That is, cost savings, arising from reduced hospital services utilisa- tion which went to offset the intervention implementation costs, and health gains, contributed to the dominant cost effectiveness point estimates. Notably, uncertainty in the analysis, and particularly the issue relating to postal ques- tionnaire non-response and resulting missing data at follow up, were non trivial factors, and should be carefully consid- ered when interpreting our findings.

The incremental costs and QALYs estimates were not individually statistically significant in the complete case or multiple imputation analysis, and were therefore consistent with chance findings. That said, the SPPiRE RCT was not

powered to specifically detect differences in costs or QALYs. Indeed, trial-based economic evaluation is often faced with inappropriate sample size constraints [44]; thereby raising the possibility that important differences between treatment arms cannot be detected at conventional levels of power and significance. To address this concern, it is recommended that the evidence should be presented in the form of expected cost effectiveness probabilities, rather than by relying solely on showing significance at conventional levels [44]. In this case, we report the estimated probabilities for the SPPiRE intervention and find them to be in the region of 90% across a range of potential cost effectiveness threshold values for Ireland, and this remained consistent in a series of sensitiv- ity analysis.

Importantly however, given the extent of the missing data on health outcomes, the expected cost effectiveness results were based on data generated from the multiple imputation analysis. While missing data did not appear to be systemati- cally different in nature across treatment arms, it was sub- stantial in totality, with only 57% of postal questionnaire data available for analysis. Further, the observed pattern of results for the patient cohort with data collection occurring post the onset of COVID-19 poses additional questions that require further scrutiny and analysis. While this may be a chance finding, it raises legitimate concerns regarding the interpretation of our expected cost effectiveness results, and whether or not they should be used for healthcare resource allocation decisions.

Taken all of the above together, we conclude that the evi- dence for the cost effectiveness of the SPPiRE intervention should be considered promising but equivocal. That said, it is ultimately the remit of decision makers to determine whether the level of evidence presented is sufficient to justify the adoption of the SPPiRE intervention in clinical practice. These findings supplement those from the parallel clinical effectiveness study which found that the SPPiRE interven- tion resulted in a statistically significant, but small reduction in the number of medicines and in a weakly significant effect on PIP [20]. Our findings also reflect those from the existing evidence base for the cost effectiveness of interventions tar- geting multimorbidity and polypharmacy, and support calls for further studies to explore the health and economic impli- cations of interventions targeting the multimorbid patient population [16–19].

Strengths and limitations

This study had a number of strengths and limitations. The trial recruited to target a vulnerable group of patients with substantial disease and treatment burden and a high base- line prevalence of potentially inappropriate prescribing, and the SPPiRE intervention was both safe and feasible [24]. There were a number of limitations relating to the conduct

441Cost effectiveness of a GP delivered medication review to reduce polypharmacy and potentially…

Table 5 SPPiRE prescriptions and potentially inappropriate prescribing (PIP) criteria

Drug group PIP Reason

Drug groups frequently associated with preventable drug related morbidity NSAIDS with diuretic and ACEi/ARB (1) Risk of renal impairment

with chronic kidney disease (eGFR < 50) (1, 2)

for ≥ 12 weeks with no gastroprotection (1) Risk of GI bleed that is not COX 2 selective, with a history of

PUD with no gastroprotection (2) and antiplatelet with no gastroprotection (2) with an anticoagulant (2, 3) with severe hypertension or heart failure (2) Risk of hypertension/ heart failure exacerbation COX-2 selective with concurrent cardiovascu-

lar disease (2) Increased risk of MI/CVA

Antiplatelets and history of PUD with no gastroprotection (1, 3)

Risk of GI bleed

and anticoagulant with no gastroprotection (1, 3)

dual antiplatelet therapy with no gastroprotec- tion (1)

consider intended duration of treatment if taking dual anti-platelet therapy for over one year post PCI (2)

Not usually indicated

Anticoagulants for first uncomplicated DVT for > 6 months duration (2)

Not indicated

for first uncomplicated PE for > 12 months duration (2)

dabigatran (Pradaxa) if eGFR < 30 ml/min/ 1.73m2 or if renal function is unknown (2)

Risk of bleeding

rivaroxaban (Xarelto)or apixaban (Eliquis) if eGFR < 15 ml/min/ 1.73m2 or if renal func- tion is unknown (2)

Diuretics and no U&E check in the last 48 weeks (1) Risk of renal impairment and electrolyte abnormalityloop diuretic and thiazide diuretic and no

U&E in the last 24 weeks (1) loop diuretic for dependent oedema and

no heart failure, liver failure or nephrotic syndrome (2)

Risks usually out-weigh benefits

thiazide diuretic with a history of gout (2) Risk of precipitating gout Drugs groups associated with morbidity in the elderly Anticholinergic drugs With comorbidities (3) Exacerbation of co-morbidity

Dementia Narrow angle glaucoma Cardiac conduction abnormalities Chronic prostatism Concomitant use of two or more drugs with

anticholinergic properties (2) Risk of anticholinergic toxicity

tricyclic antidepressant as first line antidepres- sant (2)

Increased risk of adverse effects in older patients and alternatives available

antimuscarinic antihistamine (2) Benzodiazepines OR Z drugs for longer than 4 weeks (2) (1) Risk of sedation, confusion, impaired balance,

falls NNT 13 and NNH 6 when used for insomnia

(4)

442 P. Gillespie et al.

of the RCT, as outlined in the main trial publication, which also apply to the economic evaluation. For example, only a quarter of invited patients agreed to participate, although no other intervention study with this degree of polyphar- macy as an inclusion criterion could be identified for com- parison. Further, a chance imbalance in the number of days from baseline to follow-up between groups was identified.

However, a sensitivity analysis including the number of days to follow-up as a covariate revealed that there was no signifi- cant effect on the results.

The sample size of the trial was based on the clinical primary endpoint and may have been insufficient to detect statistically significant changes in the health economic out- comes. Further, outcome measures were assessed at just

Abbreviations: NSAID non-steroidal anti-inflammatory drug, ACEi angiotensin converting enzyme inhibitor, ARB aldosterone receptor blocker, eGFR estimated glomerular filtration rate, PUD peptic ulcer disease, GI gastro-intestinal, MI myocardial infarction, CVA cerebrovascular acci- dent, COX-2 cyclooxygenase-2, DVT deep vein thrombosis, PCI percutaneous coronary intervention, PE pulmonary embolism, NNT number needed to treat, NNH number needed to harm 1. Dreischulte T, Grant AM, McCowan C, McAnaw JJ, Guthrie B. Quality and safety of medication use in primary care: consensus validation of a new set of explicit medication assessment criteria and prioritisation of topics for improvement. BMC clinical pharmacology. 2012;12:5 2. O'Mahony D, O'Sullivan D, Byrne S, O'Connor MN, Ryan C, Gallagher P. STOPP/START criteria for potentially inappropriate prescribing in older people: version 2. Age and ageing. 2015;44(2):213–8 3. Clyne B, Bradley MC, Hughes CM, Clear D, McDonnell R, Williams D, et al. Addressing potentially inappropriate prescribing in older patients: development and pilot study of an intervention in primary care (the OPTI-SCRIPT study). BMC health services research. 2013;13:307 4. Glass J, Lanctot KL, Herrmann N, Sproule BA, Busto UE. Sedative hypnotics in older people with insomnia: meta-analysis of risks and ben- efits. Bmj. 2005;331(7526):1169 5. Ballard CG, Waite J, Birks J. Atypical antipsychotics for aggression and psychosis in Alzheimer's disease. Cochrane Database of Systematic Reviews. 2006(1)

Table 5 (continued)

Drug group PIP Reason

Antipsychotics with dementia and no psychosis (1, 2) Increased risk of stroke, only use when all other means have failed and shortest possible dose for shortest duration (5)

Miscellaneous drug groups; included because of prevalence or high risk Methotrexate not prescribed as weekly (1) Increased risk of potentially fatal medication

errorsprescribed > 1 strength tablet (1) Opioids used regularly with no laxative (2) Risk of severe constipation Corticosteroids use ≥ 12 weeks with no bone protection (2) Risk of fracture PPI for uncomplicated PUD/erosive pep-

tic oesophagitis at full therapeutic dose ≥ 8 weeks (2)

Not indicated

Metformin with eGFR < 30 ml/min/ 1.73m2 (2) Risk of lactic acidosis

Table 6 Lost to Follow Up Analysis 1: Comparison of participants lost to follow up and followed up

Abbreviations: I intervention, C control, SD standard deviation, PIP potentially inappropriate prescription, EQVAS EQ – 5D visual analogue scale, GMS general medical services

Characteristic All participants Lost to follow up Followed up

Intervention (N = 208)

Control (N = 196)

Intervention (N = 16)

Control (N = 19)

Intervention (N = 192)

Control (N = 177)

Mean age (SD) 76.67 (6.80) 76.33 (6.88) 81.93 (7.57) 79.78 (7.35) 76.24 (6.57) 75.95 (6.74) % Female 57.21 57.14 37.50 78.95 58.85 54.80 Mean number medicines at

baseline (SD) 16.96 (3.25) 17.82 (3.71) 18.06 (3.99) 18.74 (4.82) 16.87 (3.17) 17.72 (3.57)

Mean PIP baseline (SD) 2.50 (1.53) 2.53 (1.40) 2.69 (1.49) 2.63 (1.64) 2.47 (1.52) 2.55 (1.43) % with ≥ 1 PIP 93.24 92.82 93.75 93.18 93.19 88.69 Mean EQVAS (SD) 59.63 (20.24) 59.75 (22.09) 48.67 (15.75) 58.42 (19.44) 60.54 (20.34) 59.90 (22.45) % with GMS card 82.00 89.78 73.33 100.00 82.70 88.69

443Cost effectiveness of a GP delivered medication review to reduce polypharmacy and potentially…

one-time point, 6 months after intervention completion. There is a possibility that the full effect of the interven- tion may not have been captured by assessing outcomes at just one point in time. Significantly, only 57% of patients reported patient-reported outcome measures at follow-up, which the QALYs gained variable was based upon. Missing data was therefore an important and significant consideration and details on missing data at each data collection point are presented. After consideration of missing data patterns, we proceeded with the assumption that the data were missing at random and multiple imputation was undertaken to impute missing values using the MI command in Stata 15. The vari- ables included in the imputation models were pragmatically chosen by the study team. This approach may be criticised on the basis that values for resource cost and EQ-5D-5L scores were imputed independently. Furthermore, general practice surgery was included as a fixed effect in the impu- tation model, reflecting recent guidance that the imputation model should be compatible with the analysis model: that is, both should reflect the multilevel nature of the data [30]. The

approach of including the cluster variable as a fixed effect in the imputation model may be problematic in some cases [45]; however, we deemed it to be appropriate. Finally, given the concerns raised above, it may be argued that the multiple imputation methods adopted have produced artificially low estimates of uncertainty in this case.

In terms of the health economic evaluation, the time horizon of the economic evaluation was limited to the trial follow-up period of 6 months; thereby excluding costs and benefits that arise beyond 6 months and over the remainder of the patients’ lifetime. This is likely be particularly rel- evant in the context of chronic disease, for which short term interventions may have long term implications. However, the concept of modelling long term health outcomes and costs for multimorbidity is an important area of future study.

While the cost analysis was conducted from the health service perspective and included an extensive range of resource use activities, certain resource items were not captured. For example, other medications costs to those started and stopped, community care costs, private

Table 7 Lost to follow up analysis 1: comparison of participants with and without patient questionnaire data at follow up

Abbreviations: SD standard deviation, PIP potentially inappropriate prescription, SEG socio-economic group

Characteristic Patient questionnaire data at follow up n = 230 (57%)

No patient questionnaire data at follow up n = 174 (43%)

Intervention n = 118 (56.7%)

Control n = 111 (56.9%)

Intervention n = 90 (43.3%)

Control n = 84 (43.1%)

Male (%) 57 (48.3) 53 (47.7) 32 (35.6) 30 (35.7) Female (%) 61 (51.7) 58 (52.3) 58 (64.4) 54 (64.3) SEG 1 or 2 (%) 30 (25.4) 21(18.9) 21 (23.3) 15 (17.9) Other SEG (%) 88 (74.6) 89 (80.1) 68 (75.6) 68 (81.0) Mean age (SD) 76.30 (6.24) 75.30 (6.87) 77.15 (7.47) 77.70 (6.74) Mean no. meds (SD) 17.00 (3.14) 17.72 (3.49) 16.91 (3.41) 17.98 (4.00) Mean no. PIP (SD) 2.49 (1.54) 2.49 (1.41) 2.48 (1.50) 2.64 (1.51)

Table 8 Unit cost estimates in 2022 € prices

Unit costs inflated using the health component of the consumer price index (CPI) from the Irish Central Statistics Office (CSO)

Healthcare Resource Item Details Unit Cost € Unit Cost Estimate Source

SPPiRE Intervention (i.e. SPPiRE videos, educational ses-

sions and materials; SPPiRE medication reviews)

Total Per Patient

53,492 257

Study Records

Prescription Medications Per Drug Dose n/a Primary Care Reimburse- ment Service (PCRS)

Other Healthcare Services General Practitioner (GP) Per Visit 53 (Smith et al., 2021) Practice Nurse (PN) Per Visit 42 Study Records Emergency Department Per Visit 297 Hospital Pricing Office Outpatient Clinic Per Visit 144 Hospital Pricing Office Hospital Inpatient Nights Per Night 988 Hospital Pricing Office

444 P. Gillespie et al.

out-of-pocket patient costs such as private health insurance premiums, and broader costs to society such as produc- tivity losses were not captured in the analysis. Nonethe- less, there is little evidence to suggest that the inclusion of these resources categories would fundamentally change the results presented. The resource utilisation data was collected directly from practice records and was compiled to a high standard of completeness and precision. Given the volume of prescription data, we assumed that follow up period of 6 months for all medication costings. In addition, the calculation approach of the total health care cost vari- able is open to criticism and it resulted in 17 participants having negative costs, since the savings from stopped medications outweighed their other cost outlays. The sen- sitivity analysis indicated that these assumptions had no bearing on the overall results. The process of conducting cost analysis in Ireland is also compromised by the lack of nationally available unit cost data. In estimating unit

costs for individual resource activities, we endeavoured at all times to be conservative in any assumptions adopted.

We employed appropriate methods for the statisti- cal analysis of cost and effect data collected alongside cluster RCTs. To account for potential covariate imbal- ances between treatment arms at baseline [28], we esti- mated separate multilevel regression models for costs and QALYs, controlling for baseline costs and health outcome covariates. To jointly account for correlation and clustering, we adopted a two-stage non-parametric bootstrapping technique [43]. While the methods adopted were appropriate, arguments could be made for a num- ber of alternative approaches. For comparative purposes, probabilistic results for the complete case analysis from the Monte Carlo simulation approach were presented, and generated lower probability estimates than the nonpara- metric and parametric methods. Notably, this method does not account for both clustering and correlation as per the

Table 9 Summary data for EQ-5D-5L health outcome at baseline and follow up

Dimension Level SPPiRE intervention Control

Baseline: N / % Follow up: N / % Baseline: N / % Follow up: N / %

Mobility 197 112 187 110 None 14.21 21.43 17.65 10.91 Slight 24.37 16.96 18.72 21.82 Moderate 31.98 34.82 35.29 33.64 Severe 25.38 25.00 24.60 29.09 Unable 4.06 1.79 3.74 4.55

Self-care 197 115 186 108 None 54.82 58.26 57.53 54.63 Slight 18.27 16.52 17.20 18.52 Moderate 17.26 17.39 16.67 20.37 Severe 5.58 6.96 5.91 4.63 Unable 4.06 0.87 2.69 1.85

Usual activities 198 113 186 109 None 23.74 23.01 23.66 19.27 Slight 22.73 22.12 25.81 18.35 Moderate 30.30 27.43 20.97 33.03 Severe 11.11 19.47 20.97 22.94 Unable 12.12 7.96 8.60 6.42

Pain/discomfort 197 114 186 110 None 7.11 5.26 10.75 5.45 Slight 25.89 22.81 16.67 48.18 Moderate 35.53 43.86 41.94 48.18 Severe 26.40 23.68 24.19 25.45 Extreme 5.08 4.39 6.45 6.36

Anxiety/depression 194 109 183 105 None 44.33 44.04 38.25 39.05 Slight 28.87 29.36 26.78 27.62 Moderate 23.20 19.27 27.32 27.62 Severe 1.55 7.34 6.01 3.81 Extreme 2.06 0.00 1.64 1.90

445Cost effectiveness of a GP delivered medication review to reduce polypharmacy and potentially…

recommendations, and may be less precise in quantifying uncertainty in cost effectiveness results [46].

In conclusion, due to ageing population dynamics, the provision of safe, effective, equitable and efficient health- care services for those with complex multimorbidity will

become an ever pressing challenge. Given the equivocal nature of the cost effectiveness results presented, imple- mentation of the SPPiRE intervention cannot be recom- mended. Further research is needed to support health pol- icy and healthcare decision makers to address the issue

Table 10 Incremental analysis of healthcare costs and health outcome variables at 6 months follow up

Estimates for mean differences generated using GEE regression models controlling for clustering and treatment arm

Healthcare Resource Items Intervention Control Univariate Mean Difference Estimate

6 months Follow Up—Mean (SD) / % Beta Coefficient (SE) (p-value) (95% CI)

Usage Cost € Usage Cost €

Intervention SPPiRE 100% 257 0% 0 257 (n/a) (n/a) (n/a) Medication Prescriptions Number of medicines stopped 3.97 (3.15) 598.80 (1063.18) 2.92 (3.17) 448.29 (1026.34) 99.08 (133.10) (0.457) (− 161.81,

359.97) Number of medicines started 3.02 (3.03) 487.01 (763.99) 2.67 (2.91) 519.98 (1443.07) 50.58 (106.32) (0.634) (− 157.80,

258.96) Other Healthcare Services GP Consultations 4.42 (3.51) 234.32 (186.10) 3.84 (3.27) 203.32 (173.39) 24.17 (26.77) (0.367) (− 28.29,

76.63) GP Phone Consultations 1.55 (2.12) 82.29 (112.33) 1.42 (2.21) 75.14 (117.27) 11.82 (20.11) (0.590) (− 27.89,

51.24) GP House Call Consultations 0.20 (0.83) 10.60 (44.02) 0.13 (0.68) 6.78 (36.07) 3.35 (5.05) (0.508) (− 6.55, 13.24) GP Out of Hours Consultations 0.33 (1.15) 17.29 (60.68) 0.16 (0.44) 8.68 (23.50) 9.00 (5.57) (0.106) (− 1.92, 19.91) GP Prescription Consultations 2.51 (2.55) 132.78 (134.96) 2.40 (2.58) 127.08 (136.46) 4.89 (25.19) (0.190) (− 44.49,

54.27) Practice Nurse Consultations 1.76(1.98) 74.00 (83.12) 1.90 (2.96) 79.83 (124.40) − 2.13 (20.16) (0.916) (− 41.64,

37.38) Outpatient Clinic Visits 2.62(5.54) 458.99 (971.62) 2.39 (2.29) 419.06 (400.95) 38.50 (87.93) (0.661) (− 133.83,

210.83) Hospital Inpatient Nights 2.43(6.19) 2399.43 (6110.99) 3.09 (9.81) 3052.21 (9695.97) − 645.11 (869.85) (0.458)

(2349.98,1059.75) Emergency Department Visits 0.46 (1.01) 140.37 (309.91) 0.33 (0.85) 102.18 (259.69) 34.76 (38.34) (0.365) (− 40.38,

109.90) Total Healthcare Cost 3744.80 (6440.69) 4137.97 (10,347.80) − 368.81 (997.16) (0.711)

(− 2323.21, 1585.59) Health Outcomes EQ-5D-5L Index Score Baseline 0.496 (0.362) 0.471 (0.383) 0.031 (0.047) (0.514) (− 0.061,

0.123) EQ-5D-5L Index Score Follow

UP 0.517 (0.382) 0.452 (0.357) 0.068 (0.054) (0.207) (− 0.038,

0.174) QALYs Gained 0.261 (0.171) 0.234 (0.167) 0.031 (0.027) (0.254) (− 0.022,

0.083)

Table 11 Detailed descriptive statistics for total healthcare cost and QALYs gained dependent variables

Variable N Mean Median Standard Deviation IQR (Q1-Q3) Min Max Skew Kurtosis ICC

QALYs gained 196 0.248 0.290 0.169 0.145 to 0.376 − 0.201 0.500 − 0.836 2.975 0.044 Total Healthcare Cost 343 3925.91 1090.46 8456.46 669.62 to 3007.58 − 13,747.64 89,276.93 5.140 41.037 0.031

446 P. Gillespie et al.

of excess polypharmacy in multimorbidity. Future studies should carefully consider the design of data collection methods for patient reported outcomes among patients with multimorbidity.

Appendix

See Fig. 1.

SPPiRE Intervention—Proctor implementation strategy

1. Specify and operationalize implementation strategies:

• Training videos were created to explain the importance of the topic and the approach for the medication review. A training manual for GPs provided similar information in written format, including background evidence for the inclusion of the relevant potentially inappropriate prescriptions (PIPs).

2. Tailor strategies to context:

• The study manager tracked the progress of reviews by checking data on the SPPiRE website. Practices that were behind schedule were contacted to encourage them to perform the reviews and to offer further infor- mation or support if needed. Practices had the flexibil- ity to adopt scheduling strategies that suited their spe- cific context. Some practices performed opportunistic reviews instead of scheduled ones, this was not part of the original intervention plan. Modifications included conducting phone reviews in response to COVID-19, although this only affected one practice.

3. Engage stakeholders:

• The trial management committee (TMC) included GP input, and the study manager was a GP, ensuring aware- ness of the context. The trial steering committee (TSC) had patient and public involvement (PPI) input to ensure that patient-facing materials were appropriate.

Table 12 Data to inform choice of multilevel generalised estimation equation (GEE) regression models

Dependent variable QALYs gained Total healthcare cost

Family/link Gaussian/identity Gaussian/iden Modified park test coefficient − 0.737709 3.558763 Modified Park Test Chi_Squared Test—p-value 0.0187 0.1244 Pearson Correlation Test—p-value 1.0000 1.0000 Pregibon Link Test—p-value 0.0120 0.2579 Modified Hosmer and Lemeshow—p-value 0.9646 0.8346 Family/Link Gamma/Log Modified Park Test Coefficient 2.427507 Modified Park Test Chi_Squared Test—p-value 0.8110 Pearson Correlation Test—p-value 0.0008 Pregibon Link Test—p-value 0.1040 Modified Hosmer and Lemeshow—p-value 0.9057 Family/Link Gamma/Iden Modified Park Test Coefficient 1.466235 Modified Park Test Chi_Squared Test—p-value 0.5495 Pearson Correlation Test—p-value 0.0303 Pregibon Link Test—p-value 0.1078 Modified Hosmer and Lemeshow—p-value 0.9684 Family/Link Gamma/1.25 Modified Park Test Coefficient 1.406432 Modified Park Test Chi_Squared Test—p-value 0.4896 Pearson Correlation Test—p-value 0.0611 Pregibon Link Test—p-value 0.1209 Modified Hosmer and Lemeshow—p-value 0.9547

447Cost effectiveness of a GP delivered medication review to reduce polypharmacy and potentially…

4. Training and education:

• The training videos and manuals served as the primary educational tools for GPs, covering both the importance of the intervention and the practical steps for conducting medication reviews.

5. Implementation planning:

• A trial Gantt chart was used for planning, but the inter- vention period had to be extended due to slow progress. Adjustments were made in consultation with the TMC and TSC.

6. Monitor and evaluate implementation:

• A parallel mixed methods process evaluation was con- ducted, including quantitative data from website usage and qualitative data from semi-structured interviews with a purposive sample of intervention GPs and patients.

7. Continuous quality improvement:

• Although not directly related to implementation strat- egies, safety data were collected on an ongoing basis. GPs were given a process for reporting any adverse drug withdrawal events (ADEs).

8. Sustainability planning:

• There were no specific strategies planned or implemented for sustainability beyond the intervention period.

SPPiRE Intervention—TIDieR Checklist for Intervention

1. Brief name:

• SPPiRE

Table 13 GEE regression incremental analysis at 6 months follow up – complete case analysis – subgroup analysis –post-COVID follow up group only

QALYs Analysis: GEE regression model, with iden link function, Gaussian variance function, estimated controlling for treatment group and baseline EQ-5D-5L Cost Analysis: GEE regression model, with iden link function, Gamma variance function, estimated controlling for treatment group and base- line total cost Probabilistic Analysis: Based on 2000 simulations from Monte Carlo Simulation assuming normal distributions for the incremental cost esti- mate and incremental qaly estimate from the independent GEE regressions Missing Data: Intervention: 52 or 91% for Total Cost variables, 44 or 57% for baseline EQ-5D-5L index score, 31 or 60% for follow up EQ-5D-5L index score, and 25 or 44% for QALYs gained Control: 39 or 80% for Total Cost variables, 44 or 88% for baseline EQ-5D-5L index score, 25 or 51% for follow up EQ-5D-5L index score, and 24 or 49% for QALYs gained

Variable/analysis Incremental analysis (Intervention minus control)

Treatment Arm Intervention Control N 57 49

Incremental analysis (Intervention minus Control)

Total Healthcare Cost Analysis € Beta Coefficient (SE) (95% CI) (p-value) N = 89 QIC = 6291.999

-6084.41 (2644.88) (− 11,268.28, − 900.54) (0.021)

QALYs gained Beta Coefficient (SE) (95% CI) (p-value) N = 49 QIC = 8.595

0.019 (0.019) (− 0.017, 0.055) (0.308)

Probability (%) that the SPPiRE Intervention is Cost Effective for Threshold Value (λ) tc_f λ = €0 λ = €10,000 λ = €20,000 λ = €30,000 λ = €40,000 λ = €45,000 0.991 0.992 0.994 0.995 0.996 0.997 λ = €50,000 λ = €60,000 λ = €70,000 λ = €80,000 λ = €90,000 λ = €100,000 0.997 0.996 0.997 0.997 0.997 0.997

448 P. Gillespie et al.

Ta bl

e 14

st- CO

V ID

fo llo

w u

p an

al ys

H ea

lth ca

re R

es ou

rc e

It em

s Fo

llo w

U p—

6 m

on th

s— M

ea n

(S D

) / %

In te

rv en

tio n

N =

57 C

on tro

l N

= 49

M ea

n di

ffe re

nc e

C os

t € C

os t €

p- va

lu e

(9 5%

C I)

In te

rv en

tio n

SP Pi

R E

25 7

0 n/

a M

ed ic

at io

n Pr

es cr

ip tio

ns N

um be

r o f m

ed ic

in es

st op

pe d

49 9.

38 (1

22 9.

52 )

63 6.

36 1

88 5.

88 )

0. 02

5 (5

6. 67

,8 38

.1 2)

N um

be r o

f m ed

ic in

es st

ar te

d 67

0. 59

(8 22

.2 9)

28 0.

75 (3

43 .6

8) 0.

69 1

(- 83

.7 0,

51 9.

10 )

O th

er H

ea lth

ca re

S er

vi ce

s G

P C

on su

lta tio

ns 15

9. 98

(1 39

.0 8)

15 0.

37 (1

61 .0

6) 0.

86 5

(- 10

1. 60

, 8 5.

35 )

G P

Ph on

e C

on su

lta tio

ns 13

1. 52

(1 54

.3 5)

15 2.

69 (1

53 .1

6) 0.

76 4

(- 93

.7 3,

6 8.

86 )

G P

H ou

se C

al l C

on su

lta tio

ns 11

.7 8

(3 5.

17 )

1. 23

(8 .0

8) 0.

16 1

(- 4.

03 , 2

4. 23

) G

P O

ut o

f H ou

rs C

on su

lta tio

ns 2.

94 (1

2. 25

) 7.

40 (2

1. 89

) 0.

00 8

(- 11

.1 3,

-1 .7

1) G

P Pr

es cr

ip tio

n C

on su

lta tio

ns 13

9. 37

(1 43

.8 8)

12 5.

72 (9

8. 15

) 0.

39 8

(- 49

.8 9,

1 25

.4 9)

Pr ac

tic e

N ur

se C

on su

lta tio

ns 52

.1 1

(6 6.

25 )

18 .5

6 (3

5. 85

) 0.

00 1

(1 3.

51 , 5

1. 31

) O

ut pa

tie nt

C lin

ic V

is its

31 5.

20 (3

20 .4

5) 43

6. 64

(3 68

.0 1)

0. 17

0 (-

28 0.

09 , 4

9. 25

) H

os pi

ta l I

np at

ie nt

N ig

ht s

10 81

.2 1

(4 34

2. 47

) 64

79 .4

4 (1

6, 80

8. 87

) 0.

00 4

(- 88

27 .1

1, -1

68 3.

69 )

Em er

ge nc

y D

ep ar

tm en

t V is

its 34

.0 6

(1 14

.0 3)

13 5.

45 (2

04 .0

3) 0.

00 0

( - 15

9. 00

, - 71

.7 3)

To ta

l H ea

lth ca

re C

os t

20 59

.9 7

(4 78

4. 11

) 81

46 .2

7 (1

7, 89

9. 62

) H

ea lth

O ut

co m

es EQ

-5 D

-5 L

In de

x Sc

or e

B as

el in

e 0.

47 8

(0 .3

31 )

0. 41

4 (0

.3 75

) 0.

27 4

(- 0.

05 9,

0 .2

07 )

EQ -5

D -5

L In

de x

Sc or

e Fo

llo w

U p

0. 51

8 (0

.3 42

) 0.

44 2

(0 .3

58 )

0. 50

3 (-

0. 14

0, 0

.2 84

) Q

A LY

s G ai

ne d

0. 24

1 (0

.1 61

) 0.

22 8

(0 .1

83 )

Va ri

ab le

/ A na

ly sis

In cr

em en

ta l A

na ly

sis (I

nt er

ve nt

io n

m in

us C

on tro

l) Tr

ea tm

en t A

rm In

te rv

en tio

n C

on tr

ol N

57 49

In cr

em en

ta l A

na ly

sis (I

nt er

ve nt

io n

m in

us C

on tro

l) To

ta l H

ea lth

ca re

C os

t A na

ly sis

€ B

et a

C oe

ffi ci

en t (

SE ) (

95 %

C I)

(p -v

al ue

) N

= 89

Q IC

= 62

91 .9

99 –6

08 4.

41 (2

64 4.

88 ) (

– 11

,2 68

.2 8,

– 90

0. 54

) ( 0.

02 1)

Q A

LY s g

ai ne

d B

et a

C oe

ffi ci

en t (

SE ) (

95 %

C I)

(p -v

al ue

) N

= 49

Q IC

= 8.

59 5

0. 01

9 (0

.0 19

) ( –

0. 01

7, 0

.0 55

) ( 0.

30 8)

Pr ob

ab ili

ty (%

) t ha

t t he

S PP

iR E

In te

rv en

tio n

is Co

st Eff

ec tiv

e fo

r T hr

es ho

ld V

al ue

(λ )

λ = €0

λ = €1

0, 00

0 λ =

€2 0,

00 0

λ = €3

0, 00

0 λ =

€4 0,

00 0

λ = €4

5, 00

0 0.

99 1

0. 99

2 0.

99 4

0. 99

5 0.

99 6

0. 99

449Cost effectiveness of a GP delivered medication review to reduce polypharmacy and potentially…

2. Why:

• To improve the quality and safety of prescribing and reduce treatment burden for patients with complex mul- timorbidity (defined as ≥ 15 repeat medicines).

3. What (materials):

• Web platform: Guided GPs through the medication review process.

• Training materials: Videos and manuals provided to GPs for preparation.

• Patient materials: Instructions for patients to bring all their medications ("brown bag").

4. What (procedures):

• Session length: 30 min, conducted once. • Components of the review:

1. Check for PIP: The GP identified any potentially inappropriate prescriptions.

2. Brown bag review: The GP and patient reviewed each medication to check for actual drug utilisation, side effects, and effectiveness.

3. Discuss treatment priorities: The GP asked patients about their current treatment priorities.

4. Shared decision-making: The GP recorded all data and worked with the patient to reach a shared deci- sion on any medication changes.

5. Who provided:

• General practitioners (GPs) in Irish general practice set- tings.

6. How:

• The intervention was delivered face-to-face between the GP and the patient, guided by a web-based platform.

7. Where:

• Conducted in primary care settings across Ireland.

8. When and how much:

• The intervention consisted of a single 30-min session.

9. Tailoring:

• The review length was suggested but not strictly instructed, and most sessions took longer in practice. M

iss in

g D

at a:

In te

rv en

tio n:

5 2

or 9

1% fo

r T ot

al C

os t v

ar ia

bl es

, 4 4

or 5

7% fo

r b as

el in

e EQ

-5 D

-5 L

in de

x sc

or e,

3 1

or 6

0% fo

r f ol

lo w

u p

EQ -5

D -5

L in

de x

sc or

e, a

nd 2

5 or

4 4%

fo r Q

A LY

s g ai

ne d

C on

tr ol

: 3 9

or 8

0% fo

r T ot

al C

os t v

ar ia

bl es

, 4 4

or 8

8% fo

r b as

el in

e EQ

-5 D

-5 L

in de

x sc

or e,

2 5

or 5

1% fo

r f ol

lo w

u p

EQ -5

D -5

L in

de x

sc or

e, a

nd 2

4 or

4 9%

fo r Q

A LY

s g ai

ne d

Ta bl

e 14

( co

nt in

ue d)

H ea

lth ca

re R

es ou

rc e

It em

s Fo

llo w

U p—

6 m

on th

s— M

ea n

(S D

) / %

In te

rv en

tio n

N =

57 C

on tro

l N

= 49

M ea

n di

ffe re

nc e

C os

t € C

os t €

p- va

lu e

(9 5%

C I)

λ = €5

0, 00

0 λ =

€6 0,

00 0

λ = €7

0, 00

0 λ =

€8 0,

00 0

λ = €9

0, 00

0 λ =

€1 00

,0 00

0. 99

7 0.

99 6

0. 99

7 0.

99 7

0. 99

7 0.

99 7

450 P. Gillespie et al.

Ta bl

e 15

H EE

R S

20 22

c he

ck lis

To pi

c N

o Ite

m Lo

ca tio

n w

he re

it em

is re

po rte

Ti tle

1 Id

en tif

y th

e stu

dy a

s a n

ec on

om ic

e va

lu at

io n

an d

sp ec

ify th

e in

te rv

en tio

ns b

ei ng

c om

pa re

d Ye

s, th

e tit

le o

f t he

p ap

er c

le ar

ly se

ts o

ut th

is in

fo rm

at io

A bs

tra ct

2 Pr

ov id

e a

str uc

tu re

d su

m m

ar y

th at

h ig

hl ig

ht s c

on te

xt , k

m et

ho ds

, r es

ul ts

, a nd

a lte

rn at

iv e

an al

ys es

Ye s,

th e

ab str

ac t p

ro vi

de s t

hi s i

nf or

m at

io n

In tro

du ct

io n

B ac

kg ro

un d

an d

ob je

ct iv

es 3

G iv

e th

e co

nt ex

t f or

th e

stu dy

, t he

st ud

y qu

es tio

n, a

its p

ra ct

ic al

re le

va nc

e fo

r d ec

is io

n m

ak in

g in

p ol

ic y

pr ac

tic e

Ye s,

th is

is o

ut lin

ed in

th e

B ac

kg ro

un d

se ct

io n

of th

e pa

pe r

M et

ho ds

H ea

lth e

co no

m ic

a na

ly si

s p la

n 4

In di

ca te

w he

th er

a h

ea lth

e co

no m

ic a

na ly

si s p

la n

w as

d ev

el -

op ed

a nd

w he

re av

ai la

bl e

Th e

he al

th e

co no

m ic

s p la

n w

as o

ut lin

ed in

th e

stu dy

p ro

to co

l. Th

is is

m en

tio ne

d in

th e

M et

ho ds

:O ve

rv ie

w se

ct io

n St

ud y

po pu

la tio

n 5

D es

cr ib

e ch

ar ac

te ris

tic s o

f t he

st ud

y po

pu la

tio n

(s uc

h as

a ge

ng e,

d em

og ra

ph ic

s, so

ci oe

co no

m ic

, o r c

lin ic

al c

ha ra

c- te

ris tic

Ye s,

th is

is d

es cr

ib ed

in th

e M

et ho

ds : R

an do

m is

ed C

on tro

lle d

Tr ia

l ( RC

T) se

ct io

n an

d in

T ab

le 1

Se tti

ng a

nd lo

ca tio

n 6

Pr ov

id e

re le

va nt

c on

te xt

ua l i

nf or

m at

io n

th at

m ay

in flu

en ce

fin

di ng

s Ye

s, th

is is

d es

cr ib

ed in

th e

M et

ho ds

: R an

do m

is ed

C on

tro lle

d Tr

ia l (

RC T)

se ct

io n

C om

pa ra

to rs

7 D

es cr

ib e

th e

in te

rv en

tio ns

o r s

tra te

gi es

b ei

ng c

om pa

re d

an d

w hy

c ho

se n

Ye s,

th is

is d

es cr

ib ed

in th

e M

et ho

ds : R

an do

m is

ed C

on tro

lle d

Tr ia

l ( RC

T) se

ct io

n Pe

rs pe

ct iv

e 8

St at

e th

e pe

rs pe

ct iv

e( s)

a do

pt ed

b y

th e

stu dy

a nd

w hy

os en

Ye s,

th is

is d

es cr

ib ed

in th

e M

et ho

ds : O

ve rv

ie w

a nd

M et

ho ds

: C

os t A

na ly

si s s

ec tio

ns Ti

m e

ho riz

on 9

St at

e th

e tim

e ho

riz on

fo r t

he st

ud y

an d

w hy

a pp

ro pr

ia te

Ye s,

th is

is d

es cr

ib ed

in th

e M

et ho

ds : O

ve rv

ie w

se ct

io n

D is

co un

t r at

e 10

Re po

rt th

e di

sc ou

nt ra

te (s

) a nd

re as

on c

ho se

n Ye

s, th

is is

d es

cr ib

ed in

th e

M et

ho ds

: O ve

rv ie

w se

ct io

n Se

le ct

io n

of o

ut co

m es

11 D

es cr

ib e

w ha

t o ut

co m

es w

er e

us ed

a s t

he m

ea su

re (s

) o f

be ne

fit (s

) a nd

h ar

m (s

) Ye

s, th

is is

d es

cr ib

ed in

th e

M et

ho ds

: E ffe

ct iv

en es

s A na

ly si

s se

ct io

n M

ea su

re m

en t o

f o ut

co m

es 12

D es

cr ib

e ho

w o

ut co

m es

u se

d to

c ap

tu re

b en

efi t(s

) a nd

rm (s

) w er

e m

ea su

re d

Ye s,

th is

is d

es cr

ib ed

in th

e M

et ho

ds : E

ffe ct

iv en

es s A

na ly

si s

se ct

io n

Va lu

at io

n of

o ut

co m

es 13

D es

cr ib

e th

e po

pu la

tio n

an d

m et

ho ds

u se

d to

m ea

su re

a nd

lu e

ou tc

om es

Ye s,

th is

is d

es cr

ib ed

in th

e M

et ho

ds : E

ffe ct

iv en

es s A

na ly

si s

se ct

io n

M ea

su re

m en

t a nd

v al

ua tio

n of

re so

ur ce

s a nd

c os

ts 14

D es

cr ib

e ho

w c

os ts

w er

e va

lu ed

Ye s,

th is

is d

es cr

ib ed

in th

e M

et ho

ds : C

os t A

na ly

si s s

ec tio

n C

ur re

nc y,

p ric

e da

te , a

nd c

on ve

rs io

n 15

Re po

rt th

e da

te s o

f t he

e sti

m at

ed re

so ur

ce q

ua nt

iti es

a nd

it co

sts , p

lu s t

he c

ur re

nc y

an d

ye ar

o f c

on ve

rs io

n Ye

s, th

is is

d es

cr ib

ed in

th e

M et

ho ds

: C os

t A na

ly si

s s ec

tio n

R at

io na

le a

nd d

es cr

ip tio

n of

m od

el 16

If m

od el

lin g

is u

se d,

d es

cr ib

e in

d et

ai l a

nd w

hy u

se d.

R ep

or t

if th

e m

od el

is p

ub lic

ly av

ai la

bl e

an d

w he

re it

c an

b e

ac ce

ss ed

N /A

A na

ly tic

s a nd

a ss

um pt

io ns

17 D

es cr

ib e

an y

m et

ho ds

fo r a

na ly

si ng

o r s

ta tis

tic al

ly tr

an s-

fo rm

in g

da ta

, a ny

e xt

ra po

la tio

n m

et ho

ds , a

nd a

pp ro

ac he

s fo

r v al

id at

in g

an y

m od

el u

se d

Ye s,

th is

is d

es cr

ib ed

fo r c

os ts

. Q A

LY s a

nd N

et B

en efi

ts in

e M

et ho

ds : s

ec tio

C ha

ra ct

er is

in g

he te

ro ge

ne ity

18 D

es cr

ib e

an y

m et

ho ds

u se

d fo

r e sti

m at

in g

ho w

th e

re su

lts o

f th

e stu

dy v

ar y

fo r s

ub gr

ou ps

N /A

451Cost effectiveness of a GP delivered medication review to reduce polypharmacy and potentially…

H us

er ea

u D

, D ru

m m

on d

M , A

ug us

to vs

ki F

, e t a

l. C

on so

lid at

ed H

ea lth

E co

no m

ic E

va lu

at io

n Re

po rti

ng S

ta nd

ar ds

2 02

2 (C

H EE

R S

20 22

) E xp

la na

tio n

an d

El ab

or at

io n:

A R

ep or

t o f t

he IS

PO R

H EE

R S

II G

oo d

Pr ac

tic es

T as

k Fo

rc e.

V al

ue H

ea lth

2 02

2; 25

. h ttp

s: // d

oi . o

rg / 1

0. 10

16 /j.

jv al

. 2 02

1. 10

. 0 08

Ta bl

e 15

( co

nt in

ue d)

To pi

c N

o Ite

m Lo

ca tio

n w

he re

it em

is re

po rte

C ha

ra ct

er is

in g

di str

ib ut

io na

l e ffe

ct s

19 D

es cr

ib e

ho w

im pa

ct s a

re d

ist rib

ut ed

a cr

os s d

iff er

en t i

nd i-

vi du

al s o

r a dj

us tm

en ts

m ad

e to

re fle

ct p

rio rit

y po

pu la

tio ns

N /A

C ha

ra ct

er is

in g

un ce

rta in

ty 20

D es

cr ib

e m

et ho

ds to

c ha

ra ct

er is

e an

y so

ur ce

s o f u

nc er

ta in

in th

e an

al ys

is Ye

s, th

is is

d es

cr ib

ed fo

r c os

ts . Q

A LY

s a nd

N et

B en

efi ts

th e

M et

ho ds

: s ec

tio n

an d

in th

e D

et er

m in

ist ic

S en

si tiv

ity

A na

ly si

s s ec

tio n

A pp

ro ac

h to

e ng

ag em

en t w

ith p

at ie

nt s a

nd o

th er

s a ffe

ct ed

th e

stu dy

21 D

es cr

ib e

an y

ap pr

oa ch

es to

e ng

ag e

pa tie

nt s o

r s er

vi ce

ci pi

en ts

, t he

g en

er al

p ub

lic , c

om m

un iti

es , o

r s ta

ke ho

ld -

er s (

su ch

a s c

lin ic

ia ns

o r p

ay er

s) in

th e

de si

gn o

f t he

st ud

N /A

Re su

lts St

ud y

pa ra

m et

er s

22 Re

po rt

al l a

na ly

tic in

pu ts

(s uc

h as

v al

ue s,

ra ng

es , r

ef er

en ce

in cl

ud in

g un

ce rta

in ty

o r d

ist rib

ut io

na l a

ss um

pt io

ns W

he re

re le

va nt

, t hi

s i s r

ep or

te d

ex pl

ic itl

y in

th e

Re su

lts se

c- tio

n Su

m m

ar y

of m

ai n

re su

lts 23

Re po

rt th

e m

ea n

va lu

es fo

r t he

m ai

n ca

te go

rie s o

f c os

an d

ou tc

om es

o f i

nt er

es t a

nd su

m m

ar is

e th

em in

th e

m os

t ap

pr op

ria te

o ve

ra ll

m ea

su re

Th is

is re

po rte

d ex

pl ic

itl y

in th

e Re

su lts

se ct

io n

Eff ec

t o f u

nc er

ta in

ty 24

D es

cr ib

e ho

w u

nc er

ta in

ty a

bo ut

a na

ly tic

ju dg

m en

ts , i

np ut

s, or

p ro

je ct

io ns

a ffe

ct fi

nd in

gs . R

ep or

t t he

e ffe

ct o

f c ho

ic e

di sc

ou nt

ra te

a nd

ti m

e ho

riz on

, i f a

pp lic

ab le

Th is

is re

po rte

d ex

pl ic

itl y

in th

e Re

su lts

se ct

io n

Eff ec

t o f e

ng ag

em en

t w ith

p at

ie nt

s a nd

o th

er s a

ffe ct

ed b

y th

e stu

dy 25

Re po

rt on

a ny

d iff

er en

ce p

at ie

nt /s

er vi

ce re

ci pi

en t,

ge ne

ra l

pu bl

ic , c

om m

un ity

, o r s

ta ke

ho ld

er in

vo lv

em en

t m ad

e to

e ap

pr oa

ch o

r fi nd

in gs

o f t

he st

ud y

N /A

D is

cu ss

io n

St ud

y fin

di ng

s, lim

ita tio

ns , g

en er

al is

ab ili

ty , a

nd c

ur re

kn ow

le dg

e 26

Re po

rt ke

y fin

di ng

s, lim

ita tio

ns , e

th ic

al o

r e qu

ity c

on si

de ra

- tio

ns n

ot c

ap tu

re d,

a nd

h ow

th es

e co

ul d

aff ec

t p at

ie nt

s, po

lic y,

o r p

ra ct

ic e

Th es

e is

su es

a re

g en

er al

ly d

is cu

ss ed

in th

e D

is cu

ss io

n se

ct io

O th

er re

le va

nt in

fo rm

at io

n So

ur ce

o f f

un di

ng 27

D es

cr ib

e ho

w th

e stu

dy w

as fu

nd ed

a nd

a ny

ro le

o f t

fu nd

er in

th e

id en

tifi ca

tio n,

d es

ig n,

c on

du ct

, a nd

re po

rti ng

th e

an al

ys is

Th is

is re

po rte

d al

on gs

id e

th e

A bs

tra ct

C on

fli ct

s o f i

nt er

es t

28 Re

po rt

au th

or s c

on fli

ct s o

f i nt

er es

t a cc

or di

ng to

jo ur

na l

or In

te rn

at io

na l C

om m

itt ee

o f M

ed ic

al Jo

ur na

l E di

to rs

qu ire

m en

Th is

is re

po rte

d al

on gs

id e

th e

A bs

tra ct

https://doi.org/10.1016/j.jval.2021.10.008

452 P. Gillespie et al.

Scheduling was flexible, allowing practices to determine when and how to conduct the reviews within a given timeframe.

10. Modifications:

• Conducted phone reviews in response to COVID-19 for one practice. Some practices performed opportunistic reviews instead of scheduled ones, although this was not part of the original plan.

11. How well (planned):

• Implementation fidelity was monitored through a mixed methods process evaluation, with data collected on web- site usage and feedback from GPs and patients.

12. How well (actual):

• Adaptations were made as necessary, such as conducting phone reviews due to COVID-19. The overall fidelity to

the planned intervention was assessed through quanti- tative data (website usage) and qualitative data (semi- structured interviews).

• The "brown bag" component was most effective in result- ing in medication changes. Most GPs and patients did not engage with the priority-setting exercise.

See Tables 5, 6, 7, 8, 9, 10, 11, 12, 13, 14 and 15; Figs. 2, 3.

Acknowledgements We thank all the patients and GP practice staff who took part in this research. Membership of the wider SPPiRE Study group who contributed to this research (administration, pilot study, practice recruitment, data collection, data entry, and website develop- ment) were Tom Fahey, Derek Corrigan, Bridget Kiely, Aisling Croke, James Larkin, Oscar James, Clare Lambert, and Brenda Quigley. The independent members of the trial steering committee were Patricia Kearney (Chair), Andrew Murphy, Cathal Cadogan, Carmel Hughes, and Brid Nolan (PPI).

Funding Open Access funding provided by the IReL Consortium. This research is funded by the HRB Primary Care Clinical Trial’s Network, Ireland (Grant CTN-2021-002).

Data availability Data will be made available on reasonable request.

Declarations

Conflict of interest The authors declare no conflicts of interest.

Fig. 2 Cost effectiveness plane – SPPiRE intervention

-€3,000

-€2,000

-€1,000

€0

€1,000

€2,000

€3,000

-0.100 -0.080 -0.060 -0.040 -0.020 0.000 0.020 0.040 0.060 0.080 0.100

In cr

em en

ta l C

os t

Incremental Effect

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

€0 €10,000 €20,000 €30,000 €40,000 €50,000 €60,000 €70,000 €80,000 €90,000 €1,00,000

Pr ob

ab ili

ty C

os t E

ffe ct

iv e

Threshold Value

Fig. 3 Cost effectiveness acceptability curve– SPPiRE intervention

http://creativecommons.org/licenses/by/4.0/

453Cost effectiveness of a GP delivered medication review to reduce polypharmacy and potentially…

References

1. Fortin, M., Lapointe, L., Hudon, C., Vanasse, A., Ntetu, A.L., Maltais, D.: Multimorbidity and quality of life in primary care: a systematic review. Health Qual. Life Outcomes 2, 51 (2004). https:// doi. org/ 10. 1186/ 1477- 7525-2- 51PMC 526383

2. Mair, F.S., Gallacher, K.I.: Multimorbidity: what next? Br. J. Gen. Pract. 67(659), 248–249 (2017). https:// doi. org/ 10. 3399/ bjgp1 7X690 965

3. Nunes, B.P., Flores, T.R., Mielke, G.I., Thumé, E., Facchini, L.A.: Multimorbidity and mortality in older adults: a systematic review and meta-analysis. Arch. Gerontol. Geriatr. 67, 130–138 (2016). https:// doi. org/ 10. 1016/j. archg er. 2016. 07. 008. (Epub 2016/08/09)

4. Payne, R.A., Abel, G.A., Guthrie, B., Mercer, S.W.: The effect of physical multimorbidity, mental health conditions and socioeco- nomic deprivation on unplanned admissions to hospital: a retro- spective cohort study. CMAJ 185(5), E221–E228 (2013). https:// doi. org/ 10. 1503/ cmaj. 121349. (Epub 2013/02/21)

5. Kongkaew, C., Noyce, P.R., Ashcroft, D.M.: Hospital admissions associated with adverse drug reactions: a systematic review of prospective observational studies. Ann. Pharmacother. 42(7), 1017–1025 (2008). https:// doi. org/ 10. 1345/ aph. 1L037. (Epub 2008/07/03)

6. Spinewine, A., Schmader, K.E., Barber, N., Hughes, C., Lapane, K.L., Swine, C., et al.: Appropriate prescribing in elderly people: how well can it be measured and optimised? Lancet 370(9582), 173–184 (2007). https:// doi. org/ 10. 1016/ S0140- 6736(07) 61091-5. (Epub 2007/07/17)

7. Wallace, J., Paauw, D.S.: Appropriate prescribing and important drug interactions in older adults. Med. Clin. North Am. 99(2), 295–310 (2015). https:// doi. org/ 10. 1016/j. mcna. 2014. 11. 005. (Epub 2015/02/24)

8. Guthrie, B., McCowan, C., Davey, P., Simpson, C.R., Dreischulte, T., Barnett, K.: High risk prescribing in primary care patients particularly vulnerable to adverse drug events: cross sectional population database analysis in Scottish general practice. BMJ 342, d3514 (2011). https:// doi. org/ 10. 1136/ bmj. d3514. (Epub 2011/06/23)

9. Payne, R.A., Abel, G.A., Avery, A.J., Mercer, S.W., Roland, M.O.: Is polypharmacy always hazardous? A retrospectivecohort analysis using linked electronic health records from primary and secondary care. Br. J. Clin. Pharmacol. 77(6), 1073–1082 (2014). https:// doi. org/ 10. 1111/ bcp. 12292

10. Gnjidic, D., Le Couteur, D.G., Kouladjian, L., Hilmer, S.N.: Deprescribing trials: methods to reduce polypharmacy and the impact on prescribing and clinical outcomes. Clin. Geriatr. Med. 28(2), 237–253 (2012). https:// doi. org/ 10. 1016/j. cger. 2012. 01. 006

11. McCarthy, C., Sheeran-Purcell, P., Fitzgerald, L., Cafferty, O., Sheeran-Purcell, L.: Medication Review: A Guide for GPs: Quick Reference Guide. ICGP Quality in Practice Committee, Dublin (2020)

12. Mair, F.S., May, C.R.: Thinking about the burden of treatment. BMJ 349, g6680 (2014). https:// doi. org/ 10. 1136/ bmj. g6680

13. Reeve, E., Gnjidic, D., Long, J., Hilmer, S.: A systematic review of the emerging definition of “deprescribing” with network analysis: implications for future research and clinical practice. Br. J. Clin. Pharmacol. (2015). https:// doi. org/ 10. 1111/ bcp. 12732

14. Skou, S.T., Mair, F.S., Fortin, M., Guthrie, B., Nunes, B.P., Miranda, J.J., Boyd, C.M., Pati, S., Mtenga, S., Smith, S.M.: Multimorbidity. Nat. Rev. Dis. Primers. 8(1), 48 (2022). https:// doi. org/ 10. 1038/ s41572- 022- 00376-4

15. Clyne, B., Smith, S.M., Hughes, C.M., Boland, F., Bradley, M.C., Cooper, J.A., et al.: Effectiveness of a multifaceted intervention for potentially inappropriate prescribing in older

patients in primary care: a cluster-randomized controlled trial (OPTI-SCRIPT Study). Ann. Fam. Med. 13(6), 545–553 (2015). https:// doi. org/ 10. 1370/ afm. 1838. (Epub 2015/11/11)

16. Smith, S.M., Wallace, E., Clyne, B., Boland, F., Fortin, M.: Interventions for improving outcomes in patients with multi- morbidity in primary care and community setting: a system- atic review. Syst. Rev. 10, 271 (2021). https:// doi. org/ 10. 1186/ s13643- 021- 01817-z

17. Rankin, A., Cadogan, C.A., Patterson, S.M., Kerse, N., Card- well, C.R., Bradley, M.C., et al.: Interventions to improve the appropriate use of polypharmacy for older people. Cochrane Database Syst. Rev. (2018). https:// doi. org/ 10. 1002/ 14651 858. CD008 165. pub4

18. Laberge, M., Sirois, C., Lunghi, C., Gaudreault, M., Nakamura, Y., Bolduc, C., Laroche, M.L.: Economic evaluations of interven- tions to optimize medication use in older adults with polyphar- macy and multimorbidity: a systematic review. Clin. Interv. Aging 5(16), 767–779 (2021). https:// doi. org/ 10. 2147/ CIA. S3040 74

19. Romano, S., Figueira, D., Teixeira, I., Perelman, J.: Deprescribing interventions among community-dwelling older adults: a system- atic review of economic evaluations. Pharmacoeconomics 40(3), 269–295 (2022). https:// doi. org/ 10. 1007/ s40273- 021- 01120-8. (Epub 2021 Dec 16)

20. McCarthy, C., Clyne, B., Boland, F., Moriarty, F., Flood, M., Wal- lace, E., Smith, S.: GP-delivered medication review of polyphar- macy, deprescribing, and patient priorities in older people with multimorbidity in Irish primary care (SPPiRE Study): a cluster randomised controlled trial. PLoS Med. 19(1), e1003862 (2022). https:// doi. org/ 10. 1371/ journ al. pmed. 10038 62

21. Drummond, M.F., Sculpher, M.J., Claxton, K., Stoddart, G.L., Torrance, G.W.: Methods for the Economic Evaluation of Health Care Programmes. Oxford University Press (2015)

22. Thorn, J., Man, M.S., Chaplin, K., Bower, P., Brookes, S., Gaunt, D., Fitzpatrick, B., Gardner, C., Guthrie, B., Holling- hurst, S., Lee, V., Mercer, S.W., Salisbury, C.: Cost-effective- ness of a patient-centred approach to managing multimorbid- ity in primary care: a pragmatic cluster randomised controlled trial. BMJ Open 10(1), e030110 (2020). https:// doi. org/ 10. 1136/ bmjop en- 2019- 030110

23. Salari, P., O’Mahony, C., Henrard, S., Welsing, P., Bhadhuri, A., Schur, N., Roumet, M., Beglinger, S., Beck, T., Jungo, K.T., Byrne, S., Hossmann, S., Knol, W., O’Mahony, D., Spinewine, A., Rodondi, N., Schwenkglenks, M.: Cost-effectiveness of a structured medication review approach for multimorbid older adults: Within-trial analysis of the OPERAM study. PLoS One 17(4), e0265507 (2022). https:// doi. org/ 10. 1371/ journ al. pone. 02655 07

24. Health, Information and Quality Authority (HIQA). Guidelines for the Economic Evaluation of Health Technologies in Ireland. 2020. http:// www. hiqa. ie/ publi cation/ guide lines- econo mic- evalu ation- health- techn ologi es- irela nd

25. Husereau, D., Drummond, M., Augustovski, F., et al.: Consoli- dated Health Economic Evaluation Reporting Standards 2022 (CHEERS 2022) statement: updated reporting guidance for health economic evaluations. BMC Med. 20, 23 (2022). https:// doi. org/ 10. 1186/ s12916- 021- 02204-0

26. Campbell, M.K., Elbourne, D.R., Altman, D.G.: CONSORT statement: extension to cluster randomised trials. BMJ 328, 702–708 (2004)

27. Gomes, M., Ng, E.S.W., Grieve, R., Nixon, R., Carpenter, J., Thompson, S.G.: Developing appropriate methods for cost- effectiveness analysis of cluster randomized trials. Med. Decis. Making 32(2), 350–361 (2012)

28. Gomes, M., Grieve, R., Nixon, R., Ng, E.S., Carpenter, J., Thompson, S.G.: Methods for covariate adjustment in

https://doi.org/10.1186/1477-7525-2-51PMC526383

https://doi.org/10.3399/bjgp17X690965

https://doi.org/10.1016/j.archger.2016.07.008

https://doi.org/10.1503/cmaj.121349

https://doi.org/10.1345/aph.1L037

https://doi.org/10.1016/S0140-6736(07)61091-5

https://doi.org/10.1016/j.mcna.2014.11.005

https://doi.org/10.1136/bmj.d3514

https://doi.org/10.1111/bcp.12292

https://doi.org/10.1016/j.cger.2012.01.006

https://doi.org/10.1136/bmj.g6680

https://doi.org/10.1111/bcp.12732

https://doi.org/10.1038/s41572-022-00376-4

https://doi.org/10.1370/afm.1838

https://doi.org/10.1186/s13643-021-01817-z

https://doi.org/10.1002/14651858.CD008165.pub4

https://doi.org/10.2147/CIA.S304074

https://doi.org/10.1007/s40273-021-01120-8

https://doi.org/10.1371/journal.pmed.1003862

https://doi.org/10.1136/bmjopen-2019-030110

https://doi.org/10.1371/journal.pone.0265507

http://www.hiqa.ie/publication/guidelines-economic-evaluation-health-technologies-ireland

https://doi.org/10.1186/s12916-021-02204-0

454 P. Gillespie et al.

cost-effectiveness analysis that use cluster randomised trials. Health Econ. 21(9), 1101–1118 (2012)

29. Ng, E.S., Diaz-Ordaz, K., Grieve, R., Nixon, R.M., Thompson, S.G., Carpenter, J.R.: Multilevel models for cost-effectiveness analyses that use cluster randomised trial data: An approach to model choice. Stat. Methods Med. Res. (2013). https:// doi. org/ 10. 1177/ 09622 80213 511719

30. Gomes, M., Díaz-Ordaz, K., Grieve, R., Kenward, M.: Multiple imputation methods for handling missing data in cost-effective- ness analyses that use data from hierarchical studies an applica- tion to cluster randomized trials. Med. Decis. Making 33(8), 1051–1063 (2013)

31. Rubin, D.: Multiple Imputation for Nonresponse in Surveys. Wiley, Chichester (1987)

32. McCarthy, C., Clyne, B., Corrigan, D., Boland, F., Wallace, E., Moriarty, F., et al.: Supporting prescribing in older people with multimorbidity and significant polypharmacy in primary care (SPPiRE): a cluster randomized controlled trial protocol and pilot. Implement. Sci. 12(1), 99 (2017). https:// doi. org/ 10. 1186/ s13012- 017- 0629-1. (Epub 2017/08/03)

33. Central Statistics Office. Dublin (www. cso. ie). (Accessed June 2022)

34. Guidelines for Inclusion of Drug Costs in Pharmacoeconomic Evaluations v3.2

35. Hardin, J.W., Hilbe, J.M.: Generalised Estimating Equations. Chapman and Hall/CRC Press, London (2003)

36. Rodríguez, G.: Multilevel generalized linear models. In: de Leeuw, J., Meijer, E. (eds.) Handbook of Multilevel Analysis. Springer, New York (2008). https:// doi. org/ 10. 1007/ 978-0- 387- 73186-5_9

37. Thompson, S.G., Nixon, R.M., Grieve, R.: Addressing the issues that arise in analysing multicentre cost data with application to a multinational study. J. Health Econ. 25, 1015–1028 (2006)

38. EuroQol—a new facility for the measurement of health-related quality of life. Health Policy. 16(3):199–208 (1990). https:// doi. org/ 10. 1016/ 0168- 8510(90) 90421-9 (Epub 1990/11/05)

39. Hobbins, A., Barry, L., Kelleher, D., Shah, K., Devlin, N., Goni, J.M.R., O’Neill, C.: Utility values for health states in Ireland: a

value set for the EQ-5D-5L. Pharmacoeconomics 36(11), 1345– 1353 (2018). https:// doi. org/ 10. 1007/ s40273- 018- 0690-x

40. Orenstein, D., Kaplan, R.: Measuring the quality of well-being in cystic fibrosis and lung transplantation: the importance of the area under the curve. Chest 100, 1016–1018 (1991)

41. Hoch, J., Rock, M., Krahn, A.: Using the net benefit regression framework to construct cost-effectiveness acceptability curves: an example using data from a trial of external loop recorders ver- sus Holter monitoring for ambulatory monitoring of “community acquired” syncope. BMC Health Serv. Res. (2006). https:// doi. org/ 10. 1186/ 1472- 6963-6-6

42. Fenwick, E., O’Brien, B., Briggs, A.: Cost-effectiveness accept- ability curves: facts, fallacies and frequently asked questions. Health Econ. 13, 405–415 (2004)

43. Ng, E.S., Grieve, R., Carpenter, J.: Two-stage non-parametric bootstrap sampling with shrinkage correction for clustered data. Stand. Genomic Sci. 13(1), 141–164 (2013)

44. Briggs, A.H.: A bayesian approach to stochastic cost-effectiveness analysis: an illustration and application to blood pressure control in type 2 diabetes. Int. J. Technol. Assess. Health Care 17(1), 69–82 (2001)

45. Diaz-Ordaz, K., Kenward, M.G., Grieve, R.: Handling missing values in cost-effectiveness analyses that use data from cluster randomised trials. 2012. J. R. Stat. Soc. Ser. A. http:// araiv. org/ 1206. 6070v1 [stat.ME]

46. Bachmann, M.O., Fairall, L., Clark, A., Mugford, M.: Methods for analyzing cost effectiveness data from cluster randomized trials. Cost Eff. Resour. Alloc. 6(5), 12 (2007). https:// doi. org/ 10. 1186/ 1478- 7547-5- 12

Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

https://doi.org/10.1177/0962280213511719

https://doi.org/10.1186/s13012-017-0629-1

http://www.cso.ie

https://doi.org/10.1007/978-0-387-73186-5_9

https://doi.org/10.1016/0168-8510(90)90421-9

https://doi.org/10.1007/s40273-018-0690-x

https://doi.org/10.1186/1472-6963-6-6

http://araiv.org/1206.6070v1

https://doi.org/10.1186/1478-7547-5-12

Cost effectiveness of a GP delivered medication review to reduce polypharmacy and potentially inappropriate prescribing in older patients with multimorbidity in Irish primary care: the SPPiRE cluster randomised controlled trial

Abstract

Background
Objective
Methods
Results
Conclusions
Trial registration

Introduction
Methods

Overview
Randomised controlled trial (RCT)
Cost analysis
Effectiveness analysis
Cost effectiveness analysis
Deterministic sensitivity analysis

Results
Discussion

Strengths and limitations

Appendix
SPPiRE Intervention—Proctor implementation strategy
SPPiRE Intervention—TIDieR Checklist for Intervention
Acknowledgements
References

image8.emf

Patients’ experiences of a patient-centred polypharmacy medication review intervention: a mixed-methods study

Lorna J Duncan, Deborah McCahon, Barbara Caddick, Roxanne M Parslow, Katrina Turner, Carolyn A Chew-Graham, Bruce Guthrie and Rupert A Payne

Abstract

Background People prescribed multiple medications need regular review to optimise medication and improve health outcomes. Although the most effective method to improve inappropriate polypharmacy remains unclear, medication reviews incorporating patient-centred care and shared decision making are believed to be key to achieving optimal outcomes.

Aim To explore patients’ experiences of an intervention to deliver patient- centred polypharmacy medication review in primary care.

Design and setting A mixed-methods process evaluation was undertaken within the Improving Medicines use in People with Polypharmacy in Primary Care (IMPPP) randomised controlled trial conducted in the Bristol and West Midlands regions of England.

Method Participants receiving the intervention were invited to complete a patient experience survey. Additionally, participants were purposively sampled and invited to participate in an interview and/or audio-recording of their medication review, to explore views and experiences.

Results Survey response rate was 72.5% (n = 556/767); 28 patients were interviewed, 27 reviews were recorded. Overall, 73.2% (n = 407) were satisfied with their review, strongly associated with shared decision making, and 79.9% (n = 444) expressed satisfaction with how the review was delivered (primarily pharmacist- led telephone consultations). Most audio-recordings of reviews demonstrated collaborative decision making. Patients valued reviews when they felt well-informed, prepared, and

received clear follow-up. Pharmacist- led delivery was acceptable, although unfamiliarity with the reviewer and concerns about prescribing authority were perceived negatively.

Conclusion Participants were satisfied with their review, although this may be contingent on preparation and support. Findings highlight the importance of communication throughout, and how the clinician role and familiarity shape patient experience. A person-centred review approach has the potential to improve patient experience, satisfaction, and engagement.

Keywords clinical trial; general practice; medicines optimisation; polypharmacy; decision making, shared; qualitative evaluation.

Introduction Polypharmacy is the use of multiple medications by a single individual, and is common among older adults or those with multiple long-term conditions.1 Although polypharmacy can be appropriate, benefits are not always realised and there is the potential for drug-related harms.2 Regular medication reviews, involving the patient and a prescriber, may offer opportunities to optimise medicines’ use and consequent

health outcomes. Reviews should involve a comprehensive assessment to identify therapeutic priorities, discuss best available evidence, and reach agreement on what medicines to use.3,4

Although the most effective method to optimise medicines in people with potentially problematic polypharmacy remains uncertain, integrating holistic, patient-centred care (that is, focusing on the patient’s needs, preferences,

and values) and shared decision making (that is, involving collaborative discussion between practitioner and patient) within the process may have value. Within the context of long-term condition management, these approaches may help patients and practitioners reach agreement on treatment plans that align with patient preferences, improve patient satisfaction, and enhance treatment adherence.5,6 However, research suggests

Research

e946 | RESEARCH British Journal of General Practice, January 2026

that patients often feel unprepared or are not offered an opportunity to actively engage during medication reviews.7 Furthermore, many healthcare professionals lack the necessary skills for conducting reviews tailored to individual patients’ needs.7–9

The Improving Medicines use in People with Polypharmacy in Primary Care (IMPPP) trial evaluated a complex intervention delivering person-centred medication reviews to improve inappropriate polypharmacy in English general practice.10 As patient experience is an indicator of quality of care and positively associated with safety and clinical effectiveness,11 the aim of the current analysis was to explore patients’ experiences of medication review in the intervention arm.

Method The IMPPP trial was a cluster randomised two-arm parallel design, conducted in English general practices, comparing a complex intervention against usual care. The conceptual framework incorporated principles of pharmaceutical care, shared decision making, behaviour change theory, and integrated primary care practice. The protocol has been published.10

Thirty-seven practices (19 intervention) participated in the main trial, and a further five practices (three intervention) participated in an external pilot. The primary outcome

was rate of potentially inappropriate prescribing (PIP; defined based on clinical code-based rules, as listed in the trial protocol),10 with secondary outcomes including other prescribing measures (for example, adherence and treatment burden), health service use, and mortality. A mixed-methods process evaluation, including a post-intervention patient experience survey, qualitative interviews, and audio-recordings of medication reviews, was included to understand the experiences of patients and clinicians. Clinicians’ perceptions of the intervention are reported elsewhere.12 The main trial effectiveness results are not yet published.

Trial population, eligibility criteria, and recruitment Potentially eligible trial participants aged ≥18 years prescribed ≥5 regular medicines, with at least one indicator of PIP, were identified pre-randomisation via an automated search of their electronic primary healthcare records. PIP indicators were defined using a set of clinical code-based rules, incorporating prescribing, diagnostic, and test codes as detailed in the trial protocol.10 Exclusion criteria, applied by a clinician at each site, included those deemed clinically inappropriate for contact (for example, end-of-life care) or unable to complete surveys without support. Eligible patients were invited by post before randomisation. Individuals interested in participating completed a postal consent form and a baseline survey. Each practice was asked to deliver the medication review to up to 50 consented patients over a 6-month period.

Intervention description The intervention involved a structured process for medication review with a comprehensive clinical encounter aimed at improving the clinical effectiveness and safety of the medication regimen (for example, through stopping or starting drugs, dose adjustment, and appropriate monitoring) tailored to the individual’s clinical circumstances (such as, age, frailty, and comorbidities) and personal values and preferences. The process comprised: automated case-finding; a pharmacist-conducted case-note review; an interprofessional collaborative discussion between pharmacist and GP; a patient-facing medication review

(led by either the pharmacist or GP); and subsequent follow-up as deemed clinically necessary.

This was supported by intervention components designed to enhance engagement, including financial incentives for practices, performance feedback, and clinician training. Participating GPs and clinical pharmacists completed a comprehensive training programme covering topics essential for patient-centred medication reviews, including key consultation and communication skills (for example, relating to shared decision making).

To enhance patient engagement in the review, participating practices were asked to provide patients with a pre-review leaflet, containing information about the purpose of a review and how to prepare for it, alongside a list of their current medicines (see Supplementary Information S1).

Audio-recording of reviews and patient interviews On providing consent at baseline, participants could optionally choose to

LJ Duncan (ORCID: 0000-0002-9629-030X), PhD, research fellow; D McCahon (ORCID: 0000-0003-2768-293X), PhD, senior research fellow; B Caddick (ORCID: 0000-0002-3032-0430), PhD, senior research associate; RM Parslow (ORCID: 0000-0002-3612-7121), PhD, research fellow; K Turner (ORCID: 0000-0002-6375-2918), PhD, professor of primary care research, Centre for Academic Primary Care, Population Health Sciences, Bristol Medical School, University of Bristol, Bristol, UK. CA Chew-Graham (ORCID: 0000-0002-9722-9981), MD, FRCGP, professor of general practice, School of Medicine, Keele University, Keele, UK. B Guthrie (ORCID: 0000-0003-4191-4880), FRSE, FMedSci, professor of general practice, Advanced Care Research Centre, University of Edinburgh, Edinburgh, UK. RA Payne (ORCID: 0000-0002-5842-4645), PhD, FBPhS, professor of primary care and clinical pharmacology, Exeter Collaboration for Academic Primary Care, University of Exeter, Exeter, UK.

CORRESPONDENCE

Deborah McCahon Centre for Academic Primary Care, Population Health Sciences, Bristol Medical School, University of Bristol, Canynge Hall, 39 Whatley Road, Bristol BS8 2PS, UK. Email: [email protected]

Submitted: 23 January 2025; Editor’s response: 10 March 2025; final acceptance: 10 June 2025.

©The Authors This is the full-length article (published online 18 November 2025) of an abridged version published in print. Cite this version as: Br J Gen Pract 2025; DOI: https://doi.org/10.3399/BJGP.2025.0052

How this fits in

Medication reviews are a central part of the process of medication optimisation, and a key policy strategy aiming to improve safe and effective prescribing and health outcomes, despite a relative lack of evidence for their effectiveness. Holistic patient-centred care can help align treatment with patients’ values and preferences, improve patient satisfaction, enhance treatment, and improve outcomes, but is not always achieved in current practice. The Improving Medicines use in People with Polypharmacy in Primary Care model of medication review delivery, including appropriate clinician training and pre-review patient preparation, achieves patient-centred care and good patient satisfaction, and may have the potential to enhance patient engagement and improve experience of care.

Research

British Journal of General Practice, January 2026 RESEARCH | e947

https://doi.org/10.3399/BJGP.2025.0052

participate in an audio-recording of their review and/or a post-review interview (see Supplementary Figure S1). Those individuals were purposely sampled to maximise variation in age, sex, and education, with no specific exclusion criteria applied. Invitations to participate in audio-recordings were mailed out before the review.

Selection and invitation for interview was undertaken post-review without knowledge of post-review survey completion. Interview times considered participant preference and included weekends and evenings to maximise inclusivity. Additional explicit consent for these activities was obtained. Interviews were conducted via telephone or video-call by two experienced qualitative researchers (the first and fourth authors) between February 2021 and June 2023. Reviews and interviews were audio-recorded and transcribed verbatim.

Interviews were guided by a topic guide developed in collaboration with lay IMPPP study advisors and revised as data collection progressed to reflect early findings (see Supplementary Information S2). During interviews, participants were invited to share their experiences of the review, including their expectations and preparations before the review. Data collection continued until no new insights or themes were identified.

Analysis of interview and audio- recorded reviews was ongoing and iterative, facilitated using NVivo (version 11) software. Thematic analysis was undertaken, employing deductive and inductive approaches to scrutinise the data for anticipated and emergent themes, and contextual factors that enhanced understanding of the survey findings. Anticipated themes included concepts in the patient experience survey and elements of shared decision- making models identified by Makoul and Clayman,13 summarised in the SHARE (seek, help, assess, reach, evaluate) approach.14 Analysis was led by the first author, with a subset of transcripts independently read and coded by the joint first author. A coding framework was refined through consensus, and key themes discussed with the wider team (the fourth, fifth, and sixth authors) to ensure credibility and external validity.

Survey Patient experience surveys were administered via post or electronically (in accordance with patient preferences expressed at recruitment) within 4 weeks of receiving the patient- facing review. Pre-paid envelopes were provided to encourage return of completed surveys. The survey (see Supplementary Information S3) captured the characteristics of the review (reviewer’s profession, duration, how the review was delivered, and medication changes made); overall satisfaction with the review and confidence in the reviewer (5-point Likert scales); perceptions of the extent to which the reviewer involved the patient in shared decision making (9-item Shared Decision Making Questionnaire [SDM-Q-9]15 and three components of the CollaboRATE scale);16 alongside patient-centredness, and processes and outcomes from the patient’s perspective (four items from the Consultation And Relational Empathy measure17 and four items from the Patient Assessment of Chronic Illness Care measure).18 Patient sociodemographics were captured separately pre-randomisation.10

Survey results were reported according to relevant patient experience themes used in the qualitative analysis of interviews using descriptive statistics. Responses to Likert scales were recoded where appropriate to facilitate analysis and interpretation. Associations between responder satisfaction and sociodemographic factors, review context, and shared decision making were examined using mixed-effects logistic regression (see Supplementary Table S1). Analyses were conducted using Stata (version 18). Survey results using original coding and including figures for missing data, grouped by instrument, are included in Supplementary Table S2.

Findings from the survey, audio- recordings, and interviews were triangulated to provide context and a comprehensive understanding of patient experiences.

Results

Participant characteristics A total of 767 participants (36 pilot study and 731 main trial) receiving the patient-facing review were invited to complete the experience survey (February 2021–April 2023), with 471 (61.4%) opting for a postal (rather than an online) survey, and 556 (72.5%) responding across all intervention practices. The median age of responders was 73 (interquartile range 66–79) years, 259 (46.6%) were female, and 531 (95.5%) identified as White British ethnicity (Table 1).

Survey responders and non- responders were broadly comparable, although responders tended to be older, retired, less socioeconomically deprived, hold higher educational qualifications (Table 1), and were more likely to have requested a paper survey (n = 410, 73.7% versus n = 146, 26.3% online survey).

Interviews were conducted with 28 participants (nine pilot and 19 main trial) from nine practices. All interviewees were of White British ethnicity, 17 were male, and age ranged from 47–90 years. Interview duration ranged from 12–53 mins (mean 29 mins).

Twenty-seven reviews were audio- recorded (10 pilot and 17 main trial) within eight practices, including 19 males, and age ranged from 51–92 years.

Overall satisfaction and perceptions of patient-centred care Overall, 407, 73.2% of survey responders were satisfied with their medication review (Table 2), with no evidence of an association between satisfaction and sociodemographic factors (see Supplementary Table S1). Most responders reported receiving patient-centred care (Table 2). Many responders rated the reviewer as good– excellent at ‘being interested in them as a whole person’ (n = 405, 72.8%), ‘understanding their concerns’ (n = 408, 73.4%), and ‘helping them take control’ (n = 399, 71.8%). Views were more mixed about reviewers’ consideration of responders’ ‘values and traditions’ (n = 289, 52.0% yes and n = 237, 42.6% no/not sure) or ‘treatment plans aligning with daily life’ (n = 257, 46.2% yes and n = 271, 48.7% no/not sure).

Research

e948 | RESEARCH British Journal of General Practice, January 2026

Responses to the SDM-Q-9 instrument are reported in Figure 1. Two-thirds (n = 365, 65.6%) of survey responders rated shared decision making above the midpoint of the total SDM-Q-9 score suggesting that patients felt their reviewer engaged them in treatment-related decision making. After adjustment for sociodemographic factors and aspects of the review context, there was strong evidence (P<0.001) that the likelihood of being satisfied was positively associated with a perception of shared decision making

(odds ratio [OR] 2.16, 95% confidence interval [CI] = 1.58 to 2.94 per quintile increase in SDM-Q-9 score; Supplementary Table S1).

Analysis of the interviews and audio-recordings of reviews provided insights into:

• why patient expectations were met (or not);

• how patient views and experiences were affected by how the review was delivered;

• the patient–practitioner relationship; and

• key elements of shared decision making (for example, seeking patient participation, information provision, discussion of patient priorities, agreement on decisions, and summarising/follow-up).

Table 1. Survey responder characteristics (N = 767)a

Characteristic Survey non-responders

(N = 211), n (%) Survey responders

(N = 556), n (%) Interview participants (N = 28), n (%)

Age (years)

≤50 18 (8.5) 16 (2.9) 2 (7.1)

51-65 53 (25.1) 111 (20.0) 5 (17.9)

66-80 106 (50.2) 325 (58.5) 18 (64.3)

≥81 34 (16.1) 104 (18.7) 3 (10.7)

Sex

Male 100 (47.4) 297 (53.4) 17 (60.7)

Female 111 (52.6) 259 (46.6) 11 (39.3)

Index of Multiple Deprivation quintile

1 (most deprived) 32 (15.2) 58 (10.4) 1 (3.6)

2 47 (22.3) 98 (17.6) 7 (25.0)

3 39 (18.5) 101 (18.2) 4 (14.3)

4 53 (25.1) 153 (27.5) 6 (21.4)

5 (least deprived) 39 (18.5) 145 (26.1) 10 (35.7)

Education level

No educational qualifications 62 (29.4) 123 (22.1) 8 (28.6)

GCSE, A -Levels, or equivalent 86 (40.8) 160 (28.8) 4 (14.3)

University degree or higher 25 (11.8) 110 (19.8) 8 (28.6)

Other 19 (9.0) 118 (21.2) 5 (17.9)

Employment

Full or part-time work 39 (18.5) 76 (13.7) 6 (21.4)

Retired from work 117 (55.5) 386 (69.4) 17 (60.7)

Other 38 (18.0) 58 (10.4) 3 (10.7)

Living situation

Own home (rented or owned) 187 (88.6) 506 (91.0) 27 (96.4)

Other 17 (8.1) 34 (6.1) 1 (3.6)

Live alone

No 137 (64.9) 384 (69.1) 20 (71.4)

Yes 61 (28.9) 148 (26.6) 7 (25.0)

Ethnicity

White British 190 (90.0) 531 (95.5) 28 (100)

Other ethnic group 14 (6.6) 12 (2.2) 0 (0.0)

aPercentages based on participant total for each column; percentages do not sum to 100% owing to missing data. GCSE = General Certificate of Secondary Education.

Research

British Journal of General Practice, January 2026 RESEARCH | e949

These themes are discussed below using both qualitative and survey data.

Patient expectations Many interviewees valued their review because it enabled them to gain a better understanding of their medications, receive reassurance that their treatment was effective and appropriate, and/or discuss changes to medications. Some, however, described the review as

Table 2. Overall satisfaction and perceptions of patient- centred care (N = 556)a

Survey question n (%)

Satisfaction

Overall satisfaction with review

Dissatisfied or very dissatisfied 50 (9.0)

Neither dissatisfied nor satisfied 88 (15.8)

Satisfied or very satisfied 407 (73.2)

Patient-centred care

Interest in patient as a whole person (CARE instrument)

Poor or fair 114 (20.5)

Good 120 (21.6)

Very good or excellent 285 (51.3)

Fully understanding patient’s concerns (CARE instrument)

Poor or fair 101 (18.2)

Good 128 (23.0)

Very good or excellent 280 (50.4)

Helping patient take control (CARE instrument)

Poor or fair 101 (18.2)

Good 120 (21.6)

Very good or excellent 279 (50.2)

Reviewer considered patient’s values and traditions (PACIC instrument)

No 88 (15.8)

Yes 289 (52.0)

Not sure 149 (26.8)

Treatment plan that aligned with daily life (PACIC instrument)

No 189 (34.0)

Yes 257 (46.2)

Not sure 82 (14.7)

aPercentages based on total survey responders; percentages do not sum to 100% owing to missing data. CARE = Consultation And Relational Empathy measure. PACIC = Patient Assessment of Chronic Illness Care measure.

unnecessary because their medication regimen was simple, easy to manage, and working well for them. A few described the review as impersonal or unhelpful, seeing it as a procedural requirement that offered nothing new:

‘If you see a marked change in things, then it’s worth it. If it’s ... just a lip service review … to meet a quota, then it’s pointless … they could be more constructive with their time, putting it into other things ... for other people.’ (Pt25)

There was evidence that the trial ‘pre-review leaflet’ was not implemented as planned in some practices and this had a negative impact on patient experience. Some interviewees who had not received the leaflet expressed a desire for information in advance of the review to improve understanding of the process and support active involvement, including information regarding what issues might be considered during the review or provision of a list of their medicines to act as an aide memoire. Prior information and time to reflect and formulate questions in advance was important for ensuring the review was productive and beneficial:

‘It would have been nice if I had a form beforehand that might have said the format of the discussion … so I could have written down a few things to remind myself.’ (Pt13)

How reviews were delivered Table 3 reports survey findings related to how the reviews were delivered. Electronic health records documented mode of review delivery for 465 survey responders, of which 347 (74.6%) were telephone consultations. In total, 444 (79.9%) responders reported they were happy with where the review took place. Responders reported 51.3% (n = 285) and 22.3% (n = 124) of reviews lasting 10–20 min and >20 min, respectively, with 453 (81.5%) considering the time they had adequate. The odds of being satisfied with the review were strongly associated with the duration (OR 32.2, 95% CI = 16.8 to 61.4) and location (OR 13.8, 95% CI = 8.2 to 23.0) of the review being considered acceptable (see Supplementary Table S1).

There were 21 interviewees who indicated the review was conducted via telephone, with views on this mode of delivery varying. Several felt that telephone was more convenient than attending the practice and did not see a need for a discussion focused on medication to be conducted in-person:

‘Over the phone … saves me a lot of time and energy. I don't think the outcome is any different.’ (Pt3)

Others expressed a strong preference for in-person reviews, feeling these provided an opportunity for more open communication, and the clinician to assess symptoms:

‘When you go down there, you do get the opportunity then if there’s anything that you wish to mention … from the doctor’s point of view, he’s trying to work out a diagnosis without seeing the patient … It’s difficult.’ (Pt36)

Patient–clinician relationship Overall, 381 (68.5%) survey responders (and all interviewees) reported that their reviewer was a pharmacist; 77 (13.8%) were unsure. All audio-recordings were of a pharmacist reviewer. A majority (n = 447, 80.4%) were ‘reasonably’, ‘mostly’, or ‘very confident’ that their reviewer knew enough about them and their health (Table 3).

The odds of being satisfied with the review were strongly associated with the patient having received previous care from the reviewer (OR 3.00, 95% CI = 1.66 to 5.42) and having confidence in the reviewer (OR 16.9, 95% CI = 9.8 to 29.1) (see Supplementary Table S1). During the interviews, most participants described clinical pharmacists as experts in medication. Several considered a pharmacist-led review a good use of resources because it freed up GP time for other activities. Others described a lack of confidence in the pharmacist’s ability to enact decisions to change prescribing or expressed a preference for a GP-led review, as they perceived GPs to have greater clinical knowledge and expertise than pharmacists:

‘… he’s not an actual doctor though, is [he]? … I’m not too sure how he’s got to change my medication, if he has got the

Research

e950 | RESEARCH British Journal of General Practice, January 2026

power. I thought the doctor was the one who does that.’ (Pt2)

Some expressing this view described feeling reassured by pharmacists acknowledging uncertainty, and that the pharmacist had spoken with a GP before the review or would do so if necessary:

‘She was quite open and honest to say, “Well, I’ll speak to doctor on that and we’ll get back to you”, which I thought was reassuring … They’d … done their homework prior because she’d said the doctor and herself had … reviewed my medications separately and then met up, which I thought was good.’ (Pt39)

Seeking patient participation from the outset Around two-thirds of survey responders agreed that the reviewer made it ‘clear that a decision needed to be made’ (n = 368, 66.2%) and wanted to know

‘how the patient wanted to be involved in decision making’ (n = 359, 64.6%) (Figure 1).

In the interviews, participants did not specifically talk about whether their pharmacists had asked how they wanted to be involved in decision making. However, audio-recordings of reviews showed that pharmacists sought patient involvement from the outset, with some using the pre-review leaflet (which practices were asked to provide to patients) as a mechanism of encouraging patient participation:

‘The plan is for us to chat around your drugs and just to see how you’re getting on and whether you’ve got anything you would like to get out of chatting to me … I’ve got a few ideas. I’ve had a chat to [GP name] and we looked at your meds but I’m happy for you to drive it

… how you wish.’ (Pharmacist [Ph2], audio-recording of review)

Provision of information Nearly four-fifths (n = 439, 79.0%) of survey responders reported that their reviewer was good–excellent at ‘explaining things clearly’ (Table 4). Reviewers were generally considered to have helped the patient understand the information provided during the review (n = 387, 69.6% agree and n = 116, 20.9% disagree), but there was less agreement that reviewers explained different treatment options available (n = 298, 53.6% agree and n = 202, 36.3% disagree) and the corresponding advantages and disadvantages (n = 324, 58.3% agree and n = 176, 31.7% disagree) (Figure 1).

Interviewees spoke mostly of wanting information about the effectiveness and need for their medication, and

1. Made clear that a decision needed made

2. Asked re. involvement in decision making

3. Different treatment options provided

4. Explained advantages/disadvantages

5. Helped to understand all information

6. Asked which treatment option preferred

7. Weighed different treament options

8. Selected treatment option together

9. Agreement on how to proceed

0% 20% 40% 60%

Survey responders, %

Completely disagree Strongly disagree Somewhat disagree

Strongly agree Completely agreeSomewhat agree

Missing

80% 100%

Figure 1. Patient-reported shared decision making (SDM-Q-9 instrument) original wording of questions is provided in Supplementary Information S3. re. = regarding. SDM-Q-9 = 9-item Shared Decision Making Questionnaire.

Research

British Journal of General Practice, January 2026 RESEARCH | e951

to better understand their medications and health. These participants welcomed the openness of the pharmacist to their questions and felt they received comprehensive and reassuring responses:

‘Information is the thing that I got. I just thought it was a good idea to make sure … that I’m not … taking any medica- tion that I don’t need to … also that the medications that I’m taking don’t become a cocktail in any way.’ (Pt39)

Discussion of patient priorities A majority of responders felt that ‘a lot’ or ‘every’ effort was made to listen to what mattered to them (n = 334, 60.1%), to include this within

Table 3. How the medication review was delivered (N = 556)a

Characteristic n (%)

Duration of review, min

<10 130 (23.4)

10 to 20 285 (51.3)

>20 124 (22.3)

Was the IMPPP medication review long enough?

No 33 (5.9)

Yes 453 (81.5)

Not sure 54 (9.7)

Happy with where review appointment took place?

No 54 (9.7)

Yes 444 (79.9)

Not sure 37 (6.7)

Professional conducting the review

GP 71 (12.8)

Pharmacist 381 (68.5)

Don’t know 77 (13.8)

Previously received care from the reviewer

No 413 (74.3)

Yes 126 (22.7)

Level of confidence in reviewer

Not at all or a little confident 96 (17.3)

Reasonably confident 144 (25.9)

Mostly or very confident 303 (54.5)

aPercentages based on total survey responders; percentages do not sum to 100% owing to missing data. IMPPP = Improving Medicines use in People with Polypharmacy in Primary Care.

the decision-making process (n = 318, 57.2%), and that they had been asked what they wanted to discuss during the review (n = 341, 61.3%) (Table 4).

Audio-recordings of reviews showed that throughout the review, pharmacists actively sought to elicit patient preferences in relation to each medication by checking whether the medication was being taken regularly, alongside when and how, and the patient’s understanding of what it was for. Many interviewees appreciated this approach, describing feeling heard and understood, and comfortable to express concerns and preferences:

‘She listened to me … took everything on board that I said … so the two things I had queries about we got aired at the very beginning.’ (Pt22)

Several others, however, expressed frustration when they felt the pharmacist had not fully listened or understood their concerns, and failed to consider their personal circumstances. As a result, they felt they lacked the advice and support they needed, particularly regarding alternative treatment and non-pharmacological options:

‘I feel that they could’ve done more ... whether there was something else I could do … different that would help me … but there weren’t … none of that. It was just literally, why are you taking this … carry on with that.’ (Pt23)

Agreement on decisions Just over half of survey responders (n = 296, 53.2%) reported that the review led to a decision to change medications, with three-quarters (n = 222/296, 75.0%) satisfied with the change. There was no evidence (P = 0.90) that overall satisfaction was associated with whether or not a change had been made (see Supplementary Table S1). Around half of responders agreed that their reviewer had asked them what their preferred treatment option was and had weighed the different treatment options (n = 300, 54.0%) (Figure 1). Just over half (n = 315, 56.7%) agreed that they had selected the treatment option together with the reviewer, with around two-thirds (n = 384, 69.1%) reporting reaching an agreement on how to proceed.

Collaborative decision making was observed within most of the audio- recordings of reviews, which showed that following proposals to change medicines pharmacists tended to seek patient involvement using questions such as ‘how does that sound?’ (Ph2/ Ph21), ‘what would you like to do?’ (Ph1/ Ph19/Ph14/Ph10), or ‘shall we see how that goes?’ (Ph3).

Many interviewees valued the opportunity to discuss the pros and cons of medication changes and fully participate in decisions, including whether to continue or modify their current treatment regimen:

‘I’m always open to discussion and debate, which is … why I’m going to go onto this new type of statin which [the pharmacist] thinks is more effective. You listen and learn … It’s a two-way conversation all the time.’ (Pt25)

A few, however, conveyed a preference for expert-driven decision making and did not feel the need to ask questions as they viewed their reviewer as the ‘expert’ (Pt27).

When a medication change was suggested by the pharmacist, interviewees spoke of needing to fully understand the reasons behind the proposal before accepting it. Those who accepted a therapeutic change felt reassured that they would receive ongoing support from the pharmacist and could reverse the decision later if they wished:

‘I could have said no. She asked me, not told me. She said can I reduce [esome- prazole] … and try it out for a month? But I don’t mind … if they try it out, because sometimes you get used to a medicine and it don’t do you any good anyway.’ (Pt28)

Those who did not want to make changes to their current treatment regimen felt confident rejecting the pharmacist-recommended changes because they saw no need for adjustments or feared changes might disrupt their health or routine. Audio-recordings showed that when patients declined these changes, pharmacists respected their decision and recommended revisiting this discussion later:

Research

e952 | RESEARCH British Journal of General Practice, January 2026

‘We’re going to reduce the [zopiclone quantity] but we’re not reducing the dose … and we just made a note there that we’ll discuss the zopiclone again in the future.’ (Ph3, audio-recording of review)

Summary provision and follow-up arrangements Only a small minority of survey responders reported receiving a written treatment plan (n = 44, 7.9%) or having a follow-up appointment arranged (n = 115, 20.7%) (Table 4).

Although few interviewees mentioned receiving a treatment plan, a verbal summary of next steps, including discussion of follow-up, was observed within most of the recorded reviews. Additionally, audio-recordings demonstrated that pharmacists checked patient understanding and reaffirmed decision making before concluding the review:

‘We’ve come up with ... an action plan ... You told me that you weren’t getting on [with drug brand] … We talked

about [alternative dosing strategy] … We checked your blood pressure and it’s lovely [and have decided to stop one drug and arrange further monitor- ing] ... we’ll meet again or perhaps do it over the phone, if that’s alright?’ (Ph4, audio-recording of review)

Several interviewees talked about planned next steps for implementing medication changes and/or monitoring progress, and described feeling reassured by the ongoing support offered by the pharmacist:

Table 4. Review decision-making process (N = 556)a

Survey question n (%)

Information provision

Explaining things clearly (CARE instrument)

Poor or fair 77 (13.8)

Good 119 (21.4)

Very good or excellent 320 (57.6)

Discussion of priorities

Effort to listen to things that matter most to patient (CollaboRATE instrument)

No effort 21 (3.8)

A little or some effort 165 (29.7)

A lot or every effort 334 (60.1)

Effort to include what matters to patient in decision-making (CollaboRATE instrument)

No effort 40 (7.2)

A little or some effort 159 (28.6)

A lot or every effort 318 (57.2)

Reviewer asked what patient would like to discuss (PACIC instrument)

No 129 (23.2)

Yes 341 (61.3)

Not sure 67 (12.1)

Agreeing on decisions

Decisions to change medicines was made during review

No 243 (43.7)

Yes 296 (53.2)

Satisfaction with changes madeb

Dissatisfied or very dissatisfied 23 (7.8)

Neither dissatisfied nor satisfied 48 (16.2)

Satisfied or very satisfied 222 (75.0)

Summarising and follow-up

Patient provided with a copy of my treatment plan (PACIC instrument)

No 450 (80.9)

Yes 44 (7.9)

Has a follow-up appointment been arranged?

No 374 (67.3)

Yes 115 (20.7)

Not sure 53 (9.5)

aPercentages based on total survey responders; percentages do not sum to 100% owing to missing data. bPercentages reported only for responders who reported a change in medication being made, n = 296. CARE = Consultation And Relational Empathy measure. PACIC = Patient Assessment of Chronic Illness Care measure.

Research

British Journal of General Practice, January 2026 RESEARCH | e953

‘She made a note for me and gave it to me at the end sort of saying … what we discussed like taking my blood pressure, not taking that tablet … I’m gonna get a follow-up … she’s going to phone me in three weeks … to see how my blood pressure is. I think that is quite impres- sive.’ (Pt10)

Discussion

Summary Overall, survey feedback reflected high levels of patient satisfaction that was positively associated with perceptions of shared decision making. In general, key elements of shared decision making were evident in all three data sources, although some participants expressed frustration that their concerns were not fully addressed. Patients described feeling comfortable to decline pharmacist-recommended changes, and pharmacists respected patient autonomy in doing so. A desire to avoid disruption, or lack of compelling justification for change, were key reasons cited for declining reviewers’ recommendations.

Patient preparation for the review was key to the value patients placed on it and patients’ engagement; practices’ failure to implement the pre-review leaflet as intended had a negative impact on patient experience. Reviews delivered by telephone were generally considered acceptable, albeit with some disadvantages.

Delivery of reviews by pharmacists was generally well received, despite patients often not knowing the clinician and concerns about pharmacists’ authority to alter prescribing; prior discussion between pharmacist and GP was considered reassuring, and familiarity with the clinician was valued. Although verbal summaries of reviews were provided, written treatment plans were not usually given to patients and follow-up was infrequent.

Strengths and limitations The mixed-methods approach and triangulation of different datasets provided in-depth understanding of patient experience and satisfaction with the IMPPP medication review intervention. The findings provide valuable insights pertinent

to intervention implementation, independent of the trial effectiveness outcomes. However, several factors might limit generalisability of the findings.

The survey response rate was high at 72.5%, but findings may suffer from response bias if individuals with negative experiences were less likely to participate. Additionally, individuals who declined or did not attend the patient- facing review appointments may differ systematically from those included in the survey.

Survey responders and interviewees predominantly identified as White British with lower levels of socioeconomic deprivation, meaning that diverse demographic and cultural perspectives were not fully represented. Furthermore, the data were gathered as part of a process evaluation in a trial of a complex intervention comprising a four-step medication review process, and not within usual general practice. Although interviewees commented on the convenience of telephone consultations, data on mode of delivery were incomplete and potentially unreliable, so a formal analysis of its association with satisfaction was not conducted. As with all interview studies, it is not possible to fully exclude bias introduced by the interview process itself. However, efforts were made to minimise this risk (for example, experienced interviewers, standardised interview protocols, and dual-coding of transcripts) and we are confident the insights obtained are authentic.

Comparison with existing literature A recent review of telehealth medication reviews shows high levels of patient satisfaction, suggesting these services are feasible, save costs, and improve care.19 In the current study, satisfaction with largely pharmacist-led reviews, which were mostly conducted via telephone, were similarly high, suggesting that patient needs and expectations were met. Nevertheless, limitations of telephone consultations are well recognised20 and this was reflected in the current study where, despite acknowledging the convenience of telephone consultation, some patients noted the lost opportunity for more open and comprehensive assessment.

Structured medication reviews are a central element of UK medicines optimisation policy, but patients do not always experience these reviews as patient centred.21 We have shown that reviews delivered in the context of support to deliver patient-centred care (such as relevant clinician training and pre-review preparation) are perceived positively and as patient centred by patients.

Patients in the current study were more satisfied when they experienced shared decision making, and valued being adequately informed, prepared for the interaction, and receiving patient- centred communication incorporating their priorities and preferences. This aligns with a recent scoping review showing considerable variation in the level of information patients desire about their medications, underlying the need to personalise communication and information during medication reviews.22

Moreover, inclusion of patients' personal priorities and preferences in care enhances effective patient–practitioner communication making patients feel heard, understood, respected, and involved in their treatment.23 This leads to increased confidence in treatment plans, greater satisfaction with care, and higher levels of trust in healthcare providers.5,24–27

Implications for practice Within the broader strategy of medicines optimisation, medication reviews are one of several approaches aimed at ensuring that medicines use is safe, effective, and aligned with patient priorities, although evidence of clinical benefit remains limited.28–30 Existing guidelines for structured medication review in England emphasise that reviews should be conducted by trained clinicians working within their scope of practice using the principles of shared decision making to underpin conversations with patients.3,31,32 Our findings suggest that the IMPPP model of medication review delivery helps achieve patient-centred care, shared decision making, and effective communication, and thus has the potential to improve experience of care and meet policy objectives.

Our findings point to the importance of patients having confidence in their clinician, yet a substantial minority

Research

e954 | RESEARCH British Journal of General Practice, January 2026

of patients report this is lacking and interviews additionally pointed towards uncertainty regarding pharmacists’ clinical authority. Improving patient awareness about the role of the growing clinical pharmacist workforce may thus lead to better patient experience, and is to be encouraged. How this should be achieved remains uncertain, although our results suggest improved collaborative working between GPs and pharmacists may be one approach of value.

To enhance patient experience and satisfaction, it is also important that shared decision-making principles are integrated throughout the review process. The current study shows patients value being properly prepared but also suggests that the conclusion of reviews is suboptimal, lacking written action plans or clear follow-up. Practice-level policies should therefore include standards for sharing of pre-review information and providing post-review written summaries and follow-up. Pre-review notifications should detail the reviewer’s scope of practice and the review agenda, and prompt patients to reflect on their medications and identify what they want to discuss. Post-review summaries should clearly document agreed modifications, therapeutic goals, and guidance for monitoring treatment effectiveness, managing issues, and attending follow-up appointments.

By taking these issues into consideration, medication review policy and practice has the potential to improve shared decision making and patient satisfaction, empowering patients to more actively engage in their review and enhancing patient experience.

Funding

The Improving Medicines use in People with Polypharmacy in Primary Care study was funded by the National Institute for Health and Care Research (NIHR) under its Health and Social Care Delivery Research programme (grant reference number: 16/118/14). Carolyn A Chew-Graham is part- funded by NIHR Applied Research Collaboration West Midlands. The views expressed are those of the

authors and not necessarily those of the NIHR or the Department of Health and Social Care. The funders had no role in study design, data collection, and analysis, decision to publish, or preparation of the manuscript.

Ethical approval

Ethical approval for the study was obtained from the Wales NHS Research Ethics Committee (REC) 6 and the Health Research Authority (REC reference number: 19/WA/0090)

Data

All data requests should be submitted to the corresponding author for consideration. Access to anonymised data may be granted following review via the University of Bristol Research Data Storage Facility.

Provenance

Freely submitted; externally peer reviewed.

Competing interests

The authors have declared no competing interests.

Contributors Lorna J Duncan and Deborah McCahon are joint first authors.

Open access

This article is Open Access: CC BY 4.0 licence (http://creativecommons.org/ licences/by/4.0/).

Discuss this article: bjgp.org/letters

References 1. Payne RA. The epidemiology of

polypharmacy. Clin Med (Lond) 2016; 16(5): 465–469. DOI: https://doi.org/10.7861/ clinmedicine.16-5-465

2. Rochon PA, Petrovic M, Cherubini A, et al. Polypharmacy, inappropriate prescribing, and deprescribing in older people: through a sex and gender lens. Lancet Healthy Longev 2021; 2(5): e290–e300. DOI: https://doi. org/10.1016/S2666-7568(21)00054-4

3. National Institute for Health and Care Excellence (NICE). Medicines optimisation: the safe and effective use of medicines to enable the best possible outcomes. NG5. London: NICE, 2015. https://www.nice.org. uk/guidance/ng5 (accessed 10 Nov 2025).

4. Picton C, Wright H. Medicines optimisation: helping patients to make the most of medicines. 2013. https://www.rpharms.com/ Portals/0/RPS%20document%20library/ Open%20access/Policy/helping-patients- make-the-most-of-their-medicines.pdf (accessed 10 Nov 2025).

5. Joosten EA, DeFuentes-Merillas L, de Weert GH, et al. Systematic review of the effects of shared decision-making on patient satisfaction, treatment adherence and health status. Psychother Psychosom 2008; 77(4): 219–226. DOI: https://doi.org/10.1159/ 000126073

6. Loh A, Leonhart R, Wills CE, et al. The impact of patient participation on adherence and clinical outcome in primary care of depression. Patient Educ Couns 2007; 65(1): 69–78. DOI: https://doi.org/10.1016/j.pec. 2006.05.007

7. McCahon D, Duncan P, Payne R, Horwood J. Patient perceptions and experiences of medication review: qualitative study in general practice. BMC Prim Care 2022; 23(1): 293.

8. Duncan P, Cabral C, McCahon D, et al. Efficiency versus thoroughness in medication review: a qualitative interview study in UK primary care. Br J Gen Pract 2019; DOI: https://doi.org/10.3399/bjgp19X701321

9. Stewart D, Madden M, Davies P, et al. Structured medication reviews: origins, implementation, evidence, and prospects. Br J Gen Pract 2021; DOI: https://doi.org/10. 3399/bjgp21X716465

10. Payne RA, Blair PS, Caddick B, et al. Improving Medicines use in People with Polypharmacy in Primary Care (IMPPP): protocol for a multicentre cluster randomised trial comparing a complex intervention for medication optimization against usual care. NIHR Open Res 2022; 2:54. DOI: https://doi.org/10.3310/ nihropenres.13285.1

11. Doyle C, Lennox L, Bell D. A systematic review of evidence on the links between patient experience and clinical safety and effectiveness. BMJ Open 2013; 3(1): e001570. DOI: https://doi.org/10.1136/ bmjopen-2012-001570

12. Parslow RM, Duncan LJ, Caddick B, et al. Collaborative discussions between gps and pharmacists to optimise patient medication: a qualitative study within a UK primary care clinical trial. Br J Gen Pract 2024; DOI: https: //doi.org/10.3399/BJGP.2024.0190

13. Makoul G, Clayman ML. An integrative model of shared decision making in medical encounters. Patient Educ Couns 2006; 60(3): 301–312. DOI: https://doi.org/10.1016/j.pec. 2005.06.010

14. Agency for Healthcare Research and Quality. The SHARE approach. 2024. https://www. ahrq.gov/sdm/share-approach/index.html (accessed 10 Nov 2025).

15. Kriston L, Scholl I, Hölzel L, et al. The 9-item Shared Decision Making Questionnaire (SDM-Q-9). Development and psychometric properties in a primary care sample. Patient Educ Couns 2010; 80(1): 94–99. DOI: https:/ /doi.org/10.1016/j.pec.2009.09.034

16. Barr PJ, Thompson R, Walsh T, et al. The psychometric properties of CollaboRATE: a fast and frugal patient-reported measure of the shared decision-making process. J Med Internet Res 2014; 16(1): e2. DOI: https:// doi.org/10.2196/jmir.3085

17. Mercer SW, Maxwell M, Heaney D, Watt GC. The Consultation And Relational Empathy (CARE) measure: development and

Research

British Journal of General Practice, January 2026 RESEARCH | e955

https://bjgp.org/letters

https://doi.org/10.7861/clinmedicine.16-5-465

https://doi.org/10.1016/S2666-7568(21)00054-4

https://www.nice.org.uk/guidance/ng5

https://www.rpharms.com/Portals/0/RPS%20document%20library/Open%20access/Policy/helping-patients-make-the-most-of-their-medicines.pdf

https://doi.org/10.1159/000126073

https://doi.org/10.1016/j.pec.2006.05.007

https://doi.org/10.3399/bjgp19X701321

https://doi.org/10.3399/bjgp21X716465

https://doi.org/10.3310/nihropenres.13285.1

https://doi.org/10.1136/bmjopen-2012-001570

https://doi.org/10.3399/BJGP.2024.0190

https://doi.org/10.1016/j.pec.2005.06.010

https://www.ahrq.gov/sdm/share-approach/index.html

https://doi.org/10.1016/j.pec.2009.09.034

https://doi.org/10.2196/jmir.3085

preliminary validation and reliability of an empathy-based consultation process measure. Fam Pract 2004; 21(6): 699–705. DOI: https://doi.org/10.1093/fampra/ cmh621

18. Glasgow RE, Wagner EH, Schaefer J, et al. Development and validation of the patient assessment of chronic illness care (PACIC). Med Care 2005; 43(5): 436–444. DOI: https://doi.org/10.1097/01.mlr.0000160375. 47920.8c

19. Shafiee Hanjani L, Caffery LJ, Freeman CR, et al. A scoping review of the use and impact of telehealth medication reviews. Res Social Adm Pharm 2020; 16(8): 1140–1153. DOI: https://doi.org/10.1016/j.sapharm.2019.12. 014

20. McKinstry B, Hammersley V, Burton C, et al. The quality, safety and content of telephone and face-to-face consultations: a comparative study. Qual Saf Health Care 2010; 19(4): 298–303. DOI: https://doi.org/ 10.1136/qshc.2008.027763

21. Madden M, Mills T, Atkin K, et al. Early implementation of the structured medication review in England: a qualitative study. Br J Gen Pract 2022; DOI: https://doi. org/10.3399/BJGP.2022.0014

22. Kusch MK, Haefeli WE, Seidling HM. How to meet patients’ individual needs for drug

information - a scoping review. Patient Prefer Adherence 2018; 12:2339–2355. DOI: https: //doi.org/10.2147/PPA.S173651

23. Street RJ, Makoul G, Arora NK, Epstein RM. How does communication heal? Pathways linking clinician-patient communication to health outcomes. Patient Educ Couns 2009; 74(3): 295–301. DOI: https://doi.org/10. 1016/j.pec.2008.11.015

24. Fogarty LA, Curbow BA, Wingard JR, et al. Can 40 seconds of compassion reduce patient anxiety? J Clin Oncol 1999; 17(1): 371–379. DOI: https://doi.org/10.1200/JCO. 1999.17.1.371

25. Świątoniowska-Lonc N, Polański J, Tański W, Jankowska-Polańska B. Impact of satisfaction with physician-patient communication on self-care and adherence in patients with hypertension: cross-sectional study. BMC Health Serv Res 2020; 20(1): 1046. DOI: https://doi.org/10.1186/s12913-020-05912- 0

26. Ha JF, Longnecker N. Doctor-patient communication: a review. Ochsner J 2010; 10(1): 38–43.

27. Chandra S, Mohammadnezhad M, Ward PR. Trust and communication in a doctor– patient relationship: a literature review. J Healthc Commun 2018; 3(3). DOI: https:// doi.org/10.4172/2472-1654.100146

28. Cole JA, Gonçalves-Bradley DC, Alqahtani M, et al. Interventions to improve the appropriate use of polypharmacy for older people. Cochrane Database Syst Rev 2023; 10(10): CD008165. DOI: https://doi.org/10. 1002/14651858.CD008165.pub5

29. Wallerstedt SM, Kindblom JM, Nylén K, et al. Medication reviews for nursing home residents to reduce mortality and hospitalization: systematic review and meta- analysis. Br J Clin Pharmacol 2014; 78(3): 488–497. DOI: https://doi.org/10.1111/bcp. 12351

30. McCahon D, Denholm RE, Huntley AL, et al. Development of a model of medication review for use in clinical practice: Bristol medication review model. BMC Med 2021; 19(1): 262. DOI: https://doi.org/10.1186/ s12916-021-02136-9

31. NHS England. Network Contract Directed Enhanced Service. Structured medication reviews and medicines optimisation: guidance. 2021. https://www.england.nhs.uk/wp- content/uploads/2021/03/B0431-network- contract-des-smr-and-mo-guidance-21-22. pdf (accessed 10 Nov 2025).

32. NICE. Shared decision making. NG197. London: NICE, 2021. https://www.nice.org. uk/guidance/ng197 (accessed 10 Nov 2025).

Research

e956 | RESEARCH British Journal of General Practice, January 2026

https://doi.org/10.1093/fampra/cmh621

https://doi.org/10.1097/01.mlr.0000160375.47920.8c

https://doi.org/10.1016/j.sapharm.2019.12.014

https://doi.org/10.1136/qshc.2008.027763

https://doi.org/10.3399/BJGP.2022.0014

https://doi.org/10.2147/PPA.S173651

https://doi.org/10.1016/j.pec.2008.11.015

https://doi.org/10.1200/JCO.1999.17.1.371

https://doi.org/10.1186/s12913-020-05912-0

https://doi.org/10.4172/2472-1654.100146

https://doi.org/10.1002/14651858.CD008165.pub5

https://doi.org/10.1111/bcp.12351

https://doi.org/10.1186/s12916-021-02136-9

https://www.england.nhs.uk/wp-content/uploads/2021/03/B0431-network-contract-des-smr-and-mo-guidance-21-22.pdf

https://www.nice.org.uk/guidance/ng197

Patients’ experiences of a patient-centred polypharmacy medication review intervention:

Introduction
Method

Trial population, eligibility criteria, and recruitment
Intervention description
Audio-recording of reviews and patient interviews
Survey

Results

Participant characteristics
Overall satisfaction and perceptions of patient-centred care
Patient expectations
How reviews were delivered
Patient–clinician relationship
Seeking patient participation from the outset
Provision of information
Discussion of patient priorities
Agreement on decisions
Summary provision and follow-up arrangements

Discussion

Summary
Strengths and limitations
Comparison with existing literature
Implications for practice

image9.emf

RESEARCH

the bmj | BMJ 2023;381:e074054 | doi: 10.1136/bmj-2022-074054 1

Optimising prescribing in older adults with multimorbidity and polypharmacy in primary care (OPTICA): cluster randomised clinical trial Katharina Tabea Jungo,1 Anna-Katharina Ansorg,1 Carmen Floriani,1 Zsofia Rozsnyai,1 Nathalie Schwab,1,2 Rahel Meier,3 Fabio Valeri,3 Odile Stalder,4 Andreas Limacher,4 Claudio Schneider,2 Michael Bagattini,5 Sven Trelle,4 Marco Spruit,6,7,8 Matthias Schwenkglenks,9,10 Nicolas Rodondi,1,2 Sven Streit1

AbstrAct Objective To study the effects of a primary care medication review intervention centred around an electronic clinical decision support system (eCDSS) on appropriateness of medication and the number of prescribing omissions in older adults with multimorbidity and polypharmacy compared with a discussion about medication in line with usual care. Design Cluster randomised clinical trial. setting Swiss primary care, between December 2018 and February 2021. ParticiPants Eligible patients were ≥65 years of age with three or more chronic conditions and five or more long term medications. interventiOn The intervention to optimise pharmacotherapy centred around an eCDSS was conducted by general practitioners, followed by shared decision making between general practitioners and patients, and was compared with a discussion about medication in line with usual care between patients and general practitioners. Main OutcOMe Measures Primary outcomes were improvement in the Medication Appropriateness Index (MAI) and the Assessment of Underutilisation (AOU) at 12 months. Secondary outcomes included number of medications,

falls, fractures, and quality of life. results In 43 general practitioner clusters, 323 patients were recruited (median age 77 (interquartile range 73-83) years; 45% (n=146) women). Twenty one general practitioners with 160 patients were assigned to the intervention group and 22 general practitioners with 163 patients to the control group. On average, one recommendation to stop or start a medication was reported to be implemented per patient. At 12 months, the results of the intention-to-treat analysis of the improvement in appropriateness of medication (odds ratio 1.05, 95% confidence interval 0.59 to 1.87) and the number of prescribing omissions (0.90, 0.41 to 1.96) were inconclusive. The same was the case for the per protocol analysis. No clear evidence was found for a difference in safety outcomes at the 12 month follow-up, but fewer safety events were reported in the intervention group than in the control group at six and 12 months. cOnclusiOns In this randomised trial of general practitioners and older adults, the results were inconclusive as to whether the medication review intervention centred around the use of an eCDSS led to an improvement in appropriateness of medication or a reduction in prescribing omissions at 12 months compared with a discussion about medication in line with usual care. Nevertheless, the intervention could be safely delivered without causing any harm to patients. trial registratiOn NCT03724539Clinicaltrials.gov NCT03724539

Introduction Inappropriate polypharmacy in older adults is a major driver of healthcare related harm.1 2 It is associated with negative health outcomes, such as adverse drug events, falls, and functional decline in activities of daily living.3-6 Patients with multiple chronic conditions (multimorbidity7) and polypharmacy, defined as the use of five or more drugs,8 are at an increased risk of inappropriate polypharmacy, such as inappropriate prescribing and prescribing omissions.9-12 This highlights the need for reducing inappropriate polypharmacy, and medication reviews represent one approach to this. Primary care settings, characterised by long term patient-provider relationships, lend themselves as ideal settings for medication reviews. Conducting medication reviews is, however, complex

For numbered affiliations see end of the article Correspondence to: S Streit [email protected] (or @Sven_Streit on Twitter; ORCID 0000-0002-3813-4616) Additional material is published online only. To view please visit the journal online. cite this as: BMJ 2023;381:e074054 http://dx.doi.org/10.1136/ bmj-2022-074054

Accepted: 06 April 2023

WhAt Is AlreAdy knoWn on thIs topIc Inappropriate prescribing is highly prevalent in older adults with multimorbidity and polypharmacy and has been associated with adverse health outcomes Medication review interventions might contribute to reducing inappropriate prescribing The evidence on medication review interventions based on electronic clinical decision support systems in primary care settings is limited

WhAt thIs study Adds The structured medication review intervention based on an electronic clinical decision support system led to the implementation of certain prescribing recommendations However, the findings as to whether the intervention led to a greater appropriateness of patients’ prescriptions overall were inconclusive

P ro

tected b

y co p

yrig h

t, in clu

d in

g fo

r u ses related

to text an

d d

ata m in

in g

, A I train

in g

, an d

sim ilar tech

n o

lo g

ies. .

b y g

u est

n 18 S

ep tem

b er 2026

ttp s://w

w w

.b m

j.co m

/ D

o w

n lo

ad ed

fro m

24 M

ay 2023. 10.1136/b

m j-2022-074054 o

B M

J: first p u

b lish

ed as

mailto:[email protected]

https://twitter.com/Sven_Streit

https://orcid.org/0000-0002-3813-4616

https://dx.doi.org/10.1136/bmj-2022-074054

https://crossmark.crossref.org/dialog/?doi=10.1136/bmj-2022-074054&domain=pdf&date_stamp=2023-05-16

https://www.bmj.com/

RESEARCH

2 doi: 10.1136/bmj-2022-074054 | BMJ 2023;381:e074054 | the bmj

and time consuming, and the evidence for medication review interventions is mixed.13-15

Medication review interventions, such as those based on the Screening Tool of Older Persons potentially inappropriate Prescription (STOPP) and Screening Tool to Alert doctors to the Right Treatment (START) criteria—evidence based criteria to inform prescribing in older adults16 17—can support general practitioners to optimise prescribing. These criteria have been shown to be effective in improving quality of prescribing and some patient outcomes.18 19 In the context of digitalisation, use of electronic clinical decision support systems (eCDSS) based on the STOPP/ START criteria, such as the Systematic Tool to Reduce Inappropriate Prescribing Assistant (STRIPA), is a promising way forward.20-22 In recent years, different eCDSS for optimising drug use have been tested.23-25 Systems based on the STOPP/START criteria were used in two randomised clinical trials conducted in the inpatient setting.19 26 However, evidence from primary care settings is lacking.

The Optimising PharmacoTherapy In the multimorbid elderly in primary CAre (OPTICA) trial tested the hypothesis that, in older adults with multimorbidity and polypharmacy, the use of an eCDSS for optimising drug therapy by general practitioners improves appropriateness of medication and reduces prescribing omissions compared with standard care.

Methods trial design The protocol of the OPTICA trial was published previously.27 We conducted a cluster randomised controlled trial with 43 general practitioners as clusters.28 The trial was conducted from 2018 to 2021.

recruitment and participants General practitioners had to be participating in the Family medicine ICPC Research using Electronic medical records (FIRE) project throughout the trial.29 This allowed electronic data exports from their practices to the trial database and the eCDSS.

General practitioners recruited eight to 10 eligible patients (supplementary figure A) by using random screening lists generated from their practice’s electronic health record data.30 If needed, more than one screening list with 20 patients each was provided to general practitioners. General practitioners obtained written informed consent from all participants or their legal representatives before enrolment in the study. For patients with cognitive impairment, written informed consent was obtained from their legal representative.

Patients were aged ≥65 years, were taking five or more long term medications (≥90 days) and had at least three chronic conditions on the basis of on ICPC-2 (international classification of primary care, 2nd edition) coding or general practitioners’ clinical judgement. To maximise the generalisability of the study population, we kept exclusion criteria to a minimum. Exclusion criteria were inability to provide consent and participation in a different intervention study.

randomisation Clusters were randomised after enrolment of patients for each cluster was completed. Participating general practitioners were randomised centrally in a web based system (REDCap).31 32 We used a one to one ratio with unstratified block randomisation and randomly varying block sizes of two and four.

trial procedures Intervention group As previously reported,27 general practitioners used the intervention at the individual patient level. The intervention consisted of a structured six step medication review using STRIPA, a web based electronic clinical decision support system based on the STOPP/ START criteria version 2 (appendix 1).17 33 For the purpose of the OPTICA trial, STRIPA was adapted to the primary care setting (for example, use of ICPC-2 codes instead of ICD (international classification of diseases) codes for the coding of diagnoses). In addition to detecting potential overuse, underuse, and misuse of drugs, STRIPA generated recommendations to prevent drug-drug interactions and inappropriate dosages. The one time intervention consisted of six steps. (1) Data on medications, chronic conditions, laboratory values, and vital data were imported to STRIPA. (2) General practitioners verified and adapted the recorded information. (3) General practitioners used the drag/ drop function to link medications and conditions. (4) General practitioners ran the medication review. (5) General practitioners decided which recommendations to move forward with. (6) At the next appointment, general practitioners implemented shared decision making with patients. General practitioners in the intervention group received a training video and written material on how to use STRIPA and conduct the shared decision making.

Control group Patients in the control group had a discussion about medication with their general practitioner in line with usual care. General practitioners were asked not to deviate from their usual practice.

blinding General practitioners were blinded during the screening and recruitment of patients to limit biased selection of patients. General practitioners in the control group remained partially blinded, as they did not know the intervention procedure. Patients in the control group remained blinded owing to the discussion with their general practitioner. The data collectors and study assessors were fully blinded. Blinding of the trial statistician was not feasible because the data export contained information on the study groups.

Outcomes Primary outcome measures Appropriateness of medication was the primary outcome. To account for the multi-dimensionality of this construct and to capture both over-prescribing

P ro

tected b

y co p

yrig h

t, in clu

d in

g fo

r u ses related

to text an

d d

ata m in

in g

, A I train

in g

, an d

sim ilar tech

n o

lo g

ies. .

b y g

u est

n 18 S

ep tem

b er 2026

ttp s://w

w w

.b m

j.co m

/ D

o w

n lo

ad ed

fro m

24 M

ay 2023. 10.1136/b

m j-2022-074054 o

B M

J: first p u

b lish

ed as

https://www.bmj.com/

RESEARCH

the bmj | BMJ 2023;381:e074054 | doi: 10.1136/bmj-2022-074054 3

and under-prescribing, we used two primary outcome measures: the Medication Appropriateness Index (MAI) and the Assessment of Underutilisation (AOU) (both at 12 months). The MAI allows assessment of the appropriateness of prescriptions, and the AOU measures the number of prescribing omissions.34-36 We assessed the AOU for each non-acute condition of the patients and the MAI for each long term medication. We used the 10 item version of the MAI. However, we excluded the MAI’s cost effectiveness item for feasibility reasons (appendix 2). This resulted in a score from 0 to 17 for each medication (with a higher score representing greater inappropriateness). We defined improvement in the MAI conservatively as a decrease of ≥1 points for all medications used by the patient, which represents an increase in appropriateness of medication. For non-acute diagnoses, the AOU assessed no prescribing omission, a marginal omission (for example, use of non-drug treatment), or an omission of an indicated medication.37-39 Improvement in the AOU was a reduction of ≥1 prescribing omissions considering all chronic conditions of the patient. The inter-rater reliability assessments showed moderate agreement regarding the MAI assessments (agreement of ratings=69%, Kendall’s coefficient of concordance W40=0.61, Cohen’s κ=0.54) and substantial agreement regarding the AOU assessments (agreement of ratings=94%, Kendall’s coefficient of concordance W=0.77), which validated the use of both assessments.40

Secondary outcomes Main secondary outcomes were patients’ long term medications (degree of polypharmacy), appropriateness of medication (measured by the ordinal MAI), the number of prescribing omissions, the number of falls and fractures, and quality of life as measured by the telephone version of the EQ-5D- 5L questionnaire.41 42 The percentage of prescribing recommendations accepted was a secondary outcome reported to describe fidelity of implementation. Next, we assessed patients’ willingness to have medications deprescribed at baseline and reported it in this manuscript. The secondary outcomes for the health economic analyses conducted alongside the trial and reported separately were health services use (formal care received—for example, number of general practitioner and specialist visits), informal care received (for example, unpaid care work by relatives/ friends), survival, quality adjusted life years,43 and direct medical costs accrued in one year.

Data collection We collected data at baseline, six months, and 12 months. The data on medications, diagnoses, laboratory values, and vital signs were imported from the electronic health records of patients through the FIRE database.29 Owing to variation in data exports from different electronic health records software programs used by general practitioners (reporting of medications and diagnoses at every encounter versus reporting

only when a change was made in the record), fewer medications and diagnoses were recorded for some patients for the observed study period. However, this does not mean that patients did not fulfil the inclusion and exclusion criteria, as general practitioners had screened and verified those as part of the recruitment process. Owing to the substantial amount of missing data in the variables needed to assess the primary outcomes (~35%), the study team collected missing information from participating general practitioners. After this additional data collection, 13% and 18% of patients had missing information on MAI and AOU improvement, respectively, between baseline and follow-up 2 (supplementary tables A and B). We collected data on quality of life, health services use, and falls and fractures through phone calls with patients or legal representatives. General practitioners reported safety information on adverse events including death. We asked general practitioners in the intervention group to report information on the implementation (or not) of prescribing recommendations. All data were coded and kept confidential.

sample size calculation We calculated the sample size needed to test for superiority of the two primary outcome measures and used the Bonferroni approach to account for multiple testing. We assumed that 35% and 60% of patients would have an improvement in the MAI and that 10% and 30% would have an improvement in the AOU in the control and intervention group, respectively. On the basis of a two sample comparison of proportions, a pre-specified number of general practitioner clusters of 40 (20 per arm), and a conservative intracluster correlation coefficient of 0.05 (values of 0.01-0.05 are typically found for binary outcomes in older people44), we needed seven patients per cluster to detect a difference in the proportion of improvement in the MAI score of 25% between the two groups with a power of 90% at a two sided α level of 0.025. Using the same assumption for the AOU, we also needed seven patients per cluster to detect a difference of 20%. This results in a total sample size of 280 patients (140 per arm). This sample size provides 81% power to detect a significant improvement in both the MAI score and the AOU index. To account for attrition due to dropout or death (15% estimated), we enlarged the number of patients per cluster to eight to 10, with a final sample size of 320 patients total (160 per group).

statistical analysis We described the sociodemographic characteristics of general practitioners. We described the characteristics of patients and presented them by group. We analysed the number of prescribing recommendations generated and implemented descriptively. In all the model based analyses, we used multiple imputed data. We used multiple imputation by chained equations to impute missing values in co-primary outcomes (MAI score and prescribing omissions measured using the AOU index) at baseline, six months, and 12 months. Imputation

P ro

tected b

y co p

yrig h

t, in clu

d in

g fo

r u ses related

to text an

d d

ata m in

in g

, A I train

in g

, an d

sim ilar tech

n o

lo g

ies. .

b y g

u est

n 18 S

ep tem

b er 2026

ttp s://w

w w

.b m

j.co m

/ D

o w

n lo

ad ed

fro m

24 M

ay 2023. 10.1136/b

m j-2022-074054 o

B M

J: first p u

b lish

ed as

https://www.bmj.com/

RESEARCH

4 doi: 10.1136/bmj-2022-074054 | BMJ 2023;381:e074054 | the bmj

models were based on all baseline characteristics of the patients (age, gender, education level, smoking status, alcohol consumption, permanent nursing home

stay, number of falls, number of hospital admissions, number of chronic medications, number of chronic conditions, quality of life predicted utilities at baseline,

Number of GPs (clusters) recruited for OPTICA trial

Excluded Did not meet inclusion criteria Declined to participate Other reasons

509 58

143

Number of GPs

Received allocated intervention Did not receive allocated intervention/ receipt of intervention unknown*

133 27

Number of patients 160

Withdrawal from study Lost to follow-up Patient deceased Other reason Opted out from phone follow-up

0 1 2 3 1

Lost to follow-up at 6 months

6 month follow-up completed by phone‡ 147

Number of GPs

Received usual care review Did not receive usual care review Receipt of usual care review unknown†

150 1

Number of patients 163

710

Number of patients assessed for eligibility

Withdrawal from study Lost to follow-up Patient deceased Other reason Opted out from phone follow-up

0 5 2 3 0

Lost to follow-up at 6 months

6 month follow-up completed by phone‡ 150

1033

Randomised 323

Withdrawal from study Lost to follow-up Patient deceased Other reason Opted out from phone follow-up

0 4 4 0 4

Lost to follow-up at 12 months

12 month follow-up completed by phone‡ 147

Withdrawal from study Lost to follow-up Patient deceased Other reason Opted out from phone follow-up

1 2 4 1 0

Lost to follow-up at 12 months

12 month follow-up completed by phone‡ 147

Did not receive allocated intervention Number of recorded chronic conditions <3 Number of recorded medications <5

13 27 47

Did not receive allocated intervention Cluster size <4 Number of recorded chronic conditions <3 Number of recorded medications <5

13 5

23 36

Strict per protocol analysis§

Intention-to-treat analysis 160

Strict per protocol analysis§

Intention-to-treat analysis 163

9797

12 8

Fig 1 | cOnsOrt patient flowchart. *For these patients, drag/drop function in systematic tool to reduce inappropriate Prescribing assistant (striPa) had not been used or was reset after intervention. †reasons are that patients did not see their general practitioner (gP) or had other urgent healthcare needs that had to be prioritised. ‡referring to follow-up calls by phone. time windows for these phone calls were +15 days at baseline, +/–30 days at six month follow-up, and +/–30 days at 12 month follow-up. For all patients, except those who withdrew from study, available data from Fire database could be used (provided that patient continued seeing same gP). §Multiple criteria can apply. Multiple imputed data were used for analyses

P ro

tected b

y co p

yrig h

t, in clu

d in

g fo

r u ses related

to text an

d d

ata m in

in g

, A I train

in g

, an d

sim ilar tech

n o

lo g

ies. .

b y g

u est

n 18 S

ep tem

b er 2026

ttp s://w

w w

.b m

j.co m

/ D

o w

n lo

ad ed

fro m

24 M

ay 2023. 10.1136/b

m j-2022-074054 o

B M

J: first p u

b lish

ed as

https://www.bmj.com/

RESEARCH

the bmj | BMJ 2023;381:e074054 | doi: 10.1136/bmj-2022-074054 5

table 1 | characteristics of general practitioners (gPs) and patients at baseline. values are numbers (percentages) unless stated otherwise

all clusters control group intervention group General practitioners (clusters) (n=43) (n=22) (n=21) No of patients per cluster: <4 2 (5) 2 (9) 0 (0) 4-7 9 (21) 3 (14) 6 (29) ≥8 32 (74) 17 (77) 15 (71) Practice location: Rural 21 (49) 12 (55) 9 (43) Urban/suburban 22 (51) 10 (45) 12 (57) Median (IQR) age, years 53 (44-58) 54 (44-55) 51 (44-58) Gender: Male 34 (79) 16 (73) 18 (86) Female 9 (21) 6 (27) 3 (14) Median (IQR) work experience as GP, years 14 (7-21) 16 (7-21) 12 (8-22) Median (IQR) No of consultations per workday 25 (20-28) 25 (22-30) 25 (20-25) Practice form: Individual practice 7 (16) 5 (23) 2 (10) Group practice 36 (84) 17 (77) 19 (90) Median (IQR) No of GPs in group practices 3 (2-4) 3 (2-4) 4 (3-5) Patients (n=323) (n=163) (n=160) Age categories: 65-74 years 111 (34) 58 (36) 53 (33) 75-84 150 (46) 75 (46) 75 (47) ≥85 62 (19) 30 (18) 32 (20) Median (IQR) age, years 77 (73-83) 77 (73-83) 77 (73-83) Gender: Male 177 (55) 88 (54) 89 (56) Female 146 (45) 75 (46) 71 (44) Highest education level: Less than mandatory schooling 4 (1) 1 (1) 3 (2) Mandatory schooling 118 (37) 62 (38) 56 (35) High school degree or apprenticeship 145 (45) 70 (43) 75 (47) University or equivalent 45 (14) 23 (14) 22 (14) Other 2 (1) 0 (0) 2 (1) Smoking status: Smoker 33 (10) 14 (9) 19 (12) Past smoker 140 (43) 65 (40) 75 (47) Non-smoker 142 (44) 77 (47) 65 (41) Alcohol consumption in 6 months before study enrolment 196 (61) 97 (60) 99 (62) Median (IQR) alcohol consumption in 6 months before study enrolment, units per week

2 (1-7) 2 (1-7) 2 (1-7)

Permanent stay in nursing home 24 (7) 9 (6) 15 (9) Cognitive impairment 12 (4) 7 (4) 5 (3) Falls in 6 months before study enrolment 58 (18) 27 (17) 31 (19) Median (IQR) No of falls in 6 months before study enrolment 1 (1-1) 1 (1-1) 1 (1-1) Median (IQR) No of hospital admissions in 6 months before study enrolment

0 (0-0) 0 (0-0) 0 (0-0)

Median (IQR) No of long term medications at baseline 7 (4-10) 8 (5-10) 7 (4-10) Median (IQR) No of chronic conditions at baseline 7 (4-10) 7 (4-10) 6 (4-9) Median (IQR) quality of life* 75 (60-80) 70 (60-80) 75 (60-80) Median (IQR) EQ-5D-5L utilities at baseline 0.89 (0.80-0.96) 0.89 (0.81-0.94) 0.89 (0.77-0.97) Unwilling to have medications deprescribed† 36 (11) 19 (12) 17 (11) Median (IQR) total MAI‡ score at baseline 12 (2-38) 5 (0-38) 15 (4-38) Median (IQR) averaged MAI score at baseline§ 1.9 (0.2-5.2) 0.6 (0.0-4.9) 3.0 (0.5-5.4) Median (IQR) total No of prescribing omissions¶ 1 (0-2) 1 (0-2) 1 (0-1) Median (IQR) averaged No of prescribing omissions** 0.1 (0.0-0.2) 0.1 (0.0-0.2) 0.1 (0.0-0.2) IQR=interquartile range; MAI=Medication Appropriateness Index. Missing values: 0 for all GP characteristics, except for number of consultations per day (2%); 0 for all patient characteristics except for education level (3%), smoking status (2%), smoking consumption (2%), living situation (2%), falls (4%), number of hospital admissions (3%), quality of life (8%), both primary outcomes MAI and Assessment of Underutilisation (AOU) at baseline (7%). *Measured by visual analogue scale of European Quality of Life-5 Dimensions questionnaire (EQ-VAS). Values range from 0 to 100, with higher values indicating higher quality of life. †Measured by agreement with statement “I would be willing to stop one or more of my medicines if my doctor said it was possible” of revised Patients’ Attitudes Towards Deprescribing (rPATD).50 Dichotomized by agree/strongly agree (yes) versus don’t know/disagree/strongly disagree (no). Results are presented only for patients for whom patient version of rPATD was used. ‡Adapted from Samsa, et al.36 Sum of all MAI scores of each medication. Each individual MAI score ranges from 0 to 17, with higher values indicating greater inappropriateness. §Averaged by number of chronic medications. ¶Measured by AOU based on Jeffery et al.38 Count of prescribing omissions for chronic conditions at baseline. **Averaged by the number of chronic conditions.

P ro

tected b

y co p

yrig h

t, in clu

d in

g fo

r u ses related

to text an

d d

ata m in

in g

, A I train

in g

, an d

sim ilar tech

n o

lo g

ies. .

b y g

u est

n 18 S

ep tem

b er 2026

ttp s://w

w w

.b m

j.co m

/ D

o w

n lo

ad ed

fro m

24 M

ay 2023. 10.1136/b

m j-2022-074054 o

B M

J: first p u

b lish

ed as

https://www.bmj.com/

RESEARCH

6 doi: 10.1136/bmj-2022-074054 | BMJ 2023;381:e074054 | the bmj

EQ-5D visual analogue scale, and patients’ willingness to have medications deprescribed) and the general practitioners’ characteristics (randomisation group, practice form, practice size, canton of the practice, general practitioners’ work experience in years, and practice location). We did not account for within cluster correlation in the imputation model. We used these imputed variables to impute all the secondary outcomes at six and 12 months. We used predictive mean matching (pmm) and logit models to impute non- binary and binary variables, respectively. We generated 50 imputed datasets, which we analysed using Rubin’s rules to combine results across datasets.45

We tested the primary outcomes separately, declaring success if at least one was statistically significant at the Bonferroni corrected, two sided α level of 0.025. We used generalised estimating equation models with robust standard errors to account for clustered data, which yields population averaged effects.46-48 For binary outcomes, we calculated odds ratios by using a binomial distribution and a logit link. For count outcomes, we calculated incidence rate ratios by using a negative binomial distribution and a log link. For continuous outcomes, we calculated mean differences by using a Gaussian distribution and an identity link. We did several pre-specified subgroup comparisons at general practitioner and patient level through the models described above (appendix 3), with additional interaction terms between randomisation group and binary subgroup indicators. Odds ratios, incidence rate ratios, or mean differences for randomised comparisons are shown for each subgroup together with a P value for interaction. For the comparison of safety outcomes, we used a robust generalised estimating equation model with a binomial distribution and a logit link.

In secondary analyses, we re-ran all models with the per protocol set of patients. Here, we excluded all patients for whom fewer than three chronic conditions and fewer than five long term medications had been recorded during the study period, as well as all clusters

that included fewer than four participants. In addition, we did a post hoc relaxed per protocol analysis in which only patients with no chronic condition and no long term medication recorded and clusters with fewer than four patients were excluded. In an additional sensitivity analysis, we adjusted models for potential confounders because cluster randomisation may lead to imbalances in baseline characteristics between groups. Finally, we did an aggregated data analysis at the general practitioner level.

Owing to overdispersion of the secondary count outcomes, a deviation from the protocol had to be made and was prespecified in the statistical analysis plan. We used a robust generalised estimating equation model with a negative binomial distribution and a log link rather than the too restrictive random effects Poisson model. Furthermore, the generalised estimating equation approach is more consistent with the other analyses than is the random effects model. We considered continuous secondary outcomes (MAI score, AOU index) as count data and analysed them using a generalised estimating equation model with a negative binomial distribution and a log link owing to their skewed distribution. As the MAI score consists only of integers with a high proportion of zeros, its distribution cannot be considered as gaussian. Further minor deviations from the statistical analysis plan are reported in appendix 4. We used Stata version 17.0 for all analyses. We report the results in line with the Consolidated Standards of Reporting Trials extension for cluster trials.49

Patient and public involvement No patients were involved in setting the research question or the outcome measures. General practitioners and patients aged ≥65 years with multimorbidity and polypharmacy were represented in the Safety and Data Monitoring Board. General practitioners and patients who participated in the trial received newsletters throughout the trial.

table 2 | recommendations generated by systematic tool to reduce inappropriate Prescribing assistant (striPa) to optimise prescribing in older adults (n=133)*

estimate range Generated prescribing recommendations (n=704) No (%) patients with ≥1 prescribing recommendation† 130/133 (98) - Mean (SD) No of prescribing recommendations per patient† 5.4 (3.2) 1-21 Mean (SD) No of START or STOPP recommendations per patient‡ 3.7 (1.8) 0-11 STOPP recommendations 2.3 (1.3) 0-7 START recommendations 1.3 (1.2) 0-6 Implementation of prescribing recommendations§ At patient level: No (%) patients with ≥1 prescribing recommendation reported to have been implemented† 31/53 (58) - Mean (SD) No of recommendations reported to have been implemented per patient† 1.0 (1.2) - At recommendation level: No (%) STOPP recommendations implemented 31/112 (28) - No (%) START recommendations implemented 11/77 (14) - SD=standard deviation; START=Screening Tool to Alert to Right Treatment; STOPP=Screening Tool of Older Persons’ Prescriptions. *Information collected from STRIPA for 133/160 patients in intervention group for whom drag/drop function in STRIPA had been used as part of intervention. All patients for whom information could be retrieved from STRIPA had ≥1 recommendation to start or stop one or several of their medications. †Includes recommendations to stop and start medications, adapt dosage of potentially inappropriate prescriptions, or flag drug-drug interactions. ‡All patients for whom information could be retrieved from STRIPA had ≥1 recommendation to start or stop one or several of their medications. §This information was reported by 7 general practitioners from OPTICA intervention group about 53 patients, which explains lower denominator.

P ro

tected b

y co p

yrig h

t, in clu

d in

g fo

r u ses related

to text an

d d

ata m in

in g

, A I train

in g

, an d

sim ilar tech

n o

lo g

ies. .

b y g

u est

n 18 S

ep tem

b er 2026

ttp s://w

w w

.b m

j.co m

/ D

o w

n lo

ad ed

fro m

24 M

ay 2023. 10.1136/b

m j-2022-074054 o

B M

J: first p u

b lish

ed as

https://www.bmj.com/

RESEARCH

the bmj | BMJ 2023;381:e074054 | doi: 10.1136/bmj-2022-074054 7

results Forty three general practitioners working in Swiss primary care practices participated in the OPTICA trial (supplementary figure A).30 Between January 2019 and February 2020, 323 older adults with multimorbidity and polypharmacy (median age 77 (interquartile range 73-83) years; 45% women) gave their informed consent. Twenty one general practitioners with 160 patients were randomised to the intervention group and 22 general practitioners with 163 patients to the control group (fig 1). During the 12 month follow-up period, 12 (4%) patients died, 12 (4%) patients were lost to follow-up and could not be reached by phone for the data collection, five (2%) patients opted out of the data collection by phone but agreed to be followed-up via the FIRE database, and one patient withdrew from the study.

Participating general practitioners had a median age of 53 (44-58) years and median work experience of 14 (7-21) years, and 21% were female. At baseline, patients had a median number of long term medications of 7 (4-10) and a median number of chronic diagnoses of 7 (4-10). The baseline characteristics of patients were similar in the two groups, except for appropriateness of medication, which was greater in the control group (as shown by the lower total MAI) (table 1).

An average of 5.4 prescribing recommendations per patient were generated, of which an average of 3.7 were STOPP or START recommendations (table 2). On average, 1.0 START or STOPP recommendations were implemented per patient in the intervention group, with 58.5% of patients having had at least one recommendation implemented. Table 3 shows the most common prescribing recommendations made by STRIPA. General practitioners most commonly reported the following reasons why prescribing recommendations were not implemented: belief that the current prescriptions were beneficial, the recommendation was not suitable for patients, and bad experiences with previous medication changes (supplementary table C).

Primary outcome measures In all, 82% (n=265) of patients had information on the MAI both at baseline and at the 12 month follow- up, and 87% (282) had information on prescribing omissions as assessed using the AOU for both time points (supplementary table A). Thirty seven per cent (119) of patients had an improvement in the MAI between baseline and the 12 month follow-up, and 12% (38) had an improvement in the number of prescribing omissions (supplementary table B).

table 3 | Most common prescribing recommendations generated by systematic tool to reduce inappropriate Prescribing assistant of 704 recommendations made recommendation type Description16 Frequency (%) START I1 Seasonal trivalent influenza vaccine annually 121 (17.2) STOPP B6 Loop diuretic as first line treatment for hypertension (safer, more effective alternatives available) 38 (5.4) START A2 Aspirin (75-160 mg once daily) in presence of chronic atrial fibrillation, in which vitamin K antagonists or direct

thrombin inhibitors or factor Xa inhibitors are contraindicated 33 (4.7)

START E3 Vitamin D supplement in patients with known osteoporosis and previous fragility fracture(s) and/or bone mineral density T scores more than −2.0 at multiple sites

21 (3.0)

START E4 Bone antiresorptive or anabolic therapy (eg, bisphosphonate, strontium ranelate, teriparatide, denosumab) in patients with documented osteoporosis, in whom no pharmacological or clinical status contraindication exists (bone mineral density T scores >2.5 at multiple sites) and/or previous history of fragility fracture(s)

18 (2.6)

START A3 Antiplatelet therapy (aspirin or clopidogrel or prasugrel or ticagrelor) with documented history of coronary, cerebral, or peripheral vascular disease

17 (2.4)

START B1 Regular inhaled β2 agonist or antimuscarinic bronchodilator (eg, ipratropium, tiotropium) for mild to moderate asthma or COPD

15 (2.1)

START A7 β blocker with ischaemic heart disease. 14 (2.0) STOPP C3 Aspirin, clopidogrel, dipyridamole, vitamin K antagonists, direct thrombin inhibitors, or factor Xa inhibitors with

concurrent significant bleeding risk (ie, uncontrolled severe hypertension, bleeding diathesis, recent non-trivial spontaneous bleeding) (high risk of bleeding)

14 (2.0)

START A6 ACE inhibitor with systolic heart failure and/or documented coronary artery disease 12 (1.7) STOPP B11 ACE inhibitors or ARBs in patients with hyperkalaemia 9 (1.3) STOPP B9 Aldosterone antagonists (eg, spironolactone, eplerenone) with concurrent potassium conserving drugs (eg, ACE

inhibitors, ARBs, amiloride, triamterene) without monitoring of serum potassium (risk of dangerous hyperkalaemia ( >6.0 mmol/L); serum potassium should be monitored regularly (ie, at least every 6 months))

8 (1.1)

START H2 Laxatives in patients receiving opioids regularly 8 (1.1) STOPP H2 Non-steroidal anti-inflammatory drug in patients with established hypertension (risk of exacerbation of hypertension) or

heart failure (risk of exacerbation of heart failure) 7 (1.0)

START E2 Bisphosphonates and vitamin D and calcium in patients taking long term systemic corticosteroid therapy 7 (1.0) START B3 Home continuous oxygen with documented chronic hypoxaemia (ie, pO2<8.0 kPa or 60 mm Hg or SaO2<89). 6 (0.9) STOPP C6 Antiplatelet agents with vitamin K antagonist, direct thrombin inhibitor, or factor Xa inhibitors in patients with stable

coronary, cerebrovascular, or peripheral arterial disease without clear indication for anticoagulant therapy (no added benefit from dual therapy)

5 (0.7)

STOPP B4 β blocker with symptomatic bradycardia (<50/min), type II heart block, or complete heart block (risk of profound hypotension, asystole)

5 (0.7)

STOPP F3 Drugs likely to cause constipation (eg, antimuscarinic/anticholinergic drugs, oral iron, opioids, verapamil, aluminium antacids) in patients with chronic constipation in whom non-constipating alternatives are appropriate (risk of exacerbation of constipation)

5 (0.7)

ACE=angiotensin converting enzyme; ARB=angiotensin receptor blocker; COPD=chronic obstructive pulmonary disease. Recommendations that were generated ≥5 times during study are shown.

P ro

tected b

y co p

yrig h

t, in clu

d in

g fo

r u ses related

to text an

d d

ata m in

in g

, A I train

in g

, an d

sim ilar tech

n o

lo g

ies. .

b y g

u est

n 18 S

ep tem

b er 2026

ttp s://w

w w

.b m

j.co m

/ D

o w

n lo

ad ed

fro m

24 M

ay 2023. 10.1136/b

m j-2022-074054 o

B M

J: first p u

b lish

ed as

https://www.bmj.com/

RESEARCH

8 doi: 10.1136/bmj-2022-074054 | BMJ 2023;381:e074054 | the bmj

The development of the MAI score and the number of prescribing omissions are shown by time point in supplementary figures B-D and supplementary table D.

The analyses compared the intervention and control groups. In the intention-to-treat analysis, the odds ratio for an improvement in the MAI score (decrease by ≥1 point) between baseline and the 12 month follow- up was 1.05 (95% confidence interval 0.59 to 1.87) and the odds ratio for an improvement in the AOU (≥1 prescribing omission less) was 0.90 (0.41 to 1.96) (table 4). The sensitivity analysis adjusting for baseline MAI and AOU values (supplementary table E), the sensitivity analysis adjusting for subgroup variables (supplementary table E), the subgroup analyses (supplementary figure E), the aggregated data analysis (supplementary table F), and the analysis based on the available case data (supplementary table G) also showed inconclusive results.

In the strict per protocol analysis, the adjusted odds ratio for an improvement in the MAI score between baseline and the 12 month follow-up was 1.03 (0.59 to 1.77) and the odds ratio for an improvement in the AOU was 1.25 (0.44 to 3.56) (supplementary tables H and I). The relaxed per protocol analysis produced similar results.

secondary outcomes The results on whether the medication review intervention centred around the use of an eCDSS led to an improvement in the secondary outcomes, in the primary outcomes at the six month follow-up (table 4), in the per protocol analyses (supplementary table J), or in additional medication related outcomes (supplementary table K) were inconclusive. We did not find clear evidence for a difference in safety outcomes at the 12 month follow-up, but fewer safety events occurred in the intervention group at six and 12 months (supplementary table L).

discussion In this cluster randomised clinical trial evaluating the effect of a structured medication review intervention supported by an electronic decision support system in older adults with multimorbidity and polypharmacy, 58.5% of patients had at least one prescribing recommendation implemented. Despite the implementation of one STOPP or START recommendation per patient on average, the results on the appropriateness of medication and the number of prescribing omissions at the 12 month follow-up compared with usual care were inconclusive, with

table 4 | comparison of outcomes between intervention and control groups control (n=163) intervention (n=160) effect size (95% ci) P value

Primary outcomes No (%) with improvement in MAI score between baseline and 12 month follow-up

67 (41); multiple imputation used for 20/163 patients

68 (43); multiple imputation used for 21/160 patients

OR 1.05 (0.59 to 1.87) 0.87

No (%) with improvement in number of prescribing omissions between baseline and 12 month follow-up

28 (17); multiple imputation used for 30/163 patients

24 (15); multiple imputation used for 28/160 patients

OR 0.90 (0.41 to 1.96) 0.79

Secondary outcomes Medication related secondary outcomes: No (%) with improvement in MAI score between baseline and 6 month follow-up

64 (39) 68 (43) OR 1.14 (0.60 to 2.14) 0.69

No (%) with improvement in number of prescribing omissions between baseline and 6 month follow-up

28 (17) 19 (12) OR 0.67 (0.31 to 1.46) 0.31

Mean (95% CI) MAI total score at 6 month follow-up 22 (17 to 27) 25 (20 to 29) IRR 1.36 (0.89 to 2.08)* 0.15 Mean (95% CI) MAI total score at 12 month follow-up 25 (19 to 30) 26 (21 to 30) IRR 1.15 (0.74 to 1.79)* 0.53 Mean (95% CI) total number of prescribing omissions at 6 month follow-up

1.1 (0.9 to 1.3) 1.1 (0.9 to 1.3) IRR 1.06 (0.83 to 1.36)* 0.64

Mean (95% CI) total number of prescribing omissions at 12 month follow-up

1.2 (1.0 to 1.3) 0.9 (0.7 to 1.1) IRR 0.83 (0.65 to 1.07)* 0.15

Mean (95% CI) number of medications at 6 month follow-up 7.6 (7.0 to 8.2) 7.2 (6.6 to 7.9) MD 0.05 (−0.87 to 0.97)* 0.91 Mean (95% CI) number of medications at 12 month follow-up 8.0 (7.4 to 8.7) 7.8 (7.2 to 8.4) MD 0.26 (−0.64 to 1.16)* 0.58 Patient reported secondary outcomes: Mean (95% CI) number of falls at 6 month follow-up 0.3 (0.1 to 0.4) 0.3 (0.2 to 0.4) IRR 0.96 (0.50 to 1.84) 0.89 Mean (95% CI) number of falls at 12 month follow-up 0.2 (0.1 to 0.3) 0.2 (0.1 to 0.3) IRR 0.90 (0.50 to 1.64) 0.74 No (%) with any fracture(s) between baseline and 6 month follow-up†

4 (3) 3 (2) OR 0.72 (0.17 to 3.10) 0.66

No (%) with any fracture(s) between baseline and 12 month follow-up†

2 (1) 3 (2) OR 1.51 (0.27 to 8.50) 0.64

Quality of life: Mean (95% CI) EQ-5D-5L utilities at 6 month follow-up‡ (inverse predicted utility)

0.2 (0.2 to 0.2) 0.2 (0.1 to 0.2) MD −0.03 (−0.07 to 0.01)* 0.13

Mean (95% CI) EQ-5D-5L utilities at 12 month follow-up‡ (inverse predicted utility)

0.1 (0.1 to 0.2) 0.1 (0.1 to 0.2) MD 0.00 (−0.04 to 0.03)* 0.93

Mean (95% CI) VAS at 6 month follow-up§ 71 (68 to 74) 72 (69 to 74) MD 0.53 (−2.99 to 4.06)* 0.77 Mean (95% CI) VAS at 12 month follow-up§ 73 (70 to 75) 72 (70 to 75) MD −0.42 (−3.77 to 2.93)* 0.81 AOU=Assessment of Underutilisation; CI=confidence interval; IRR=incident rate ratio; MAI=Medication Appropriateness Index; MD=mean difference; OR=odds ratio; VAS=visual analogue scale. This table is based on multiple imputed data. Additional descriptive information on secondary outcomes by study time point can be found in supplementary table M. *Adjusted for respective baseline score. †Owing to low number of fractures, only binary variable (yes/no) was considered. ‡Calculated based on German value set for EQ-5D-5L by Ludwig et al.51. §Measured by VAS of European Quality of Life-5 Dimensions questionnaire (EQ-VAS); values range from 0 to 100 with higher values indicating higher quality of life.

P ro

tected b

y co p

yrig h

t, in clu

d in

g fo

r u ses related

to text an

d d

ata m in

in g

, A I train

in g

, an d

sim ilar tech

n o

lo g

ies. .

b y g

u est

n 18 S

ep tem

b er 2026

ttp s://w

w w

.b m

j.co m

/ D

o w

n lo

ad ed

fro m

24 M

ay 2023. 10.1136/b

m j-2022-074054 o

B M

J: first p u

b lish

ed as

https://www.bmj.com/

RESEARCH

the bmj | BMJ 2023;381:e074054 | doi: 10.1136/bmj-2022-074054 9

odds ratios for the MAI and the AOU being close to one and confidence intervals being wide. However, the intervention could be safely delivered without causing harm to patients. The results on whether the medication review intervention led to an improvement in the secondary outcomes, such as the number of falls and fractures, were also inconclusive. We emphasise that given the age characteristics and the comorbidities of the study population, not all falls and fractures can be assumed to be an (adverse) effect of polypharmacy despite older adults with polypharmacy being at a higher risk for such events.52-54

strengths and limitations of study The OPTICA trial has several strengths. Recruitment of patients was based on screening lists with random samples of each general practitioner’s patient population. The trial had minimal exclusion criteria, which resulted in the recruitment of patients who were comparable to other older patients with multimorbidity and polypharmacy in Swiss primary care.30 The randomisation of general practitioner clusters once the recruitment per cluster was completed helped to limit differential, selective recruitment. However, owing to the Hawthorne effect, we cannot exclude the possibility that general practitioners recruited patients with more appropriate medication regimens or more favourable unmeasurable characteristics (for example, better patient-provider relationship). A low number of patients were lost to follow-up owing to the pragmatic data collection design consisting of phone calls and electronic health record data imports. The pragmatic nature of the trial in primary care settings provides real world insights. Finally, the collection of information related to the implementation of prescribing recommendations was insightful. However, we were unable to collect this information from all general practitioners.

The trial also has several limitations. Owing to a requirement from the regulatory authority, general practitioners in the control group had to have a discussion about medication with their enrolled patients in line with usual care. General practitioners were explicitly asked not to deviate from their usual practice. Despite this, we cannot rule out the possibility that some of them diverged from their usual prescribing practice owing to the Hawthorne effect. The number of medications, however, did not differ between the two groups at any time point. The quality of the data exports from the FIRE database was challenging because of differences in how electronic health record software programs exported data. Therefore, the study team had to manually collect missing information from practices and use multiple imputation methods. As the intervention occurred at a single time point, its effectiveness may have been diluted over time. The implementation of an average of one STOPP/START recommendation per patient did not seem to affect the primary outcomes. Next, the different barriers faced by patients and providers related to the implementation of prescribing recommendations,

which are numerous and well documented in the literature (for example, fears of negative health outcomes, alert fatigue),55 56 may have led to a low implementation of recommendations. An imbalance in the MAI existed at baseline, which we attribute to chance. Therefore, we also present adjusted outcome models in the supplementary material (supplementary table E). However, all other baseline variables, including the other medication related variables, were balanced, which indicates that randomisation of the clusters worked as intended. Next, in the sample size calculation, we chose a conservative intracluster correlation coefficient to be on the safe side. Finally, STRIPA used clinical information but did not consider patients’ preferences, which may have contributed to patients’ reluctance to implement recommendations.

comparison with other studies Our null findings are in line with the literature on previous clinical trials testing the effect of medication review interventions on appropriateness of medication in primary care settings. For instance, the PRIMUM trial, which randomised 72 primary care practices and 505 patients with multimorbidity and polypharmacy, did not find an improvement in the MAI after a medication review based on an eCDSS for general practitioners.57 This trial, however, did not assess under-prescribing and excluded patients with cognitive impairment. The results of the PRIMA-eDS trial, with 359 primary care practices and 3904 adults aged ≥75 years using multiple medications, showed that the use of an eCDSS did not lead to evidence for a between group difference in mortality or unplanned hospital admissions after a 24 month follow-up period.25 This study was strengthened by its longer observation period. However, it did not study medication underuse. In the OPTIMIZE trial, in which 3012 patients with dementia or mild cognitive impairment from 19 primary care clinics were randomised, the number of (potentially inappropriate) medications was similar in the two groups at the end of the six month follow-up period.58 These findings show that meaningful, sustainable improvements are difficult to achieve by the one-time use of an eCDSS to support medication reviews in primary care settings.

Our findings are in line with the results of previous multicentre trials evaluating an eCDSS based on the STOPP/START criteria. The SENATOR trial, in which 1536 inpatients were randomised, did not find any evidence for a between group difference in the recurrence of adverse drug events within 14 days of randomisation.26 Similarly, in the OPERAM trial, in which 2008 patients from 110 clusters were randomised and STRIPA was also used, the results were inconclusive as to whether the medication review intervention led to a reduction in readmission to hospital at the 12 month follow-up despite a trend towards a reduction in readmission and other clinical outcomes.19 The intervention came with potential cost savings of CHF3588 (£3229; €3666; $4065) per patient and a gain of 0.025 (95% confidence interval

P ro

tected b

y co p

yrig h

t, in clu

d in

g fo

r u ses related

to text an

d d

ata m in

in g

, A I train

in g

, an d

sim ilar tech

n o

lo g

ies. .

b y g

u est

n 18 S

ep tem

b er 2026

ttp s://w

w w

.b m

j.co m

/ D

o w

n lo

ad ed

fro m

24 M

ay 2023. 10.1136/b

m j-2022-074054 o

B M

J: first p u

b lish

ed as

https://www.bmj.com/

RESEARCH

10 doi: 10.1136/bmj-2022-074054 | BMJ 2023;381:e074054 | the bmj

–0.002 to 0.052) quality adjusted life years per patient.59 Despite the more user-friendly application of the STOPP/START criteria as an eCDSS, difficulties related to implementing prescribing recommendations exist.

The low implementation rate of prescribing recommendations could, in part, account for the negative trial. However, the implementation rate was similar in previous trials. In the SENATOR trial, 15% of the overall recommendations were implemented.26 In the OPERAM trial, 62% of patients had at least one recommendation successfully implemented at two months.19 Expecting every prescribing recommendation to be implementable would be unreasonable, but the partial uptake of prescribing recommendations points to many factors that affect the overall effectiveness of medication review interventions. Our results show that the most common reasons why general practitioners reported not implementing the prescribing recommendations were that general practitioners thought that patients’ current medicines were beneficial and that recommended changes were not suitable. These findings are in line with the literature, which has shown that prescribing recommendations can be difficult to implement owing to many barriers faced by patients and prescribers.55 60 Furthermore, the types of prescribing recommendations generated by the eCDSS were relevant for older community dwelling adults, according to two recent systematic reviews.11 61 In the OPTICA trial, the recommendation to vaccinate patients against influenza was the most common recommendation. This was probably because influenza vaccines were not commonly recorded in patients’ medication lists.

implications of findings The STRIPA web based tool was not integrated into the electronic health record software programs used by general practitioners, which could explain the challenges in implementation. The future widespread and successful use of eCDSS in primary care settings requires that limitations, such as the lack of integration of eCDSS into existing practice software and clinical workflows, must be overcome. Furthermore, in Swiss primary care settings, it will be crucial to establish an industry standard that will allow reliable exports/ imports of data from electronic health record software to eCDSS, which was also one of the main challenges observed during the OPTICA trial and led to an increased expenditure of time by general practitioners who had to update the data manually in the eCDSS.

Future trials on interventions to optimise medication would benefit from interventions focusing on overcoming challenges to implementation. This not only requires preparatory qualitative studies to better understand the challenges but also use of implementation science strategies to integrate interventions into clinical workflows, tailoring interventions to users’ needs, and piloting interventions. Future interventions may benefit from being designed as repeated interventions to

accommodate the dynamic prescribing practices and frequent medication changes in older patients with multimorbidity and polypharmacy.

conclusions In this primary care based trial, 58.5% of patients had at least one prescribing recommendation implemented. Despite this, the results as to whether the medication review intervention centred around the use of an eCDSS led to an improvement in appropriateness of medication or a reduction in prescribing omissions at 12 months compared with a discussion about medication in line with usual care were inconclusive. Nevertheless, the intervention could be safely delivered without causing any harm to patients.

authOr aFFiliatiOns 1Institute of Primary Health Care (BIHAM), University of Bern, Bern, Switzerland 2Department of General Internal Medicine, Inselspital, Bern University Hospital, University of Bern, Bern, Switzerland 3Institute of Primary Care, University of Zurich and University Hospital Zurich, Zurich, Switzerland 4CTU Bern, University of Bern, Bern, Switzerland 5mfe Haus and Kinderärzte Schweiz, Bern, Switzerland 6Department of Information and Computing Sciences, Utrecht University, Utrecht, Netherlands 7Public Health and Primary Care (PHEG), Leiden University Medical Center, Leiden University, Leiden, Netherlands 8Leiden Institute of Advanced Computer Science (LIACS), Faculty of Science, Leiden University, Leiden, Netherlands 9Institute of Pharmaceutical Medicine (ECPM), University of Basel, Basel, Switzerland 10Epidemiology, Biostatistics and Prevention Institute (EBPI), University of Zurich, Zurich, Switzerland We thank the general practitioners who participated in the OPTICA trial for their contributions, without which conducting this study would have been impossible. We also thank the participants for consenting to participate in our study. We thank Sophie Mantelli, Axel Löwe, Fanny Lindemann, Jeanne Moor, Heinz Bhend, and Lamia Elloumi for their support with the conduct of the OPTICA trial; Shana Bertato, Rebekah Hoeks, Dominic Salvisberg, Renata Lüthold, Yasmin Mohd Yatim, Melisa Harmanci, and Linda Chau for their support with the data entry; and Kristie Weir for her editorial suggestions. KTJ is a member of the Junior Investigator Intensive Program of the US Deprescribing Research Network, which is funded by the National Institute on Ageing. Contributors: NR, SS, MSchwenkglenks, ST, MS, CS, and KTJ designed the trial. KTJ, AKA, CF, ZR, RM, NS, FV, MB, ST, MSpruit, MSchwenkglenks, NR, and SS contributed to the acquisition, analysis, and interpretation of data. KTJ and OS wrote the first draft of the manuscript. All other authors provided feedback and approved the final version of the manuscript. OS, AL, and ST provided statistical support and did the statistical analyses. NR, MSchwenkglenks, and SS obtained the funding for the OPTICA trial. NS and FV provided administrative and technical support. NR and SS supervised the conduct of the trial. KTJ and OS had full access to all the data in the study and are the guarantors. The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted. Funding: The OPTICA trial was funded by the Swiss National Science Foundation, within the framework of the National Research Programme 74 “Smarter Health Care” (NRP74) under contract number 407440_167465 (to SS, NR, and MSchwenkglenks). The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Competing interests: All authors have completed the ICMJE uniform disclosure form at https://www.icmje.org/disclosure-of-interest/ and declare: financial support from grants from the Swiss National Science Foundation; MSpruit reports a settlement agreement between Spru IT and Utrecht University, in which all intellectual property

P ro

tected b

y co p

yrig h

t, in clu

d in

g fo

r u ses related

to text an

d d

ata m in

in g

, A I train

in g

, an d

sim ilar tech

n o

lo g

ies. .

b y g

u est

n 18 S

ep tem

b er 2026

ttp s://w

w w

.b m

j.co m

/ D

o w

n lo

ad ed

fro m

24 M

ay 2023. 10.1136/b

m j-2022-074054 o

B M

J: first p u

b lish

ed as

https://www.icmje.org/disclosure-of-interest/

https://www.bmj.com/

RESEARCH

the bmj | BMJ 2023;381:e074054 | doi: 10.1136/bmj-2022-074054 11

related to the Systematic Tool to Reduce Inappropriate Prescribing Assistant (STRIPA) is transferred to Utrecht University, in exchange for obtaining a free but non-exclusive right to provide STRIPA consultancy or support services, both on a commercial basis, and to update the STRIPA until June 2023; ST, AL, and OS are affiliated with CTU Bern, University of Bern, which has a staff policy of not accepting honorariums or consultancy fees; however, CTU Bern is involved in design, conduct, or analysis of clinical studies funded by not-for-profit and for profit organisations; in particular, pharmaceutical and medical device companies provide direct funding to some of these studies (for an up-to-date list of CTU Bern’s conflicts of interest see https://www. ctu.unibe.ch/research_projects/declaration_of_interest/index_eng. html); no other relationships or activities that could appear to have influenced the submitted work. Ethical approval: The study protocol of the OPTICA trial and other documentation was approved by the competent ethics committee of the canton of Bern (KEK), Switzerland, and the Swiss regulatory authority (Swissmedic) (BASEC ID: 2018–00914). The KEK and Swissmedic received annually safety reports and were informed about the end of the study. The OPTICA trial was performed in accordance with relevant regulations and guidelines. General practitioners and the study team obtained written informed consent from all participants or their legal representatives before study enrolment. Data sharing: The data for this study are available to other researchers on request. The data will be made available for scientific research purposes, after the proposed analysis plan has been approved. Data and documentation will be made available through a secure file exchange platform after approval of the proposal. In addition, a data transfer agreement must be signed (which defines obligations that the data requester must adhere to with regard to privacy and data handling). Deidentified participant data limited to the data used for the proposed project will be made available, along with a data dictionary and annotated case report forms. For data access, please contact the corresponding author. The lead authors (the manuscript’s guarantors) affirm that the manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as originally planned (and, if relevant, registered) have been explained. Dissemination to participants and related patient and public communities: The results of the OPTICA trial were presented at a national press conference in January 2023. The results of the OPTICA trial will be published in German and French in a journal that is commonly read by Swiss general practitioners. The submission of this article is planned for 2023. A lay summary of the OPTICA trial in German, French, and English can be found on the website of the NRP74 (https://www.nfp74.ch/de/3R0qxh8zc3RXcvXm/projekt/ projekt-streit). Provenance and peer review: Not commissioned; externally peer reviewed. This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/.

1 Thomas RE, Thomas BC. A Systematic Review of Studies of the STOPP/START 2015 and American Geriatric Society Beers 2015 Criteria in Patients ≥ 65 Years. Curr Aging Sci 2019;12:121-54. doi:1 0.2174/1874609812666190516093742

2 Maher RL, Hanlon J, Hajjar ER. Clinical consequences of polypharmacy in elderly. Expert Opin Drug Saf 2014;13:57-65. doi:10.1517/1474 0338.2013.827660

3 Xing XX, Zhu C, Liang HY, et al. Associations Between Potentially Inappropriate Medications and Adverse Health Outcomes in the Elderly: A Systematic Review and Meta-analysis. Ann Pharmacother 2019;53:1005-19. doi:10.1177/1060028019853069

4 Davies LE, Spiers G, Kingston A, Todd A, Adamson J, Hanratty B. Adverse Outcomes of Polypharmacy in Older People: Systematic Review of Reviews. J Am Med Dir Assoc 2020;21:181-7. doi:10.1016/j.jamda.2019.10.022

5 Mekonnen AB, Redley B, de Courten B, Manias E. Potentially inappropriate prescribing and its associations with health-related and system-related outcomes in hospitalised older adults: A systematic review and meta-analysis. Br J Clin Pharmacol 2021;87:4150-72. doi:10.1111/bcp.14870

6 Liew TM, Lee CS, Goh Shawn KL, Chang ZY. Potentially Inappropriate Prescribing Among Older Persons: A Meta-Analysis of Observational Studies. Ann Fam Med 2019;17:257-66. doi:10.1370/afm.2373

7 Johnston MC, Crilly M, Black C, Prescott GJ, Mercer SW. Defining and measuring multimorbidity: a systematic review of systematic reviews. Eur J Public Health 2019;29:182-9. doi:10.1093/eurpub/cky098

8 Masnoon N, Shakib S, Kalisch-Ellett L, Caughey GE. What is polypharmacy? A systematic review of definitions. BMC Geriatr 2017;17:230. doi:10.1186/s12877-017-0621-2

9 Nothelle SK, Sharma R, Oakes A, Jackson M, Segal JB. Factors associated with potentially inappropriate medication use in community-dwelling older adults in the United States: a systematic review. Int J Pharm Pract 2019;27:408-23. doi:10.1111/ijpp.12541

10 Tian F, Chen Z, Wu J. Prevalence of Polypharmacy and Potentially Inappropriate Medications Use in Elderly Chinese Patients: A Systematic Review and Meta-Analysis. Front Pharmacol 2022;13:862561. doi:10.3389/fphar.2022.862561

11 Tommelein E, Mehuys E, Petrovic M, Somers A, Colin P, Boussery K. Potentially inappropriate prescribing in community-dwelling older people across Europe: a systematic literature review. Eur J Clin Pharmacol 2015;71:1415-27. doi:10.1007/s00228-015-1954-4

12 Kuijpers MA, van Marum RJ, Egberts AC, Jansen PA, OLDY (OLd people Drugs & dYsregulations) Study Group. Relationship between polypharmacy and underprescribing. Br J Clin Pharmacol 2008;65:130-3. doi:10.1111/j.1365- 2125.2007.02961.x

13 Faulkner L, Hughes CM, Barry HE. Interventions to improve medicines optimisation in older people with frailty in primary care: a systematic review. Int J Pharm Pract 2022;30:297-304. doi:10.1093/ijpp/ riac036

14 Saeed D, Carter G, Parsons C. Interventions to improve medicines optimisation in frail older patients in secondary and acute care settings: a systematic review of randomised controlled trials and non- randomised studies. Int J Clin Pharm 2022;44:15-26. doi:10.1007/ s11096-021-01354-8

15 Rankin A, Cadogan CA, Patterson SM, et al. Interventions to improve the appropriate use of polypharmacy for older people. Cochrane Database Syst Rev 2018;9:CD008165. doi:10.1002/14651858. CD008165.pub4

16 O’Mahony D, O’Sullivan D, Byrne S, O’Connor MN, Ryan C, Gallagher P. STOPP/START criteria for potentially inappropriate prescribing in older people: version 2. Age Ageing 2015;44:213-8. doi:10.1093/ ageing/afu145

17 Gallagher P, Ryan C, Byrne S, Kennedy J, O’Mahony D. STOPP (Screening Tool of Older Person’s Prescriptions) and START (Screening Tool to Alert doctors to Right Treatment). Consensus validation. Int J Clin Pharmacol Ther 2008;46:72-83. doi:10.5414/ CPP46072

18 Hill-Taylor B, Walsh KA, Stewart S, Hayden J, Byrne S, Sketris IS. Effectiveness of the STOPP/START (Screening Tool of Older Persons’ potentially inappropriate Prescriptions/Screening Tool to Alert doctors to the Right Treatment) criteria: systematic review and meta-analysis of randomized controlled studies. J Clin Pharm Ther 2016;41:158-69. doi:10.1111/jcpt.12372

19 Blum MR, Sallevelt BTGM, Spinewine A, et al. Optimizing Therapy to Prevent Avoidable Hospital Admissions in Multimorbid Older Adults (OPERAM): cluster randomised controlled trial. BMJ 2021;374:n1585. doi:10.1136/bmj.n1585

20 Drenth-van Maanen AC, Leendertse AJ, Jansen PAF, et al. The Systematic Tool to Reduce Inappropriate Prescribing (STRIP): Combining implicit and explicit prescribing tools to improve appropriate prescribing. J Eval Clin Pract 2018;24:317-22. doi:10.1111/jep.12787

21 Meulendijk MC, Spruit MR, Drenth-van Maanen AC, et al. Computerized Decision Support Improves Medication Review Effectiveness: An Experiment Evaluating the STRIP Assistant’s Usability. Drugs Aging 2015;32:495-503. doi:10.1007/s40266- 015-0270-0

22 Meulendijk MC, Spruit MR, Willeboordse F, et al. Efficiency of Clinical Decision Support Systems Improves with Experience. J Med Syst 2016;40:76. doi:10.1007/s10916-015-0423-z

23 Kempen TGH, Bertilsson M, Hadziosmanovic N, et al. Effects of Hospital-Based Comprehensive Medication Reviews Including Postdischarge Follow-up on Older Patients’ Use of Health Care: A Cluster Randomized Clinical Trial. JAMA Netw Open 2021;4:e216303. doi:10.1001/jamanetworkopen.2021.6303

24 McDonald EG, Wu PE, Rashidi B, et al. The MedSafer Study-Electronic Decision Support for Deprescribing in Hospitalized Older Adults: A Cluster Randomized Clinical Trial. JAMA Intern Med 2022;182:265- 73. doi:10.1001/jamainternmed.2021.7429

25 Rieckert A, Reeves D, Altiner A, et al. Use of an electronic decision support tool to reduce polypharmacy in elderly people with chronic diseases: cluster randomised controlled trial. BMJ 2020;369:m1822. doi:10.1136/bmj.m1822

26 O’Mahony D, Gudmundsson A, Soiza RL, et al. Prevention of adverse drug reactions in hospitalized older patients with multi-morbidity and polypharmacy: the SENATOR* randomized controlled clinical trial. Age Ageing 2020;49:605-14. doi:10.1093/ageing/afaa072

P ro

tected b

y co p

yrig h

t, in clu

d in

g fo

r u ses related

to text an

d d

ata m in

in g

, A I train

in g

, an d

sim ilar tech

n o

lo g

ies. .

b y g

u est

n 18 S

ep tem

b er 2026

ttp s://w

w w

.b m

j.co m

/ D

o w

n lo

ad ed

fro m

24 M

ay 2023. 10.1136/b

m j-2022-074054 o

B M

J: first p u

b lish

ed as

https://www.ctu.unibe.ch/research_projects/declaration_of_interest/index_eng.html

https://www.nfp74.ch/de/3R0qxh8zc3RXcvXm/projekt/projekt-streit

http://creativecommons.org/licenses/by/4.0/

https://www.bmj.com/

RESEARCH

12 doi: 10.1136/bmj-2022-074054 | BMJ 2023;381:e074054 | the bmj

27 Jungo KT, Rozsnyai Z, Mantelli S, et al. ‘Optimising PharmacoTherapy In the multimorbid elderly in primary CAre’ (OPTICA) to improve medication appropriateness: study protocol of a cluster randomised controlled trial. BMJ Open 2019;9:e031080. doi:10.1136/ bmjopen-2019-031080

28 Jungo KT, Ansorg A-K, Floriani C, et al. Optimizing Prescribing in Older Adults with Multimorbidity and Polypharmacy in Primary Care: A Cluster Randomized Clinical Trial (OPTICA Trial).medRxiv 2022:2022.10.31.22281164. doi:10.1101/2022.10.31.22281164 .

29 Chmiel C, Bhend H, Senn O, Zoller M, Rosemann T, FIRE study- group. The FIRE project: a milestone for research in primary care in Switzerland. Swiss Med Wkly 2011;140:w13142.

30 Jungo KT, Meier R, Valeri F, et al. Baseline characteristics and comparability of older multimorbid patients with polypharmacy and general practitioners participating in a randomized controlled primary care trial. BMC Fam Pract 2021;22:123. doi:10.1186/s12875-021- 01488-8

31 Harris PA, Taylor R, Minor BL, et al, REDCap Consortium. The REDCap consortium: Building an international community of software platform partners. J Biomed Inform 2019;95:103208. doi:10.1016/j.jbi.2019.103208

32 Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)--a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform 2009;42:377-81. doi:10.1016/j. jbi.2008.08.010

33 Huibers CJA, Sallevelt BTGM, de Groot DA, et al. Conversion of STOPP/START version 2 into coded algorithms for software implementation: A multidisciplinary consensus procedure. Int J Med Inform 2019;125:110-7. doi:10.1016/j. ijmedinf.2018.12.010

34 Hanlon JT, Schmader KE, Ruby CM, Weinberger M. Suboptimal prescribing in older inpatients and outpatients. J Am Geriatr Soc 2001;49:200-9. doi:10.1046/j.1532-5415.2001. 49042.x

35 Hanlon JT, Schmader KE, Samsa GP, et al. A method for assessing drug therapy appropriateness. J Clin Epidemiol 1992;45:1045-51. doi:10.1016/0895-4356(92)90144-C

36 Samsa GP, Hanlon JT, Schmader KE, et al. A summated score for the medication appropriateness index: development and assessment of clinimetric properties including content validity. J Clin Epidemiol 1994;47:891-6. doi:10.1016/0895-4356(94) 90192-9

37 Dove S, Ruby C, Hanlon J, et al. The effect of hospitalization on suboptimal prescribing in elderly patients. American Society of Consultant Pharmacists 27th Annual Meeting and Exhibition. Nashville, Tenn, 1996.

38 Jeffery S, Ruby C, Hanlon J, et al. The impact of an interdisciplinary team on suboptimal prescribing in a long term care facility. Consult Pharm 1999;14:1386-91.

39 Wright RM, Sloane R, Pieper CF, et al. Underuse of indicated medications among physically frail older US veterans at the time of hospital discharge: results of a cross-sectional analysis of data from the Geriatric Evaluation and Management Drug Study. Am J Geriatr Pharmacother 2009;7:271-80. doi:10.1016/j. amjopharm.2009.11.002

40 Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159-74. doi:10.2307/2529310

41 EuroQol Group. EuroQol--a new facility for the measurement of health-related quality of life. Health Policy 1990;16:199-208. doi:10.1016/0168-8510(90)90421-9

42 Janssen MF, Pickard AS, Golicki D, et al. Measurement properties of the EQ-5D-5L compared to the EQ-5D-3L across eight patient groups: a multi-country study. Qual Life Res 2013;22:1717-27. doi:10.1007/s11136-012-0322-4

43 Weinstein MC, Torrance G, McGuire A. QALYs: the basics. Value Health 2009;12(Suppl 1):S5-9. doi:10.1111/j.1524- 4733.2009.00515.x

44 Smeeth L, Ng ES. Intraclass correlation coefficients for cluster randomized trials in primary care: data from the MRC Trial of the Assessment and Management of Older People in the Community. Control Clin Trials 2002;23:409-21. doi:10.1016/S0197-2456(02)00208-8

45 Rubin D. Multiple Imputation for Nonresponse in Surveys. John Wiley & Sons, 2004.

46 White H. A Heteroskedasticity-Consistent Covariance Matrix Estimator and a Direct Test for Heteroskedasticity. Econometrica 1980;48:817- 38. doi:10.2307/1912934 .

47 White H. Maximum Likelihood Estimation of Misspecified Models. Econometrica 1982;50:1-25. doi:10.2307/1912526 .

48 Huber PJ. The behavior of maximum likelihood estimates under nonstandard conditions. In: Le Cam LM, Neyman J, eds. Berkeley Symposium on Mathematical Statistics and Probability. University of California Press, 1967.

49 Campbell MK, Piaggio G, Elbourne DR, Altman DG, CONSORT Group. Consort 2010 statement: extension to cluster randomised trials. BMJ 2012;345:e5661. doi:10.1136/bmj.e5661

50 Reeve E, Low LF, Shakib S, Hilmer SN. Development and Validation of the Revised Patients’ Attitudes Towards Deprescribing (rPATD) Questionnaire: Versions for Older Adults and Caregivers. Drugs Aging 2016;33:913-28. doi:10.1007/s40266-016-0410-1

51 Ludwig K, Graf von der Schulenburg JM, Greiner W. German Value Set for the EQ-5D-5L. Pharmacoeconomics 2018;36:663-74. doi:10.1007/s40273-018-0615-8

52 Ming Y, Zecevic A. Medications & Polypharmacy Influence on Recurrent Fallers in Community: a Systematic Review. Can Geriatr J 2018;21:14-25. doi:10.5770/cgj.21.268

53 Pan H-H, Li C-Y, Chen T-J, Su TP, Wang KY. Association of polypharmacy with fall-related fractures in older Taiwanese people: age- and gender-specific analyses. BMJ Open 2014;4:e004428. doi:10.1136/ bmjopen-2013-004428

54 Dhalwani NN, Fahami R, Sathanapally H, Seidu S, Davies MJ, Khunti K. Association between polypharmacy and falls in older adults: a longitudinal study from England. BMJ Open 2017;7:e016358. doi:10.1136/bmjopen-2017-016358

55 Doherty AJ, Boland P, Reed J, et al. Barriers and facilitators to deprescribing in primary care: a systematic review. BJGP Open 2020;4:bjgpopen20X101096. doi:10.3399/bjgpopen20X101096

56 Kouri A, Yamada J, Lam Shin Cheung J, Van de Velde S, Gupta S. Do providers use computerized clinical decision support systems? A systematic review and meta-regression of clinical decision support uptake. Implement Sci 2022;17:21. doi:10.1186/s13012-022- 01199-3

57 Muth C, Uhlmann L, Haefeli WE, et al. Effectiveness of a complex intervention on Prioritising Multimedication in Multimorbidity (PRIMUM) in primary care: results of a pragmatic cluster randomised controlled trial. BMJ Open 2018;8:e017740. doi:10.1136/ bmjopen-2017-017740

58 Bayliss EA, Shetterly SM, Drace ML, et al. Deprescribing Education vs Usual Care for Patients With Cognitive Impairment and Primary Care Clinicians: The OPTIMIZE Pragmatic Cluster Randomized Trial. JAMA Intern Med 2022;182:534-42. doi:10.1001/ jamainternmed.2022.0502

59 Salari P, O’Mahony C, Henrard S, et al. Cost-effectiveness of a structured medication review approach for multimorbid older adults: Within-trial analysis of the OPERAM study. PLoS One 2022;17:e0265507. doi:10.1371/journal.pone.0265507

60 Wu H, Kouladjian O’Donnell L, Fujita K, Masnoon N, Hilmer SN. Deprescribing in the Older Patient: A Narrative Review of Challenges and Solutions. Int J Gen Med 2021;14:3793-807. doi:10.2147/IJGM. S253177

61 Hill-Taylor B, Sketris I, Hayden J, Byrne S, O’Sullivan D, Christie R. Application of the STOPP/START criteria: a systematic review of the prevalence of potentially inappropriate prescribing in older adults, and evidence of clinical, humanistic and economic impact. J Clin Pharm Ther 2013;38:360-72. doi:10.1111/jcpt.12059

Web appendix: Supplementary materials

P ro

tected b

y co p

yrig h

t, in clu

d in

g fo

r u ses related

to text an

d d

ata m in

in g

, A I train

in g

, an d

sim ilar tech

n o

lo g

ies. .

b y g

u est

n 18 S

ep tem

b er 2026

ttp s://w

w w

.b m

j.co m

/ D

o w

n lo

ad ed

fro m

24 M

ay 2023. 10.1136/b

m j-2022-074054 o

B M

J: first p u

b lish

ed as

https://www.bmj.com/

image10.emf

RESEARCH

Implementation evaluation of a pharmacist-led complex intervention: A mixed-methods analysis embedded within the ASPIRE randomized controlled trial

Laura Hellemans, Julie Hias, Leen Haegemans, Karolien Walgraeve, Astrid Liesenborghs, Astrid Lammens, Lorenz Van der Linden, Mieke Deschodt, Jos Tournoy *

a r t i c l e i n f o

Article history: Received 6 May 2025 Accepted 16 September 2025 Available online 22 September 2025

a b s t r a c t

Background: Geriatric patients with multimorbidity and polypharmacy are at high risk for medication-related harm. Complex interventions including medication reconciliation and medication review can improve outcomes, though their effectiveness in older patients re- mains unclear due to implementation failure in recent trials. The ASPIRE (The effect of a trAnSitional Pharmacist Intervention in geRiatric inpatients on hospital visits after dischargE) project evaluates a complex intervention aimed at reducing unplanned hospital revisits in geriatric patients while simultaneously maximizing intervention implementation. Objectives: To perform an implementation evaluation by describing the implementation strategies, evaluating fidelity, feasibility, and acceptability and documenting process out- comes of the ASPIRE intervention. Methods: A convergent parallel mixed-methods design was used with quantitative data from all intervention patients and qualitative data from semi-structured interviews and one focus group. The implementation strategies were designed based on the Expert Recommendations for Implementing Change guidelines. Fidelity to each intervention component was quantitatively measured for all intervention patients, using an 80% threshold to define successful imple- mentation. Feasibility and acceptability were qualitatively assessed using Flottorp's contextual analysis framework. Intervention duration and process outcomes were reported descriptively. Results: The ASPIRE trial enrolled 415 intervention participants aged 85.9 (±5.78) years. The complex intervention was successfully implemented with 87% of patients receiving all inter- vention components and considered feasible and acceptable by the majority of stakeholders. Key factors for successful implementation with the greatest impact on fidelity, feasibility, and acceptability included adequate time and workforce allocation, additional training, established working relationships, a shared information system and a detailed intervention guide. Median intervention duration was 77.3 (interquartile range (IQR) 65.8-93.5) minutes and 64.5 (IQR 54.2- 78.8) minutes for patients discharged home or to a nursing home, respectively. Conclusion: The ASPIRE intervention was successfully implemented and considered feasible and acceptable by the stakeholders, highlighting its potential to improve care for geriatric patients. The high level of implementation provide a strong basis for the further evaluation of its effectiveness. © 2025 American Pharmacists Association ® . Published by Elsevier Inc. All rights are reserved,

including those for text and data mining, AI training, and similar technologies.

Disclosure: The authors declare no relevant conflicts of interest or financial relationships. Funding: This work was supported by Research Foundation-Flanders (FWO) (11K7822N), Clinical Research Fund of UZ Leuven (KOOR), and an unre- stricted Pfizer Global Medical grant (number 53876865). Pfizer had no part in collection, management, analysis, and interpretation of the data, nor in writing and reporting study conclusions. * Correspondence: Jos Tournoy, MD, PhD, Department of Public Health and

Primary Care, KU Leuven, Herestraat 49, Leuven 3000, Belgium.

E-mail address: [email protected] (J. Tournoy).

ORCID Laura Hellemans: http://orcid.org/0000-0002-6787-9430 Mieke Deschodt: http://orcid.org/0000-0003-1560-2277 Jos Tournoy: http://orcid.org/0000-0002-0265-9154

Contents lists available at ScienceDirect

Journal of the American Pharmacists Association

journal homepage: www.japha.org

https://doi.org/10.1016/j.japh.2025.102928 1544-3191/© 2025 American Pharmacists Association ® . Published by Elsevier Inc. All rights are reserved, including those for text and data mining, AI training, and similar technologies.

Journal of the American Pharmacists Association 66 (2026) 102928

SCIENCE AND PRACTICE

Delta:1_given-name

Delta:1_surname

Delta:1_given-name

Delta:1_surname

Delta:1_given-name

Delta:1_surname

Delta:1_given-name

Delta:1_surname

Delta:1_given-name

mailto:[email protected]

http://orcid.org/0000-0002-6787-9430

http://orcid.org/0000-0003-1560-2277

http://orcid.org/0000-0002-0265-9154

http://crossmark.crossref.org/dialog/?doi=10.1016/j.japh.2025.102928&domain=pdf

www.sciencedirect.com/science/journal/15443191

www.japha.org

https://doi.org/10.1016/j.japh.2025.102928

Background

Older adults with multimorbidity and polypharmacy face high risks for medication-related harm, including hospital admissions and mortality. 1-4 These risks are particularly pronounced among frail, geriatric patients with increased susceptibility to adverse drug events. Inappropriate pre- scribing and poor adherence contribute significantly to these harms, creating opportunities for pharmacists to get involved. Pharmacist-delivered interventions including medication reconciliation, medication review, patient counseling, and transitional care optimization have been shown to improve outcomes in general populations, though evidence in geriatric patients remains limited. 5-9 Combining these interventions into a complex intervention may reduce medication-related harm in this high-risk group.

Complex interventions are shaped by their components, the flexibility of said components, and their interaction with the broader clinical context, which together influence both the likelihood of success and the mechanisms through which re- sults are achieved. 10 While randomized controlled trials (RCTs) are the gold standard for evaluating efficacy and safety, implementation evaluations alongside RCTs reveal critical in- formation on implementation processes, mechanisms and contextual factors. 11 Implementation evaluations explain how

and why complex interventions succeed or fail in real-world settings and are increasingly integrated into RCTs. 11-13

Recent trials on complex interventions to reduce medication-related harm in older adults have yielded incon- clusive results. 14-16 Process and implementation evaluations of these trials suggest low to intermediate implementation rates, indicating implementation failure rather than intervention failure. 17-20 Consequently, drawing robust conclusions on their effectiveness remains challenging.

To address these gaps, effective interventions to optimize medication use in high-risk geriatric inpatients should pri- oritize implementation while simultaneously evaluating their impact in an RCT. The ASPIRE (The effect of a trAnSitional Pharmacist Intervention in geRiatric inpatients on hospital

visits after dischargE) project addresses both aspects. First, it evaluates a complex clinical pharmacy intervention aimed to reduce unplanned hospital revisits within 6 months post- discharge in geriatric patients in an RCT. 21,22 Second, it in- corporates a multifaceted implementation strategy to maximize uptake. This paper reports on the implementation evaluation of the complex clinical pharmacy intervention. Analysis of the primary outcome results on hospital revisits is still ongoing and will be reported separately to avoid inter- pretation bias, as suggested by Oakley et al. 12

Objectives

To perform an implementation evaluation by 1) describing the implementation strategies, 2) evaluating fidelity, feasi- bility, and acceptability and 3) reporting on process outcomes and intervention duration of the ASPIRE intervention.

Methods and materials

Study design

We used a convergent parallel mixed-methods design to evaluate the implementation of the ASPIRE intervention (Figure 1). We collected quantitative data on intervention fi- delity, intervention duration and process outcomes, while qualitative data from semi-structured interviews and a focus group were used to evaluate the acceptability and feasibility of the intervention and to explore the underlying mecha- nisms influencing its implementation. Both types of data were integrated at the interpretation level to enhance under- standing of implementation. 23

Study setting

The study was conducted on the 3 acute geriatric wards at the University Hospitals Leuven (UZ Leuven), a large tertiary hospital with 1764 beds located in northern Belgium.

The ASPIRE intervention

The pharmacist-led complex intervention implemented in the ASPIRE study comprised the following components: 1) medication reconciliation at hospital admission, including the patient's and/or caregiver's concerns and goals; 2) medication review and reconciliation at discharge; followed by 3) coun- seling; 4) a follow-up call with the general practitioner (GP); 5) a discharge medication list for the patient, caregivers and all community-based healthcare providers (including the GP, community pharmacist, home care nurse, or care home nurse); 6) if applicable, a call with the home nurse or the nurse from

the care home; and 7) a follow-up call 1 week after discharge with the patient, caregiver or care home nurse. 8) Finally, clinical pharmacists remained available post-discharge through a telepharmacology service. Detailed information on the intervention is available in the protocol paper. 22 For simplicity, this complex intervention is referred to as the ‘ASPIRE intervention’ throughout the manuscript. Patients in the control group received usual care consisting of compre- hensive geriatric care without pharmacist involvement.

Key points

What was already known:

• Complex interventions could improve medication-

related harm in older adults but often face imple-

mentation failure in recent trials.

• Implementation failure rather than intervention fail-

ure hinders effectiveness' evaluations of these com-

plex interventions.

What this study adds:

• The ASPIRE complex intervention was feasible and

acceptable.

• Intervention implementation was successful for a

vast majority of patients.

• The implementation strategy is easily transferrable

to other settings.

L. Hellemans et al. / Journal of the American Pharmacists Association 66 (2026) 102928

SCIENCE AND PRACTICE

The first intervention component was executed by trained master students in biomedical sciences due to limited avail- ability of clinical pharmacists for the purpose of this study. Intervention components 2—6 were performed by the geri- atric clinical pharmacy team who was working at the geriatric wards prior to the study and who received at least 4 months of additional training in geriatric clinical pharmacy. Compo- nents 7 and 8 were carried out by 2 members of the research team: LHe who had previously collaborated with the geriatric clinical pharmacy team and who also received additional training in geriatric clinical pharmacy and JH who was also a member of the geriatric clinical pharmacy team. The geriatric clinical pharmacy team was present on the ward for 4 out of 5 working days. In case of time constraints or absence of the geriatric clinical pharmacy team, steps 1—6 were completed by the research team (trained clinical pharmacists). This occurred approximately once every 2 weeks.

Implementation strategies

The implementation strategy consisted of 10 distinct ele- ments, outlined in Table 1. These elements are described and mapped according to the Expert Recommendations for Implementing Change (ERIC) categories. 24,25 Due to financial constraints and the monocentric nature of our study, we prioritized non-financial and medium-scale implementation elements. Overall, we applied the following ERIC categories: a) develop educational materials, b) conduct ongoing training, c) assess for readiness and identify barriers and facilitators, d) build a coalition, e) mandate change, f) provide clinical su- pervision, g) distribute educational materials, h) remind cli- nicians, i) change physical structure and equipment, j) audit and provide feedback, k) conduct educational meetings, l) facilitation, and m) provide ongoing consultation.

Study sample

Quantitative data were used from all 415 intervention patients in the RCT. Eligible patients were those admitted to

the study ward and supervised by a geriatrician (fully trained physician who has completed additional specialized training in geriatric medicine). Patients who were admitted for only 1 day, unable to understand Dutch, in a palliative stage or discharged to another ward or hospital were excluded. For the collection of qualitative data, we applied purposive sampling to select a diverse group of participants of the ASPIRE inter- vention aiming to capture a range of positive, negative, or mixed experiences pertaining to the ASPIRE intervention. 26

We aimed to interview both intervention deliverers and re- cipients. Members of all the different healthcare providers exposed to the intervention were recruited by LHe. The 3 clinical pharmacists participated in a focus group, while the following participants were interviewed individually: 3 ward- based geriatricians-in-training (physicians specializing in geriatric medicine through fellowship training) with mini- mum 4 weeks intervention experience, four GP's (fully trained physicians with additional specialized training in family medicine), 3 community pharmacists, 3 ward-based nurses with minimum 4 weeks intervention experience, 3 nursing home nurses, and 2 home care nurses. Additionally, 2 Dutch-speaking intervention patients and 2 family members from intervention patients, who varied in functional status and reasons for hospital admission, were interviewed as well. Patients and family members were selected based on their recall of the intervention during the follow-up phone call 1 week postdischarge.

Data collection

Quantitative data collection spanned from February 25th 2021 to April 4th 2024 and was collected on paper forms or from the electronic health record. The research team trans- ferred all data to REDCap (Research Electronic Data Capture), a secure data management system. 27,28 Intervention fidelity was evaluated by tracking the delivery of components 1—7 of the ASPIRE intervention. 11 Component 8 was considered optional as community-based health providers could choose whether to contact the hospital pharmacist team; hence,

Quantitative study Sample  All intervention patients (N=415)

Source  Paper forms, electronic health record

Outcome  Descriptive numerical data on intervention fidelity, duration and outcomes

Qualitative study Sample  Actors involved in the intervention across

different settings (N=25)

Source  Semi-structured interview or focus group

Outcome  Opinion and experiences of different actors of the ASPIRE intervention

Mixed methods analysis Sample  Joint display and narrative (contiguous

approach)

Outcome  Fidelity, feasibility and acceptability of the ASPIRE intervention and contextual factors impacting its implementation

Figure 1. Design of the convergent parallel mixed-methods study.

Implementation evaluation of the ASPIRE trial

SCIENCE AND PRACTICE

Table 1

Implementation strategy

elements Objective

of strategy element ERIC

categories Timepoint

1. Patient-specific intervention protocol guide

The research

team

developed

a comprehensive

protocol

guide, which was printed and

provided

to the clinical

pharmacists for each enrolled patient. This guide

was

intended to be completed

by researchers and clinical

pharmacists upon conducting any

study-related

procedures.

To enhance

adherence

to all different study components. •

Develop

educational materials Prior to study

start

During study

2. Initial study training The

protocol guide was discussed

by the research

team

with the clinical pharmacists during a

1 h initial training

workshop

To improve

understanding

of the ASPIRE

intervention. •

Conduct ongoing training Prior to study

start

3. Workshop on motivational interviewing

Each clinical pharmacist who

conducted

discharge

counseling

sessions completed

a 2-h group

session

motivational interviewing, provided by a specialist

nurse associated

with

the hospital.

To enhance

the effectiveness of the counseling

sessions in

influencing the behavior of patients and

their family

members, thereby promoting

the successful

implementation of therapy changes.

• Conduct ongoing training Prior to study

start

4. Geriatrician and nurse

involvement

The research

team

discussed

the study, the intervention

and

designated

responsibilities in detail with

the head

the geriatric department, all attending geriatricians and

the head

nurses of study wards prior to its start.

To foster a coalition

and

secured

the support of the

hospital-based healthcare

professionals involved.

• Assess for readiness and identify

barriers and facilitators

• Build

a coalition

• Mandate change

• Provide clinical supervision

Prior to study start

5. Flyer distribution The

research

team

distributed

informational flyers

outlining the study

objectives, intervention and

contact

information of the research

team. These flyers were

displayed

in the nursing

stations and

doctor's offices of

the study

wards in the

hospital.

To support awareness of the intervention

and

its different

components. • Distribute educational materials

• Remind

clinicians

Prior to study start

During study

6. Distribution of study materials Relevant study documents and materials were

placed

designated desk

organizers on the

study

wards by the

research

team.

To facilitate

structured

access and efficient distribution of

informational flyers and medication boxes to improve

patient empowerment and therapy

adherence.

• Change physical structure and equipment Prior to study

start

7. Clinical case discussions Every

other wk

a 1 h meeting

was held between

the clinical

pharmacist team

and the research

team

to discuss

clinical cases.

To facilitate

a more

standardized

execution

of medication

reviews and supported

the ongoing

development of

expertise in geriatric pharmacotherapy

of the clinical

pharmacists.

• Audit and provide feedback

• Build

a coalition

• Conduct ongoing training

During study

8. Educational meetings for nurses and ward-based

physicians

Physicians and nurses working on one

of the study

wards

who were

unfamiliar with

the ASPIRE

intervention

received

a 30-min

educational meeting

from

one

of the

members of the research team.

To enhance

familiarity

and

understanding

of the

intervention. • Conduct educational meetings Prior to study

start

During study

9. Telepharmacology service

Once a y as long as patients were

recruited, information

the telepharmacology

service

was sent out via email to

community-based health

providers working

in the

wider Leuven region.

To spread

and

increase

awareness of the study

and

the

possibility to contact hospital-based clinical pharmacists.

• Distribute educational materials Prior to study

start

During study

10. Coordination and continuity

of intervention

Research members LHe and

alternated absences to

guarantee

100%

availability

of a research member during

weekdays between 8h30

and

17h00.

To ensure

effective

coordination

and

continuity

in the

implementation of the intervention, thereby optimizing

its

uptake.

• Facilitation

• Provide ongoing consultation

During study

Abbreviation used: ERIC, Expert Recommendations for Implementing

Change.

L. H ellem

ans et al. / Journal of the A m erican

Pharm acists A

ssociation 66

(2026) 102928

SC IEN

C E A N D PR

A C TIC

fidelity for this component was not measured. We measured the duration of each intervention component and reported this separately for patients discharged home or to a long-term

care facility. Process outcome measures included: 1) total medication

discrepancies on admission, 2) patients with ≥1 medication discrepancy (with/without validated medication list), 3) total medication review recommendations, 4) patients with ≥1 recommendation, 5) implementation rate of recommenda- tions, 6) agreement rate of recommendations, 7) patients with ≥1 implemented recommendation, and 8) types of pharmacist recommendations. We also collected age, sex, cognitive function 29,30 , Body Mass Index, Clinical frailty scale 31 , length of hospital stay, preadmission medications, living situation prior to admission, and after discharge as patient characteristics.

Semi-structured interviews and one focus group were conducted using tailored interview guides for each partici- pant group. The interview guides, developed in Dutch, considered the ASPIRE intervention, prior interview guides used by Kempen et al., guidance from Boyce et al and assessed feasibility and acceptability of the intervention and its different components. 32-34 Draft guides were reviewed by the research team and adjusted accordingly (Appendix A). The interviews were held online or at a location of choice in a neutral and low-stimulus environment, between December 2022 and November 2023. The focus group was moderated by LHe with ASVH being the observer. Each participant received a voucher of 30 euro as compensation for their time. All ses- sions were audio-recorded and transcribed verbatim. Identi- fiable information was anonymized and recordings were deleted afterward.

Analysis

Quantitative data were descriptively analyzed in R version 4.4.2 and Microsoft Excel (version 2409). Normality was assessed using the Shapiro—Wilk test and visual inspection of Q-Q plots. Mean and standard deviation (SD) were reported for normally distributed data; otherwise, median and inter- quartile range (IQR) were provided. Fidelity was calculated as the proportion of patients who received each intervention component relative to the number of patients who were eligible for that component and was reported for intervention components 1—7. A threshold of 80% was used as a cutoff for successful implementation.

Missing data on intervention duration were presumed to be missing at random, most likely due to occasional docu- mentation lapses. These cases were excluded from the anal- ysis, given their limited number and the assumption that missingness was not systematically related to the outcome. Available cases were used to calculate the duration per intervention component (n = 400 for medication reconcilia- tion upon admission, n = 397 for medication review, n = 383 for medication reconciliation at discharge, n = 391 for communication with health care providers, n = 249 for pa- tient counseling, n = 387 for follow-up call). Conversely, median duration of the entire intervention was calculated on the complete cases (n = 342). The rate of recommendation implementation was calculated as the number of immediately implemented recommendations divided by the total number

of provided recommendations. The agreement rate was calculated as the sum of immediately implemented recom- mendations combined with recommendations adopted in the electronic patient record by the ward-based physician divided by the total number of provided recommendations.

Qualitative data were thematically analyzed using the contextual analysis framework by Flottorp et al. 35 This framework consists of 7 domains and 57 determinants to identify determinants of practice that could facilitate or hinder implementation of complex interventions. 35 LHe and LHa independently read and coded the first 3 transcripts deductively using NVivo Release 1.7.1. After reaching consensus on the coding framework, they continued to code the remaining interviews. Additional interviews were per- formed until data saturation was reached, meaning no new

themes emerged. Results were then discussed within the research team to reach consensus on the interpretation, in- clusion, and adjustment of determinants to better reflect factors relevant to the ASPIRE complex intervention.

Data integration

The quantitative and qualitative data were analyzed separately as per the convergent parallel design and inte- grated at the interpretation phase using joint display and a contiguous narrative approach (Figure 1). 23

Results

Demographics of participants

Intervention patients had a mean age of 85.9 (±5.8) years and 40.7% were male (Table 2). Intervention patients took a median of 10 [IQR: 7—12] medications upon admission and the majority (80%) were living at home prior to hospital admission.

Each eligible participant approached, agreed to be inter- viewed. In total, we interviewed 21 healthcare providers, the majority of whom were female (90%), with work experience ranging between a few months to over 41 years. We inter- viewed 2 patients and 2 caregivers, half of them being female. Interviews lasted 30.4 min (SD 9.9 min) on average. After analysis of the first 18 interviews and the focus group, we conducted 3 more interviews where no new themes emerged.

Fidelity

Overall, 87% (n = 361) of patients received every inter- vention component and 97% (n = 403) of patients received at least all but one intervention component. Fidelity was also high for individual intervention components both within and outside the predefined timeframes with the majority of re- ported reasons for protocol deviation beyond control of the research team (Table 3).

Intervention duration

Median duration of the entire intervention was 64.5 min (IQR 54.2—78.8) for nursing home residents and 77.3 min (IQR 65.8—93.5) for patients living at home.

Implementation evaluation of the ASPIRE trial

SCIENCE AND PRACTICE

Process outcomes

Medication reconciliation upon admission detected a total of 1364 medication discrepancies, with a median of 3 (IQR: 2—5) per patient. In total, 85.8% of patients had at least one medication discrepancy and 38.2% had their medication list from admission already wrongfully validated by another healthcare professional. Clinical pharmacists provided a total of 1171 medication recommendations during the medication review, with a median of 3 (IQR 1—4) recommendations per patient. The majority of patients (91.6%) received at least 1 recommendation. The agreement rate was high as ward-based physicians implemented 72.5% of recommendations immedi- ately in the discharge medication list and documented 17% of recommendations in the patient file or discharge letter. This resulted in an overall agreement rate of 89.5%. A total of 82.7% patients had at least one recommendation immediately implemented. The types of pharmacist recommendations included: discontinuation (35.7%), initiation (20.2%), providing information and follow-up (12.0%), optimization of therapy (11.1%), dosage reduction (9.8%), medication switch (5.6%), dosage increase (3.6%), other (1.5%) and therapeutic drug monitoring or follow-up parameters (0.5%).

Acceptability and feasibility of the intervention

The overall intervention was accepted and deemed valu- able by all participants involved with an expected positive impact on medication- and patient-related outcomes. The- matic analysis of interview data further revealed 32 de- terminants that hindered or facilitated successful implementation of our intervention. These implementation determinants were categorized into 5 domains of the frame- work for contextual analysis by Flottorp et al. 35 No themes emerged from domain 6 (Capacity for organizational change) and domain 7 (social, political, and legal factors) during the interviews and these were not included in the contextual

analysis. Appendix B reports on a detailed description of the determinants and a selection of quotes while Appendix C contains a joint display with contextual factors and duration related to the individual intervention components.

The first domain addresses guideline factors. The ASPIRE intervention was seen as valuable for most geriatric patients. Conversely, some healthcare providers' unfamiliarity with clinical pharmacists' role hindered collaboration, whereas those familiar with their clinical role viewed their involve- ment as crucial for improving medication management. Clinical pharmacists' proactive attitude, knowledge, accessi- bility, strong collaboration, and communication skills facili- tated implementation. Deliverers of the intervention considered it feasible to perform, but barriers including competing patient priorities, unpredictable physician avail- ability and increased pharmacist workload were mentioned. Integration of pharmacist recommendations into hospital software improved visibility and counseling and phone calls supported medication inquiries. Despite its perceived value, the telepharmacology service was used for only 23 (5.5%) patients due to limited awareness of the service and phar- macists' roles.

The second domain covers individual health care profes- sional factors. Health care providers felt confident in their knowledge but acknowledged gaps in specific areas and the need for multidisciplinary collaboration. Clinical pharmacists were recognized for their expertise and skills and the inter- vention was widely supported, especially medication recon- ciliation upon admission. Ward-based physicians prioritized evidence-based practices while general practitioners favored patient-specific approaches, occasionally leading to disagreements with medication recommendations. Adher- ence was supported by positivity, motivation and teamwork, but hindered by time constraints and workflow disruptions. Clear multidisciplinary communication, particularly in discharge planning, was crucial, with clinical pharmacists ensuring continuity of care.

The third domain includes patient factors. Patients valued clinical pharmacists' role in optimizing therapy, reducing polypharmacy and improving understanding through coun- seling and follow-up. Key to safety and engagement were accurate medication lists, shared decision-making and strong health care provider communication. The intervention increased patients' confidence and sense of empowerment, encouraging greater involvement in medication management after discharge.

The fourth domain contains professional interactions. Inter- disciplinary communication was essential, but challenged by timing issues, workflow disruptions and technical limitations. The absence of a shared electronic patient record and involvement of numerous healthcare providers hindered seamless communication. GP's preferred phone contact over electronic messaging such as SIILO, despite its nondisruptive nature. 36 Facilitators included strong working relationships, awareness of interteam biases, tailored communication and pharmacists participation in hospital team meetings. Sug- gested improvements include enhancing discharge letters, expanding shared electronic patient records and involving GPs earlier, during the medication review rather than afterward.

Finally, the fifth domain covers incentives and resources. Implementation of the ASPIRE intervention was hindered by

Table 2 Characteristics of intervention patients of the ASPIRE trial (n = 415)

Characteristics Intervention group

Age in y (mean (±SD); minimum-maximum) 85.9 (±5.8); 70-102 Male sex, n (%) 169 (40.7%)

Cognitive function (median [IQR]) MMSE (maximum score = 30) (n = 308) 23 [19-26] MoCA (maximum score = 30) (n = 40) 19 [15-23.5]

Body mass index (median [IQR]) a 24.8 [21.9-28.6] Clinical frailty scale (median [IQR]) b 6 [5.8-7] Length of stay in d (median [IQR]) 12 [8-17] Number of medications (admission)

(median [IQR]) 10 [7-12]

Living situation upon admission, n (%) Home (incl. Service flats) 332 (80%) Nursing home 81 (19.5%) Other long-term care facility 2 (0.5%)

Living situation after discharge, n (%) Home (incl. Service flats) 258 (62.2%) Nursing home 154 (37.1%) Other long-term care facility 3 (0.7%)

Abbreviations used: MMSE, mini-mental state examination; MoCa, montreal cognitive assessment.

a 4 Missing. b 153 Missings.

L. Hellemans et al. / Journal of the American Pharmacists Association 66 (2026) 102928

SCIENCE AND PRACTICE

limited time, staffing and organizational support, high- lighting the need for additional funding. Digitization could improve efficiency, but some healthcare providers, such as care home nurses, lacked necessary tools like SIILO. Financial concerns and some GP's reluctance to collaborate without adequate compensation were also barriers. Key motivators included personal contact and reducing polypharmacy in geriatric patients. Suggested solutions involved a universal communication platform, shared medication records, and teleconsultations.

Discussion

In this study, we showed successful implementation of the ASPIRE intervention, reflected by an overall fidelity rate of 87% for the entire intervention and even higher fidelity rates for individual intervention components. These rates exceeded the 80% threshold for successful implementation. The process outcomes offer a comprehensive assessment of the return on investment for the approximately 70 min of time spent by the clinical pharmacists, highlighting its implications for both patient care and collaboration with

other healthcare professionals. Our qualitative findings confirmed the feasibility and acceptability of the ASPIRE intervention, with all actors involved in the intervention highlighting its potential to improve care for geriatric pa- tients and describing it as practical and valuable in routine practice.

To the best of our knowledge, no previous trials investi- gating complex interventions aimed at optimizing medication management have reported fidelity rates as high as those observed in our study. 18-20,37-39 Contextual analysis revealed five key implementation factors: 1) lack of time and personnel with competing patient care priorities, 2) additional training, 3) previously established working relationships, 4) common information system and 5) availability of a detailed study protocol. These implementation factors had the most signifi- cant influence on the fidelity, feasibility, and acceptability of the intervention.

First, lack of time and personnel, as well as competing patient care priorities have been identified as barriers in prior trials with lower fidelity rates. 18,20,39,40 We ensured 100% availability of a clinical pharmacist during weekdays by backing up any absences of the geriatric clinical pharmacist

Table 3 Fidelity and reasons for protocol deviation

Intervention component (applicable sample)

Timeframe Fidelity % (n) Reason for no execution (n)

Fidelity within timeframe % (n) Reason for outside timeframe (n)

Medication reconciliation (n = 415) Admission 100% (415) 100% (415) Medication review (n = 415) 72h before discharge

at the earliest 99.8% (414) Discharge missed (1)

97.3% (404) Discharge postponed (6) Discharge missed (4)

Medication reconciliation (n = 415) Discharge 99% (411) 99% (411) Discharge missed (4) Discharge missed (4)

Medication list for community pharmacist (n = 415)

Discharge 98.3% (408) 98.3% (408)

Discharged missed (4) Discharged missed (4) Reason not reported (3) Reason not reported (3)

Contact with GP (n = 414) a 72h postdischarge at the latest

96.4% (399) 82.4% (342)

Unreachable after 7 d (7) GP unavailable (46) Patient readmitted (3) Discharge before weekend/holiday (5) Patient without GP (2) Absence research team (4) Discharge missed (2) Patient readmitted (1)

New GP (1) Contact care home or long-term

care facility (n = 157) b 96h postdischarge 99.4% (156) 98.1% (154)

Discharge missed (discovered after 96 h) (1)

Unavailability of nurse (3)

Patient and relative counseling (n = 258)b

Discharge 98.4% (254) 98.4% (254)

Discharge missed (3) Discharge missed (3) No perceived added value (1) No perceived added value (1)

Contact home nurse (n = 80) b Postdischarge 57.5% (46) 57.5% (46) Shared responsibility with counseled relative (27)

Shared responsibility with counseled relative (27)

No nurse assigned at discharge (6) No nurse assigned at discharge (6) Nurse responsible but accidently relative counseled (1)

Nurse responsible but accidently relative counseled (1)

Follow-up call (n = 410) a 10 d postdischarge 95.6% (392) 85.3% (354) Patient readmitted (12) Patient/relative unavailable (18) Unreachable after 7 d (2) Absence research team (10) No longer wants to be contacted (2) Reason not reported (6) Reason not reported (2) Unavailability of nurse (2)

Abbreviation used: GP, general practitioner. a Patients deceased after discharge but before contact with GP were excluded. b Applicable sample according to discharge location.

Implementation evaluation of the ASPIRE trial

SCIENCE AND PRACTICE

team with the presence of a pharmacist of the research team, resulting in minimal nondelivery cases.

Additionally, intervention deliverers should receive a minimum of 4 months additional training in geriatric phar- macotherapy and pathophysiology and have experience in performing parts of the intervention. This was observed in a dual-site analysis where better-trained pharmacists achieved improved intervention adoption and sustainability. 41 Our team was highly trained, experienced in medication reviews, and reconciliation and participated in biweekly case discus- sions with a geriatrician. Their training was appreciated by ward-based physicians as a facilitator of intervention delivery.

Moreover, previously established working relationships facilitated intervention delivery. 18,41 Pharmacists integrated into the ward team prior to the study achieved higher inter- vention delivery. 33,40 Our geriatric clinical pharmacist team

had well-established relationships for nearly 10 years with ward-based physicians who valued their input, as reflected in the high medication recommendation agreement rate. Nevertheless, the quality of the working relationship is prioritized over its duration. Clinical pharmacists also noted that effective interprofessional collaboration was crucial, as they felt demotivated to conduct medication reviews for pa- tients under the care of physicians skeptical of their role.

Furthermore, the absence of a common information sys- tem to efficiently communicate across care settings was a barrier. 20,37,39 Time spent attempting to contact other health providers and coordinating schedules, though unrecorded, significantly impacted the clinical pharmacists' workload and intervention delivery efficiency. Despite past initiatives, this remains a major barrier that still warrants urgent optimiza- tion. However, we believe the lack of a common information system should not be seen as an insurmountable obstacle to implementing the ASPIRE intervention in daily practice as demonstrated by our high fidelity rates.

At last, the availability of a detailed study protocol facili- tated implementation, as evidenced by the successful adher- ence to the ASPIRE intervention. 20,37,40,41 Moreover, Nurjono et al reported high fidelity rates specifically for these com- ponents that had a detailed intervention protocol guide. 37 Our patient-specific protocol guide included detailed steps for each intervention component, preventing inadvertent omissions.

Evidence on intervention duration and its components is scarce. Studies on similar interventions often use different components or methodologies, which complicates compari- sons, and report intervention durations varying between 66.9 min and 114 min. 38,42-44 The intervention durations re- ported in our study are on the shorter end, likely due to different measurement methods. Research team members and clinical pharmacists were asked to record time spent on intervention components, but this likely underestimates the total time, as we did not account for preparation, searching for contact details or failed calls which are included in time- and-motion studies. Nonetheless, our results are compara- ble to the overall median duration of the time-and-motion study of the PHARM-DC trial whose intervention consisted of 12 best practices across 3 domains (medication reconcili- ation, medication review and medication adherence) and closely aligned with ours, thereby confirming the validity of our measurements. 38 Subsequently, with an average duration

of approximately 70 min, the ASPIRE intervention is consid- ered feasible for integration into routine clinical practice.

Implications of findings

The study's implications for future research are threefold. First, the high fidelity rate allows robust conclusions on the interventions' effectiveness as any neutral or negative effects will not be due to implementation failure. Secondly, involving GP's during the medication review could reduce disagree- ments with pharmacist recommendations. Therefore, accep- tance and implementation of the medication changes by GP's should be further evaluated to draw conclusions on the timing of the medication review. Thirdly, we identified key de- terminants for successful implementation of complex in- terventions in medication management. We recommend our multifaceted implementation strategy for other researchers seeking to implement complex interventions in high-risk geriatric patients.

Limitations

Several limitations should be considered. First, LHe had multiple roles, including performing parts of the intervention, recruiting participants, conducting interviews and analyzing data. We ensured methodologic rigor and validity 1) in data collection using structured interview guides, 2) in data interpretation by involving multiple researchers (investigator triangulation) to decrease bias, and 3) in data integration using both qualitative and quantitative data. Moreover, the number of interviewed patients and family members was low. We envisioned to include more but many had no recall of the intervention as it was perceived as standard-of-care and were unable to provide new themes. The analysis of the first 3 in- terviews with patients and family members revealed a near- complete overlap in themes and answers, which influenced the decision to limit further interviews in these groups. Furthermore, feasibility and acceptability were not assessed using quantitative methods due to the limited number of hospital-based health professionals involved, which would have resulted in an insufficient sample size for meaningful analysis. Instead, we opted for a qualitative evaluation through interviews and a focus group, allowing a more in- depth exploration of participants' perspectives on the feasi- bility and acceptability of the intervention. In addition, there was a potential risk of bias in responses from health pro- fessionals who had prior working relationships with the re- searchers or clinical pharmacy team. To mitigate this risk and ensure a broader range of perspectives, more community- based health care professionals, who had no working re- lationships with the research team, were interviewed, compared to hospital-based healthcare professionals. More- over, while interviews were conducted both online and in person, we do not expect this to significantly affect the con- tent due to the non-sensitive nature of the questions. Finally, patients and caregivers were recruited for interviews if they agreed to participate in the RCT, limiting the sample. Lastly, most interviewed healthcare providers were female, reflect- ing the gender distribution in the involved professions, and we do not expect gender to influence their experiences.

L. Hellemans et al. / Journal of the American Pharmacists Association 66 (2026) 102928

SCIENCE AND PRACTICE

Conclusion

We successfully implemented a feasible and acceptable pharmacist-led complex intervention to optimize medication therapy in geriatric patients. For implementation in daily practice, we recommend adequate time and workforce allo- cation, additional training, well-established working re- lationships, an efficient common information system, and a detailed intervention guide. The absence of implementation failure will allow to draw robust conclusions on the effec- tiveness of the ASPIRE intervention in optimizing outcomes for geriatric patients.

Author contributions

Laura Hellemans: Conceptualization, Investigation, Formal analysis, Writing — original draft, Writing — review & editing. Julie Hias: Conceptualization, Writing — review & editing. Leen Haegemans: Formal analysis, Writing — review & edit- ing. Karolien Walgraeve: Writing — review & editing. Astrid Liesenborghs: Writing — review & editing. Astrid Lammens: Writing — review & editing. Lorenz Van der Linden: Concep- tualization, Writing — review & editing. Mieke Deschodt: Conceptualization, Writing — review & editing. Jos Tournoy: Conceptualization. Writing — review & editing.

Ethics approval

Ethical approval was obtained from the ethics committee UZ/KU Leuven prior to study start (S64758). The clinical trial was registered with trial registration number NCT04617340 (ClinicalTrials.gov) and date of registration 2020-10-29.

Consent to participate

All participants provided written informed consent for participation.

Acknowledgments

The authors thank all the participants for their engagement in the study. The authors acknowledge the contributions of Anne-Sophie Vanhoonacker (ASVH), Jolien Broekmans (JB), Isabel Spriet (IS), and Johan Flamaing (JF).

Supplementary data

Supplementary data related to this article can be found at https://doi.org/10.1016/j.japh.2025.102928.

References

1. Yadesa TM, Kitutu FE, Deyno S, Ogwang PE, Tamukong R, Alele PE. Prevalence, characteristics and predicting risk factors of adverse drug reactions among hospitalized older adults: a systematic review and meta-analysis. SAGE Open Med. 2021;9:20503121211039099. https:// doi.org/10.1177/20503121211039099.

2. El Morabet N, Uitvlugt EB, van den Bemt BJF, van den Bemt PMLA, Janssen MJA, Karapinar-Çarkit F. Prevalence and preventability of drug- related hospital readmissions: a systematic review. J Am Geriatr Soc. 2018;66(3):602—608. https://doi.org/10.1111/jgs.15244.

3. Gustafsson M, Sj €olander M, Pfister B, Jonsson J, Schneede J, L €ovheim H. Drug-related hospital admissions among old people with dementia. Eur J

Clin Pharmacol. 2016;72(9):1143—1153. https://doi.org/10.1007/s00228- 016-2084-3.

4. Muhlack DC, Hoppe LK, Weberpals J, Brenner H, Sch€ ottker B. The asso- ciation of potentially inappropriate medication at older age with cardiovascular events and overall mortality: a systematic review and meta-analysis of cohort studies. J Am Med Dir Assoc. 2017;18(3): 211—220. https://doi.org/10.1016/j.jamda.2016.11.025.

5. Mekonnen AB, McLachlan AJ, Brien JA. Effectiveness of pharmacist-led medication reconciliation programmes on clinical outcomes at hospi- tal transitions: a systematic review and meta-analysis. BMJ Open. 2016;6 (2):e010003. https://doi.org/10.1136/bmjopen-2015-010003.

6. Bonetti AF, Reis WC, Mendes AM, et al. Impact of pharmacist-led discharge counseling on hospital readmission and emergency depart- ment visits: a systematic review and meta-analysis. J Hosp Med. 2020;15 (1):52—59. https://doi.org/10.12788/jhm.3182.

7. De Oliveira GS, Castro-Alves LJ, Kendall MC, McCarthy R. Effectiveness of pharmacist intervention to reduce medication errors and health-care resources utilization after transitions of care: a meta-analysis of ran- domized controlled trials. J Patient Saf. 2021;17(5):375—380. https://doi. org/10.1097/PTS.0000000000000283.

8. Bülow C, Clausen SS, Lundh A, Christensen M. Medication review in hospitalised patients to reduce morbidity and mortality. Cochrane Data- base Syst Rev. 2023;1(1):CD008986. https://doi.org/10.1002/14651858. CD008986.pub4.

9. Dautzenberg L, Bretagne L, Koek HL, et al. Medication review in- terventions to reduce hospital readmissions in older people. J Am Geriatr Soc. 2021;69(6):1646—1658. https://doi.org/10.1111/jgs.17041.

10. Skivington K, Matthews L, Simpson SA, et al. A new framework for developing and evaluating complex interventions: update of medical research council guidance. Int J Nurs Stud. 2024;154:104705. https://doi. org/10.1016/j.ijnurstu.2024.104705.

11. Moore GF, Audrey S, Barker M, et al. Process evaluation of complex in- terventions: medical research council guidance. BMJ. 2015;350:h1258. https://doi.org/10.1136/bmj.h1258.

12. Oakley A, Strange V, Bonell C, Allen E, Stephenson J, Team RS. Process evaluation in randomised controlled trials of complex interventions. BMJ. 2006;332(7538):413—416. https://doi.org/10.1136/bmj.332. 7538.413.

13. Craig P, Dieppe P, Macintyre S, Michie S, Nazareth I, Petticrew M. Developing and Evaluating Complex Interventions: New Guidance; 2019. Available at: https://www.unisante.ch/sites/default/files/inline-files/ Complex%20Interventions%20Guidance%2029-9-08_0.pdf. Accessed January 13, 2025.

14. Kempen TGH, Bertilsson M, Hadziosmanovic N, et al. Effects of hospital- based comprehensive medication reviews including postdischarge follow-up on older patients' use of health care: a cluster randomized clinical trial. JAMA Netw Open. 2021;4(4):e216303. https://doi.org/ 10.1001/jamanetworkopen.2021.6303.

15. Johansen JS, Halvorsen KH, Svendsen K, et al. Interdisciplinary collabo- ration across secondary and primary care to improve medication safety in the elderly (The IMMENSE study) - a randomized controlled trial. BMC Health Serv Res. 2022;22(1):1290. https://doi.org/10.1186/s12913-022- 08648-1.

16. Uitvlugt EB, Heer SE, van den Bemt BJF, et al. The effect of a transitional pharmaceutical care program on the occurrence of ADEs after discharge from hospital in patients with polypharmacy. Res Social Adm Pharm. 2022;18(4):2651—2658. https://doi.org/10.1016/j.sapharm.2021.05.009.

17. Rychetnik L, Frommer M, Hawe P, Shiell A. Criteria for evaluating evi- dence on public health interventions. J Epidemiol Community Health. 2002;56(2):119—127. https://doi.org/10.1136/jech.56.2.119.

18. Kempen TGH, Cam H, K€ alvemark A, et al. Intervention fidelity and process outcomes of medication reviews including post-discharge follow-up in older hospitalized patients: process evaluation of the MedBridge trial. J Clin Pharm Ther. 2020;45(5):1021—1029. https://doi. org/10.1111/jcpt.13128.

19. Johansen JS, Halvorsen KH, Havnes K, Wetting HL, Svendsen K, Garcia BH. Intervention fidelity and process outcomes of the IMMENSE study, a pharmacist-led interdisciplinary intervention to improve medication safety in older hospitalized patients. J Clin Pharm Ther. 2022;47(5):619—627. https://doi.org/10.1111/jcpt.13581.

20. En-Nasery-de Heer S, Uitvlugt EB, Bet PM, et al. Implementation of a pharmacist-led transitional pharmaceutical care programme: pro- cess evaluation of medication actions to reduce hospital admissions through a collaboration between community and hospital pharma- cists (MARCH). J Clin Pharm Ther. 2022;47(7):1049—1069. https://doi. org/10.1111/jcpt.13645.

21. Van der Linden L, Hias J, Walgraeve K, Flamaing J, Tournoy J, Spriet I. Clinical pharmacy services in older inpatients: an evidence-based re- view. Drugs Aging. 2020;37(3):161—174. https://doi.org/10.1007/ s40266-019-00733-1.

Implementation evaluation of the ASPIRE trial

SCIENCE AND PRACTICE

http://ClinicalTrials.gov

https://doi.org/10.1016/j.japh.2025.102928

https://doi.org/10.1177/20503121211039099

https://doi.org/10.1111/jgs.15244

https://doi.org/10.1007/s00228%2D016%2D2084%2D3

https://doi.org/10.1016/j.jamda.2016.11.025

https://doi.org/10.1136/bmjopen%2D2015%2D010003

https://doi.org/10.12788/jhm.3182

https://doi.org/10.1097/PTS.0000000000000283

https://doi.org/10.1002/14651858.CD008986.pub4

https://doi.org/10.1111/jgs.17041

https://doi.org/10.1016/j.ijnurstu.2024.104705

https://doi.org/10.1136/bmj.h1258

https://doi.org/10.1136/bmj.332.7538.413

https://www.unisante.ch/sites/default/files/inline-files/Complex%20Interventions%20Guidance%2029-9-08_0.pdf

https://doi.org/10.1001/jamanetworkopen.2021.6303

https://doi.org/10.1186/s12913%2D022%2D08648%2D1

https://doi.org/10.1016/j.sapharm.2021.05.009

https://doi.org/10.1136/jech.56.2.119

https://doi.org/10.1111/jcpt.13128

https://doi.org/10.1111/jcpt.13581

https://doi.org/10.1111/jcpt.13645

https://doi.org/10.1007/s40266%2D019%2D00733%2D1

22. Hias J, Hellemans L, Laenen A, et al. The effect of a trAnSitional phar- macist intervention in geRiatric inpatients on hospital visits after dischargE (ASPIRE): protocol for a randomized controlled trial. Contemp Clin Trials. 2022;119:106853. https://doi.org/10.1016/j.cct.2022.106853.

23. Fetters MD, Curry LA, Creswell JW. Achieving integration in mixed methods designs-principles and practices. Health Serv Res. 2013;48(6 Pt 2):2134—2156. https://doi.org/10.1111/1475-6773.12117.

24. Proctor EK, Powell BJ, McMillen JC. Implementation strategies: recom- mendations for specifying and reporting. Implement Sci. 2013;8:139. https://doi.org/10.1186/1748-5908-8-139.

25. Powell BJ, Waltz TJ, Chinman MJ, et al. A refined compilation of imple- mentation strategies: results from the expert recommendations for implementing change (ERIC) project. Implement Sci. 2015;10:21. https:// doi.org/10.1186/s13012-015-0209-1.

26. Patton M. Qualitative Research & Evaluation Methods. 3rd ed. Sage Pub- lications, Inc.; 2002:598.

27. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)–a metadata-driven methodology and workflow process for providing translational research informatics sup- port. J Biomed Inform. 2009;42(2):377—381. https://doi.org/10.1016/j. jbi.2008.08.010.

28. Harris PA, Taylor R, Minor BL, et al. The REDCap consortium: building an international community of software platform partners. J Biomed Inform. 2019;95:103208. https://doi.org/10.1016/j.jbi.2019.103208.

29. Folstein MF, Folstein SE, McHugh PR. "Mini-mental state". A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12(3):189—198. https://doi.org/10.1016/0022-3956 (75)90026-6.

30. Nasreddine ZS, Phillips NA, B� edirian V, et al. The montreal cognitive assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc. 2005;53(4):695—699. https://doi.org/10.1111/j.1532- 5415.2005.53221.x.

31. Rockwood K, Song X, MacKnight C, et al. A global clinical measure of fitness and frailty in elderly people. CMAJ. 2005;173(5):489—495. https://doi.org/10.1503/cmaj.050051.

32. Kempen TGH, K€ alvemark A, Gillespie U, Stewart D. Comprehensive medication reviews by ward-based pharmacists in Swedish hospitals: what does the patient have to say? J Eval Clin Pract. 2020;26(1): 149—157. https://doi.org/10.1111/jep.13121.

33. Kempen TGH, K€ alvemark A, Sawires M, Stewart D, Gillespie U. Facili- tators and barriers for performing comprehensive medication reviews and follow-up by multiprofessional teams in older hospitalised patients. Eur J Clin Pharmacol. 2020;76(6):775—784. https://doi.org/10.1007/ s00228-020-02846-8.

34. Boyce C, Neale P. Conducting In-Depth Interviews: A Guide for Designing and Conducting In-Depth Interviews for Evaluation Input. Watertown, MA, USA: Pathfinder International; 2006.

35. Flottorp SA, Oxman AD, Krause J, et al. A checklist for identifying de- terminants of practice: a systematic review and synthesis of frame- works and taxonomies of factors that prevent or enable improvements in healthcare professional practice. Implement Sci. 2013;8:35. https:// doi.org/10.1186/1748-5908-8-35.

36. Siilo. https://www.siilo.com/nl/; 2016. Accessed January 31, 2025. 37. Nurjono M, Shrestha P, Ang IYH, et al. Implementation fidelity of a

strategy to integrate service delivery: learnings from a transitional care program for individuals with complex needs in Singapore. BMC Health Serv Res. 2019;19(1):177. https://doi.org/10.1186/s12913-019-3980-x.

38. Nuckols TK, Berdahl CT, Henreid AJ, et al. Comprehensive pharmacist- led transitions-of-care medication management around hospital

discharge adds modest cost relative to usual care: time-and-motion cost analysis. Inquiry. 2023;60:469580231218625. https://doi.org/10.1177/ 00469580231218625.

39. Guilcher SJT, Fernandes O, Luke MJ, et al. A developmental evaluation of an intraprofessional pharmacy communication partnership (PROMPT) to improve transitions in care from hospital to community: a mixed- methods study. BMC Health Serv Res. 2020;20(1):99. https://doi.org/ 10.1186/s12913-020-4909-0.

40. Oche O, Murry LT, Keller MS, et al. Pharmacist, nurse, and physician per- spectives on the implementation of the pharmacist discharge care (pharm-dc) intervention: a qualitative study. Res Social Adm Pharm. 2024;20(8):740—746. https://doi.org/10.1016/j.sapharm.2024.04.009.

41. Murry LT, Keller MS, Pevnick JM, Schnipper JL, Kennelty KA, PHARM- DC Group. A qualitative dual-site analysis of the pharmacist discharge care (PHARM-DC) intervention using the CFIR framework. BMC Health Serv Res. 2022;22(1):186. https://doi.org/10.1186/ s12913-022-07583-5.

42. Ravn-Nielsen LV, Duckert ML, Lund ML, et al. Effect of an in-hospital multifaceted clinical pharmacist intervention on the risk of read- mission: a randomized clinical trial. JAMA Intern Med. 2018;178(3): 375—382. https://doi.org/10.1001/jamainternmed.2017.8274.

43. Meguerditchian AN, Krotneva S, Reidel K, Huang A, Tamblyn R. Medi- cation reconciliation at admission and discharge: a time and motion study. BMC Health Serv Res. 2013;13:485. https://doi.org/10.1186/1472- 6963-13-485.

44. Alix L, Bajeux E, Hubert J, et al. Medication reconciliation in hos- pital patients over the age of 65: how long does it take and how

much does it cost? A time-motion study in an internal medicine ward. Eur J Intern Med. 2020;73:100—102. https://doi.org/10.1016/j. ejim.2019.12.003.

Laura Hellemans, PharmD, Research Foundation Flanders (FWO), Brussels, Belgium; Department of Pharmaceutical and Pharmacological Sciences, KU Leuven, Leuven, Belgium; Hospital Pharmacy Department, University Hospital Leuven, Leuven, Belgium

Julie Hias, PharmD, PhD, Department of Pharmaceutical and Pharmacological Sciences, KU Leuven, Leuven, Belgium; Hospital Pharmacy Department, Uni- versity Hospital Leuven, Leuven, Belgium

Leen Haegemans, Faculty of Medicine, KU Leuven, Leuven, Belgium

Karolien Walgraeve, PharmD, Hospital Pharmacy Department, University Hospital Leuven, Leuven, Belgium

Astrid Liesenborghs, PharmD, Hospital Pharmacy Department, University Hospital Leuven, Leuven, Belgium

Astrid Lammens, PharmD, Hospital Pharmacy Department, University Hospital Leuven, Leuven, Belgium

Lorenz Van der Linden, PharmD, PhD, Professor, Department of Pharmaceu- tical and Pharmacological Sciences, KU Leuven, Leuven, Belgium; Hospital Pharmacy Department, University Hospital Leuven, Leuven, Belgium

Mieke Deschodt, RN, PhD, Professor, Department of Public Health and Primary Care, KU Leuven, Leuven, Belgium; Competence Center of Nursing, University Hospital Leuven, Belgium

Jos Tournoy, MD, PhD, Professor, Department of Public Health and Primary Care, KU Leuven, Leuven, Belgium; Department of Geriatric Medicine, Univer- sity Hospital Leuven, Belgium

L. Hellemans et al. / Journal of the American Pharmacists Association 66 (2026) 102928

SCIENCE AND PRACTICE

https://doi.org/10.1016/j.cct.2022.106853

https://doi.org/10.1111/1475%2D6773.12117

https://doi.org/10.1186/1748%2D5908%2D8%2D139

https://doi.org/10.1186/s13012%2D015%2D0209%2D1

http://refhub.elsevier.com/S1544-3191(25)00607-7/sref26

https://doi.org/10.1016/j.jbi.2008.08.010

https://doi.org/10.1016/j.jbi.2019.103208

https://doi.org/10.1016/0022%2D3956%2875%2990026%2D6

https://doi.org/10.1111/j.1532%2D5415.2005.53221.x

https://doi.org/10.1503/cmaj.050051

https://doi.org/10.1111/jep.13121

https://doi.org/10.1007/s00228%2D020%2D02846%2D8

http://refhub.elsevier.com/S1544-3191(25)00607-7/sref34

https://doi.org/10.1186/1748%2D5908%2D8%2D35

https://www.siilo.com/nl/

https://doi.org/10.1186/s12913%2D019%2D3980%2Dx

https://doi.org/10.1177/00469580231218625

https://doi.org/10.1186/s12913%2D020%2D4909%2D0

https://doi.org/10.1016/j.sapharm.2024.04.009

https://doi.org/10.1186/s12913%2D022%2D07583%2D5

https://doi.org/10.1001/jamainternmed.2017.8274

https://doi.org/10.1186/1472%2D6963%2D13%2D485

https://doi.org/10.1016/j.ejim.2019.12.003

Implementation evaluation of a pharmacist-led complex intervention: A mixed-methods analysis embedded within the ASPIRE ran ...

Background
Objectives
Methods and materials

Study design
Study setting
The ASPIRE intervention

What was already known
What this study adds

Implementation strategies
Study sample
Data collection
Analysis
Data integration

Results

Demographics of participants
Fidelity
Intervention duration
Process outcomes
Acceptability and feasibility of the intervention

Discussion

Implications of findings
Limitations

Conclusion
Author contributions
Ethics approval
Consent to participate
Acknowledgments
Supplementary data
References

image11.emf

Journal of Evaluation in Clinical Practice

ORIGINAL PAPER

Effectiveness of Pharmacist Interventions in Improving Medication Use in Hospitalised Older Patients Diagnosed With Cardiovascular Diseases: INFAR Before‐and‐After Study Romana Santos Gama1 | Luiz Carlos Passos2 | Welma Wildes Amorim3 | Renato Morais Souza4 | Marcio Galvão Oliveira1,5

1Postgraduate Program in Pharmaceutical Services and Policies, Federal University of Bahia, Salvador, Bahia, Brazil | 2Department of Internal Medicine,

Postgraduate Program in Medicine and Health, Federal University of Bahia, Salvador, Bahia, Brazil | 3Department of Health Sciences, State University of

Southwest Bahia, Vitória da Conquista Campus, Vitória da Conquista, Bahia, Brazil | 4Sarah Kubitschek Rehabilitation Center, Brasilia, Distrito Federal,

Brazil | 5Multidisciplinary Institute in Health–Anísio Teixeira Campus, Federal University of Bahia, Vitória da Conquista, Bahia, Brazil

Correspondence: Marcio Galvão Oliveira ([email protected])

Received: 3 February 2025 | Revised: 7 May 2025 | Accepted: 14 June 2025

Funding: This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior ‐ Brasil (CAPES) – Finance Code 001.

ABSTRACT Objectives: This study aimed to assess the effectiveness of pharmacist interventions to reduce the omission of evidence‐based cardiovascular medication, as well as polypharmacy and promote the deprescribing of potentially inappropriate medications in

hospitalised older patients diagnosed with cardiovascular diseases.

Methods: This before‐and‐after study was conducted among patients aged ≥ 60 years (n=319) at a cardiovascular hospital in Brazil.

Pharmacists conducted medication reviews for these patients. The first prescription on hospital admission and that at discharge were

collected and compared for prescribing omission, polypharmacy, and prescribing of potentially inappropriate medications.

Results: The mean patient age was 68.9 (±6.2) years. The mean incidences of prescribing potentially inappropriate medications

and omissions decreased from 0.90 at admission to 0.10 at discharge (p< 0.001) and from 0.65 to 0.30 (p< 0.001), respectively. The

number of potentially inappropriate medications prescribed decreased significantly, from 291 at admission to 28 at discharge,

reflecting a 90% reduction. Additionally, the mean number of medications prescribed decreased from 9.8 to 6.5 (p< 0.001).

Conclusion: This study emphasises the role of medication reviews by clinical pharmacists in reducing polypharmacy, pre-

scribing omissions, and inappropriate prescribing in older adults with cardiovascular diseases, demonstrating that targeted

pharmacist interventions improve medication safety.

Trial Registration: Registered on ClinicalTrials.gov (NCT04800900).

1 | Introduction

Cardiovascular diseases (CVDs) are significant health problems in older adults [1]. Notably, they are a leading cause of death, accounting for 17.9 million deaths annually, among

noncommunicable diseases, which cause 74% of global deaths [2]. People aged 65 years and older with CVDs experience a considerable reduction in ‘successful’ life expectancy, defined as a life with good health and functionality [3]. It has been pro- jected that 70% of individuals over 70 will develop CVDs, with

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly

cited.

1 of 10Journal of Evaluation in Clinical Practice, 2025; 31:e70190 https://doi.org/10.1111/jep.70190

https://doi.org/10.1111/jep.70190

https://orcid.org/0000-0002-1677-1251

https://orcid.org/0000-0001-7608-1482

https://orcid.org/0000-0001-6964-892X

https://orcid.org/0000-0003-1075-6410

https://orcid.org/0000-0001-5281-7889

mailto:[email protected]

http://creativecommons.org/licenses/by/4.0/

https://doi.org/10.1111/jep.70190

more than 66% also experiencing additional non‐cardiovascular health issues. Moreover, these individuals are likely to present with multiple chronic health issues that contribute to a sub- stantial medication burden.

Although pharmacological treatments can markedly improve the functionality of physiological systems, they carry a substantial risk of harm [4]. More than half of all preventable harm in medical care worldwide is attributable to medication‐related issues [5]. The prescribing process is associated with the highest prevalence of preventable medication‐related harm [6]. Older patients and those in specialised or emergency care settings are particularly vul- nerable because they are exposed to a high risk of prevent- able medication‐related harm, such as inappropriate prescribing, prescribing omission, and polypharmacy‐related harm. Consequently, integrating pharmacists' medication reviews is a promising strategy for improving the detection of such harm [7].

Studies have indicated that many geriatric patients receive medications without proper indications, at excessive doses, or for prolonged periods, whereas others may not receive the necessary treatment. These issues, characterised by over- prescribing and prescribing omission, increase the risk of adverse health outcomes [8]. To mitigate these risks, medication reviews and deprescribing, i.e., the reduction or discontinuation of medications when their risks outweigh the benefits, should be integral to patient care [8]. Clinical pharmacists play a cru- cial role in this process because their extensive knowledge of medications and regular patient interactions enable them to optimise prescriptions, reduce polypharmacy, and improve overall clinical outcomes [9]. Evidence strongly supports phar- macists’ involvement in the care of hospitalised patients, highlighting their capacity to markedly improve patient safety and treatment effectiveness [8]. In this context, interventions conducted with the participation of clinical pharmacologists or hospital pharmacists (especially with adequately trained spe- cialists) can significantly enhance medication review and the effectiveness of the deprescribing process [10].

This study was aimed at evaluating the effectiveness of phar- macist interventions to reduce the omission of evidence‐based cardiovascular medications as well as polypharmacy and pro- mote the deprescribing of potentially inappropriate medications in hospitalised older patients diagnosed with cardiovascular diseases.

2 | Methods

2.1 | Procedures

2.1.1 | Study Design

INFAR (Effectiveness of Pharmaceutical Interventions in Hos- pitalized Patients; ClinicalTrials.gov identifier NCT04800900) was a before‐and‐after uncontrolled study. In this study, med- ication reviews were conducted by pharmacists at hospital admission and during the hospitalisation of the patients included in the study.

2.1.2 | Population

The INFAR study was conducted from March 2021 to January 2022 at the Ana Nery Hospital as a reference for CVD man- agement in the public health system of Bahia, Brazil. The inclusion criteria were patients aged 60 years or older with a diagnosis of CVD who were hospitalised for more than 24 h. Patients who died in the hospital were excluded from the study because the study focused on evaluating admission and dis- charge prescriptions. Participants were recruited exclusively from the hospital's medical wards and invited to participate in the study after 24 h of hospitalization. Data collection en- compassed the medications prescribed at both hospital admis- sion and discharge.

2.1.3 | Outcomes

Omissions were identified using the START criteria translated into Portuguese and included 22 indicators of potential pre- scribing omissions in older adults, grouped by physiological systems [11]. Polypharmacy was defined as the prescription of 5 or more medications, and excessive polypharmacy as the pre- scription of 10 or more medications. The primary outcome of this study was potentially inappropriate prescription prevalence, assessed using the MPI Brasil application (app) [12]. This application was developed based on the Brazilian Consensus on Potentially Inappropriate Medications for Older People [13], aiming to validate the Beers and STOPP Criteria (2012 and 2006, respectively) contents to establish Brazilian‐specific criteria for older adults. In this study, we used the second version of the app, which incorporates the updates from the STOPP and Beers criteria (2014 and 2019, respectively). The MPI Brasil app is a comprehensive resource for evaluating potentially inappropriate medications and was designed to facilitate access to this infor- mation. It outlines the rationale for classifying a drug as poten- tially inappropriate and specifies exceptional situations in which a medication may not be considered potentially inappropriate. Furthermore, the app includes a convenient search function and suggests safer therapeutic alternatives, deprescribing strategies, and monitoring protocols for cases in which potentially inappropriate medication use is unavoidable [14].

Data on prescribing omissions, polypharmacy, and potentially inappropriate prescribing were collected twice, i.e., in the form of the first prescription 24 h after hospital admission and then the discharge prescription. One prescription per patient was analysed at each stage, ensuring that the same number of patients and prescriptions were included in both time points for comparison. Patient information and details were collected through interviews. Clinical data and prescriptions were obtained from medical records. The questionnaire and tools used for data collection were entered into the KoBoToolbox® online data collection platform.

2.1.4 | Variables of Interest

The age‐adjusted Charlson Comorbidity Index (ACCI) was chosen to associate the comorbidity burden because this index

2 of 10 Journal of Evaluation in Clinical Practice, 2025

incorporates age into the Charlson Comorbidity Index to predict mortality and survival. The result is determined by the sum of the comorbidity weights plus an additional score for every 10‐year period starting at the age of 50 years [15]. The higher the ACCI score, the lower the estimated 10‐year survival, with scores of 1, 2, 3, 4, 5, 6, and ≥ 7 corresponding to survival rates of 96%, 90%, 77%, 53%, 21%, 2%, and 0%, respectively [15].

Cognitive impairment was assessed using the Mini‐Mental State Examination (MMSE), a screening tool for evaluating the pro- gression of cognitive disorders, the scores of which range from 0 to 30. Cutoff points were adjusted according to participants’ educational levels, because the threshold for abnormality varies with educational attainment [16]. Daily activities and inde- pendence were evaluated using the adapted Katz Index of Independence in Activities of Daily Living, which analyses in- dividuals’ abilities to perform daily tasks such as dressing, mobility, feeding, and attending to physiological needs [17].

2.1.5 | Sampling

To calculate the sample of prescriptions verified before and after the intervention, the following factors were considered: an ex- pected pre‐intervention proportion of potentially inappropriate prescribing of 50%, an expected difference before and after the intervention of −30%, a significance level of 5%, and a test power of 80%. A 20% loss of prescriptions was also considered, resulting in a calculated sample size of 319 prescriptions. The samples were collected using a non‐probabilistic approach.

2.1.6 | Study Procedures

The hospital's pharmacy staff was trained to manage older pa- tients more effectively and conduct medication reviews. Addi- tionally, they were trained to use tools to improve the evaluation of appropriate prescribing for older patients, including a checklist with the START criteria translated into Portuguese, and the MPI Brasil app [14].

2.1.7 | Medication Review

The objective of the medication review was to assess and opti- mise the patient's entire medication regimen. Initially, all medications taken by the patients were identified, including prescription drugs, over‐the‐counter medications, herbal sup- plements, and vitamins. This information was collected through patient interviews, medical records, and occasional discussions with family members or caregivers. Subsequently, each medi- cation was evaluated for appropriateness, effectiveness, safety, and potential adverse effects, considering the patient's health condition, age‐related changes, and other relevant factors. Special emphasis was placed on identifying medications that could exacerbate heart conditions or adversely interact with other treatments.

In addition, a checklist with the START criteria was used to identify any potential prescribing omissions, while the MPI

Brasil app was used to detect potentially inappropriate medi- cations [14, 17]. Based on this comprehensive evaluation, rec- ommendations were made to attending medical teams. These recommendations included necessary interventions, such as the introduction of new therapies, deprescribing of unnecessary or potentially harmful medications, and dosage adjustments to optimise treatment efficacy and safety. Patients were monitored continuously throughout their hospitalisation, with adjust- ments made as needed to align with their evolving clinical conditions, ensuring that the interventions remained both safe and effective.

2.1.8 | Statistical Analysis

Data were analysed using SPSS Statistics for Windows, version 28.0 (IBM Corp., Armonk, NY). Sociodemographic and disease variables have been described as absolute and relative frequen- cies. Differences between the mean values for polypharmacy, potentially inappropriate medication, and prescribing omissions before and after the interventions were assessed using the Wil- coxon nonparametric test, because the Shapiro‐Wilk test indi- cated that the numerical variables were not normally distributed.

2.1.9 | Ethics Approval

This study was approved by the Ethics Committee on Research in Human Beings of Ana Nery Hospital, Salvador, Bahia (opinion number: 4.572.301). All the participants provided written informed consent.

3 | Results

Of the 332 older patients interviewed, 13 were excluded because of death, resulting in a final sample of 319 individuals. The mean patient age was 68.9 (±6.2) years; 60.2% of the patients were male and 85.9% self‐identified as having black or brown skin colour. Based on the ACCI, a high comorbidity index (ACCI > 2) was identified in 71.5% of the patients. Of these patients, 40.1% had an estimated 10‐year survival rate of ≤ 53%. The mean number of comorbidities was 4.56 (±1.76), and the top five medical diagnoses were hypertension (80.6%), heart attack (47.6%), coronary artery disease (42%), heart failure (35.1%), and diabetes (34.5%) (Table 1).

A total of 3109 medications were prescribed during the initial prescription stage and 2060 medications were prescribed at discharge. During the initial prescription stage, the incidence of prescribing omissions was 47.3%; potentially inappropriate prescribing was identified in 79.3% of cases; and polypharmacy was noted in 98.5% of cases, 52.4% of which were classified as excessive polypharmacy. Among patients with polypharmacy (n= 314), 249 (79.3%) displayed at least one potentially inappropriate medication prescribed. Metoclopramide was the most frequently prescribed potentially inappropriate medication upon admission (220 prescriptions); however, the number of instances in which it was classified as such at discharge was 2, a significant decrease. Similarly, omeprazole was the most

3 of 10

frequently prescribed potentially inappropriate medication at discharge (5 prescriptions); the number of instances in which it was classified as such at admission was 9 (Table 2).

Furthermore, the number of potentially inappropriate medica- tions decreased significantly from 291 at admission to 28 at discharge, reflecting an approximately 90% reduction through deprescribing. A total of 207 prescribing omissions were observed at admission and 97 at discharge. In both instances, the highest frequencies of omissions were associated with criteria involving the cardiovascular system (Table 3 and 4).

Additionally, the omission of metformin, a necessary prescrip- tion for patients with type 2 diabetes or metabolic syndrome, was the most common prescribing omission, with 105 instances at admission and 48 at discharge (Tables 3 and 4).

There were significant reductions in the rates of potentially inappropriate prescribing, prescribing omissions, and poly- pharmacy between admission and discharge (Table 5). Specifi- cally, the incidence of potentially inappropriate prescribing was reduced from a mean of 0.90 (±0.7) at admission to 0.10 (±0.4) at discharge (p< 0.001). Similarly, the mean and maximum num- bers of omissions decreased from 0.65 (±0.86) at admission to 0.30 (±0.65) at discharge (p< 0.001) and from 6 at admission to 4 at discharge, respectively. We observed a significant reduction in the proportion of prescription omissions (47.3% vs. 22.9%; p< 0.001), potentially inappropriate prescriptions (79.3% vs. 6.9%; p< 0.001), and polypharmacy (98.4% vs. 85.6%; p< 0.001) from admission to discharge. In addition, among the 273 patients with poly- pharmacy, 20 (7.3%) had at least one potentially inappropriate prescribed medication.

Furthermore, the mean and maximum numbers of drugs pre- scribed decreased from 9.8 (±2.5) at admission to 6.5 (±2.1) at discharge (p< 0.001) and from 19 at admission to 14 at dis- charge, respectively.

TABLE 1 | Sociodemographic and clinical characteristics of the

study population (n= 319).

Characteristics N (%)

Sex

Male 192 (60.2)

Female 127 (39.8)

Literacy

Illiterate 75 (23.5)

Literate 244 (76.5)

Skin colour

White 45 (14.1)

Other (such as black and brown) 274 (85.9)

Smoking

Nonsmoker 105 (32.9)

Smoker and former smoker 214 (67.1)

Alcohol consumption

No 98 (30.7)

Yes 221 (69.3)

Cognitive impairment

No 274 (85.9)

Yes 45 (14.1)

Poor self‐rated health

No 90 (28.2)

Yes 229 (71.8)

Poor self‐rated memory

No 226 (70.8)

Yes 93 (29.2)

ADL impairment

No 285 (89.3)

Yes 34 (10.7)

Delirium

No 314 (98.4)

Yes 5 (1.6)

Charlson Comorbidity Index age‐adjusted 96% survival in 10 years 20 (6.3)

90% survival in 10 years 71 (22.3)

77% survival in 10 years 100 (31.3)

53% survival in 10 years 59 (18.5)

21% survival in 10 years 34 (10.7)

2% survival in 10 years 17 (5.3)

0% survival in 10 years 18 (5.6)

Medication use before admission

No 34 (10.7)

Yes 285 (89.3)

Abbreviation: ADL, Activities of daily living.

TABLE 2 | Top 10 potentially inappropriate medications

prescribed.

Admission prescription (n= 291)

Discharge prescription (n= 28)

Drug N (%) Drug N (%)

Metoclopramide 220 (75.6)

Omeprazole 5 (17.9)

Clonazepam 28 (9.6)

Quetiapine 3 (10.7)

Omeprazole 9 (3.1) Mineral oil 2 (7.1)

Quetiapine 6 (2.1) Clonazepam 3 (10.7)

Mineral oil 5 (1.7) Metoclopramide 2 (7.1)

Dimenhydrinate 4 (1.4) Clonidine 3 (7.1)

Digoxin 2 (0.7) Amitriptyline 4 (7.1)

Clonidine 1 (0.3) Risperidone 5 (7.1)

Hydralazine 2 (0.3) Furosemide 1 (3.6)

Carvedilol 3 (0.3) Hydralazine 2 (3.6)

4 of 10 Journal of Evaluation in Clinical Practice, 2025

T A B L E 3

| P re sc ri bi n g om

is si on

s at

ad m is si on

D ru

g C li n ic al

st at u s

N u m be

r of

p at ie n ts

w it h in d ic at io n

N u m be

r of

om is si on

s F re q u en

cy of

om is si on

s (%

)

M et fo rm

in C as es

of ty pe

2 di ab

et es

or m et ab

ol ic

sy n dr om

e in

th e ab

se n ce

of re n al

dy sf u n ct io n

11 0

10 5

95 .5

A n gi ot en

si n ‐c on

ve rt in g en

zy m e

in h ib it or s

A ft er

ac u te

m yo ca rd ia l in fa rc ti on

18 6

24 12 .9

B et a‐ bl oc ke

rs C as es

of st ab

le ch

ro n ic

an gi n a

19 4

16 8. 2

A n gi ot en

si n ‐c on

ve rt in g en

zy m e

in h ib it or s

C as es

of ch

ro n ic

h ea rt

fa il u re

10 5

10 9. 5

St at in s

C as es

of pe

ri ph

er al

or ce re br al

va sc u la r di se as e

w h en

th e pa

ti en

t's ‘fu

n ct io n al

st at u s’ re m ai n s

‘in de

pe n de

n ce

in da

il y ac ti vi ti es ’a n d li fe

ex pe

ct an

cy is

> 5 ye ar s

24 2

9 3. 7

A n ti pl at el et

th er ap

y C as es

of di ab

et es

m el li tu s if 1 or

m or e

ca rd io va sc u la r ri sk

fa ct or s ar e pr es en

t (h yp

er te n si on

, h yp

er ch

ol es te ro la em

ia , or

h is to ry

of sm

ok in g)

11 2

9 8. 0

St at in s

C as es

of di ab

et es

m el li tu s w it h 1 or

m or e

ca rd io va sc u la r‐a

ss oc ia te d fa ct or

11 0

8 7. 3

A sp ir in

or cl op

id og re l

C as es

w it h a h is to ry

of co ro n ar y ar te ry

di se as e,

ce re br al ,o

r pe

ri ph

er al

va sc u la r di se as e w it h a si n u s

rh yt h m

20 8

7 3. 4

B et a‐ 2 ag on

is ts

or an

ti ch

ol in er gi c

ag en

ts R eg u la r in h al at io n in

ca se s of

m il d‐ to ‐m

od er at e

as th m a or

ch ro n ic

ob st ru ct iv e pu

lm on

ar y di se as e

12 5

41 .7

A n ti de

pr es sa n t dr u gs

C as es

w it h m od

er at e‐ to ‐s ev er e de

pr es si ve

sy m pt om

s in

th e pr ev io u s 3 m on

th s

8 4

A n ti h yp

er te n si ve

th er ap

y C as es

in w h ic h th e sy st ol ic

bl oo

d pr es su re

is co n si st en

tl y h ig h er

th an

16 0 m m H g

26 5

2 0. 8

W ar fa ri n

C as es

of ch

ro n ic

at ri al

fi br il la ti on

53 2

3. 8

A n gi ot en

si n ‐c on

ve rt in g en

zy m e

in h ib it or

or an

gi ot en

si n re ce pt or

bl oc ke

C as es

of di ab

et es

w it h n ep

h ro pa

th y,

i.e ., ov er t

pr ot ei n u ri a,

m ic ro al bu

m in u ri a on

u ri n al ys is

(a lb u m in

le ve l h ig h er

th an

30 m g/ 24

h ), or

re n al

im pa

ir m en

37 2

5. 4

C or ti co st er oi ds

R eg u la r in h al at io n in

ca se s of

m od

er at e‐ to ‐s ev er e

as th m a or

ch ro n ic

ob st ru ct iv e pu

lm on

ar y di se as e

w h en

th e fo rc ed

ex pi ra to ry

vo lu m e in

1 s is

< 50 %

7 2

28 .6

(C on

ti n u es )

5 of 10

4 | Discussion

This before‐and‐after study evaluated the impact of pharmacist interventions on polypharmacy, potentially inappropriate medications, and prescribing omissions. Despite patients being cared for by cardiovascular specialists, polypharmacy, poten- tially inappropriate medications, and prescribing omissions were noted, highlighting the need for comprehensive medica- tion reviews. By incorporating interventions, such as depre- scribing and addressing omissions, we observed a reduction in the overall incidence of potentially inappropriate medication, prescribing omissions, and polypharmacy, underscoring the potential of clinical pharmacy to improve patient outcomes through targeted medication management.

One of the key strengths of this study is the use of an appro- priate sample size drawn from a representative population, which enhanced the generalizability of the findings. Addition- ally, the use of patient medical records allowed for a more accurate assessment of inappropriate prescribing compared with reliance on patient surveys or pharmacy dispensing records. One notable strength is that this is the first study of its kind to evaluate the effectiveness of pharmacist interventions, specifically in older patients with CVDs, making this study unique.

However, this study has some limitations. It did not include an assessment of the acceptability of pharmacist interventions among physicians, which could have provided valuable insights into the practical implementation of these interventions. Additionally, potential confounders were not fully explored, such as interviews with prescribing physicians to understand the rationale behind potentially inappropriate medication or monitoring related adverse events. The impact of inappropriate prescribing on clinical outcomes was also not measured. Fur- thermore, the before‐and‐after study design has inherent methodological limitations, including the absence of blinding, potentially introducing bias. Moreover, we did not adjust for key confounding variables such as baseline disease severity, hospital stay duration, or treating clinical team‐related differ- ences between the two periods.

Despite not being a primary objective, evaluating the effects of deprescribing at discharge could be beneficial. This is particu- larly relevant given that clinical condition relapse is among the concerns associated with deprescription and the prevalence of such relapse [18]. In addition, assessing these effects is impor- tant as it reportedly exerts a protective effect of medication review and deprescribing on rehospitalisation, especially within the first three months after hospital discharge [19].

Furthermore, this being a quasi‐experimental study, the design has inherent limitations, including the lack of random sampling and absence of a comparator group for participants receiving interventions. Additionally, the study design introduced the risk of unrecognised confounding factors and potential biases. Finally, owing to updates to the STOPP/START criteria, future research should explore whether similar results would be obtained using the latest version of the criteria. Unfortunately, the collected data could not be reanalysed using the third ver- sion of the STOPP/START criteria because the pharmacist teamT

A B L E 3

| (C

on ti n u ed

)

D ru

g C li n ic al

st at u s

N u m be

r of

p at ie n ts

w it h in d ic at io n

N u m be

r of

om is si on

s F re q u en

cy of

om is si on

s (%

)

H om

e co n ti n u ou

s ox yg en

th er ap

y C as es

in w h ic h ch

ro n ic

ty pe

1 re sp ir at or y fa il u re

(p O 2 < 60

m m H g;

pC O 2 < 48 .7 5 m m H g)

or ty pe

2 re sp ir at or yf ai lu re

(p O 2 < 60

m m H g;

pC O 2 > 48 .7 5 m m H g)

h as

be en

do cu

m en

te d

1 1

10 0

C al ci u m

an d vi ta m in

D D ia gn

os ed

ca se s of

os te op

or os is

(r ad

io lo gi ca l

ev id en

ce or

fr ac tu re

du e to

pr ev io u s fr ag il it y or

ac qu

ir ed

do rs al

ky ph

os is )

1 1

10 0

A bb

re vi at io n s: pC

O 2, pa

rt ia l pr es su re

of ca rb on

di ox id e;

pO 2, pa

rt ia l pr es su re

of ox yg en

6 of 10 Journal of Evaluation in Clinical Practice, 2025

T A B L E 4

| P re sc ri bi n g om

is si on

s at

di sc h ar ge .

D ru

g C li n ic al

st at u s

N u m be

r of

p at ie n ts

w it h in d ic at io n

N u m be

r of

om is si on

s F re q u en

cy of

om is si on

(% )

M et fo rm

in C as es

of ty pe

2 di ab

et es

or m et ab

ol ic

sy n dr om

e in

th e

ab se n ce

of re n al

dy sf u n ct io n

11 1

48 43 .2

A n gi ot en

si n ‐c on

ve rt in g

en zy m e in h ib it or s

A ft er

ac u te

m yo ca rd ia l in fa rc ti on

19 3

9 4. 7

A n gi ot en

si n ‐c on

ve rt in g

en zy m e in h ib it or s

C as es

of ch

ro n ic

h ea rt

fa il u re

11 3

8 7. 1

A n ti pl at el et

th er ap

y C as es

of di ab

et es

m el li tu s in

w h ic h 1 or

m or e

ca rd io va sc u la r ri sk

fa ct or s ar e pr es en

t (h yp

er te n si on

, h yp

er ch

ol es te ro la em

ia , or

a h is to ry

of sm

ok in g)

11 2

6 5. 4

St at in s

C as es

of di ab

et es

m el li tu s w it h 1 or

m or e ca rd io va sc u la r

as so ci at ed

fa ct or

11 1

6 5. 4

St at in s

C as es

of pe

ri ph

er al

or ce re br al

va sc u la r di se as e w h en

th e pa

ti en

t's ‘fu

n ct io n al

st at u s’ re m ai n s ‘in

de pe

n de

n ce

in da

il y ac ti vi ti es ’a

n d li fe

ex pe

ct an

cy is

> 5 ye ar s

24 7

6 2. 4

B et a‐ 2 ag on

is ts

or an

ti ch

ol in er gi c ag en

ts R eg u la r in h al at io n in

ca se s of

m il d‐ to ‐m

od er at e as th m a

or ch

ro n ic

ob st ru ct iv e pu

lm on

ar y di se as e

11 4

36 .4

C or ti co st er oi ds

R eg u la r in h al at io n in

ca se s of

m od

er at e‐ to ‐s ev er e

as th m a or

ch ro n ic

ob st ru ct iv e pu

lm on

ar y di se as e w h en

th e fo rc ed

ex pi ra to ry

vo lu m e in

1 s is

< 50 %

7 3

42 .9

A n ti de

pr es sa n ts

C as es

w it h m od

er at e‐ to ‐s ev er e de

pr es si ve

sy m pt om

s du

ri n g th e pr ev io u s 3 m on

th s

8 2

A sp ir in

or cl op

id og re l

C as es

w it h a h is to ry

of co ro n ar y ar te ry

di se as e, ce re br al ,

or pe

ri ph

er al

va sc u la r di se as e w it h a si n u s rh yt h m

22 0

2 0. 9

P ro to n pu

m p in h ib it or s

C as es

of se ve re

ga st ro es op

h ag ea l ac id

re fl u x or

pe pt ic

st ri ct u re

re qu

ir in g di la ta ti on

. 4

1 25

W ar fa ri n

C as es

of ch

ro n ic

at ri al

fi br il la ti on

60 1

1. 7

A n ti h yp

er te n si ve

th er ap

y C as es

in w h ic h th e sy st ol ic

bl oo

d pr es su re

is co n si st en

tl y h ig h er

th an

16 0 m m H g

27 5

1 0. 4

7 of 10

no longer had access to medical records. One study suggested that applying the STOPP/START criteria without clinical information may overestimate potentially inappropriate medi- cation rates and underestimate prescribing omissions [20].

In our study, nearly all the individuals experienced poly- pharmacy during hospitalisation. Specifically, excessive poly- pharmacy was observed in 52.4% of the first prescriptions evaluated. A retrospective study involving older patients ad- mitted to the cardiology department reported similar findings, with 95% of the patients experiencing polypharmacy and 69% experiencing excessive polypharmacy [21]. Additionally, a study conducted in 5 hospitals in Spain found polypharmacy and excessive polypharmacy rates of 93.8% and 58.8%, respectively [22]. The high incidences of polypharmacy and excessive polypharmacy observed in our study could be attributed to several factors.

First, the individuals presented a high comorbidity index, which, combined with prescribing omissions and the use of inappropriate medications, likely contributed to the increased incidence of polypharmacy [23]. Additionally, polypharmacy tends to increase during hospitalisation, because patients often require intensive treatment regimens to manage their acute conditions [24]. Moreover, older patients with CVDs are com- monly prescribed multiple medications for primary and sec- ondary prevention, symptom relief, slowing disease progression, and improving overall health outcomes [24]. Patients with a history of cardiovascular events are at an even greater risk of polypharmacy, given the need for comprehensive management to prevent recurrence [25].

Regarding the prescription of potentially inappropriate medi- cation, 79.3% of the patients had at least 1 such prescription at baseline; this finding was consistent with those of similar studies conducted in Brazil (80.2%) and Spain (81.5%) [24, 26]. This high incidence was expected, given the high rates of polypharmacy and comorbidities in the study population. Specifically, in our study, the most frequently prescribed inappropriate medication on admission was metoclopramide, likely because of the prevalence of nausea and vomiting as symptoms in CVDs [27]. Notably, there was a considerable reduction in the number of potentially inappropriate medica- tions from admission to discharge, indicating the effectiveness of deprescribing efforts. Omeprazole was the most frequently

prescribed potentially inappropriate medication at discharge. However, the deprescribing of proton pump inhibitors such as omeprazole is often a gradual process because abrupt dis- continuation can lead to rebound symptoms [28]. This reaction is an example of an adverse drug withdrawal event, which could typically be minimised, or even avoided, through a structured, patient‐centred deprescribing process involving planning, dose tapering, and close monitoring [18]. These findings highlight the importance of careful medication management to reduce the risk of adverse withdrawal effects when deprescribing potentially inappropriate medications.

Similar to other studies, in this study, there were frequent medication omissions during the admission and discharge stages, particularly concerning cardiovascular and endocrine treatments [29]. Almost half of the participants had at least 1 omitted medication upon admission. In addition, the omission of metformin was common at admission. At discharge, the same trends persisted, with cardiovascular and endocrine medications being the most omitted, with metformin remaining the most omitted medication. Metformin omission may be associated with the higher prescribing omissions observed in patients with diabetes [26]. The high frequency of omissions in these patients may be attributed to physicians’ intentions to minimise therapies for patients with CVDs owing to their co- morbidities [30]. Moreover, efforts to reduce polypharmacy among older adults may result in prescribing omissions [29].

Medication reviews among older adults represent a critical component of healthcare management aimed at optimising pharmacotherapy and reducing medication‐related problems [31]. Our results indicate an improvement in prescribing prac- tices at discharge after medication review and pharmacist interventions. This improvement was notable when this study was contrasted with another study, which reported an average of 11.6 (±4.5) medications prescribed for home use, without the implementation of pharmacist interventions [21]. The reduction in polypharmacy observed in our study can be attributed to the deprescribing of potentially inappropriate medications and the reduction in omissions.

Additionally, the proportion of patients taking at least 1 potentially inappropriate medication decreased significantly at discharge. This observation is in good agreement with a systematic review and meta‐analysis of 30 studies, most involving clinical pharma- cologists or hospital pharmacists, in which 15 studies reported a reduction in potentially inappropriate medication at discharge following medication review and deprescribing interventions [10]. These finding reinforces the involvement of clinical pharmacists in significantly improving prescription quality by improving the de- prescribing of inappropriate medications. Furthermore, these results are consistent with findings from systematic reviews and meta‐analyses, which indicate that pharmacist interventions can substantially reduce the incidence of potentially inappropriate medications in older adults [32].

Similarly, there was a significant reduction in drug omissions from admission. In contrast, a study conducted in New Zeal- and reported that the rate of prescribing omissions remained nearly unchanged from admission to discharge (40% vs. 39%) [33]. Similarly, a Brazilian study reported 39.6% omission at

TABLE 5 | Patients with polypharmacy, potentially inappropriate

prescribing, and prescribing omission before and after pharmacist

intervention.

Mean

Admission Discharge p‐value

Omission 0.65 (±0.86) 0.3 (±0.65) < 0.001

Potentially inappropriate prescribing

0.9 (±0.7) 0.1 (±0.4) < 0.001

Number of prescribed drugs

9.8 (±2.5) 6.5 (±2.1) < 0.001

8 of 10 Journal of Evaluation in Clinical Practice, 2025

discharge [29]. None of these studies evaluated the impact of pharmacist interventions on prescribing omissions. This un- derscores the critical role of pharmacist oversight in ensuring that essential evidence‐based treatments for cardiovascular conditions are not omitted during care transitions, which is vital for improving clinical outcomes. The implementation of medication reviews and pharmacist interventions to address omissions likely contributed to optimised treatment regimens and improved patient safety.

5 | Conclusion

This study highlights the role of clinical pharmacists in the management of medication therapy for older adults with CVDs. The significant reductions in polypharmacy, poten- tially inappropriate medications, and prescribing omissions underscore the effectiveness of targeted pharmacist inter- ventions in improving medication safety and quality. Given the complexity of managing medicine in this population, integrating clinical pharmacists into healthcare teams is es- sential. Future research should focus on longitudinal out- comes to further assess the impact of these interventions on patient health beyond hospitalisation and explore their applicability across diverse healthcare settings. Ultimately, enhancing medication management through pharmacist involvement may lead to better health outcomes and im- proved quality of life in older adults with chronic health issues.

Acknowledgements

The authors thank the MPI Brasil Project team for their work on the MPI Brasil app, in memory of Daniel Porto, who contributed tire- lessly to its development. This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior ‐ Brasil (CAPES) – Finance Code 001. The Article Processing Charge for the publication of this research was funded by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior ‐ Brasil (CAPES) (ROR identifier: 00x0ma614).

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

Data cannot be shared for ethical/privacy reasons. The data underlying this article cannot be shared publicly owing to the sensitive nature of the topic and the risk of identification of participants.

Data were stored on a shared, password‐protected channel of a Teams site accessible to all authors until the end of the project.

References

1. E. J. Benjamin, P. Muntner, A. Alonso, et al., “Heart Disease and Stroke Statistics; 2019. Update: A Report From the American Heart Association,” Circulation 10 (2019): 139, https://www.ahajournals.org/ doi/10.1161/CIR.0000000000000659.

2. “Noncommunicable Diseases,” accessed September 21, 2024, https:// www.who.int/news-room/fact-sheets/detail/noncommunicable-diseases.

3. “Successful Aging: Effect of Subclinical Cardiovascular Disease | Cardiology | JAMA Internal Medicine,” JAMA Network, accessed Septem- ber 21, 2024, https://jamanetwork.com/journals/jamainternalmedicine/ fullarticle/216217.

4. R. C. de Andrade, M. M. dos Santos, E. E. Ribeiro, J. D. O. dos Santos Júnior, H. L. M. Campos, and E. B. de Leon Poli- farmácia, medicamentos potencialmente inapropriados e a vulnerabil- idade de pessoas idosas, accessed September 21, 2024, https://www. scielo.br/j/rbgg/a/Fj83KxWRbk7wwHBNDq7tP9v/?lang=en.

5. A. Hodkinson, N. Tyler, D. M. Ashcroft, et al., “Preventable Medi- cation Harm Across Health Care Settings: A Systematic Review and Meta‐Analysis,” BMC Medicine 18 (2020): 313, https://doi.org/10.1186/ s12916-020-01774-9.

6. “Medication Error Definition,” accessed September 21, 2024, https:// www.nccmerp.org/about-medication-errors.

7. “Preventable Medication Harm Across Health Care Settings: A Sys- tematic Review and Meta‐Analysis | BMC Medicine,” Full Text, ac- cessed September 22, 2024, https://bmcmedicine.biomedcentral.com/ articles/10.1186/s12916-020-01774-9.

8. I. A. Scott, S. N. Hilmer, E. Reeve, et al., “Reducing Inappropriate Polypharmacy: The Process of Deprescribing,” JAMA Internal Medicine 175 (2015): 827–834, https://doi.org/10.1001/jamainternmed.2015.0324.

9. K. Cardwell, S. M. Smith, B. Clyne, et al., “Evaluation of the General Practice Pharmacist (GPP) Intervention to Optimise Pre- scribing in Irish Primary Care: A Non‐Randomised Pilot Study,” BMJ Open 10 (2020): e035087, https://doi.org/10.1136/bmjopen- 2019-035087.

10. M. Carollo, S. Crisafulli, G. Vitturi, et al., “Clinical Impact of Medication Review and Deprescribing in Older Inpatients: A Systematic Review and Meta‐Analysis,” Journal of the American Geriatrics Society 72, no. 10 (2024): 3219–3238, https://doi.org/10.1111/jgs.19035.

11. A. C. Luz, M. G. Oliveira, and L. Noblat, “Cross‐Cultural Adaptation and Content Validation of Start,” Sao Paulo Medical Journal 134 (2016): 20–27, https://doi.org/10.1590/1516-3180.2014.00303101.

12. MPI Brasil, accessed May 2, 2025. https://www.socialtech.org.br/ mpibrasil/.

13. M. G. Oliveira, W. W. Amorim, C. R. B. Oliveira, H. L. Coqueiro, L. C. Gusmão, and L. C. Passos Geriatrics, Gerontology and Aging, accessed September 21, 2024, http://www.ggaging.com/details/397/pt- BR/brazilian-consensus-of-potentially-inappropriate-medication-for- elderly-people.

14. W. W. Amorim, L. C. Passos, R. S. Gama, et al., “Using a Mobile Application to Reduce Potentially Inappropriate Prescribing for Older Brazilian Adults in Primary Care: A Triple‐Blind Randomised Clinical Trial,” BMC Geriatrics 24 (2024): 35, https://doi.org/10.1186/s12877- 023-04645-z.

15. M. Charlson, T. P. Szatrowski, J. Peterson, and J. Gold, “Vali- dation of a Combined Comorbidity Index,” Journal of Clinical Epidemiology 47 (1994): 1245–1251, https://doi.org/10.1016/0895- 4356(94)90129-5.

16. S. M. D. Brucki, R. Nitrini, P. Caramelli, P. H. F. Bertolucci, and I. H. Okamoto SUGEST'es para O uso do mini‐EXAME do estado mental no brasil.

17. Y. A. O. Duarte, C. L. de Andrade, and M. L. Lebrão, “O Índex De Katz Na Avaliação Da Funcionalidade Dos Idosos,” Revista da Escola de Enfermagem da USP 41 (2007): 317–325, https://doi.org/10.1590/s0080- 62342007000200021.

18. E. Reeve, F. Moriarty, R. Nahas, J. P. Turner, L. Kouladjian O'Donnell, and S. N. Hilmer, “A Narrative Review of the Safety

9 of 10

https://www.ahajournals.org/doi/10.1161/CIR.0000000000000659

https://www.who.int/news-room/fact-sheets/detail/noncommunicable-diseases

https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/216217

https://www.scielo.br/j/rbgg/a/Fj83KxWRbk7wwHBNDq7tP9v/?lang=en

https://doi.org/10.1186/s12916-020-01774-9

https://www.nccmerp.org/about-medication-errors

https://bmcmedicine.biomedcentral.com/articles/10.1186/s12916-020-01774-9

https://doi.org/10.1001/jamainternmed.2015.0324

https://doi.org/10.1136/bmjopen-2019-035087

https://doi.org/10.1111/jgs.19035

https://doi.org/10.1590/1516-3180.2014.00303101

https://www.socialtech.org.br/mpibrasil/

http://www.ggaging.com/details/397/pt-BR/brazilian-consensus-of-potentially-inappropriate-medication-for-elderly-people

https://doi.org/10.1186/s12877-023-04645-z

https://doi.org/10.1016/0895-4356(94)90129-5

https://doi.org/10.1590/s0080-62342007000200021

Concerns of Deprescribing in Older Adults and Strategies to Mitigate Potential Harms,” Expert Opinion on Drug Safety 17, no. 1 (2018): 39–49, https://doi.org/10.1080/14740338.2018.1397625.

19. A. Fontana, M. Carollo, S. Crisafulli, and G. Trifirò, “Reply To: Deprescribing Is Associated With Reduced Readmission to Hospital: An Updated Meta‐Analysis of Randomized Controlled Trials,” Journal of the American Geriatrics Society 73, no. 1 (2025): 306–311, https://doi. org/10.1111/jgs.19169.

20. C. Ryan, D. O'Mahony, D. Ó. O'Donovan, et al., “A Comparison of the Application of STOPP/START to Patients’ Drug Lists With and Without Clinical Information,” International Journal of Clinical Pharmacy 35 (2013): 230–235, https://doi.org/10.1007/s11096- 012-9733-0.

21. M. Sheikh‐Taha and M. Asmar, “Polypharmacy and Severe Potential Drug‐Drug Interactions Among Older Adults With Cardiovascular Disease in the United States,” BMC Geriatrics 21 (2021): 233, https://doi. org/10.1186/s12877-021-02183-0.

22. M. Baré, M. Lleal, S. Ortonobes, et al., “Factors Associated to Potentially Inappropriate Prescribing in Older Patients According to STOPP/START Criteria: MoPIM Multicentre Cohort Study,” BMC Geriatrics 22 (2022): 44, https://doi.org/10.1186/s12877-021-02715-8.

23. G. Doumat, D. Daher, M. Itani, L. Abdouni, K. El Asmar, and G. Assaf, “The Effect of Polypharmacy on Healthcare Services Utiliza- tion in Older Adults With Comorbidities: A Retrospective Cohort Study,” BMC Primary Care 24 (2023): 120, https://doi.org/10.1186/ s12875-023-02070-0.

24. J. Tamargo, K. P. Kjeldsen, E. Delpón, et al., “Facing the Challenge of Polypharmacy When Prescribing for Older People With Cardiovas- cular Disease. A Review by the European Society of Cardiology Working Group on Cardiovascular Pharmacotherapy,” European Heart Journal ‐ Cardiovascular Pharmacotherapy 8 (2022): 406–419, https://doi.org/10. 1093/ehjcvp/pvac005.

25. M. Volpe, D. Chin, and F. Paneni, “The Challenge of Polypharmacy in Cardiovascular Medicine,” Fundamental & Clinical Pharmacology 24 (2010): 9–17, https://doi.org/10.1111/j.1472-8206.2009.00757.x.

26. A. L. P. M. Mori, R. C. Carvalho, P. M. Aguiar, et al., “Potentially Inappropriate Prescribing and Associated Factors in Elderly Patients at Hospital Discharge in Brazil: A Cross‐Sectional Study,” International Journal of Clinical Pharmacy 39 (2017): 386–393, https://doi.org/10. 1007/s11096-017-0433-7.

27. Z. Guo, X. Yang, M. Chen, J. Liu, L. Xu, and Y. Zhang, “Impact of Cardiogenic Vomiting in Patients With Stemi: A Study From China,” Medical Science Monitor 21 (2015): 3792–3797, https://doi.org/10.12659/ msm.895451.

28. B. Farrell, K. Pottie, W. Thompson, et al., “Deprescribing Proton Pump Inhibitors: Evidence‐Based Clinical Practice Guideline,” Canadian Family Physician 63 (2017): 354–364.

29. A. C. Luz, M. G. Oliveira, and L. Noblat, “Potential Prescribing Omissions According to START Criteria at the Time of Hospital Dis- charge,” Brazilian Journal of Pharmaceutical Sciences 57: e181060, https://doi.org/10.1590/s2175-979020200004181060.

30. M. M. Murillo‐Muñoz, A. Gaviria‐Mendoza, and J. E. Machado‐ Alba, “Potential Prescribing Omissions in Patients With Cardiovascular Disease,” International Journal of Clinical Practice 73 (2019): e13428, https://doi.org/10.1111/ijcp.13428.

31. N. L. C. Yaacob, M. Loganathan, N. A. Hisham, et al., “The Impact of Pharmacist Medication Reviews on Geriatric Patients: A Scoping Review,” Korean Journal of Family Medicine 45 (2024): 125–133, https:// doi.org/10.4082/kjfm.23.0220.

32. S. Zhou, R. Li, X. Zhang, et al., “The Effects of Pharmaceutical Interventions on Potentially Inappropriate Medications in Older Pa- tients: A Systematic Review and Meta‐Analysis,” Frontiers in Public Health 11 (2023): 1154048, https://doi.org/10.3389/fpubh.2023.1154048.

33. M. A. Mohammed, A. H. Y. Chan, S. Chan, et al., “Potentially Inappropriate Medications and Potential Prescribing Omissions in Hospitalised Older Adults in New Zealand: A Retrospective Study,” Australasian Journal on Ageing 42 (2023): 668–674, https://doi.org/10. 1111/ajag.13210.

10 of 10 Journal of Evaluation in Clinical Practice, 2025

https://doi.org/10.1080/14740338.2018.1397625

https://doi.org/10.1111/jgs.19169

https://doi.org/10.1007/s11096-012-9733-0

https://doi.org/10.1186/s12877-021-02183-0

https://doi.org/10.1186/s12877-021-02715-8

https://doi.org/10.1186/s12875-023-02070-0

https://doi.org/10.1093/ehjcvp/pvac005

https://doi.org/10.1111/j.1472-8206.2009.00757.x

https://doi.org/10.1007/s11096-017-0433-7

https://doi.org/10.12659/msm.895451

https://doi.org/10.1590/s2175-979020200004181060

https://doi.org/10.1111/ijcp.13428

https://doi.org/10.4082/kjfm.23.0220

https://doi.org/10.3389/fpubh.2023.1154048

https://doi.org/10.1111/ajag.13210

Effectiveness of Pharmacist Interventions in Improving Medication Use in Hospitalised Older Patients Diagnosed With Cardiovascular Diseases: INFAR Before-and-After Study

1 Introduction
2 Methods

2.1 Procedures

2.1.1 Study Design
2.1.2 Population
2.1.3 Outcomes
2.1.4 Variables of Interest
2.1.5 Sampling
2.1.6 Study Procedures
2.1.7 Medication Review
2.1.8 Statistical Analysis
2.1.9 Ethics Approval

3 Results
4 Discussion
5 Conclusion
Acknowledgements
Conflicts of Interest
Data Availability Statement
References

image12.emf

Deprescribing Medications Among Older Adults From End of Hospitalization Through Postacute Care A Shed-MEDS Randomized Clinical Trial Eduard E. Vasilevskis, MD, MPH; Avantika Saraf Shah, MPH; Emily Kay Hollingsworth, MSW; Matthew Stephen Shotwell, PhD; Sunil Kripalani, MD, MSc; Amanda S. Mixon, MD, MS, MSPH; Sandra F. Simmons, PhD

IMPORTANCE Deprescribing is a promising approach to addressing the burden of polypharmacy. Few studies have initiated comprehensive deprescribing in the hospital setting among older patients requiring ongoing care in a postacute care (PAC) facility.

OBJECTIVE To evaluate the efficacy of a patient-centered deprescribing intervention among hospitalized older adults transitioning or being discharged to a PAC facility.

DESIGN, SETTING, AND PARTICIPANTS This randomized clinical trial of the Shed-MEDS (Best Possible Medication History, Evaluate, Deprescribing Recommendations, and Synthesis) deprescribing intervention was conducted between March 2016 and October 2020. Patients who were admitted to an academic medical center and discharged to 1 of 22 PAC facilities affiliated with the medical center were recruited. Patients who were 50 years or older and had 5 or more prehospital medications were enrolled and randomized 1:1 to the intervention group or control group. Patients who were non–English speaking, were unhoused, were long-stay residents of nursing homes, or had less than 6 months of life expectancy were excluded. An intention-to-treat approach was used.

INTERVENTIONS The intervention group received the Shed-MEDS intervention, which consisted of a pharmacist- or nurse practitioner–led comprehensive medication review, patient or surrogate-approved deprescribing recommendations, and deprescribing actions that were initiated in the hospital and continued throughout the PAC facility stay. The control group received usual care at the hospital and PAC facility.

MAIN OUTCOMES AND MEASURES The primary outcome was the total medication count at hospital discharge and PAC facility discharge, with follow-up assessments during the 90-day period after PAC facility discharge. Secondary outcomes included the total number of potentially inappropriate medications at each time point, the Drug Burden Index, and adverse events.

RESULTS A total of 372 participants (mean [SD] age, 76.2 [10.7] years; 229 females [62%]) were randomized to the intervention or control groups. Of these participants, 284 were included in the intention-to-treat analysis (142 in the intervention group and 142 in the control group). Overall, there was a statistically significant treatment effect, with patients in the intervention group taking a mean of 14% fewer medications at PAC facility discharge (mean ratio, 0.86; 95% CI, 0.80-0.93; P < .001) and 15% fewer medications at the 90-day follow-up (mean ratio, 0.85; 95% CI, 0.78-0.92; P < .001) compared with the control group. The intervention additionally reduced patient exposure to potentially inappropriate medications and Drug Burden Index. Adverse drug event rates were similar between the intervention and control groups (hazard ratio, 0.83; 95% CI, 0.52-1.30).

CONCLUSIONS AND RELEVANCE Results of this trial showed that the Shed-MEDS patient-centered deprescribing intervention was safe and effective in reducing the total medication burden at PAC facility discharge and 90 days after discharge. Future studies are needed to examine the effect of this intervention on patient-reported and long-term clinical outcomes.

TRIAL REGISTRATION ClinicalTrials.gov Identifier: NCT02979353

JAMA Intern Med. 2023;183(3):223-231. doi:10.1001/jamainternmed.2022.6545 Published online February 6, 2023.

Visual Abstract

Supplemental content

Author Affiliations: Author affiliations are listed at the end of this article.

Corresponding Author: Eduard E. Vasilevskis, MD, MPH, Vanderbilt University, 2525 West End Ave, Ste 450, Nashville, TN 37203 ([email protected]).

Research

JAMA Internal Medicine | Original Investigation | LESS IS MORE

(Reprinted) 223

Downloaded from jamanetwork.com by SCELC - Azusa Pacific Univ user on 09/19/2026

https://beta.clinicaltrials.gov/study/NCT02979353?distance=50&term=NCT02979353&rank=1

https://jamanetwork.com/journals/jama/fullarticle/10.1001/jamainternmed.2022.6545?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

https://jamanetwork.com/journals/imd/fullarticle/10.1001/jamainternmed.2022.6545?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

mailto:[email protected]

P olypharmacy is prevalent among older hospitalized patients and is associated with adverse postdischarge outcomes.1-4 Patients and caregivers, overwhelm-

ingly, are willing to deprescribe (ie, stop or reduce) 1 or more of their medications if their physicians agree.5-7 Although deprescribing is effective, substantial evidence gaps remain. For example, the majority of deprescribing interventions have been limited to specific drug classes or medical conditions.8-10

Few deprescribing interventions have considered the en- tirety of the medication list across the breadth of medical conditions. Furthermore, fewer trials have initiated depre- scribing in the hospital, and, to our knowledge, none of these trials have included patients needing continuing care in a post- acute care (PAC) facility.11

There may be advantages to deprescribing in both the acute care and PAC settings for older adults. First, the high preva- lence of polypharmacy and multimorbidity suggests a likeli- hood of benefit.4 Second, older patients may be admitted for conditions associated with specific medications (eg, syn- cope), which may enhance the receptiveness of patients and clinicians to reducing medications.12 Third, both care set- tings provide an opportunity to supervise and monitor poten- tial deprescribing-related adverse drug withdrawal events (ADWEs).13

The purpose of this randomized clinical trial was to evaluate the efficacy of a patient-centered deprescribing intervention (Shed-MEDS [Best Possible Medication History, Evaluate, Deprescribing Recommendations, and Synthesis]) among hospitalized older adults transitioning or being dis- charged to a PAC facility. The primary objective of the inter- vention was to reduce overall medication count at hospital discharge and PAC facility discharge and to assess the main- tenance of intervention effects 90 days after PAC facility dis- charge. The secondary objective was to identify the inter- vention effects on the number of potentially inappropriate medications (PIMs)14-16 and the anticholinergic and sedative drug burden.17

Methods Trial Design, Site, and Population The Shed-MEDS randomized clinical trial 18,19 was approved by the institutional review board at Vanderbilt University Medical Center (VUMC) and a National Institute on Aging– appointed Data Safety Monitoring Board. The trial protocol is provided in Supplement 1. All patients or surrogates pro- vided written informed consent for their participation. We fol- lowed the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline.

Between March 2016 and October 2020, we recruited pa- tients 50 years or older who were admitted to VUMC and had 5 or more prehospital medications. These patients were re- ferred to 1 of 22 PAC facilities within a 9-county area surround- ing VUMC in Nashville, Tennessee. The geographic inclusion area enabled a home visit by the research team at the 90-day follow-up. We excluded patients who were non–English speaking, were unhoused, were long-stay residents of nurs-

ing homes, or had less than 6 months of life expectancy. Participants were identified through a daily physical therapy report of patients who had been recommended for PAC place- ment and subsequently confirmed in the case management system as having a referral to a partner PAC facility.

Measures and Randomization Demographic characteristics, comorbidities, and medication measures of participants (Table 1) were extracted from their medical record or assessed by study personnel via a standard- ized interview. Demographic characteristics were collected to characterize the trial population and included age, sex, race and ethnicity (including Asian; Black or African American; Hispanic or non-Hispanic; Native American or Alaska Native; Pacific Islander or Native Hawaiian; White; and other, un- known, or declined/refused to answer), and educational level (≤high school or > high school diploma). Medical diagnoses (using International Classification of Diseases, Ninth Revision, and International Statistical Classification of Diseases and Re- lated Health Problems, Tenth Revision, codes) were used to cal- culate the Charlson Comorbidity Index, which ranged from 0 to 37, with higher scores indicating greater number of comorbidities.20 We assessed cognitive status using the Brief Interview for Mental Status, which had scores ranging from 0 to 15, with scores below 13 indicating cognitive impairment.21

We conducted a Best Possible Medication History (BPMH) pro- cess, supplementing the initial electronic medical record medi- cation list with patient or surrogate interview, pharmacy re- fill history, review of outside medical records (eg, Medication Admission Record from a transferring facility), and the Ten- nessee Controlled Substance Monitoring Database to deter- mine the total number of medications for each patient.18,22 We also collected the number of outpatient prescribers and phar- macies attributed to the medications present at hospital ad- mission. Finally, we asked patients if they received assis- tance with medication management at home (yes or no), such as picking up refills from the pharmacy, organizing pills, and/or being reminded to take their medications.

We randomized participants in a 1:1 ratio using permuted blocks to the Shed-MEDS intervention (intervention group) or

Key Points Question Does a pharmacist- or nurse practitioner–led patient-centered deprescribing intervention reduce or stop medications for older adults at hospital discharge, postacute care (PAC) facility discharge, and 90-day follow-up?

Findings In this randomized clinical trial that included 372 older adults with polypharmacy who were transitioning from hospitalization to PAC, those who received a patient-centered deprescribing intervention had significantly fewer medications compared with the control group who received usual care at PAC facility discharge and at the 90-day follow-up.

Meaning The findings suggest the safety and effectiveness of the deprescribing intervention in reducing the total medication burden at PAC facility discharge, which was sustained 90 days after this discharge.

Research Original Investigation Deprescribing Medications Among Older Adults Through Postacute Care

224 JAMA Internal Medicine March 2023 Volume 183, Number 3 (Reprinted) jamainternalmedicine.com

Downloaded from jamanetwork.com by SCELC - Azusa Pacific Univ user on 09/19/2026

https://jamanetwork.com/journals/jama/fullarticle/10.1001/jamainternmed.2022.6545?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

https://www.equator-network.org/reporting-guidelines/consort/

http://www.jamainternalmedicine.com?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

to BPMH plus usual care at the hospital and PAC facility (control group). Block size was either 2 or 4 and was selected uniformly at random. A total of 9283 hospitalized patients were screened for eligibility. Figure 1 indicates the frequency of exclusions, with living outside the geographic area being the most common reason.

Shed-MEDS Intervention A description of the intervention protocol has been previ- ously published.19,22 In brief, the hospital intervention phase included 4 steps that were conducted by trained research cli- nicians (ie, pharmacists or nurse practitioners). First, a re- search clinician reviewed all BPMH-listed medications, in- cluding over-the-counter (OTC) medications and supplements, for deprescribing (defined as stopping or reducing medica- tions) indications. Second, research clinicians led a depre- scribing conversation with the patient and/or surrogate, in- cluding medications that were identified in the first step along with medications that the patient or surrogate had an inter- est in deprescribing. This semistructured interview included medication-specific questions about medication knowledge, adherence, adverse effects, and efficacy.23

Third, we discussed potential deprescribing actions with the outpatient prescribers, including dose reductions or titra- tions. We synthesized deprescribing targets into final actions and discussed these targets with the inpatient treatment team,

who implemented them at their own discretion. Fourth, within 48 hours of the transfer to the PAC facility, the research clini- cian contacted via telephone the admission or charge nurse at the receiving PAC facility to review the medication list and agreed-on deprescribing recommendations. During the PAC facility stay, research clinicians called weekly to review the medication administration record and monitor deprescribing actions with the facility’s prescribing authority, including docu- mentation of recommended dose reductions or titrations and related symptom monitoring. At the time of PAC facility dis- charge, research clinicians sent the reconciled PAC discharge list and ongoing deprescribing recommendations to relevant outpatient prescribers.

The Shed-MEDS deprescribing intervention ended at PAC facility discharge. However, we followed up with partici- pants for up to 90 days after discharge, including performing an in-person home visit at the final follow-up time point.

Outcome and Safety Measures The primary outcome was the total medication count at hos- pital discharge and PAC facility discharge, with follow-up as- sessments during the 90 days after PAC facility discharge. Medi- cations included all prescribed or OTC drugs that were scheduled or as needed with oral, intravenous, subcutane- ous, rectal, transdermal, and ophthalmic administrations, while topicals (eg, lotions) and otic medications were excluded.

Table 1. Baseline Characteristics and Length of Stay by Group

Variable

Median (IQR) Standardized mean difference

Control group (n = 186)

Intervention group (n = 186)

Demographic characteristics

Age, y 76.9 (67.0-84.6) 76.4 (69.5-85.3) –0.11

Sex –0.08

Female, No. (%) 112 (60) 117 (63)

Male, No. (%) 74 (40) 69 (37)

Race and ethnicity, No. (%) –0.04

Asian 1 (1) 1 (1)

Black or African American 30 (16) 27 (15)

White, non-Hispanic 155 (83) 158 (85)

≤High school educational level, No. (%) 88 (48) 76 (42) 0.12

Living situation prior to hospitalization, No. (%) 0.26

Home, alone or with family 158 (85) 145 (78)

Assisted living facility 18 (10) 28 (15)

SNF or other facility 10 (5) 13 (7)

Clinical characteristics

CCI 6 (5-9) 7 (5-9) 0.09

Cognitive impairment, No. (%)a 32 (20.4) 46 (28.4) –0.19

Length of stay, d

Hospital admission to trial enrollment 5.0 (4.0-7.75) 5.0 (3.0-7.0) 0.19

Trial enrollment to hospital discharge 2.0 (1.0-4.0) 2.0 (1.0-4.0) 0.02

PAC facility length of stay 23.0 (16.0-34.0) 20.5 (15.0-34.8) 0.06

Medication management measures

No. of outpatient prescribers 3 (2-4) 3 (2-3) –0.05

No. of pharmacies (in past 3 mo) 1.5 (1-2) 1 (1-2) –0.13

Help with medication management, No. (%)b 106 (57) 106 (57) 0.14

Abbreviations: CCI, Charlson Comorbidity Index; PAC, postacute care; SNF, skilled nursing facility. a Scored 12 or lower on Brief

Interview for Mental Status, indicating moderate to severe cognitive impairment.

b Help with medication was defined as assisting the participant with picking up medications from the pharmacy, organizing pills, and/or reminders to take their medications.

Deprescribing Medications Among Older Adults Through Postacute Care Original Investigation Research

jamainternalmedicine.com (Reprinted) JAMA Internal Medicine March 2023 Volume 183, Number 3 225

Downloaded from jamanetwork.com by SCELC - Azusa Pacific Univ user on 09/19/2026

http://www.jamainternalmedicine.com?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

Enrollment (baseline) medication counts were based on the prehospital medication count (per the BPMH) and any newly prescribed in-hospital medications.

The secondary outcomes were the total number of PIMs at each time point, the Drug Burden Index (DBI), and adverse events. A PIM was defined as any medication on 1 of 3 previ- ously published lists, including the Beers Criteria,14 the STOPP (Screening Tool of Older Persons' Prescriptions) Criteria,15 and the RASP (Rationalization of Home Medication by an Ad- justed STOPP in Older Patients) List.16 The PIM classification assessed only for presence or absence on any list regardless of the clinical context or indication. Additionally, we measured the DBI, a validated continuous measure of sedative and an- ticholinergic burden.17 The DBI is the sum of each medica-

tion’s daily dose divided by the minimum effective dose (as estimated by the US Food and Drug Administration [FDA] minimum recommended dose) and the patient’s daily dose. The DBI captures reductions in dose, even when the total num- ber of medications is not reduced. For both the intervention and control groups, we examined the frequency, by FDA gen- eral drug categories, of dose-reduced or stopped medications at PAC facility discharge.24

We measured adverse events throughout the trial (from en- rollment to 90 days after PAC facility discharge) in both groups. An adverse event was defined as any unplanned emergency department visit, an intensive care unit transfer among cur- rently hospitalized patients, a new hospitalization, or death, with the latter 3 events being categorized as serious. All ad- verse events were assessed by a trained physician reviewer using the 10-point Naranjo Scale to determine whether events were possibly related (score >1) to an adverse drug event (ADE) or ADWE.25,26 Physician reviewers were blinded to the group assignment. Further details on adverse event determination are provided in the eMethods in Supplement 2.

Blinding Due to the in-person nature of the Shed-MEDS intervention, the clinical research staff conducting the intervention and collecting data at the follow-up time point could not be blinded to the intervention status. The primary investigators and trained phy- sician reviewers (including A.S.M. and S.K.) for safety measures (eg, ADEs) were blinded to the group assignment.

Statistical Analysis The effect of the intervention on the total number of medica- tions and PIMs at hospital discharge, PAC facility discharge, and 90 days after PAC facility discharge was quantified using mixed- effects Poisson regression, adjusting for measurement time point (as a categorical covariate), the enrollment medication count for each outcome, and the interaction of intervention and time point. The within-participant association among repeated mea- surements was modeled using a random intercept indexed by participant. The overall statistical significance of the treat- ment effect was evaluated using a Wald-type multiple degree of freedom, which tested the null hypothesis that the interven- tion had no effect at any time point (hospital discharge, PAC facility discharge, and 90-day follow-up) after randomization. Two-sided P < .05 was considered statistically significant. No adjustments were made to control the familywise type-I error probability for multiple testing. The intervention effect at each time point was summarized using the estimated relative effect on the mean, or mean ratio, with 95% CIs.

In a sensitivity analysis, we repeated the main analysis ex- cluding OTC medications. The effect of intervention on the DBI was assessed in a similar fashion using linear mixed-effects regression rather than Poisson regression. Data for all random- ized patients who were discharged from the hospital to a part- ner PAC facility were included in the analyses. All participant measurements, regardless of subsequent attrition, were used in the analyses. The incidence of missing data due to attrition was examined by group assignment. Unadjusted compari- sons across groups were made using the Wilcoxon test or Pear-

Figure 1. Flow of Participants in Shed-MEDS

9283 Inpatients aged ≥50 y evaluated for SNF placement

1353 Patients eligible for Shed-MEDS

186 Randomized to intervention (intervention group)

372 Randomized

7930 Did not meet inclusion criteria 4378 Lived outside

geographic region 1910 Discharged within 48 h

of screening or to a non-SNF nonpartner SNF disposition

789 Less than 6-mo mortality 366 No polypharmacy 184 Homeless, incarcerated,

or no telephone 171 Admitted from long-term

care facility 68 Did not speak English 64 Participating in another

drug trial or VA insurance program

911 Declined participation 70 Not approached for

participation (eg, no surrogate available, COVID-19 restrictions)

174 Completed hospital intervention

7 Declined further participation

5 Transferred to hospice care

142 Completed PAC facility intervention 24 Discharged to a nonpartner

PAC 4 Transferred to hospice care 3 Died 1 Declined further participation

116 Completed 90-d follow-up 10 Transferred to hospice care 7 Died 9 Declined further participation

or could not be contacted

186 Randomized to usual care (control group) 181 Completed usual care

2 Declined further participation

2 Transferred to hospice care

142 Completed PAC facility usual care 29 Discharged to a nonpartner

PAC 9 Died 1 Declined further participation

115 Completed 90-d follow-up 12 Transferred to hospice care 1 Died

14 Declined further participation or could not be contacted

1 Died

PAC indicates postacute care; SNF, skilled nursing facility; VA, Veterans Affairs.

Research Original Investigation Deprescribing Medications Among Older Adults Through Postacute Care

226 JAMA Internal Medicine March 2023 Volume 183, Number 3 (Reprinted) jamainternalmedicine.com

Downloaded from jamanetwork.com by SCELC - Azusa Pacific Univ user on 09/19/2026

https://jamanetwork.com/journals/jama/fullarticle/10.1001/jamainternmed.2022.6545?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

http://www.jamainternalmedicine.com?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

son χ2 test, and summaries were generated using the sample proportion, median, and IQR as appropriate. The rates of ad- verse events, including those that reoccurred, were com- pared across groups by computing a hazard ratio (HR) with 95% CI using a shared frailty model adjusted for the assigned group (intervention vs control), with frailty term indexed by participant. This type of model is an extension of the Cox pro- portional hazards regression model, in which the hazard func- tion has a random effect that accounts for heterogeneity among participants. The frailty model is used when the event of interest may occur multiple times for any one participant (eg, rehospitalization).27

At the time of trial initiation, we expected to enroll ap- proximately 576 participants across 4 years of enrollment with 27.5% attrition, or 420 total participants. The pilot interven- tion was associated with an approximately 50% reduction in the count of total medications from enrollment to hospital discharge, whereas an approximately 25% reduction was ob- served in the control group under usual hospital care.19 Using the mixed-effects Poisson regression method, we imple- mented a simulation-based power analysis assuming that these effects would be attenuated by 20% at PAC facility discharge and again by 20% at the 90-day follow-up. A planned sample of 420 participants provided greater than 95% power to de- tect this magnitude of effect.

An intention-to-treat method was used for all statistical analyses: all participants were analyzed according to the group assignment at the time of randomization, regardless of sub- sequent compliance with the study protocol or follow-up com- pleteness. Data were managed using the REDCap platform.28

All statistical analyses were performed with R, version 4.0.3 (R Foundation for Statistical Computing).

Results Participants Of the 1353 eligible patients, 372 (28%) provided consent and were randomized to either the intervention group (n = 186) or control group (n = 186) (Figure 1). A total of 284 partici- pants (142 in the intervention group and 142 in the control group) were included in the intention-to-treat analysis. Table 1 shows participant demographic and clinical charac- teristics for the total sample and by group. Overall, partici- pants had a mean (SD) age of 76.2 (10.7) years and included 229 females (62%) and 143 males (38%), with most being of White race and ethnicity (312 [84%]) and admitted from home (303 [81%]). Participants had a median (IQR) Charlson Comorbidity Index of 6.0 (5.0-9.0), and 78 (21%) exhibited cognitive impairment. The median (IQR) hospital length of stay from enrollment to hospital discharge was 2.0 (1.0-4.0) days, and the median (IQR) PAC length of stay was 22 (15.0- 34.0) days. Patients received prescriptions from a median (IQR) of 3 (2-3) outpatient prescribers and 1 (1-2) pharmacy in the previous 3 months. A total of 212 patients (57%) reported receiving help with medication management prior to hospi- tal admission, and 123 patients (33%) reported not receiving such help.

Baseline Medications and Intervention Effects on Outcome Measures The median (IQR) number of prehospital medications was 16 (12.0-20.0) per patient. At enrollment, the median (IQR) total number of prehospital medications in addition to new hospi- tal medications was 23.0 (19.0-29.0) per patient (Table 2).

The median (IQR) total number of medications stopped at PAC facility discharge (from the time of enrollment) was 14.0 (11.0-18.0) in the intervention group and 12.0 (9.0-16.0) in the control group. The median (IQR) total number of medica- tions started at PAC facility discharge was 3.0 (1.0-5.0) in the intervention group and 3.0 (1.2-4.8) in the control group. Ad- justing for the total medication count at enrollment, the mean total medication count was similar between intervention and control groups at hospital discharge, but patients in the inter- vention group had a mean of 14% fewer medications at PAC facility discharge (mean ratio, 0.86; 95% CI, 0.80-0.93; P < .001) when the intervention ended and a mean of 15% fewer medications 90 days after PAC facility discharge (mean ratio, 0.85; 95% CI, 0.78-0.92; P < .001) (Table 2).

The magnitude of the overall treatment effect was high across the entire period of observation (P < .001, testing the null hypothesis that the intervention had no effect at any time point), with the intervention group having significantly fewer medications compared with the control group (hospital dis- charge mean ratio, 0.96 [95% CI, 0.90-1.02]; PAC discharge mean ratio, 0.86 [95% CI, 0.80-0.93]; 90-day follow-up mean ratio, 0.85 [95% CI, 0.78-0.92]) (Figure 2). Similar treatment effects were observed in sensitivity analyses that excluded all OTC medications (hospital discharge mean ratio, 0.95 [95% CI, 0.89-1.02]; PAC discharge mean ratio, 0.87 [95% CI, 0.80-0.95]; 90-day follow-up mean ratio, 0.87 [95% CI, 0.80- 0.96]) (eTable 1 in Supplement 2).

The mean (SD) number of PIMs was similar between the intervention and control groups at hospital discharge (9.3 [3.5] vs 8.8 [3.3]). The intervention group, however, was pre- scribed significantly fewer PIMs at PAC facility discharge (7.7 [3.0]) and during the 90 days after PAC facility discharge (8.9 [3.5]). The DBI was significantly lower for the intervention group at each time point (hospital discharge mean differ- ence, –0.28 [95% CI, –0.51 to –0.04]; PAC discharge mean dif- ference, –0.60 [95% CI, –0.86 to –0.34]; 90-day follow-up mean difference, –0.35 [95% CI, –0.63 to –0.07]). For both PIMs and DBI, the overall treatment effects had a great magnitude (P < .001). The PIM counts and DBI values across time points are shown in eFigures 1 and 2 in Supplement 2, respectively.

Medication Deprescribing by Drug Categories and Rates of Adverse Events The Shed-MEDS intervention deprescribed medications across numerous drug classes according to FDA general drug catego- ries. eFigure 3 in Supplement 2 displays the most frequently deprescribed medication classes (at least 40 deprescribing events), with vitamins or supplements, laxatives, and antihy- pertensives being the most frequently deprescribed medica- tions at PAC facility discharge. eTable 2 in Supplement 2 provides further description of the top 3 medications depre- scribed by drug categories.

Deprescribing Medications Among Older Adults Through Postacute Care Original Investigation Research

jamainternalmedicine.com (Reprinted) JAMA Internal Medicine March 2023 Volume 183, Number 3 227

Downloaded from jamanetwork.com by SCELC - Azusa Pacific Univ user on 09/19/2026

https://jamanetwork.com/journals/jama/fullarticle/10.1001/jamainternmed.2022.6545?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

http://www.jamainternalmedicine.com?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

The rates of overall adverse events, ADEs, and ADWEs were comparable between intervention and control groups (HR, 0.83; 95% CI, 0.52-1.30). There was a consistent pattern to- ward lower rates in the intervention group (Table 3). For ex- ample, the most frequent adverse event was hospitalization, with 84 events in the intervention group and 107 in the control group (HR, 0.76; 95% CI, 0.53-1.09).

Discussion After the patient-centered Shed-MEDS deprescribing inter- vention, participants at PAC facility discharge experienced a significant reduction in total medications across a wide range of drug categories, which was sustained up to 90 days after the end of active deprescribing, when compared with pa- tients who received usual care. Additionally, the intervention decreased PIMs and lowered the DBI across the intervention period and during the 90-day follow-up. The intervention was not associated with increased rates of overall adverse events or ADWEs. These findings demonstrated that a hospital- initiated, pharmacist- or nurse practitioner–led, patient- centered deprescribing intervention can effectively and safely

reduce the medication burden among hospitalized older pa- tients needing PAC.

Although this trial was the first, to our knowledge, to imple- ment deprescribing among hospitalized older patients transi- tioning to PAC, this trial was preceded by other inpatient depre- scribing studies.29-35 The overall effect size of deprescribing in this trial was larger at the end of the intervention (PAC fa- cility discharge) in comparison to the effect size in previous reports. Compared with other inpatient trials, in the present work, we were able to suggest deprescribing actions through- out the PAC facility stay, thus potentially enhancing the effi- cacy of the Shed-MEDS intervention. In addition, most previ- ous trials used explicit tools to target PIMs for deprescribing, most often defined by the Beers Criteria. A systematic review identified 9 such trials, 3 of which demonstrated statistically significant reductions in PIMs between hospital admission and discharge.11 The number of PIMs at enrollment in those studies was lower than the population of the present trial, a difference that may be explained by the more restrictive ap- proach taken by previous studies to classify PIMs in compari- son to this study, which included any medication on 3 PIM lists, regardless of prescribing indication or patient condition. Ad- ditionally, the BPMH process we followed was exhaustive

Table 2. Primary and Secondary Medication Outcomes by Group and Time Point

Outcome

Median (IQR) Mean ratio or mean difference (95% CI)aControl group (n = 186) Intervention group (n = 186)

Primary outcome

Total No. of medications

Prehospital and new hospital medications at trial enrollment (n = 372) 23.0 (19.0 to 29.0) 24.0 (19.2 to 29.0) NA

Prehospital 16.0 (11.2 to 20.0) 16.0 (12.0 to 20.0) NA

Hospital discharge (n = 355) 16.0 (12.0 to 19.0) 15.0 (12.0 to 18.0) 0.96 (0.9 to 1.02)

PAC facility discharge (n = 284) 15.0 (12.0 to 19.0) 13.0 (10.0 to 16.0) 0.86 (0.80 to 0.93)b

90-d Follow-up (n = 231) 14.0 (11.0 to 18.0) 12.0 (9.0 to 16.0) 0.85 (0.78 to 0.92)b

Overall treatment effect P value NA NA <.001c

Secondary outcome

No. of PIMs

Prehospital and new hospital PIMs at trial enrollment (n = 372) 13.0 (10.0 to 17.0) 13.0 (11.0 to 17.0) NA

Prehospital 10.0 (7.0 to 13.0) 10.0 (7.0 to 13.0) NA

Hospital discharge (n = 355) 9.0 (7.0 to 11.0) 9.0 (6.2 to 11.0) 0.95 (0.89 to 1.03)

PAC facility discharge (n = 284) 9.0 (7.0 to 11.0) 7.0 (6.0 to 10.0) 0.86 (0.79 to 0.93)b

90-d Follow-up (n = 231) 9.0 (6.0 to 11.5) 8.0 (5.0 to 10.0) 0.88 (0.80 to 0.97)d

Overall treatment effect P value NA NA <.001c

DBI

Prehospital and new hospital DBI at trial enrollment (n = 372) 4.5 (3.1 to 6.3) 4.1 (2.9 to 5.4) NA

Prehospital 3.2 (1.8 to 4.7) 2.9 (1.7 to 4.2) NA

Hospital discharge (n = 355) 2.9 (1.9 to 4.1) 2.4 (1.5 to 3.4) –0.28 (–0.51 to –0.04)e

PAC facility discharge (n = 284) 2.7 (1.5 to 4.3) 2.2 (1.3 to 3.1) –0.59 (–0.85 to –0.34)b

90-d Follow-up (n = 231) 2.3 (1.3 to 3.7) 2.0 (1.0 to 2.8) –0.34 (–0.63 to –0.07)e

Overall treatment effect P value NA NA <.001c

Abbreviations: DBI, drug burden index; NA, not applicable; PAC, postacute care; PIM, potentially inappropriate medication. a Mean ratios are reported from the mixed-effects Poisson regression analyses for

the total number of medications outcome and the PIMs. Mean differences are reported from the mixed-effects linear regression analysis for the DBI. All analyses were adjusted for time by group interaction and enrollment medication count (prehospital medication count and new medications at time of enrollment).

b P � .001. c Overall treatment effect tested the null hypothesis that the intervention had

no effect at hospital discharge, PAC facility discharge, or 90-day follow-up time point.

d P � .01. e P = .02.

Research Original Investigation Deprescribing Medications Among Older Adults Through Postacute Care

228 JAMA Internal Medicine March 2023 Volume 183, Number 3 (Reprinted) jamainternalmedicine.com

Downloaded from jamanetwork.com by SCELC - Azusa Pacific Univ user on 09/19/2026

http://www.jamainternalmedicine.com?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

and identified patient medications that may be missing from standard admission medications lists.18 In a meta-analysis of 25 deprescribing trials, the population effect was a mean dif- ference of 0.4 drugs between the intervention and control groups.36 Recently, a large multicenter deprescribing cluster randomized trial (OPERAM [Optimizing Therapy to Prevent Avoidable Hospital Admissions in Multimorbid Older Adults]) did not result in significant differences in PIMs between the intervention and control groups.37 The larger absolute amount of deprescribing achieved in the present trial (eg, 2 fewer PIMs at PAC facility discharge) may be attributed to a consider- ation of a broader range of medications and the continued deprescribing throughout the PAC facility stay.

The intervention group in this trial experienced signifi- cant reductions in total medications at PAC facility discharge, and those reductions were maintained 90 days after dis- charge. Medication reductions, however, were modest at the earliest postenrollment time point (hospital discharge). The smaller effect size at hospital discharge may be explained by 2 factors. First, the median time between enrollment and hospital discharge was only 2 days. This period allowed for a limited time to fully implement all deprescribing steps. The disadvantage of delayed enrollment was counterbalanced by the ability to continue the intervention in the PAC setting, where the median length of stay was 22 days. This time al- lowed for weekly opportunities to perform medication re- views, respond to clinical concerns, and reinforce deprescrib- ing recommendations with the PAC team. Without the ongoing involvement of a clinician facilitating deprescribing after PAC facility discharge, the intervention effect appeared durable but did not increase in magnitude. Hospitalized older patients commonly experience recurrent acute care episodes (ie, emer- gency department visits or rehospitalizations), which is a risk factor for initiation of new PIMs.38 Future interventions may incorporate ongoing surveillance to address potential new deprescribing indications.

The PIMs and DBI similarly decreased among partici- pants in the intervention group. This finding suggests that medication reductions include those with high clinical rel- evance, such as those associated with geriatric syndromes (eg, falls). Clearly identifying PIMs (including sedating and anticholinergic medications) according to multiple validated lists and the DBI may aid clinicians in prioritizing medica- tions for deprescribing, especially when the starting list is substantial. There are multiple drug classes that may not be included on specific PIM lists or the DBI but may cause unin- tended harm. For example, guidelines for diabetes control change among older patients; however, patients may con- tinue to be on higher medication doses for which the risk of harms may outweigh the benefits.39 In addition, although there are benefits to lower blood pressure targets for older patients, particularly among those who are frail and continu- ing care at a PAC facility, there continue to be opportunities to reduce the medication burden of patients with hyperten- sion in response to either ongoing hypotension; ongoing symptoms (eg, lightheadedness); or admissions with falls, fractures, or syncope.40-44 In this trial, the most frequently deprescribed drug classes were supplements and laxatives.

Commonly deprescribed medications for chronic disease included those for hypertension and diabetes. This finding suggests that in addition to focusing on PIMs, assessing the appropriateness of supplements and blood pressure and hemoglobin A1c targets is important in developing depre- scribing strategies to reduce the complexity and potential harm of medication regimens.

A key concern of initiating deprescribing, from both pa- tient and clinician perspectives, is developing serious ad- verse events, including hospitalization and death.45 We found no increases in emergency department visits, hospitaliza- tions, or mortality. The number of ADWEs was modest in the intervention group and did not differ significantly from that in the control group. This similarity in adverse event out- comes supports the safety of a patient-centered deprescrib- ing strategy when initiated in the hospital and continued in the PAC facility. These results can support the conversations about the safety of deprescribing. Another barrier of deprescribing, specifically in the PAC setting, is the availability and training of clinicians. The efficacy of deprescribing in PAC would suggest the benefits of engaging clinical pharmacists or nurse practitioners with specific abilities to perform comprehen- sive medication reviews, identify opportunities to depre- scribe, monitor symptoms, and provide patient and clinician education at the time of discharge.

Limitations These findings must be interpreted in consideration of trial limitations. First, patient enrollment at a single academic hos- pital may limit generalizability, although this was balanced by the inclusion of 22 PAC facilities. Second, enrolled patients may be more willing to deprescribe. This potential for enrollment bias could increase intervention effects; however, it also could increase deprescribing behavior in the control group. Third, the research clinicians provided deprescribing recommenda- tions but did not implement the actual deprescribing actions, which remained the responsibility of the primary prescrib- ers. Although this approach may have weakened the inten-

Figure 2. Total Medication Counts Over Time

To ta

l m ed

ic at

io n

co un

Prehospitalization Hospital discharge

90-d Follow-up

Trial enrollment

Control group Intervention group

PAC facility discharge

Active deprescribing began at trial enrollment and stopped at postacute care (PAC) facility discharge. Each point represents the mean, and each error bar represents the SE.

Deprescribing Medications Among Older Adults Through Postacute Care Original Investigation Research

jamainternalmedicine.com (Reprinted) JAMA Internal Medicine March 2023 Volume 183, Number 3 229

Downloaded from jamanetwork.com by SCELC - Azusa Pacific Univ user on 09/19/2026

http://www.jamainternalmedicine.com?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

sity of the deprescribing effect, we believe it was a pragmatic and acceptable approach for both clinicians and patients and may help explain the sustainability of the intervention ef- fects. Fourth, due to the large number of PAC sites and small numbers of enrolled participants at some of these sites, we did not adjust for site-level effects. Fifth, inpatient clinicians were not blinded to the intervention. Thus, it is possible that new deprescribing skills and behaviors may have been applied to patients in the control group, which would reduce the ob- served degree of differences between the intervention and control groups.

Conclusions

In this randomized clinical trial, older adults requiring PAC after hospitalization experienced high levels of polypharmacy. The patient-centered Shed-MEDS deprescribing intervention was found to be safe and effective in reducing the total medi- cation burden at PAC facility discharge, with a sustained ef- fect of up to 90 days after discharge. Future research is needed to examine the effects of deprescribing on patient-reported and additional long-term clinical outcomes.

ARTICLE INFORMATION

Accepted for Publication: December 4, 2022.

Published Online: February 6, 2023. doi:10.1001/jamainternmed.2022.6545

Author Affiliations: Center for Quality Aging, Vanderbilt University Medical Center, Nashville, Tennessee (Vasilevskis, Shah, Hollingsworth, Kripalani, Mixon, Simmons); Geriatric Research Education and Clinical Center, Veterans Affairs Tennessee Valley Healthcare System, Nashville, Tennessee (Vasilevskis, Mixon, Simmons); Section of Hospital Medicine, Division of General Internal Medicine and Public Health, Vanderbilt University Medical Center, Nashville, Tennessee (Vasilevskis, Kripalani, Mixon); Center for Health Services Research, Vanderbilt University Medical Center, Nashville, Tennessee (Vasilevskis, Kripalani, Mixon, Simmons); Department of Biostatistics, Vanderbilt University, Nashville, Tennessee (Shotwell); Division of Geriatrics, Vanderbilt University Medical Center, Nashville, Tennessee (Simmons).

Author Contributions: Drs Vasilevskis and Simmons had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Concept and design: Vasilevskis, Shah, Shotwell, Kripalani, Simmons. Acquisition, analysis, or interpretation of data: Vasilevskis, Shah, Hollingsworth, Shotwell, Mixon, Simmons. Drafting of the manuscript: Vasilevskis, Shah, Hollingsworth, Shotwell, Simmons. Critical revision of the manuscript for important intellectual content: Vasilevskis, Shah, Shotwell, Kripalani, Mixon, Simmons. Statistical analysis: Vasilevskis, Shah, Shotwell, Simmons. Obtained funding: Vasilevskis, Simmons. Administrative, technical, or material support: Shah,

Hollingsworth, Mixon. Supervision: Vasilevskis, Shah, Kripalani, Simmons. Other - I'm the principal investigator of this study: Simmons.

Conflict of Interest Disclosures: Dr Kripalani reported receiving grants from Bristol Myers Squibb/Sanofi and grants from IBM Corporation outside the submitted work. Dr Mixon reported receiving a Health Services Research and Development Service grant from the US Department of Veterans Affairs outside the submitted work. No other disclosures were reported.

Funding/Support: This study was funded by grant R01AG053264 from the National Institute on Aging of the National Institutes of Health (co–principal investigators: Drs Vasilevskis and Simmons). The use of institutional data management system REDCap was supported by a Clinical and Translational Science Awards grant UL1TR000445 from the National Center for Advancing Translational Sciences.

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data Sharing Statement: See Supplement 3.

Additional Contributions: We thank the members of the Shed-MEDS team, including the following personnel who are not listed as coauthors: Carole Bartoo, GNP; Jennifer Kim, GNP; Kanah Lewallen, GNP; Whitney Narramore, PharmD; Robin Parker, PharmD; Susan Lincoln, BS; Joanna Gupta, MEd; and Jessica Lovell, PharmD. They received no additional compensation, beyond their usual salary, for their contribution to this work.

REFERENCES

1. Huang ES, Karter AJ, Danielson KK, Warton EM, Ahmed AT. The association between the number of prescription medications and incident falls in a multi-ethnic population of adult type-2 diabetes patients: the diabetes and aging study. J Gen Intern Med. 2010;25(2):141-146. doi:10.1007/s11606-009- 1179-2

2. Shmuel S, Lund JL, Alvarez C, et al. Polypharmacy and incident frailty in a longitudinal community-based cohort study. J Am Geriatr Soc. 2019;67(12):2482-2489. doi:10.1111/jgs.16212

3. Picker D, Heard K, Bailey TC, Martin NR, LaRossa GN, Kollef MH. The number of discharge medications predicts thirty-day hospital readmission: a cohort study. BMC Health Serv Res. 2015;15:282. doi:10.1186/s12913-015-0950-9

4. Saraf AA, Petersen AW, Simmons SF, et al. Medications associated with geriatric syndromes and their prevalence in older hospitalized adults discharged to skilled nursing facilities. J Hosp Med. 2016;11(10):694-700. doi:10.1002/jhm.2614

5. Reeve E, Wolff JL, Skehan M, Bayliss EA, Hilmer SN, Boyd CM. Assessment of attitudes toward deprescribing in older Medicare beneficiaries in the United States. JAMA Intern Med. 2018;178(12):1673-1680. doi:10.1001/jamainternmed. 2018.4720

6. Weir KR, Ailabouni NJ, Schneider CR, Hilmer SN, Reeve E. Consumer attitudes towards deprescribing: a systematic review and meta-analysis. J Gerontol A Biol Sci Med Sci. 2022;77 (5):1020-1034. doi:10.1093/gerona/glab222

7. Hollingsworth EK, Shah AS, Shotwell MS, Simmons SF, Vasilevskis EE. Older patient and surrogate attitudes toward deprescribing during the transition from acute to post-acute care. J Appl

Table 3. Safety Outcomes: per Person per Month Rates of Health Care Use, Death, and ADEs by Group

Type of event

Intervention group (n = 186 [606.32 person-mo])

Control group (n = 186 [595.39 person-mo])

Hazard ratio (95% CI)a Total No. of events

Total events per person-mo, rate

Total No. of events

Total events per person-mo, rate

Overall adverse events 129 0.213 166 0.279 NA

ED visit 32 0.053 44 0.074 0.74 (0.44-1.25)

Hospitalization 84 0.139 107 0.180 0.76 (0.53-1.09)

Death 10 0.016 11 0.018 0.90 (0.38-2.12)

Transfer to ICU 3 0.005 4 0.007 0.74 (0.17-3.32)

ADEs 44 0.073 54 0.091 0.83 (0.52-1.30)

ADWEs 19 0.031 24 0.040 NA

Abbreviations: ADE, adverse drug event; ADWE, adverse drug withdrawal event; ED, emergency department; ICU, intensive care unit; NA, not applicable. a No suitable method was available to

model the overall adverse events, acompositeofoutcomesthatcanoccur multipletimesandotheroutcomesthat can occur only once. There was insufficient information in the data to estimate a hazard ratio for ADWEs. Thus, no hazard ratio is presented for these 2 outcomes.

Research Original Investigation Deprescribing Medications Among Older Adults Through Postacute Care

230 JAMA Internal Medicine March 2023 Volume 183, Number 3 (Reprinted) jamainternalmedicine.com

Downloaded from jamanetwork.com by SCELC - Azusa Pacific Univ user on 09/19/2026

https://jamanetwork.com/journals/jama/fullarticle/10.1001/jamainternmed.2022.6545?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

https://dx.doi.org/10.1007/s11606-009-1179-2

https://dx.doi.org/10.1111/jgs.16212

https://dx.doi.org/10.1186/s12913-015-0950-9

https://dx.doi.org/10.1002/jhm.2614

https://jamanetwork.com/journals/jama/fullarticle/10.1001/jamainternmed.2018.4720?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

https://dx.doi.org/10.1093/gerona/glab222

http://www.jamainternalmedicine.com?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

Gerontol. 2022;41(3):788-797. doi:10.1177/ 07334648211015756

8. Iyer S, Naganathan V, McLachlan AJ, Le Couteur DG. Medication withdrawal trials in people aged 65 years and older: a systematic review. Drugs Aging. 2008;25(12):1021-1031. doi:10.2165/0002512-200825120-00004

9. Page AT, Clifford RM, Potter K, Schwartz D, Etherton-Beer CD. The feasibility and effect of deprescribing in older adults on mortality and health: a systematic review and meta-analysis. Br J Clin Pharmacol. 2016;82(3):583-623. doi:10. 1111/bcp.12975

10. Bayliss EA, Shetterly SM, Drace ML, et al. Deprescribing education vs usual care for patients with cognitive impairment and primary care clinicians: the OPTIMIZE pragmatic cluster randomized trial. JAMA Intern Med. 2022;182(5): 534-542. doi:10.1001/jamainternmed.2022.0502

11. Thillainadesan J, Gnjidic D, Green S, Hilmer SN. Impact of deprescribing interventions in older hospitalised patients on prescribing and clinical outcomes: a systematic review of randomised trials. Drugs Aging. 2018;35(4):303-319. doi:10.1007/ s40266-018-0536-4

12. Scott S, Clark A, Farrow C, et al. Deprescribing admission medication at a UK teaching hospital; a report on quantity and nature of activity. Int J Clin Pharm. 2018;40(5):991-996. doi:10.1007/s11096- 018-0673-1

13. Scott S, Twigg MJ, Clark A, et al. Development of a hospital deprescribing implementation framework: a focus group study with geriatricians and pharmacists. Age Ageing. 2019;49(1):102-110. doi:10.1093/ageing/afz133

14. Campanelli CM; American Geriatrics Society 2012 Beers Criteria Update Expert Panel. American Geriatrics Society updated Beers Criteria for potentially inappropriate medication use in older adults. J Am Geriatr Soc. 2012;60(4):616-631. doi:10.1111/j.1532-5415.2012.03923.x

15. Gallagher P, O’Mahony D. STOPP (Screening Tool of Older Persons’ Potentially Inappropriate Prescriptions): application to acutely ill elderly patients and comparison with Beers’ Criteria. Age Ageing. 2008;37(6):673-679. doi:10.1093/ageing/ afn197

16. Van der Linden L, Decoutere L, Flamaing J, et al. Development and validation of the RASP list (Rationalization of Home Medication by an Adjusted STOPP List in Older Patients): a novel tool in the management of geriatric polypharmacy. Eur Geriatr Med. 2014;5(3):175-180. doi:10.1016/j. eurger.2013.12.005

17. Hilmer SN, Mager DE, Simonsick EM, et al. A drug burden index to define the functional burden of medications in older people. Arch Intern Med. 2007;167(8):781-787. doi:10.1001/archinte. 167.8.781

18. Shah AS, Hollingsworth EK, Shotwell MS, Mixon AS, Simmons SF, Vasilevskis EE. Sources of medication omissions among hospitalized older adults with polypharmacy. J Am Geriatr Soc. 2022; 70(4):1180-1189. doi:10.1111/jgs.17629

19. Petersen AW, Shah AS, Simmons SF, et al. Shed-MEDS: pilot of a patient-centered deprescribing framework reduces medications in

hospitalized older adults being transferred to inpatient postacute care. Ther Adv Drug Saf. 2018;9 (9):523-533. doi:10.1177/2042098618781524

20. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40 (5):373-383. doi:10.1016/0021-9681(87)90171-8

21. Saliba D, Buchanan J, Edelen MO, et al. MDS 3.0: brief interview for mental status. J Am Med Dir Assoc. 2012;13(7):611-617. doi:10.1016/j.jamda.2012. 06.004

22. Vasilevskis EE, Shah AS, Hollingsworth EK, et al; Shed-MEDS Team. A patient-centered deprescribing intervention for hospitalized older patients with polypharmacy: rationale and design of the Shed-MEDS randomized controlled trial. BMC Health Serv Res. 2019;19(1):165. doi:10.1186/ s12913-019-3995-3

23. Kim JL, Lewallen KM, Hollingsworth EK, Shah AS, Simmons SF, Vasilevskis EE. Patient-reported barriers and enablers to deprescribing recommendations during a clinical trial. Gerontologist. 2022;gnac100. doi:10.1093/ geront/gnac100

24. US Food and Drug Administration. General drug categories. November 3, 2018. Accessed April 1, 2022. https://www.fda.gov/drugs/investigational- new-drug-ind-application/general-drug-categories

25. Graves T, Hanlon JT, Schmader KE, et al. Adverse events after discontinuing medications in elderly outpatients. Arch Intern Med. 1997;157 (19):2205-2210. doi:10.1001/archinte.1997. 00440400055007

26. Naranjo CA, Busto U, Sellers EM, et al. A method for estimating the probability of adverse drug reactions. Clin Pharmacol Ther. 1981;30(2): 239-245. doi:10.1038/clpt.1981.154

27. Jazić I, Haneuse S, French B, MacGrogan G, Rondeau V. Design and analysis of nested case-control studies for recurrent events subject to a terminal event. Stat Med. 2019;38(22):4348-4362. doi:10.1002/sim.8302

28. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)–a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42(2):377-381. doi:10.1016/j.jbi.2008.08.010

29. Edey R, Edwards N, Von Sychowski J, Bains A, Spence J, Martinusen D. Impact of deprescribing rounds on discharge prescriptions: an interventional trial. Int J Clin Pharm. 2019;41(1): 159-166. doi:10.1007/s11096-018-0753-2

30. Potter EL, Lew TE, Sooriyakumaran M, Edwards AM, Tong E, Aung AK. Evaluation of pharmacist-led physician-supported inpatient deprescribing model in older patients admitted to an acute general medical unit. Australas J Ageing. 2019;38(3):206-210. doi:10.1111/ajag.12643

31. Poquet I, Tornero C. Deprescription at hospital discharge: outcomes of a deprescription promoting campaign. Eur J Intern Med. 2017;42:e22-e23. doi:10.1016/j.ejim.2017.04.008

32. Marvin V, Ward E, Poots AJ, Heard K, Rajagopalan A, Jubraj B. Deprescribing medicines in the acute setting to reduce the risk of falls. Eur J

Hosp Pharm. 2017;24(1):10-15. doi:10.1136/ ejhpharm-2016-001003

33. Garfinkel D, Mangin D. Feasibility study of a systematic approach for discontinuation of multiple medications in older adults: addressing polypharmacy. Arch Intern Med. 2010;170(18): 1648-1654. doi:10.1001/archinternmed.2010.355

34. McKean M, Pillans P, Scott IA. A medication review and deprescribing method for hospitalised older patients receiving multiple medications. Intern Med J. 2016;46(1):35-42. doi:10.1111/imj.12906

35. McDonald EG, Wu PE, Rashidi B, et al. The MedSafer study: a controlled trial of an electronic decision support tool for deprescribing in acute care. J Am Geriatr Soc. 2019;67(9):1843-1850. doi:10.1111/jgs.16040

36. Johansson T, Abuzahra ME, Keller S, et al. Impact of strategies to reduce polypharmacy on clinically relevant endpoints: a systematic review and meta-analysis. Br J Clin Pharmacol. 2016;82(2): 532-548. doi:10.1111/bcp.12959

37. Blum MR, Sallevelt BTGM, Spinewine A, et al. Optimizing Therapy to Prevent Avoidable Hospital Admissions in Multimorbid Older Adults (OPERAM): cluster randomised controlled trial. BMJ. 2021;374 (1585):n1585. doi:10.1136/bmj.n1585

38. Scales DC, Fischer HD, Li P, et al. Unintentional continuation of medications intended for acute illness after hospital discharge: a population-based cohort study. J Gen Intern Med. 2016;31(2):196-202. doi:10.1007/s11606-015-3501-5

39. Anderson TS, Lee S, Jing B, et al. Prevalence of diabetes medication intensification in older adults dicharged from US Veterans Health Administration hospitals. JAMA Netw Open. 2020;3(3):e201511. doi:10.1001/jamanetworkopen.2020.1511

40. Anderson TS, Jing B, Auerbach A, et al. Clinical outcomes after intensifying antihypertensive medication regimens among older adults at hospital discharge. JAMA Intern Med. 2019;179(11):1528-1536. doi:10.1001/jamainternmed.2019.3007

41. Boockvar KS, Song W, Lee S, Intrator O. Hypertension treatment in US long-term nursing home residents with and without dementia. J Am Geriatr Soc. 2019;67(10):2058-2064. doi:10.1111/ jgs.16081

42. Brunström M, Carlberg B. Association of blood pressure lowering with mortality and cardiovascular disease across blood pressure levels: a systematic review and meta-analysis. JAMA Intern Med. 2018; 178(1):28-36. doi:10.1001/jamainternmed.2017.6015

43. Pajewski NM, Berlowitz DR, Bress AP, et al. Intensive vs standard blood pressure control in adults 80 years or older: a secondary analysis of the systolic blood pressure intervention trial. J Am Geriatr Soc. 2020;68(3):496-504. doi:10.1111/jgs. 16272

44. Weiss J, Freeman M, Low A, et al. Benefits and harms of intensive blood pressure treatment in adults aged 60 years or older: a systematic review and meta-analysis. Ann Intern Med. 2017;166(6): 419-429. doi:10.7326/M16-1754

45. Reeve E, Low LF, Hilmer SN. Beliefs and attitudes of older adults and carers about deprescribing of medications: a qualitative focus group study. Br J Gen Pract. 2016;66(649): e552-e560. doi:10.3399/bjgp16X685669

Deprescribing Medications Among Older Adults Through Postacute Care Original Investigation Research

jamainternalmedicine.com (Reprinted) JAMA Internal Medicine March 2023 Volume 183, Number 3 231

Downloaded from jamanetwork.com by SCELC - Azusa Pacific Univ user on 09/19/2026

https://dx.doi.org/10.1177/07334648211015756

https://dx.doi.org/10.2165/0002512-200825120-00004

https://dx.doi.org/10.1111/bcp.12975

https://jamanetwork.com/journals/jama/fullarticle/10.1001/jamainternmed.2022.0502?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

https://dx.doi.org/10.1007/s40266-018-0536-4

https://dx.doi.org/10.1007/s11096-018-0673-1

https://dx.doi.org/10.1093/ageing/afz133

https://dx.doi.org/10.1111/j.1532-5415.2012.03923.x

https://dx.doi.org/10.1093/ageing/afn197

https://dx.doi.org/10.1016/j.eurger.2013.12.005

https://jamanetwork.com/journals/jama/fullarticle/10.1001/archinte.167.8.781?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

https://dx.doi.org/10.1111/jgs.17629

https://dx.doi.org/10.1177/2042098618781524

https://dx.doi.org/10.1016/0021-9681(87)90171-8

https://dx.doi.org/10.1016/j.jamda.2012.06.004

https://dx.doi.org/10.1186/s12913-019-3995-3

https://dx.doi.org/10.1093/geront/gnac100

https://www.fda.gov/drugs/investigational-new-drug-ind-application/general-drug-categories

https://jamanetwork.com/journals/jama/fullarticle/10.1001/archinte.1997.00440400055007?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

https://dx.doi.org/10.1038/clpt.1981.154

https://dx.doi.org/10.1002/sim.8302

https://dx.doi.org/10.1016/j.jbi.2008.08.010

https://dx.doi.org/10.1007/s11096-018-0753-2

https://dx.doi.org/10.1111/ajag.12643

https://dx.doi.org/10.1016/j.ejim.2017.04.008

https://dx.doi.org/10.1136/ejhpharm-2016-001003

https://jamanetwork.com/journals/jama/fullarticle/10.1001/archinternmed.2010.355?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

https://dx.doi.org/10.1111/imj.12906

https://dx.doi.org/10.1111/jgs.16040

https://dx.doi.org/10.1111/bcp.12959

https://dx.doi.org/10.1136/bmj.n1585

https://dx.doi.org/10.1007/s11606-015-3501-5

https://jamanetwork.com/journals/jama/fullarticle/10.1001/jamanetworkopen.2020.1511?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

https://jamanetwork.com/journals/jama/fullarticle/10.1001/jamainternmed.2019.3007?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

https://dx.doi.org/10.1111/jgs.16081

https://jamanetwork.com/journals/jama/fullarticle/10.1001/jamainternmed.2017.6015?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

https://dx.doi.org/10.1111/jgs.16272

https://dx.doi.org/10.7326/M16-1754

https://dx.doi.org/10.3399/bjgp16X685669

http://www.jamainternalmedicine.com?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamainternmed.2022.6545

image13.emf

Original Investigation | Pharmacy and Clinical Pharmacology

Medication Optimization Protocol Efficacy for Geriatric Inpatients A Randomized Clinical Trial Kenya Ie, MD, MPH, PhD; Masanori Hirose, MD, PhD; Tsubasa Sakai, MD; Iori Motohashi, MD, MPH; Mari Aihara, MD; Takuya Otsuki, MD, PhD; Ayako Tsuboya, PharmD, PhD; Hiroshi Matsumoto, PharmD; Hikari Hashi, PharmD; Eisuke Inoue, PhD; Masaki Takahashi, MSc; Eiko Komiya, PharmD; Yuka Itoh, PharmD; Reiko Machino, PharmD; Tomoya Tsuchida, MD, PhD; Steven M. Albert, PhD; Yoshiyuki Ohira, MD, PhD; Chiaki Okuse, MD, PhD

Abstract

IMPORTANCE There is currently no consensus on clinically effective interventions for polypharmacy among older inpatients.

OBJECTIVE To evaluate the effect of multidisciplinary team-based medication optimization on survival, unscheduled hospital visits, and rehospitalization in older inpatients with polypharmacy.

DESIGN, SETTING, AND PARTICIPANTS This open-label randomized clinical trial was conducted at 8 internal medicine inpatient wards within a community hospital in Japan. Participants included medical inpatients 65 years or older who were receiving 5 or more regular medications. Enrollment took place between May 21, 2019, and March 14, 2022. Statistical analysis was performed from September 2023 to May 2024.

INTERVENTION The participants were randomly assigned to receive either an intervention for medication optimization or usual care including medication reconciliation. The intervention consisted of a medication review using the STOPP (Screening Tool of Older Persons’ Prescriptions)/ START (Screening Tool to Alert to Right Treatment) criteria, followed by a medication optimization proposal for participants and their attending physicians developed by a multidisciplinary team. On discharge, the medication optimization summary was sent to patients’ primary care physicians and community pharmacists.

MAIN OUTCOMES AND MEASURES The primary outcome was a composite of death, unscheduled hospital visits, and rehospitalization within 12 months. Secondary outcomes included the number of prescribed medications, falls, and adverse events.

RESULTS Between May 21, 2019, and March 14, 2022, 442 participants (mean [SD] age, 81.8 [7.1] years; 223 [50.5%] women) were randomly assigned to the intervention (n = 215) and usual care (n = 227). The intervention group had a significantly lower percentage of patients with 1 or more potentially inappropriate medications than the usual care group at discharge (26.2% vs 33.0%; adjusted odds ratio [OR], 0.56 [95% CI, 0.33-0.94]; P = .03), at 6 months (27.7% vs 37.5%; adjusted OR, 0.50 [95% CI, 0.29-0.86]; P = .01), and at 12 months (26.7% vs 37.4%; adjusted OR, 0.45 [95% CI, 0.25-0.80]; P = .007). The primary composite outcome occurred in 106 participants (49.3%) in the intervention group and 117 (51.5%) in the usual care group (stratified hazard ratio, 0.98 [95% CI, 0.75-1.27]). Adverse events were similar between each group (123 [57.2%] in the intervention group and 135 [59.5%] in the usual care group).

CONCLUSIONS AND RELEVANCE In this randomized clinical trial of older inpatients with polypharmacy, the multidisciplinary deprescribing intervention did not reduce death, unscheduled

(continued)

Key Points Question What is the effect of

multidisciplinary team–based

medication optimization on clinical end

points in older inpatients with

polypharmacy?

Findings In this randomized clinical trial

that included 442 patients, the

intervention did not reduce death,

unscheduled hospital visits, or

rehospitalization. The intervention was

safe and effective in reducing the

number of medications and potentially

inappropriate medications.

Meaning These results corroborate the

safety of deprescribing interventions in

older inpatients with polypharmacy,

consistent with previous studies;

however, the inability of this health

care–centered intervention to affect

health outcomes may suggest the need

for patient-centered interventions.

+ Visual Abstract

+ Invited Commentary

+ Supplemental content

Author affiliations and article information are listed at the end of this article.

Open Access. This is an open access article distributed under the terms of the CC-BY License.

JAMA Network Open. 2024;7(7):e2423544. doi:10.1001/jamanetworkopen.2024.23544 (Reprinted) July 30, 2024 1/13

Downloaded from jamanetwork.com by guest on 09/19/2026

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamanetworkopen.2024.23544&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamanetworkopen.2024.23529&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamanetworkopen.2024.23544&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

Abstract (continued)

hospital visits, or rehospitalization within 12 months. The intervention was effective in reducing the number of medications with no significant adverse effects on clinical outcomes, even among older inpatients with polypharmacy.

TRIAL REGISTRATION UMIN Clinical Trials Registry: UMIN000035265

JAMA Network Open. 2024;7(7):e2423544.

Corrected on November 12, 2024. doi:10.1001/jamanetworkopen.2024.23544

Introduction

A substantial portion of older adults are subject to polypharmacy, defined as the concurrent use of multiple medications, a trend that has been on the rise.1,2 Use of multiple medications is associated with a high risk of physical and cognitive disability,3 emergency department visits and hospitalizations,4 and mortality.5 Furthermore, polypharmacy is associated with nearly doubling overall health care costs and tripling pharmacy expenses, as indicated by the 2017 Medical Expenditure Panel Survey data.6

Various interventions have been developed to mitigate the negative impact of polypharmacy in the last few decades. Interventions with a focus on reducing specific medications7,8 or potentially inappropriate medications (PIMs) (listed in explicit criteria such as the Beers criteria,9 which identifies PIMs in older adults, or STOPP [Screening Tool of Older Persons’ Prescriptions]/START [Screening Tool to Alert to Right Treatment] criteria,10 which outline medications to stop and start in older patients), have been a mainstay of such efforts. More recently, patient-centered deprescribing has become the focus of interventions and research in the field.11 It involves an implicit criteria-based approach to identify and discontinue drugs for which harms outweigh benefits. However, recent systematic reviews regarding the effect of interventions to reduce polypharmacy, including both the explicit and implicit criteria-based approaches, are not consistent in demonstrating their benefits on clinically important end points.12-15

Several factors may explain the inconsistent results observed in previous systematic reviews, including insufficient statistical power and heterogeneity in meta-analyses. Furthermore, the effect of interventions using explicit criteria could be potentially compromised by adverse drug events involving medications not listed in these criteria.16 A Cochrane review also highlighted methodological limitations of previous clinical trials, such as short follow-up, which might obscure the true effect of deprescribing protocols.13 Addressing these gaps, our intervention uses both explicit and implicit criteria to promote prescribing optimization, which we assessed over 12 months. This study aims to evaluate the efficacy of multidisciplinary team-based medication optimization on survival, unscheduled hospital visits, and rehospitalization in older inpatients with polypharmacy.

Methods

Study Design We conducted a single-center, open-label, randomized clinical trial with a 2-group parallel design. The study was approved by the St Marianna University School of Medicine institutional ethical committee. This manuscript was written in accordance with the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline17 and the Template for Intervention Description and Replication (TIDieR) checklist.18 Further details of the methods are available in the trial protocol (Supplement 1).19

JAMA Network Open | Pharmacy and Clinical Pharmacology Medication Optimization Protocol Efficacy for Geriatric Inpatients

JAMA Network Open. 2024;7(7):e2423544. doi:10.1001/jamanetworkopen.2024.23544 (Reprinted) July 30, 2024 2/13

Downloaded from jamanetwork.com by guest on 09/19/2026

https://rctportal.niph.go.jp/en/detail?trial_id=UMIN000035265

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamanetworkopen.2024.23544&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

http://www.equator-network.org/reporting-guidelines/consort/

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamanetworkopen.2024.23544&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

Participants Medical inpatients admitted to a community hospital were eligible to participate if they were 65 years or older, receiving 5 or more regular medications, deemed eligible to take oral medication by their attending physician, and whose duration of hospitalization was expected to be 1 week or longer. Patients were excluded if the attending physician disagreed on their participation, or if they had a life expectancy of less than 1 month based on their attending physician’s clinical judgement. A regular medication was defined as any orally administered prescription medication that was documented in the participant’s medical record with a duration of 28 days or longer at the time of admission. Medications categorized as as needed were excluded from the medication counts (eMethods in Supplement 2). We obtained written informed consent from all participants or their next of kin.

Randomization and Masking We stratified eligible participants by age group (65-74 years, 75-84 years, and �85 years) and randomly allocated them within strata to the intervention and usual care groups by block randomization with block sizes of 4. The block size was concealed until the completion of the trial and was known only to the statistical analysts who were not involved in enrollment and data acquisition. Physicians, pharmacists, patients, and/or their next of kin were not masked to the allocation. All primary outcome data and part of secondary outcomes including quality-of-life score, level of long- term care required, adverse events, and falls were collected through the bimonthly telephone interview assessments by designated research assistants who were masked to group allocation. Prescription drug data were extracted either from the participant’s medical record handbook or electronic health record by research assistants using a predefined procedure.

Trial Procedures Given the medical risk of older patients, who are at a higher risk of medication-related adverse events, we deemed it ethically unacceptable to withhold medication reconciliation. Therefore, usual care included medication reconciliation conducted by ward-based pharmacists at admission. For participants in the intervention group, a multidisciplinary deprescribing team, composed of a physician and a pharmacist, promptly initiated the intervention within 48 hours of allocation. All members of the intervention team underwent standardized guidance and training in advance. The structure of the intervention included: (1) baseline data collection, including age, sex, previous medical history, comorbid conditions, smoking status, physical measurements on admission (height, weight and vital signs), estimated glomerular filtration rate (eGFR), serum sodium and potassium level, and regularly prescribed medications, conducted through medical record review; (2) preliminary medication optimization proposal using a clinical-decision support system; baseline data were entered into a computer-based clinical-decision support system developed specifically for this trial by the deprescribing team using a Microsoft Excel spreadsheet. Using these data, the clinical- decision support system generated a preliminary list of potentially inappropriate prescriptions as well as prescribing omissions in line with STOPP/START criteria (version 2)20; and (3) medication optimization protocol-based team discussion; the deprescribing team reviewed the draft proposal and embarked on a step-by-step discussion, adhering to the medication optimization protocol; this discussion followed a specific algorithm, as proposed by Scott et al11,19: (1) assessment of medication indication; (2) balancing benefits and harms; (3) evaluation of symptomatic medications; and (4) evaluation of preventive medications.

Following these steps, the medication optimization plan was explained and discussed with the participant or their next of kin. With the participant’s consent, the team recommended the medication optimization plan, including its rationale, to the attending physician. The decision to accept or decline the proposal was at the discretion of their attending physician. A summary of the medication optimization, including reasons for modifications and relevant precautions, was conveyed to the participant’s primary care physician and community pharmacists upon discharge, ensuring continuity of care. More details are provided in the trial protocol (Supplement 1).19

JAMA Network Open | Pharmacy and Clinical Pharmacology Medication Optimization Protocol Efficacy for Geriatric Inpatients

JAMA Network Open. 2024;7(7):e2423544. doi:10.1001/jamanetworkopen.2024.23544 (Reprinted) July 30, 2024 3/13

Downloaded from jamanetwork.com by guest on 09/19/2026

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamanetworkopen.2024.23544&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

Outcomes The primary outcome was a composite of all-cause death, unscheduled hospital visits, and rehospitalization within 48 weeks of enrollment. Secondary outcomes included each of the primary outcome, number of regular medications and PIMs based on STOPP/START criteria version 2,20 level of long-term care required, health-related quality of life (measured using EuroQol 5 dimensions 3-levels [EQ5D-3L]),21,22 adverse events, falls, and all-cause death during initial hospitalization (eMethods in Supplement 2). Events were tracked in regular follow-up telephone interviews performed by trained research assistants masked to group allocation. Number of regular medications and PIMs at discharge, 24 weeks, and 48 weeks were collected using medication lists included in the participant’s medical record handbook or electronic medical record and consecutively adjudicated with medication lists sent from the relevant community pharmacist. Level of long-term care required and health-related quality of life21,22 were assessed at baseline, hospital discharge, 24 weeks, and 48 weeks based on the follow-up telephone interview according to the prespecified questionnaire. All adverse events, regardless of their potential relevance to the intervention, were recorded during initial hospital admission and follow-up period. Serious adverse events included all-cause death and events that resulted in persistent disability or hospital admission (eMethods in Supplement 2). The relevance to the intervention was assessed by consensus among the deprescribing team and reviewed by the institutional data and safety monitoring board.

Statistical Analysis With 250 participants in each group, an assumption of primary end point rates of 30% and 40% within 48 weeks in the intervention and control groups, respectively, and a true hazard ratio (HR) of 0.75,23,24 statistical power was 80% (2-sided type 1 error level of .05, with a 48-week drop-out rate of 15%). The efficacy end points were analyzed according to modified intention-to-treat principle. All study participants were included in the analysis of primary and secondary end points, except those found to be ineligible after enrollment. Survival functions of the primary end point were analyzed by Kaplan-Meier method and compared using a log-rank test stratified by age group. Two-sided P < .05 was considered statistically significant. Age group-stratified HR and 95% CI across groups was estimated using Cox proportional hazards model. We also conducted prespecified subgroup analyses of the primary end point for indicator diseases (including heart failure, pneumonia, diabetes, ischemic stroke, and urinary tract infection) and indicator drug classes (antiplatelets, antihypertensives, antidiabetics, and sedatives) for exploratory, hypothesis-generating purposes. Furthermore, a per-protocol analysis, using data from participants for whom at least 1 deprescribing was proposed and at least 1 of the proposed drugs was actually reduced at discharge, was conducted to assess the effect of the acceptance rate of the proposal. Patients who died during initial hospitalization were excluded from the per-protocol analysis, as the acceptance of the proposal could not be determined. All statistical analyses were performed using R version 4.2.2 software (R Project for Statistical Computing) from September 2023 to May 2024.

Results

Between May 21, 2019, and March 14, 2022, a total of 465 participants were enrolled and 460 randomized to the intervention or usual care. During follow-up, 18 patients withdrew from the study and 442 patients were included in the analysis (mean [SD] age, 81.8 [7.1] years; 223 [50.5%] women; 215 in the intervention group and 227 in the usual care group). Five patients were lost to follow-up, and 103 (23.3%) died within 12 months of follow-up (Figure 1). Baseline characteristics were similar in both groups, with the exception of a higher prevalence of smoking among participants in the usual care group (Table 1). At baseline, the participants had a mean (SD) of 4.2 (1.7) diagnoses, used a median (IQR) of 8 (7-10) regular medications, 187 (42.3%) had been prescribed 1 or more PIMs, and 109 (24.7%) had fallen at least once during the previous 3 months prior to study participation. Total

JAMA Network Open | Pharmacy and Clinical Pharmacology Medication Optimization Protocol Efficacy for Geriatric Inpatients

JAMA Network Open. 2024;7(7):e2423544. doi:10.1001/jamanetworkopen.2024.23544 (Reprinted) July 30, 2024 4/13

Downloaded from jamanetwork.com by guest on 09/19/2026

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamanetworkopen.2024.23544&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

medication counts, PIM use, and the distribution of drug class at baseline were similar between the groups (eTables 1 and 2 in Supplement 2).

The mean (SD) preparation time of prescribing optimization proposals took 16.0 (7.2) minutes for both a clinician and a clinical pharmacist. The intervention team proposed a total of 737 deprescribing proposals directed to attending physicians and participants. Of these proposals, 254 (34.5%) were Anatomical Therapeutic Chemical (ATC) code A (alimentary tract and metabolism), 147 (19.9%) were ATC code N (nervous system), and 142 (19.3%) were ATC code C (cardiovascular system). Among 188 participants in the intervention group who were subjected to 1 or more recommendations for deprescribing, 153 (81.4%) accepted at least 1 proposal. The overall acceptance rate of the proposal was 68.7% (eTable 3 in Supplement 2).

The primary outcome (composite of death, unscheduled hospital visits, and rehospitalization) occurred in 106 participants (49.3%) in the intervention group and 117 participants (51.5%) in the usual care group. In an intention-to-treat analysis, there were no significant differences in the primary outcome between groups (stratified HR, 0.98 [95% CI, 0.75-1.27]; P = .85) (Figure 2A). In per-protocol analysis, which was restricted to 153 participants in the intervention group who received 1 or more deprescribing recommendations and accepted at least 1 proposal, the results remained nonsignificant for the primary end point (stratified HR, 1.05 [95% CI, 0.78-1.41]; P = .70) (eTable 4 in Supplement 2).

Upon discharge, the intervention group was prescribed a mean of 0.89 (95% CI, 0.38-1.40) (P < .001) fewer medications than the usual care group, adjusted for baseline medication count and age-group strata. This difference in the total medication counts remained significant at 6 and 12 months, with adjusted mean differences of 0.92 (95% CI, 0.37-1.46) (P = .001) and 0.62 (95% CI, 0.03-1.20) (P = .04), respectively (Figure 3). We observed a significant decrease in the percentage

Figure 1. Flow of Participants Through the Study

14 Attending physicians disagree on participation

192 No response to invitation

837 Patients assessed for eligibility

229 Randomized to intervention group

460 Randomized

377 Excluded

139 Declined to participate 19 Unable to take medicines orally

8 Life expectancy <1 mo 5 Discharge before randomization

8 Declined further participation

221 Received allocated intervention

44 Discontinued study

177 Completed 6-mo follow-up

34 Died 4 Lost to follow-up 6 Withdrew

15 Discontinued study due to death

162 Completed 12-mo follow-up

215 Included in the analysis

231 Randomized to usual care group and received care

194 Completed 6-mo follow-up

172 Completed 12-mo follow-up

227 Included in the analysis

37 Discontinued study 33 Died 1 Lost to follow-up 3 Withdrew

22 Discontinued study 21 Died 1 Withdrew

JAMA Network Open | Pharmacy and Clinical Pharmacology Medication Optimization Protocol Efficacy for Geriatric Inpatients

JAMA Network Open. 2024;7(7):e2423544. doi:10.1001/jamanetworkopen.2024.23544 (Reprinted) July 30, 2024 5/13

Downloaded from jamanetwork.com by guest on 09/19/2026

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamanetworkopen.2024.23544&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

of patients with 1 or more PIMs in the intervention group compared with the usual care group at discharge (26.2% vs 33.0%; adjusted odds ratio [OR], 0.56 [95% CI, 0.33-0.94]; P = .03), at 6 months (27.7% vs 37.5%; adjusted OR, 0.50 [95% CI, 0.29-0.86]; P = .01), and at 12 months (26.7% vs 37.4%; adjusted OR, 0.45 [95% CI, 0.25-0.80]; P = .007). (eTable 5 in Supplement 2).

Other secondary efficacy outcomes, including all-cause mortality (stratified HR, 1.00 [95% CI, 0.68-1.47]; P = .98), unscheduled hospital visits (stratified HR, 0.96 [95% CI, 0.70-1.32]; P = .80), rehospitalizations (stratified HR, 0.92 [95% CI, 0.64-1.33]; P = .66), level of long-term care required, health-related quality of life, and death during the initial hospitalization were similar in both groups (Figure 2B, Figure 2C, Figure 2D; Table 2). Compared with the control group, the incidence of fall and fall-related injuries in the intervention group were not statistically significantly lower (between- group difference: stratified HR, 0.86 [95% CI, 0.56-1.32]; P = .48) (eFigure 1 in Supplement 2).

Adverse events occurred in 123 participants (57.2%) in the intervention group and 135 participants (59.5%) in the control group (Table 2). Overall, adverse events were similar between groups, and there were no safety events related to the study intervention (eTable 6 in Supplement 2).

The intervention effect on primary end point did not differ in prespecified subgroup analyses both for indicator diseases (heart failure, pneumonia, diabetes, ischemic stroke, and urinary tract infection) and indicator drug classes at baseline (antiplatelets, antihypertensives, antidiabetics, and sedatives) (eFigure 2 in Supplement 2). A post hoc subgroup analysis restricted to participants in both groups with 1 or more PIMs did not show a significant difference for the primary end point (eTable 7 in Supplement 2).

Table 1. Baseline Characteristics of Study Participants

Characteristic

No. (%)

Intervention group (n = 215) Usual care group (n = 227) Age, mean (SD), y 81.7 (6.9) 81.9 (7.4)

Sex

Female 112 (52.1) 111 (48.9)

Male 103 (47.9) 116 (51.1)

BMI, mean (SD) 22.6 (4.2) 22.3 (4.7)

Smoking status

Current or former 62 (28.8) 96 (42.3)

Never 153 (71.2) 131 (57.7)

Care levela

No long-term care insurance application 99 (46.0) 105 (46.3)

Need support level 30 (14.0) 25 (11.0)

Need care level 1-2 48 (22.3) 54 (23.8)

Need care level 3-5 36 (16.7) 41 (18.1)

Missing 2 (0.9) 2 (0.9)

EQ5D-3L score, median (IQR) 0.7 (0.5-0.8) 0.7 (0.5-0.8)

Any fall during previous 3 mo 53 (24.7) 52 (23.0)

No. of diagnoses, mean (SD) 4.0 (1.7) 4.3 (1.8)

Index diseases

Heart failure 14 (6.5) 13 (5.7)

Pneumonia 37 (17.2) 54 (23.8)

Diabetes 16 (7.4) 17 (7.5)

Ischemic stroke 11 (5.1) 7 (3.1)

Urinary tract infection 16 (7.4) 21 (9.3)

eGFR, mean (SD), mL/min/1.73 m2 49.5 (33.7) 50 (32.1)

No. of drugs, median (IQR)b 8 (7-10) 8 (7-10)

No. of PIMs, median (IQR)c 0 (0-1) 0 (0-1)

Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); EQ5D-3L, EuroQol 5 dimensions 3-levels; PIM, potentially inappropriate medication. a Care level represents the degree of caregiving

needed by older adults, ranging from those who have not applied for long-term care insurance to those requiring the highest level of care in care level 5, as defined by the Japanese long-term care system.

b Medications categorized as as needed were excluded from the medication counts.

c PIM was determined based on STOPP/START criteria version 2.20

JAMA Network Open | Pharmacy and Clinical Pharmacology Medication Optimization Protocol Efficacy for Geriatric Inpatients

JAMA Network Open. 2024;7(7):e2423544. doi:10.1001/jamanetworkopen.2024.23544 (Reprinted) July 30, 2024 6/13

Downloaded from jamanetwork.com by guest on 09/19/2026

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamanetworkopen.2024.23544&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

Discussion

This randomized clinical trial of older inpatients with polypharmacy investigated the efficacy of multidisciplinary team-based intervention integrating explicit and implicit criteria to promote

Figure 2. Survival Plot of Time to Event for Participants in the Intervention Group and Usual Care Group

1.0

0.8

0.7

0.6

0.9

0.4

No. at risk

215 227

133 148

66 70

Pr ob

ab ili

ty o

f e ve

nt −f

re e

su rv

iv al

Time, mo

0.5

Intervention group Usual care group

Primary outcomeA

0.4

No. at risk

215 227

177 195

103 114

1.0

0.8

Pr ob

ab ili

ty o

f e ve

nt −f

re e

su rv

iv al

Time, mo

0.7

0.6

0.5

Intervention group Usual care group

All-cause deathB

0.9

0.4

No. at risk

215 227

137 157

70 72

1.0

0.8

Pr ob

ab ili

ty o

f e ve

nt −f

re e

su rv

iv al

Time, mo

0.7

0.6

0.5

Intervention group Usual care group

Unscheduled hospital visitsC

0.9

0.4

No. at risk

215 227

152 160

80 88

1.0

0.8 Pr

ob ab

ili ty

o f e

ve nt

−f re

e su

rv iv

Time, mo

0.7

0.6

0.5

Intervention group Usual care group

RehospitalizationD

0.9

Usual care group Intervention group

A, Primary composite outcome of all-cause death, unscheduled hospital visits, and rehospitalization. B, All-cause death. C, Unscheduled hospital visits; D, Rehospitalizations. Tick marks indicate censoring events.

Figure 3. Total Medication Counts and Potentially Inappropriate Medications Over Time

0.8

0.7

0.4

0.3

0.6

0.5

To ta

l m ed

ic at

io n

co un

ts , m

ea n

(9 5%

C I)

Total medication counts over timeA

Admission Discharge 12 mo 0.2

Po te

nt ia

lly in

ap pr

op ria

te m

ed ic

at io

n co

un ts

, m ea

n (9

5% C

Potentially inappropriate medication counts over timeB

Admission Discharge 12 mo6 mo 6 mo

Usual care group Intervention group

Each point represents the mean, and each error bar represents the 95% CI. Medications categorized as as needed were excluded from the medication counts.

JAMA Network Open | Pharmacy and Clinical Pharmacology Medication Optimization Protocol Efficacy for Geriatric Inpatients

JAMA Network Open. 2024;7(7):e2423544. doi:10.1001/jamanetworkopen.2024.23544 (Reprinted) July 30, 2024 7/13

Downloaded from jamanetwork.com by guest on 09/19/2026

prescribing optimization. Older inpatients receiving the intervention or usual care did not differ in all-cause death, unscheduled hospital visits, or rehospitalization. However, total medication counts were reduced in the intervention group compared with usual care, and the proportion of patients prescribed PIM was significantly lower in the intervention group than in usual care. Furthermore, no excess adverse events due to the intervention were observed.

This study demonstrated the safety of deprescribing in hospitalized patients with a high mortality rate in which more than 20% died during the 1-year follow-up. Mortality rates during the study period among previous studies ranged from 2.1% to 11.3% in the primary care setting,25-29 and from 5.6% at 90 days30 to 19.2% at 1 year31 in the inpatient setting, suggesting that our study population had a higher morbidity and mortality than previously reported studies. Our results corroborate the safety of deprescribing interventions in older inpatients with polypharmacy. The effect of our intervention on drug reduction is comparable to or exceeds that of previous studies. Although there is no consensus minimal clinically important difference for polypharmacy interventions, a meta-analysis of 25 deprescribing intervention trials reported a mean drug reduction of 0.4 drugs due to intervention.14 The Medication Optimization Protocol Efficacy for Geriatric Inpatients (MPEG) trial results support a potentially positive impact of deprescribing on health care costs and treatment burden; however, evidence is still limited regarding the cost-effectiveness of deprescribing in a hospital and discharge setting.32 Furthermore, reductions in medication counts and PIMs were maintained through 12 months, suggesting that the integration of a community-based intervention including primary care physicians and community pharmacists may have contributed to maintaining the effect of the intervention.

On the other hand, even with a design intended to overcome the limitations of existing studies, such as the 12-month long-term follow-up,13 our study failed to demonstrate clinical efficacy. The Kaplan-Meier survival curve of the primary outcome revealed no trend suggesting a potential difference between groups toward the latter half of the 12-month period. This is in line with previous deprescribing studies with no demonstrable effects on clinical end points.28,33-35 There are several possible explanations why the intervention did not improve the primary end point. The high mortality of the study participants, which is a competing risk for unscheduled hospital visits and rehospitalization, may have led to an underestimation of the effect of the intervention. Furthermore, in older adults with high morbidity and mortality, the effect of polypharmacy on clinical outcomes may be relatively small compared with other competing risks, making it difficult to observe explicit benefits even when deprescribing is effective. It is also possible that clinically meaningful outcome

Table 2. Patient-Reported Outcomes by Follow-Up and Death During Initial Hospitalization

Intervention group Usual care group

Adjusted difference (95% CI) P valueTotal No.a No. (%) Total No.a No. (%) Need care level 3-5b

Baseline 215 36 (16.7) 227 41 (18.1) NA NA

6 mo 177 38 (21.5) 194 55 (28.4) 0.74 (0.40 to 1.35)c .33

12 mo 162 38 (23.5) 172 50 (29.1) 0.79 (0.43 to 1.41)c .42

Mean EQ5D-3L score (SD)

Baseline 215 0.6 (0.3) 227 0.6 (0.3) NA NA

6 mo 177 0.8 (0.2) 194 0.8 (0.3) 0.03 (−0.01 to 0.08)d .17

12 mo 162 0.8 (0.2) 172 0.8 (0.2) 0.02 (−0.03 to 0.06)d .41

Adverse events 215 123 (57.2) 227 135 (59.5) 0.91 (0.63 to 1.34)e .32

Serious adverse events 215 83 (38.6) 227 91 (40.1) 0.95 (0.65 to 1.39)e .39

Death during initial hospitalization 215 9 (4.2) 227 9 (4.0) 1.07 (0.42 to 2.75)e .45

Abbreviations: EQ5D-3L, EuroQol 5 dimensions 3-levels; NA, not applicable. a Number of participants varied as each was based on available data. b Care level represents the degree of caregiving needed by older adults, ranging from

those who have not applied for long-term care insurance to those requiring the highest level of care in care level 5, as defined by the Japanese long-term care system.

c Odds ratio from multivariable logistic regression model adjusted for baseline care level and age group.

d Risk difference from multiple regression model adjusted for baseline EQ5D-3L score and age group.

e Odds ratio from multivariable logistic regression model adjusted for age group.

JAMA Network Open | Pharmacy and Clinical Pharmacology Medication Optimization Protocol Efficacy for Geriatric Inpatients

JAMA Network Open. 2024;7(7):e2423544. doi:10.1001/jamanetworkopen.2024.23544 (Reprinted) July 30, 2024 8/13

Downloaded from jamanetwork.com by guest on 09/19/2026

improvements were not achieved because many of the medications that were actually reduced were relatively low-risk medications (Supplement 2). From the perspective of treatment burden, the MPEG trial intervention identified medications that were potentially unnecessary and showed that reducing these low-risk medications did not increase harm. Additionally, it is possible that the single- center, open-label design may have led to contamination in delivery of the intervention, which could have attenuated the effect of intervention. However, previous cluster randomized clinical trials of deprescribing also failed to demonstrate clinical effectiveness, suggesting that the challenges in establishing the clinical effectiveness of deprescribing extend beyond study design limitations.29,31

Deprescribing interventions focusing on specific drug classes, such as antihypertensive or sedative medications, may not be more effective than interventions designed to reduce the total number of medications. Subgroup analyses by baseline prescription drug class (eFigure 2 in Supplement 2) in MPEG did not suggest differences in effects by indicator drug classes. This interpretation is supported by evidence suggesting that the number of medications taken is the best predictor of adverse drug events.36 While focusing on the number of medications as a surrogate marker is considered the cornerstone of polypharmacy interventions, future efforts should shift from simply reducing the number of drugs to focusing on patient-centered interventions to change patient attitudes and health-related behavior. Recent review articles point to the potentially greater clinical effectiveness of patient-centered interventions for deprescribing over other interventions.12,37 For instance, an interventional study in hospital wards, adopting multiple motivational interviews in addition to medication review, has shown a reduction in unscheduled hospital visits and rehospitalization.24 This effect could be attributed to active involvement of patients in polypharmacy interventions, which allow patients to be more proactive38 and promote subsequent behavioral change. Future research will require development and subsequent validation of replicable patient- centered, comprehensive medication review interventions across settings.

Limitations This study has limitations. First, the lack of blinding of participants and the deprescribing team could have potentially introduced bias into the study results. To address observer bias, research assistants were masked to treatment allocation, ensuring a more objective outcome assessment even in situations where participants needed to be fully informed about their treatment.

The single-center, open-label study design with randomization at the patient level could have influenced attending physicians’ prescribing behavior in the usual care setting. As a result, their prescription practices might not accurately reflect scenarios in clinical practice, potentially making it difficult to assess the intervention’s effectiveness and generalizability. However, to mitigate the aforementioned influence, we conducted the trial across 8 internal medicine inpatient wards, aiming to ensure a diverse range of attending physicians involved in the study.

Our study was limited to participants with an estimated duration of hospitalization of 1 week or longer. This specific eligibility criteria, important for assuring patient safety, should warrant caution when applying the results to older inpatients in general. In addition, the exclusion of as-needed medications might exclude drugs at high risk of adverse events, thus attenuating the generalizability of the current study.

The impact of the relatively high number of participants with no PIMs and the use of medication reconciliation in the control group may have biased toward null results. However, our eligibility criteria utilizing simple medication count, rather than the appropriateness of medications, may reflect situations that clinicians and pharmacists face in day-to-day clinical practice.

Additionally, our medication optimization protocol-based team discussions, while aimed at patient-centeredness, remained predominantly clinician-driven. Since our intervention utilized a similar multifaceted intervention as a previous study,24 we hypothesized that our intervention may have a similar effect on clinical end points. Nevertheless, it is plausible that our intervention may have lacked an important element for clinical effectiveness, such as motivational interviewing. Recent developments in personalized medicine, including research based on pharmacogenomics, have

JAMA Network Open | Pharmacy and Clinical Pharmacology Medication Optimization Protocol Efficacy for Geriatric Inpatients

JAMA Network Open. 2024;7(7):e2423544. doi:10.1001/jamanetworkopen.2024.23544 (Reprinted) July 30, 2024 9/13

Downloaded from jamanetwork.com by guest on 09/19/2026

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamanetworkopen.2024.23544&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

heightened interest in individualized health care. In the context of polypharmacy management, there is a growing emphasis on comprehensively understanding patients from diverse biological, psychological, and social perspectives, fostering initiatives that encourage patients to take a more active role in polypharmacy management.

Conclusions

Our multidisciplinary team-based deprescribing intervention had no demonstrable effect on the composite end point of death, unscheduled hospital visits, or rehospitalization. However, the intervention was effective in reducing the number of medications and PIMs, with no increase in adverse events even in older inpatients with polypharmacy.

ARTICLE INFORMATION Accepted for Publication: May 14, 2024.

Published: July 30, 2024. doi:10.1001/jamanetworkopen.2024.23544

Correction: This article was corrected on November 12, 2024, to fix an error in Figure 1.

Corresponding Author: Kenya Ie, MD, MPH, PhD, Department of General Internal Medicine, Kawasaki Municipal Tama Hospital, 1-30-37 Shukugawara, Kawasaki, Kanagawa 214-8525, Japan ([email protected]).

Author Affiliations: Department of General Internal Medicine, St Marianna University School of Medicine, Kanagawa, Japan (Ie, Hirose, Sakai, Motohashi, Aihara, Otsuki, Tsuchida, Ohira, Okuse); Department of General Internal Medicine, Kawasaki Municipal Tama Hospital, Kanagawa, Japan (Ie, Sakai, Motohashi, Aihara, Otsuki, Machino, Okuse); Department of Pharmacy, Kawasaki Municipal Tama Hospital, Kanagawa, Japan (Tsuboya, Matsumoto, Hashi, Komiya, Itoh); Showa University Research Administration Center, Showa University, Tokyo, Japan (Inoue); Division of Medical Informatics, St Marianna University School of Medicine, Kanagawa, Japan (Takahashi); Department of Behavioral and Community Health Sciences, University of Pittsburgh Graduate School of Public Health, Pittsburgh, Pennsylvania (Albert).

Author Contributions: Drs Ie and Takahashi had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Ie, Hirose, Tsuboya, Matsumoto, Tsuchida, Albert.

Acquisition, analysis, or interpretation of data: Ie, Hirose, Sakai, Motohashi, Aihara, Otsuki, Tsuboya, Hashi, Inoue, Takahashi, Komiya, Itoh, Machino, Tsuchida, Albert, Ohira, Okuse.

Drafting of the manuscript: Ie, Inoue, Takahashi, Komiya, Tsuchida, Albert.

Critical review of the manuscript for important intellectual content: Ie, Hirose, Sakai, Motohashi, Aihara, Otsuki, Tsuboya, Matsumoto, Hashi, Inoue, Takahashi, Itoh, Machino, Tsuchida, Albert, Ohira, Okuse.

Statistical analysis: Inoue, Takahashi, Albert.

Obtained funding: Ie.

Administrative, technical, or material support: Ie, Hirose, Sakai, Motohashi, Aihara, Tsuboya, Matsumoto, Hashi, Komiya, Itoh, Machino, Tsuchida.

Supervision: Ohira, Okuse.

Conflict of Interest Disclosures: Dr Ie reported grants from Japanese Ministry of Education, Science, Sports and Culture during the conduct of the study. No other disclosures were reported.

Funding/Support: This project was funded by the Ministry of Education, Science, Sports and Culture, Grant-in-Aid for Young Scientists, 2018-2021 (grant number 18K15434).

Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data Sharing Statement: See Supplement 3.

JAMA Network Open | Pharmacy and Clinical Pharmacology Medication Optimization Protocol Efficacy for Geriatric Inpatients

JAMA Network Open. 2024;7(7):e2423544. doi:10.1001/jamanetworkopen.2024.23544 (Reprinted) July 30, 2024 10/13

Downloaded from jamanetwork.com by guest on 09/19/2026

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamanetworkopen.2024.23544&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

https://jamanetwork.com/pages/cc-by-license-permissions/?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

mailto:[email protected]

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamanetworkopen.2024.23544&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

Additional Contributions: We thank the entire participating research team: Ruriko Kashikuma, RN, BA (formerly with Kawasaki Municipal Tama Hospital), and Rei Miyazawa, BA (Kawasaki Municipal Tama Hospital), for participant recruitment and enrollment; Shota Asamizu, RN (Kawasaki Municipal Tama Hospital), Takahide Sadakata, RN (Kawasaki Municipal Tama Hospital), and Chie Sekigawa, RN (formerly with Kawasaki Municipal Tama Hospital) for supporting study conduct; Junko Mita (formerly with Kawasaki Municipal Tama Hospital), Miyuki Kondo, MA (Kawasaki Municipal Tama Hospital), Satomi Tsuchiya, BA (formerly with Kawasaki Municipal Tama Hospital), and Maki Matsui, BA (formerly with Kawasaki Municipal Tama Hospital) for their assistance with data collection and data entry; Yumiko Yamaguchi, MEng (St Marianna University School of Medicine Clinical Research Data Center) and St Marianna University School of Medicine Clinical Research Data Center for their thorough work related to data management. Junko Mita was directly employed with this study’s research funding. No other contributors were compensated.

REFERENCES 1. Midão L, Giardini A, Menditto E, Kardas P, Costa E. Polypharmacy prevalence among older adults based on the survey of health, ageing and retirement in Europe. Arch Gerontol Geriatr. 2018;78:213-220. doi:10.1016/j.archger. 2018.06.018

2. Young EH, Pan S, Yap AG, Reveles KR, Bhakta K. Polypharmacy prevalence in older adults seen in United States physician offices from 2009 to 2016. PLoS One. 2021;16(8):e0255642. doi:10.1371/journal.pone.0255642

3. Rawle MJ, Cooper R, Kuh D, Richards M. Associations between polypharmacy and cognitive and physical capability: a British birth cohort study. J Am Geriatr Soc. 2018;66(5):916-923. doi:10.1111/jgs.15317

4. Liew TM, Lee CS, Goh Shawn KL, Chang ZY. Potentially inappropriate prescribing among older persons: a meta- analysis of observational studies. Ann Fam Med. 2019;17(3):257-266. doi:10.1370/afm.2373

5. Leelakanok N, Holcombe AL, Lund BC, Gu X, Schweizer ML. Association between polypharmacy and death: a systematic review and meta-analysis. J Am Pharm Assoc (2003). 2017;57(6):729-738.e10. doi:10.1016/j.japh. 2017.06.002

6. Kwak MJ, Chang M, Chiadika S, et al. Healthcare expenditure associated with polypharmacy in older adults with cardiovascular diseases. Am J Cardiol. 2022;169:156-158. doi:10.1016/j.amjcard.2022.01.012

7. Tannenbaum C, Martin P, Tamblyn R, Benedetti A, Ahmed S. Reduction of inappropriate benzodiazepine prescriptions among older adults through direct patient education: the EMPOWER cluster randomized trial. JAMA Intern Med. 2014;174(6):890-898. doi:10.1001/jamainternmed.2014.949

8. Fraser LA, Vogt KN, Adachi JD, Thabane L. Fracture risk associated with continuation versus discontinuation of bisphosphonates after 5 years of therapy in patients with primary osteoporosis: a systematic review and meta- analysis. Ther Clin Risk Manag. 2011;7:157-166. doi:10.2147/TCRM.S19385

9. By the 2023 American Geriatrics Society Beers Criteria® Update Expert Panel. American Geriatrics Society 2023 updated AGS Beers Criteria® for potentially inappropriate medication use in older adults. J Am Geriatr Soc. 2023; 71(7):2052-2081. doi:10.1111/jgs.18372

10. O’Mahony D, Cherubini A, Guiteras AR, et al. STOPP/START criteria for potentially inappropriate prescribing in older people: version 3. Eur Geriatr Med. 2023;14(4):625-632. doi:10.1007/s41999-023-00777-y

11. Scott IA, Hilmer SN, Reeve E, et al. Reducing inappropriate polypharmacy: the process of deprescribing. JAMA Intern Med. 2015;175(5):827-834. doi:10.1001/jamainternmed.2015.0324

12. Bloomfield HE, Greer N, Linsky AM, et al. Deprescribing for community-dwelling older adults: a systematic review and meta-analysis. J Gen Intern Med. 2020;35(11):3323-3332. doi:10.1007/s11606-020-06089-2

13. Rankin A, Cadogan CA, Patterson SM, et al. Interventions to improve the appropriate use of polypharmacy for older people. Cochrane Database Syst Rev. 2018;9(9):CD008165. doi:10.1002/14651858.CD008165.pub4

14. Johansson T, Abuzahra ME, Keller S, et al. Impact of strategies to reduce polypharmacy on clinically relevant endpoints: a systematic review and meta-analysis. Br J Clin Pharmacol. 2016;82(2):532-548. doi:10.1111/bcp.12959

15. Huiskes VJB, Burger DM, van den Ende CHM, van den Bemt BJF. Effectiveness of medication review: a systematic review and meta-analysis of randomized controlled trials. BMC Fam Pract. 2017;18(1):5. doi:10.1186/ s12875-016-0577-x

16. Verdoorn S, Kwint HF, Faber A, Gussekloo J, Bouvy ML. Majority of drug-related problems identified during medication review are not associated with STOPP/START criteria. Eur J Clin Pharmacol. 2015;71(10):1255-1262. doi: 10.1007/s00228-015-1908-x

17. Schulz KF, Altman DG, Moher D; CONSORT Group. CONSORT 2010 statement: updated guidelines for reporting parallel group randomised trials. BMJ. 2010;340:c332. doi:10.1136/bmj.c332

18. Hoffmann TC, Glasziou PP, Boutron I, et al. Better reporting of interventions: template for intervention description and replication (TIDieR) checklist and guide. BMJ. 2014;348:g1687. doi:10.1136/bmj.g1687

JAMA Network Open | Pharmacy and Clinical Pharmacology Medication Optimization Protocol Efficacy for Geriatric Inpatients

JAMA Network Open. 2024;7(7):e2423544. doi:10.1001/jamanetworkopen.2024.23544 (Reprinted) July 30, 2024 11/13

Downloaded from jamanetwork.com by guest on 09/19/2026

https://dx.doi.org/10.1016/j.archger.2018.06.018

https://dx.doi.org/10.1371/journal.pone.0255642

https://dx.doi.org/10.1111/jgs.15317

https://dx.doi.org/10.1370/afm.2373

https://dx.doi.org/10.1016/j.japh.2017.06.002

https://dx.doi.org/10.1016/j.amjcard.2022.01.012

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamainternmed.2014.949&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

https://dx.doi.org/10.2147/TCRM.S19385

https://dx.doi.org/10.1111/jgs.18372

https://dx.doi.org/10.1007/s41999-023-00777-y

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamainternmed.2015.0324&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

https://dx.doi.org/10.1007/s11606-020-06089-2

https://dx.doi.org/10.1002/14651858.CD008165.pub4

https://dx.doi.org/10.1111/bcp.12959

https://dx.doi.org/10.1186/s12875-016-0577-x

https://dx.doi.org/10.1007/s00228-015-1908-x

https://dx.doi.org/10.1136/bmj.c332

https://dx.doi.org/10.1136/bmj.g1687

19. Ie K, Hirose M, Sakai T, et al. Protocol of a randomised controlled trial on the efficacy of medication optimisation in elderly inpatients: medication optimisation protocol efficacy for geriatric inpatients (MPEG) trial. BMJ Open. 2020;10(10):e041125. doi:10.1136/bmjopen-2020-041125

20. O’Mahony D, O’Sullivan D, Byrne S, O’Connor MN, Ryan C, Gallagher P. STOPP/START criteria for potentially inappropriate prescribing in older people: version 2. Age Ageing. 2015;44(2):213-218. doi:10.1093/ageing/afu145

21. EuroQol Group. EuroQol–a new facility for the measurement of health-related quality of life. Health Policy. 1990;16(3):199-208. doi:10.1016/0168-8510(90)90421-9

22. Nishimura S, Tsuchiya A, Hisashige A, Ikegami N, Ikeda S. The development of the Japanese EuroQol instrument. Iryo To Shakai. 1998;8:109-123. doi:10.4091/iken1991.8.1_109

23. Potter K, Flicker L, Page A, Etherton-Beer C. Deprescribing in frail older people: a randomised controlled trial. PLoS One. 2016;11(3):e0149984. doi:10.1371/journal.pone.0149984

24. Ravn-Nielsen LV, Duckert ML, Lund ML, et al. Effect of an in-hospital multifaceted clinical pharmacist intervention on the risk of readmission: a randomized clinical trial. JAMA Intern Med. 2018;178(3):375-382. doi:10. 1001/jamainternmed.2017.8274

25. Mortsiefer A, Löscher S, Pashutina Y, et al. Family conferences to facilitate deprescribing in older outpatients with frailty and with polypharmacy: the COFRAIL cluster randomized trial. JAMA Netw Open. 2023;6(3):e234723. doi:10.1001/jamanetworkopen.2023.4723

26. Anderson K, Freeman C, Foster M, Scott I. GP-led deprescribing in community-living older Australians: an exploratory controlled trial. J Am Geriatr Soc. 2020;68(2):403-410. doi:10.1111/jgs.16273

27. Zechmann S, Senn O, Valeri F, et al. Effect of a patient-centred deprescribing procedure in older multimorbid patients in Swiss primary care - a cluster-randomised clinical trial. BMC Geriatr. 2020;20(1):471. doi:10.1186/ s12877-020-01870-8

28. Herrinton LJ, Lo K, Alavi M, et al. Effectiveness of bundled hyperpolypharmacy deprescribing compared with usual care among older adults: a randomized clinical trial. JAMA Netw Open. 2023;6(7):e2322505. doi:10.1001/ jamanetworkopen.2023.22505

29. Rieckert A, Reeves D, Altiner A, et al. Use of an electronic decision support tool to reduce polypharmacy in elderly people with chronic diseases: cluster randomised controlled trial. BMJ. 2020;369:m1822. doi:10.1136/ bmj.m1822

30. Vasilevskis EE, Shah AS, Hollingsworth EK, et al. Deprescribing medications among older adults from end of hospitalization through postacute care: a Shed-MEDS randomized clinical trial. JAMA Intern Med. 2023;183(3): 223-231. doi:10.1001/jamainternmed.2022.6545

31. Blum MR, Sallevelt BTGM, Spinewine A, et al. Optimizing Therapy to Prevent Avoidable Hospital Admissions in Multimorbid Older Adults (OPERAM): cluster randomised controlled trial. BMJ. 2021;374(1585):n1585. doi:10. 1136/bmj.n1585

32. Curtin D, Jennings E, Daunt R, et al. Deprescribing in older people approaching end of life: a randomized controlled trial using STOPPFrail criteria. J Am Geriatr Soc. 2020;68(4):762-769. doi:10.1111/jgs.16278

33. Zhou D, Chen Z, Tian F. Deprescribing interventions for older patients: a systematic review and meta-analysis. J Am Med Dir Assoc. 2023;24(11):1718-1725. doi:10.1016/j.jamda.2023.07.016

34. McDonald EG, Wu PE, Rashidi B, et al. The MedSafer study-electronic decision support for deprescribing in hospitalized older adults: a cluster randomized clinical trial. JAMA Intern Med. 2022;182(3):265-273. doi:10.1001/ jamainternmed.2021.7429

35. Bayliss EA, Shetterly SM, Drace ML, et al. Deprescribing education vs usual care for patients with cognitive impairment and primary care clinicians: the OPTIMIZE pragmatic cluster randomized trial. JAMA Intern Med. 2022; 182(5):534-542. doi:10.1001/jamainternmed.2022.0502

36. Scott IA, Anderson K, Freeman CR, Stowasser DA. First do no harm: a real need to deprescribe in older patients. Med J Aust. 2014;201(7):390-392. doi:10.5694/mja14.00146

37. Ie K, Aoshima S, Yabuki T, Albert SM. A narrative review of evidence to guide deprescribing among older adults. J Gen Fam Med. 2021;22(4):182-196. doi:10.1002/jgf2.464

38. Michie S, Miles J, Weinman J. Patient-centredness in chronic illness: what is it and does it matter? Patient Educ Couns. 2003;51(3):197-206. doi:10.1016/S0738-3991(02)00194-5

SUPPLEMENT 1. Trial Protocol

JAMA Network Open | Pharmacy and Clinical Pharmacology Medication Optimization Protocol Efficacy for Geriatric Inpatients

JAMA Network Open. 2024;7(7):e2423544. doi:10.1001/jamanetworkopen.2024.23544 (Reprinted) July 30, 2024 12/13

Downloaded from jamanetwork.com by guest on 09/19/2026

https://dx.doi.org/10.1136/bmjopen-2020-041125

https://dx.doi.org/10.1093/ageing/afu145

https://dx.doi.org/10.1016/0168-8510(90)90421-9

https://dx.doi.org/10.4091/iken1991.8.1_109

https://dx.doi.org/10.1371/journal.pone.0149984

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamainternmed.2017.8274&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamanetworkopen.2023.4723&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

https://dx.doi.org/10.1111/jgs.16273

https://dx.doi.org/10.1186/s12877-020-01870-8

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamanetworkopen.2023.22505&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

https://dx.doi.org/10.1136/bmj.m1822

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamainternmed.2022.6545&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

https://dx.doi.org/10.1136/bmj.n1585

https://dx.doi.org/10.1111/jgs.16278

https://dx.doi.org/10.1016/j.jamda.2023.07.016

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamainternmed.2021.7429&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamainternmed.2022.0502&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2024.23544

https://dx.doi.org/10.5694/mja14.00146

https://dx.doi.org/10.1002/jgf2.464

https://dx.doi.org/10.1016/S0738-3991(02)00194-5

SUPPLEMENT 2. eMethods eTable 1. Top 10 Regular Medications at Baseline in the Intervention and Usual Care Group eTable 2. Top 10 Potentially Inappropriate Medication at Baseline in the Intervention and Usual Care Group eTable 3. Number of Recommendations and Acceptance Rate According to ATC Classification eTable 4. Per-Protocol Analysis Restricted to Participants in the Intervention Arm Who Were Subjected to One or More Recommendations for Deprescribing and Accepted the Proposal eTable 5. Proportion of Participants With Polypharmacy (5+ Medications), Hyper Polypharmacy (10+ Medications), and Potentially Inappropriate Medication in Both Groups eTable 6. Adverse Events Reported in the Intervention and Usual Care Group eTable 7. Post-Hoc Subgroup Analysis Restricted to Participants in Both Groups With One or More Potentially Inappropriate Medications eFigure 1. Kaplan-Meier Survival Plot of Time to First Fall or Fall-Related Injuries for Participants in the Intervention Group and Usual Care Group: Results From Modified Intention-to-Treat Analysis eFigure 2. Forest Plot of the Primary Outcome According to Subgroups

SUPPLEMENT 3. Data Sharing Statement

JAMA Network Open | Pharmacy and Clinical Pharmacology Medication Optimization Protocol Efficacy for Geriatric Inpatients

JAMA Network Open. 2024;7(7):e2423544. doi:10.1001/jamanetworkopen.2024.23544 (Reprinted) July 30, 2024 13/13

Downloaded from jamanetwork.com by guest on 09/19/2026

image14.emf

RESEARCH ARTICLE

Cost-effectiveness of a structured medication

review approach for multimorbid older

adults: Within-trial analysis of the OPERAM

study

Paola SalariID 1*, Cian O’MahonyID

2, Séverine Henrard3,4, Paco Welsing5, Arjun Bhadhuri1,

Nadine Schur1, Marie RoumetID 6, Shanthi BeglingerID

7,8, Thomas Beck7, Katharina

Tabea Jungo8, Stephen Byrne2, Stefanie Hossmann6, Wilma Knol9, Denis O’Mahony10,

Anne Spinewine4,11, Nicolas Rodondi7,8, Matthias Schwenkglenks1

1 Institute of Pharmaceutical Medicine (ECPM), University of Basel, Basel, Switzerland, 2 Pharmaceutical

Care Research Group, School of Pharmacy, University College Cork, Cork University Hospital, Cork, Ireland,

3 Louvain Drug Research Institute, Clinical Pharmacy Research Group, UCLouvain, Brussels, Belgium,

4 Institute of Health and Society (IRSS), UCLouvain, Brussels, Belgium, 5 Division of Internal Medicine and

Dermatology, University Medical Centre Utrecht, Utrecht, The Netherlands, 6 Clinical Trial Unit Bern,

University of Bern, Bern, Switzerland, 7 Department of General Internal Medicine, Inselspital, Bern University

Hospital, University of Bern, Bern, Switzerland, 8 Institute of Primary Health Care (BIHAM), University of

Bern, Bern, Switzerland, 9 Department of Geriatric Medicine and Expertise Centre Pharmacotherapy in Old

Persons, University Medical Centre Utrecht, Utrecht University, Utrecht, The Netherlands, 10 Department of

Medicine (Geriatrics), University College Cork, Cork University Hospital, Cork, Ireland, 11 Pharmacy

Department, CHU UCL Namur, Yvoir, Belgium

* [email protected]

Abstract

Background

Inappropriate polypharmacy has been linked with adverse outcomes in older, multimorbid

adults. OPERAM is a European cluster-randomized trial aimed at testing the effect of a

structured pharmacotherapy optimization intervention on preventable drug-related hospital

admissions in multimorbid adults with polypharmacy aged 70 years or older. Clinical results

of the trial showed a pattern of reduced drug-related hospital admissions, but without statisti-

cal significance. In this study we assessed the cost-effectiveness of the pharmacotherapy

optimisation intervention.

Methods

We performed a pre-planned within-trial cost-effectiveness analysis (CEA) of the OPERAM

intervention, from a healthcare system perspective. All data were collected within the trial

apart from unit costs. QALYs were computed by applying the crosswalk German valuation

algorithm to EQ-5D-5L-based quality of life data. Considering the clustered structure of the

data and between-country heterogeneity, we applied Generalized Structural Equation Mod-

els (GSEMs) on a multiple imputed sample to estimate costs and QALYs. We also per-

formed analyses by country and subgroup analyses by patient and morbidity

characteristics.

PLOS ONE

PLOS ONE | https://doi.org/10.1371/journal.pone.0265507 April 11, 2022 1 / 17

a1111111111

OPEN ACCESS

Citation: Salari P, O’Mahony C, Henrard S, Welsing

P, Bhadhuri A, Schur N, et al. (2022) Cost-

effectiveness of a structured medication review

approach for multimorbid older adults: Within-trial

analysis of the OPERAM study. PLoS ONE 17(4):

e0265507. https://doi.org/10.1371/journal.

pone.0265507

Editor: Shinya Tsuzuki, National Center for Global

Health and Medicine, JAPAN

Received: September 9, 2021

Accepted: March 1, 2022

Published: April 11, 2022

access article distributed under the terms of the

Creative Commons Attribution License, which

permits unrestricted use, distribution, and

reproduction in any medium, provided the original

author and source are credited.

Data Availability Statement: This study involves

human research participant data containing

sensitive patient information. In the EU Horizon

2020 grant agreement for the OPERAM study, it

had been specified that the data will be made

available upon request if the use has been

approved by an ethical committee. Therefore,

restrictions to make the underlying data directly

publicly available are both due to legal and ethical

reasons, as health data are sensitive data. Data for

this study will be made available for scientific

https://orcid.org/0000-0003-3757-6081

https://orcid.org/0000-0003-3581-5077

https://orcid.org/0000-0001-6655-1428

https://orcid.org/0000-0002-7965-4466

https://doi.org/10.1371/journal.pone.0265507

http://crossmark.crossref.org/dialog/?doi=10.1371/journal.pone.0265507&domain=pdf&date_stamp=2022-04-11

https://doi.org/10.1371/journal.pone.0265507

http://creativecommons.org/licenses/by/4.0/

Results

Trial-wide, the intervention was numerically dominant, with a potential cost-saving of CHF

3’588 (95% confidence interval (CI): -7’716; 540) and gain of 0.025 QALYs (CI: -0.002;

0.052) per patient. Robustness analyses confirmed the validity of the GSEM model. Sub-

group analyses suggested stronger effects in people at higher risk.

Conclusion

We observed a pattern towards dominance, potentially resulting from an accumulation of

multiple small positive intervention effects. Our methodological approaches may inform

other CEAs of multi-country, cluster-randomized trials facing presence of missing values

and heterogeneity between centres/countries.

Introduction

A large proportion of current healthcare spending in developed countries is on multimorbid

older adults. A significant proportion of healthcare funds may be wasted on overtreatment i.e.

unnecessary interventions and inappropriate medications. Inappropriate medications in older

adults may trigger adverse events, such as bleeding, falls and fractures, resulting in drug-related

hospital admissions (DRAs) [1], which are costly and potentially preventable [2, 3]. Few stud-

ies have assessed the costs and effectiveness of medication-optimisation interventions, and fur-

thermore the evidence from these studies is mixed. Some evidence of reduced DRAs was

found for hospital visits, emergency department visits and drug-related readmissions [4]. Two

recent Cochrane reviews, which evaluated interventions to optimise prescribing for older

adults in nursing homes, found that the effect of these interventions on drug costs was unclear

[5]. Another review on pharmacist-participated medication management for older adults in

nursing homes found favourable economic outcomes, albeit not all statistically significant [6].

Similarly, one study focussed on an intervention to reduce potentially inappropriate prescrib-

ing for older people found uncertainty with respect to its cost-effectiveness [7].

The OPERAM (OPtimising thERapy to prevent Avoidable hospital admissions in the Multi- morbid older people) trial (ClinicalTrials.gov Identifiers: main trial: NCT02986425; health eco-

nomic sub-study: NCT03108092) was a cluster-randomized trial conducted between 2016 and

2019, aimed at reducing preventable DRAs (primary endpoint) in adults aged 70 years or older

with multi-morbidity (i.e.� 3 coexistent chronic conditions defined by ICD-10 codes) and

polypharmacy (i.e.� 5 different regular drugs for more than 30 days), through improving

pharmacotherapy [8, 9].

The trial found no statistically significant clinical effect on its primary endpoint [10]. How-

ever, there were multiple, small intervention effects, consistently in the expected direction (e.g.

the hazard ratio for DRA was 0.87 and for death 0.90 when restricting intervention patients to

those who had�1 potentially inappropriate medication discontinued at 2 months). Hence, it

remained important to assess the effect of the trial intervention on (general) health relative to

its economic impact.

We performed a pre-planned, within-trial cost-effectiveness analysis (CEA) of the

OPERAM trial. This involved a series of methodological challenges, principally the clustered

structure of the trial data, possible heterogeneity between centres/countries and the unavoid-

able occurrence of missing data. Published literature offers abundant guidance on such

PLOS ONE Cost-effectiveness analysis of a structured medication review approach for multimorbid older adults

PLOS ONE | https://doi.org/10.1371/journal.pone.0265507 April 11, 2022 2 / 17

purposes upon request for researchers whose

proposed use of the data has been approved by the

OPERAM publication committee. After approval

and signing of a data transfer agreement ensuring

adherence to privacy and data handling, data and

documentation will be made available through a

secure file exchange platform. Partially deidentified

participant data, a data dictionary and annotated

case report forms will be made available. For data

access, external researchers can contact OPERAM

([email protected]).

Funding: This work is part of the project

“OPERAM: OPtimising thERapy to prevent

Avoidable hospital admissions in the Multimorbid

elderly” supported by the European Union’s

Horizon 2020 research and innovation programme

under the grant agreement No 6342388, and by the

Swiss State Secretariat for Education, Research

and Innovation (SERI) under contract number

15.0137. The opinions expressed and arguments

employed herein are those of the authors and do

not necessarily reflect the official views of the EC

and the Swiss government. The funders had no

role in study design, data collection and analysis,

decision to publish, or preparation of the

manuscript.

Competing interests: The authors have declared

that no competing interests exist.

https://doi.org/10.1371/journal.pone.0265507

mailto:[email protected]

methodological issues arising in within-trial cost-effectiveness studies [11–13]. However, these

are usually treated in isolation and there is very little guidance on how they should be dealt

with in combination. The methodological approach applied in our analysis combined all these

aspects.

Materials and methods

Design of the OPERAM trial

The overall objective of the OPERAM trial was to assess whether a software-assisted approach

to pharmacotherapy optimisation, namely the Systematic Tool to Reduce Inappropriate Pre- scribing (STRIP) based on STOPP/START criteria [14] and including STRIP assistant

(STRIPA), implemented by an interprofessional team composed of a medical doctor and a

pharmacist, led to an improvement in clinical and economic outcomes in the target popula-

tion, enrolled at the beginning of an index acute non-specific hospitalisation episode [15]. The

trial was cluster-randomised with clusters defined by a prescribing hospital physician. The

control group included patients receiving usual care. The trial was performed in four clinical

centres in Europe, namely University Hospital Bern, Switzerland; Saint-Luc University Hospi-

tal, Belgium; Cork University Hospital, Cork, Ireland; and University Medical Centre Utrecht,

The Netherlands. The planned sample size was 2,000 patients, enrolled in 80 clusters. Each

cluster represented a group of patients treated by the same consultant in the same ward. After

their initial hospitalization, patients were followed up by telephone interviews at 2, 6 and 12

months from randomization. More details are available from the trial protocol [8, 9] and trial

publication [10].

Approach to cost-effectiveness analysis

We followed a detailed health economic analysis plan (HEAP) developed before the end of the

trial and last modified before the lock of the trial database (for deviations from the HEAP, see

S1 Section in S1 File). All health economic data elements, including those covering medical

resource use and utilities, were collected within the trial, with the exception of unit cost data.

Cost-effectiveness was estimated for a one-year time horizon, aligned with the 12 month fol-

low-up period of the OPERAM trial. Due to the one-year time horizon, discounting was not

applied. We adopted a healthcare system perspective based on local unit costs for the main

analysis, with cost results expressed in Swiss francs (CHF), as Switzerland was the largest

recruiter into the OPERAM trial. We also approximated a societal perspective for a secondary

analysis, by adding the costs of informal care. Given the clustered nature of the trial data,

together with the likely presence of between-country heterogeneity, we adopted a regression-

based approach [13]. We applied Generalized Structural Equation Models (GSEMs) that allow

simultaneous estimation of costs and QALYs, and in the process account for the clustered

structure of the data (i.e. the clusters were treated as random effects) [16]. Individual patient

characteristics and country fixed effects were added to the models, the former as potential con-

founders given residual baseline imbalances, the latter to account for between-centre/country

heterogeneity [17–19]. We assessed heterogeneity between centres/countries using the method

described by Cook et al., by testing for qualitative and quantitative interaction on the key out-

comes of incremental QALYs and incremental costs [20, 21].

Calculation of quality-adjusted life years

Information on utilities was collected using the European Quality of Life-5 Dimensions (EQ-

5D-5L) instrument [22–26]. We combined the EQ-5D-5L responses with the interim

PLOS ONE Cost-effectiveness analysis of a structured medication review approach for multimorbid older adults

PLOS ONE | https://doi.org/10.1371/journal.pone.0265507 April 11, 2022 3 / 17

https://doi.org/10.1371/journal.pone.0265507

(crosswalk) valuation algorithm for Germany to calculate utility scores [27–29], to approxi-

mate a Swiss perspective given the absence of a Swiss valuation algorithm [30]. In the country-

specific analyses, we applied the German crosswalk algorithm for Switzerland, the UK cross-

walk algorithm for Ireland, and the Dutch crosswalk algorithm for Belgium and the Nether-

lands. QALYs were estimated for the one-year trial follow-up period, using standard area

under the curve methods following the trapezium rule [21]. For patients who died during the

trial, we set utility to zero from the date of death.

Calculation of costs

Based on unit cost data from non-trial sources, the following cost items were included in the

main analysis: costs of hospitalizations, rehabilitation facilities, medical visits, nursing visits at

home, nursing home care and drugs. Costs of informal care was included in a secondary analy-

sis approximating a societal perspective. We applied local unit costs for the year 2018 to each

country, similar to previous studies [13, 31–33]. We converted all local unit costs into one

common currency (Swiss Francs, CHF), using purchasing power parities (PPP) [34], as recom-

mended in the literature [13, 32, 33, 35].

Regarding sources of unit cost data, hospitalization costs were estimated using diagnosis-

related group-based (DRGs) reimbursement for Switzerland [36] and Ireland [37], retrieved

from validated hospital data for Belgium [38] and estimated using the Dutch manual for cost-

ing studies in healthcare for the Netherlands [39]. Costs of outpatient physician visits by spe-

cialty, and for visits with other healthcare providers (e.g. physiotherapists), were provided by a

provider of statutory health insurance for Switzerland, taken from the National Institute for

Health and Disability Insurance (NIHDI) website for Belgium [40], and based on national

costing studies for the Netherlands [16] and Ireland [41, 42]. Costs of nursing visits at home,

nursing home care and stays in rehabilitation facilities were drawn from national statistical

data for Switzerland [43], Ireland [44] and Belgium [40] and from national costing studies for

the Netherlands [16]. Drug costs were drawn from official data sources for Switzerland [45]

and Belgium [46], adapted from Belgium for the Netherlands [47] and estimated using a pre-

purchased wholesaler price list for Ireland. The costing of the STRIP intervention was based

on estimated cost of the software (based on similar products available on the market), and

trial-observed staff times and staff costs for all countries. Finally, costs of informal, unpaid care

provided by family caregivers were based on the average salary per hour in each country. More

detailed information on the collection of unit cost data is provided in the S2 Section in S1 File.

Missing data

Once we computed QALYs and total costs, approximately 8% of patients had a missing value

for at least one cost category (164 patients) and 26% had at least one missing element required

for QALY estimation (523 patients). More details can be found in the S3 Section in S1 File.

We assumed a missing at random (MAR) pattern of missing data. Variables were multiple

imputed using multilevel joint modelling, for any missing EQ-5D-5L score and for each cost

category rather than for aggregated measures of QALYs and total costs [48]. Patients’ personal

characteristics with no (or very few) missing values were used as the basis for imputation,

namely: age, sex, education, smoking status, quantity of alcohol consumed, number of drugs at

baseline, number of comorbidities at baseline, number of hospitalizations during twelve

months prior to baseline, being housebound at baseline, living in a nursing home at baseline,

having dementia at baseline, duration of index hospitalization, whether the index hospitaliza-

tion was in a medical or surgical ward, country, duration of patient follow-up, and whether or

not the patient died during the trial. Multiple imputations (MI) were performed separately by

PLOS ONE Cost-effectiveness analysis of a structured medication review approach for multimorbid older adults

PLOS ONE | https://doi.org/10.1371/journal.pone.0265507 April 11, 2022 4 / 17

https://doi.org/10.1371/journal.pone.0265507

treatment arm. Each imputation created five multiple imputed databases, each generated from

100 iterations [49]. We specified an MI model with random intercepts and slopes. Multiple

imputed results were estimated according to the combination rules by Rubin [50].

We challenged the MAR assumption by running the main model under several scenarios

where a MNAR structure of missing data was assumed [51]. Overall, results were robust and

not sensitive to the MAR assumption (see S4 Section in S1 File).

Robustness checks

Several robustness checks were performed to test and confirm the validity of the main model.

Firstly, we computed the simple difference between intervention and comparator arm costs

and QALYs without any regression analysis. Secondly, we compared the main GSEM model

with a simpler Seemingly Unrelated Regression (SUR) model [52] as well as separate linear

mixed models of costs and QALYs. We, then, added interaction terms between arm and coun-

try to the main model and performed a likelihood ratio test to compare the two models, with

and without the interaction terms.

In a further robustness check, we relaxed the normality assumption of the GSEM model

and assumed gamma distributed errors with a log link function. Finally, we ran a CEA includ-

ing complete observations only, excluding participants for whom incomplete cost or incom-

plete EQ-5D-5L responses were obtained.

Country-level and subgroup analyses

We performed country-specific analyses, by applying local costs (converted to CHF) and

locally relevant EQ-5D-5L valuation algorithms (as specified above) for each country in turn.

In addition, we performed subgroup analyses for the following subgroups defined in the

OPERAM trial protocol [8]: patient’s sex (female versus male), degree of independence (com-

munity-dwelling versus living in a nursing home at baseline), medical specialty of clusters

(hospitalization in a medical versus surgical ward), age (70–79 years; 80–89 years; more than

90 years), number of drugs (5–9;� 10), number of chronic comorbidities (3–6;�7).

Sensitivity analysis

We reduced and increased the unit costs for each cost category (medical visits, drugs, etc.), in

turn, by 30% to assess the impact of related uncertainty on incremental cost-effectiveness.

Finally, we combined a non-parametric bootstrap-based estimation of uncertainty ranges and

probabilistic sensitivity analysis, by drawing a vector of values from normal distributions rep-

resenting the parameter uncertainties of the cost parameters, alongside 1’000 bootstrap replica-

tions [53]. In each replication, costs were multiplied with the draws resulting from the

applicable normal distributions, the main GSEM model was re-estimated and incremental cost

and QALY results derived. We drew 1’000 bootstrap samples from each of the 5 imputed data-

sets separately, then pooled the samples together and presented them in a cost-effectiveness

plane [34].

Technical implementation

All analyses were run in STATA, version 15, apart from the multilevel joint modelling multiple

imputation, which was conducted on R software by using the R package JOMO, suitable for

cluster-specific covariance matrices (https://www.rdocumentation.org/packages/jomo, [54]).

PLOS ONE Cost-effectiveness analysis of a structured medication review approach for multimorbid older adults

PLOS ONE | https://doi.org/10.1371/journal.pone.0265507 April 11, 2022 5 / 17

https://www.rdocumentation.org/packages/jomo

https://doi.org/10.1371/journal.pone.0265507

Ethical approval

The OPERAM trial/project was approved by the independent research ethics committees at

each centre (lead ethics committee: Cantonal Ethics Committee Bern, Switzerland, ID 2016–

01200; Medical Research Ethics Committee Utrecht, Netherlands, ID 15-522/D; Comité

d’Ethique Hospitalo-Facultaire Saint-Luc-UCL: 2016/20JUL/347–Belgian registration No:

B403201629175; Cork University Teaching Hospitals Clinical Ethics Committee, Cork, Repub-

lic of Ireland; ID ECM 4 (o) 07/02/17), and by Swissmedic as the responsible regulatory

authority.

Results

Between December 2016 and October 2018, 2,008 participants were allocated to 54 clusters

(963 participants) in the intervention group, and 56 clusters (1,045 participants) in the control

group. Participants were recruited in Switzerland (822), Belgium (388), Ireland (346) and the

Netherlands (452). The median age was 79 years and 898 participants (44.7%) were women.

Ten (0.5%) participants were lost to follow-up, 118 (5.9%) participants withdrew from the

trial, and 384 (19.1%) died during follow-up.

Descriptive statistics of costs and QALYs

Descriptive statistics presented in Tables 1 and 2 are based on the observed sample (i.e. non-

imputed). Table 1 shows the descriptive statistics for QALYs, estimated using the Germany

EQ-5D-5L crosswalk valuation algorithm [25], for the one-year trial observation period for all

countries and by country. During the trial observation period, patients in the intervention

group and control groups accrued a mean of 0.649 and 0.632 QALYs, respectively. Mean

QALYs were higher for intervention patients in Switzerland and Belgium, but lower in Ireland

and the Netherlands. In contrast, median QALYs were higher for intervention patients in Swit-

zerland, Ireland and Belgium, but lower in the Netherlands.

Table 1. QALYs for all countries and by country per patient over one year.

QALYs N Mean Std. Dev. Min Max Median

All countries Control arm 765 0.632 0.307 -0.01 1 0.742

Intervention arm 720 0.649 0.312 -0.01 1 0.771

Switzerland Control arm 285 0.666 0.291 -0.01 1 0.772

Intervention arm 353 0.670 0.316 -0.01 1 0.799

Ireland Control arm 171 0.650 0.306 0.00 1 0.747

Intervention arm 110 0.625 0.348 0.00 1 0.796

Belgium Control arm 162 0.597 0.306 0.00 1 0.700

Intervention arm 120 0.697 0.220 0.05 1 0.766

The Netherlands Control arm 147 0.583 0.334 -0.01 1 0.716

Intervention arm 137 0.574 0.330 0.00 1 0.686

Note: QALYs were estimated over the one-year trial observation period.

https://doi.org/10.1371/journal.pone.0265507.t001

PLOS ONE Cost-effectiveness analysis of a structured medication review approach for multimorbid older adults

PLOS ONE | https://doi.org/10.1371/journal.pone.0265507 April 11, 2022 6 / 17

https://doi.org/10.1371/journal.pone.0265507.t001

https://doi.org/10.1371/journal.pone.0265507

Table 2 shows non-adjusted total direct medical costs per patient, which were slightly lower

for intervention patients (CHF 44’353) than for control patients (CHF 44’767). In Switzerland,

Ireland and Belgium, results showed the same trend of lower direct medical costs for interven-

tion patients, while in the Netherlands medical costs were lower for control patients.

Fig 1 presents differences by cost category between the trial arms, for all countries com-

bined. Intervention patients were on average more costly in terms of drugs (CHF +406), reha-

bilitation (CHF +2’170) and medical visits (CHF +109) but less costly in terms of

hospitalizations (CHF -1’750), nursing visits at home (CHF -1’181), and stays in nursing

Table 2. Total medical costs (CHF) per patient over one year.

Total costs (CHF) N Mean Std. Dev. Min Max Median

All countries Control arm 954 44’767 51’787 27 314’210 24’630

Intervention arm 890 44’353 50’812 94 412’074 23’976

Switzerland Control arm 345 52’904 54’674 49 314’210 34’721

Intervention arm 427 44’513 48’513 94 313’687 25’693

Ireland Control arm 202 44’872 51’871 104 257’562 23’865

Intervention arm 125 43’251 55’539 161 412’074 22’148

Belgium Control arm 204 33’181 42’351 30 307’074 16’927

Intervention arm 136 26’518 37’052 192 253’061 12’530

Netherlands Control arm 203 42’474 53’181 27 264’802 18’721

Intervention arm 202 56’703 56’892 125 256’195 34’729

Note: local costs expressed in Swiss Francs (CHF) through the PPP index.

https://doi.org/10.1371/journal.pone.0265507.t002

Fig 1. Mean cost differences per patient (CHF) between intervention and control patients, broken down by cost

categories, all countries. Note: A positive cost difference (bar on the right hand-side) means that costs were higher in

the intervention arm, while a negative cost difference (on the left hand-side), indicates that costs were higher in the

control arm.

https://doi.org/10.1371/journal.pone.0265507.g001

PLOS ONE Cost-effectiveness analysis of a structured medication review approach for multimorbid older adults

PLOS ONE | https://doi.org/10.1371/journal.pone.0265507 April 11, 2022 7 / 17

https://doi.org/10.1371/journal.pone.0265507.t002

https://doi.org/10.1371/journal.pone.0265507.g001

https://doi.org/10.1371/journal.pone.0265507

homes (CHF -1’090). In the S1 File, further details of the estimated costs are provided (S2-S9

Tables in S1 File).

Results of cost-effectiveness analyses and country-heterogeneity test

In the main, trial-wide GSEM analysis (Table 3), the STRIP intervention was estimated to gen-

erate 0.025 incremental QALYs (95% confidence interval (95% CI): -0.002; 0.052, p-value:

0.068) and reduce health care costs by CHF 3’588 (95% CI: -7’716; 540, p-value: 0.088)

(approximately EUR 3’000). Both the qualitative and quantitative heterogeneity tests indicated

the presence of between-country heterogeneity for incremental costs, and the qualitative test

also for incremental QALY estimates

Table 3 shows the results of the main, GSEM-based CEA of the overall trial, including

covariate effects. Numerically, the intervention strategy was dominant over the control strategy

but the intervention arm effects on QALYs and costs were not statistically significant. Adop-

tion of the approximated societal perspective in which informal care costs were additionally

included, resulted in incremental costs of CHF -4’214 (95% CI: -8’476; 48), in favour of the

intervention strategy, compared to CHF -3’588 (95% CI: -7’716; 540) for the healthcare system

perspective. The directions of associations of covariates with total healthcare costs and total

QALYs were all as expected. All results shown are multiple imputation-based, except where

stated otherwise.

Country-specific and subgroup analyses

Table 4 shows the incremental costs and incremental QALYs (i.e. the coefficients for interven-

tion arm) obtained in the country-specific and subgroup analyses. Full results are provided in

the S10-S16 Tables in S1 File.

The country-specific analyses yielded very similar incremental costs for Switzerland (CHF

-7’026, 822 patients), Ireland (CHF -8’963 equal to EUR -8’150, 346 patients) and Belgium

(CHF -6’081 equal to EUR -5’530, 388 patients), but very different ones for the Netherlands

(CHF 5’758 equal to EUR 5’234, 452 patients).

Overall, the results of the subgroup analyses were consistent with the main results (Table 4).

They may suggest that the potential cost savings and increase in QALYs in the intervention

arm, were higher for patients with a more serious clinical situation. For instance, having at

least seven comorbidities corresponded with an estimate of mean incremental costs of CHF

-5’964 (95% CI: -10’811; -1’117), and patients taking at least ten drugs were estimated to gener-

ate mean incremental QALYs of 0.038 (95% CI: 0.002; 0.073). The intervention was also esti-

mated to generate a statistically significant increase in QALYs for the male subgroup (0.045,

95% CI: 0.006; 0.083), and for the subgroup aged 80–89 years (0.063, 95% CI: 0.021; 0.105).

Robustness checks

Results of additional robustness checks for the main analysis are shown in the S1 File. Com-

pared to the GSEM-based analysis, the simple, non-multivariable-adjusted CEA presented

smaller incremental costs (CHF -1’486, 95% CI: -6’153; 3’180) and similar incremental QALYs

(0.026, 95% CI: -0.005; 0.057, S17 Table in S1 File).

The SUR model (incremental costs: CHF -3’822, 95% CI: -7’970; 326, and incremental

QALYs: 0.025, 95% CI: -0.001; 0.052, S19 Table in S1 File, column 1) and separate linear

mixed models of costs and QALYs (CHF -3’670, 95% CI: -7’662; 321, and incremental QALYs:

0.025, 95% CI: -0.002; 0.052, S19 Table in S1 File, column 2) yielded results similar to those of

the main GSEM model. In a further robustness check, we added an interaction term between

arm and country to the main model, which was not statistically significant (details not shown).

PLOS ONE Cost-effectiveness analysis of a structured medication review approach for multimorbid older adults

PLOS ONE | https://doi.org/10.1371/journal.pone.0265507 April 11, 2022 8 / 17

https://doi.org/10.1371/journal.pone.0265507

Table 3. Results of the main cost-effectiveness analysis, GSEM models.

Healthcare system perspective Societal perspective

Effects on costs (CHF)

Intervention arm -3’588 -4’214

[-7’716,540] [-8’476,48]

Age 800�� 946��

[445,1’156] [573,1’319]

Female -34 2’258

[-4’159,4’090] [-2’050,6’566]

Utility 6 months before -4’008 -6’650

[-17’712,9’696] [-19’862,6’561]

Utility baseline -21’033�� -21’897��

[-30’854,-11’212] [-31’496,-12’298]

Number of drugs 1’071�� 1’168��

[562,1’581] [636,1’699]

Number of comorbidities 593�� 601��

[215,971] [210,991]

Housebound 3’218 3’402

[-3’430,9’867] [-3’195,10’000]

Smoker 1’062 3’148

[-6’654,8’779] [-4’807,11’104]

High School 1’303 -235

[-3’760,6’366] [-5’452,4’982]

University 1’397 269

[-4’600,7’394] [-5’857,6’396]

Living in nursing home 49’600�� 46’151��

[40’042,59’157] [36’425,55’878]

Dementia 3’462 516

[-6’246,13’170] [-9’417,10’450]

N. of hosp. 1 year before 2’861�� 2’784��

[1’427,4’296] [1’309,4’259]

Medical ward 7’938�� 9’003��

[2’426,13’449] [3’202,14’804]

Observation time 141�� 156��

[124,159] [137,175]

Duration baseline hosp. 489�� 517��

[309,669] [333,702]

Ireland -299 11’507��

[-7’085,6’487] [4’312,18’702]

Belgium -7’789� -7’492

[-14’977,-600] [-15’027,43]

Netherlands 9’293�� 11’793��

[2’863,15’723] [4’947,18’638]

Constant -85’427�� -100’042��

[-116’461,-54’392] [-132’669,-67’414]

Effects on QALYs Intervention Arm 0.025 0.025

[-0.001,0.052] [-0.001,0.052]

Age -0.006�� -0.006��

(Continued)

PLOS ONE Cost-effectiveness analysis of a structured medication review approach for multimorbid older adults

PLOS ONE | https://doi.org/10.1371/journal.pone.0265507 April 11, 2022 9 / 17

https://doi.org/10.1371/journal.pone.0265507

Table 3. (Continued)

Healthcare system perspective Societal perspective

[-0.008,-0.004] [-0.008,-0.004]

Female -0.000 -0.000

[-0.035,0.035] [-0.035,0.035]

Utility 6 months before 0.191�� 0.191��

[0.131,0.252] [0.131,0.252]

Utility baseline 0.316�� 0.316��

[0.265,0.367] [0.265,0.367]

Number of drugs -0.007�� -0.007��

[-0.010,-0.004] [-0.010,-0.004]

Number of comorbidities -0.003� -0.003�

[-0.005,-0.000] [-0.005,-0.000]

Housebound -0.056� -0.056�

[-0.099,-0.012] [-0.099,-0.012]

Smoker 0.016 0.016

[-0.032,0.066] [-0.032,0.066]

High School 0.004 0.004

[-0.033,0.042] [-0.033,0.042]

University 0.028 0.028

[-0.015,0.071] [-0.015,0.071]

Living in nursing home 0.013 0.013

[-0.068,0.095] [-0.068,0.095]

Dementia -0.005 -0.005

[-0.082,0.070] [-0.081,0.071]

N. of hosp. 1 year before -0.013�� -0.013��

[-0.022,-0.004] [-0.022,-0.004]

Medical ward -0.065�� -0.066��

[-0.105,-0.026] [-0.105,-0.026]

Duration baseline hosp. -0.003�� -0.003��

[-0.005,-0.002] [-0.005,-0.002]

Ireland -0.029 -0.029

[-0.085,0.027] [-0.085,0.026]

Belgium -0.036 -0.036

[-0.086,0.013] [-0.086,0.013]

Netherlands -0.051� -0.051�

[-0.094,-0.008] [-0.094,-0.008]

Constant 1.037�� 1.039��

[0.844,1.231] [0.845,1.232]

Observations 2008 2008

Note: GSEM models. 95% confidence intervals in brackets.

� p<0.05

�� p<0.01

�� <0.001. Local costs are expressed in Swiss Francs (CHF) and combined using purchasing power parity indices (PPP). The main results of the GSEM-based analysis,

i.e. the incremental costs and incremental QALYs representing differences between the intervention and control arms, are equivalent to the coefficients of the variable

“intervention arm”. Results always represent average values per patient. A positive value of the coefficient for ‘intervention arm‘, for costs/QALYs, indicates that the

intervention is associated with higher average costs/QALYs per patient, and vice versa. Intervention arm, female, housebound, smoker, nursing home (i.e. living in a

nursing home at baseline), dementia and medical ward (whether the baseline hospitalization occurred in a medical vs surgical ward) are dichotomous variables.

Switzerland, Ireland, Belgium and the Netherlands form parts of a categorical variable; Switzerland serves as the reference group. Education (less than high school, high

school, university) is a categorical variable; less than high school serves as the reference group. Age is measured in years; utility 6 months before baseline and utility at

baseline are ranged from -0.2 to 1; number of drugs (at baseline), number of comorbidities (at baseline) and number of hospitalizations (in the year prior to baseline) are

integers; observation times and duration of baseline hospitalization (duration baseline hosp.) are measured in days.

https://doi.org/10.1371/journal.pone.0265507.t003

PLOS ONE Cost-effectiveness analysis of a structured medication review approach for multimorbid older adults

PLOS ONE | https://doi.org/10.1371/journal.pone.0265507 April 11, 2022 10 / 17

https://doi.org/10.1371/journal.pone.0265507.t003

https://doi.org/10.1371/journal.pone.0265507

The model with gamma distributed errors with a log link function resulted in mean incremen-

tal costs and QALYs similar to the main analysis (S18 Table in S1 File). Results obtained with

the non-imputed dataset (only observed data included) resulted in incremental costs of CHF

-4’022 (95% CI: -8404; 360) and incremental QALYs of 0.017 (95% CI: -0.007; 0,044), which

were close to the results from the main analysis (S19 Table in S1 File, column 3).

Sensitivity analyses

Deterministic sensitivity analysis of unit cost parameters, showed no substantial effect from

varying these parameters (S20 Table in S1 File). Fig 2 shows the results of the combined boot-

strap and probabilistic sensitivity analysis. The majority of the bootstrap replications (92.4%)

were in the lower right quadrant of the cost-effectiveness plane, indicating dominance of the

intervention.

Discussion and conclusion

We assessed the cost-effectiveness of the OPERAM randomized clinical trial intervention

STRIP to reduce potentially inappropriate prescribing and DRAs, using a software-based phar-

macotherapy optimization intervention based on the STOPP/START criteria [14]. By doing

this, we have built on a recent strand of literature dealing with the optimization of pharmaco-

therapy as a means of reducing hospital admissions and improving other patient-relevant out-

comes. Based on the health economic data collected in the OPERAM trial, the focus of our

analysis was on the Swiss healthcare system, with complementary analyses undertaken for the

other countries participating in OPERAM.

Table 4. Results of cost-effectiveness analyses for countries and subgroups, coefficients of main interest only, local costs (expressed in CHF).

Incremental Costs 95% CI Incremental QALYs 95% CI ICER

Analyses by country Switzerland (CHF) -7’027� [-13’130–924] 0.068 [-0.038 0.052] Dominant

Ireland (CHF) -8’963 [-20’373 24’456] -0.006 [-0.072 0.059] 1’493’833

Belgium (CHF) -6’081 [-17’073 4’910] 0.023 [-0.064 0 .111] Dominant

The Netherlands (CHF) 5’758 [5’273 16’789] 0.074 [-0.002 0.151] 77’810

Subgroup analyses Only female -3’642 [-9’983 2’699] 0.003 [-0.040 0.045] Dominant

Only male -4’270 [-9’763 1’222] 0.045 [0.006 0.083] Dominant

Community-dwelling -3’081 [-7’344 1’182] 0.024 [-0.002 0.048] Dominant

Nursing homes -902 [-15’013 13’207] 0.069 [-0.052 0.189] Dominant

Medical ward -4’615 [-9’719 488] 0.019 [-0.010 0.049] Dominant

Surgical ward -1’046 [-10’463 8’370] 0.044 [-0.021 0.108] Dominant

Age 70–79 -2’547 [-7’740 2’646] -0.001 [-0.041 0.039] 2’791’232

Age 80–89 -5’293 [-12’497 1’909] 0.063 [0.021 0.105] Dominant

Age 90+ 984 [-13’668 15’637] 0.047 [-0.078 0.173] 20’793

N. drugs: 5–9 -3’884 [-10’667 2’899] 0.009 [-0.035 0.053] Dominant

N. drugs:� 10 -3’177 [-8’548 2’192] 0.038 [0.002 0.073] Dominant

N. comorbidities: 3–6 4’388 [-6’349 15’127] -0.029 [-0.105 0.046] Dominated

N. comorbidities:� 7 -5’964 [-10’811–1’117] 0.034 [0.005 0.063] Dominant

Note: the first column characterizes the analysis performed. All analyses were run with the same set of covariates as were used for the main analysis. Column 2 and

column 4 report the coefficients of the variable “Intervention arm”, representing incremental costs and incremental QALYs respectively. ICER: incremental cost-

effectiveness ratio in CHF per QALY gained. The positive ICER value relative to the analyses for “Ireland” and “Age group 70–79” represent saving per QALY lost.

https://doi.org/10.1371/journal.pone.0265507.t004

PLOS ONE Cost-effectiveness analysis of a structured medication review approach for multimorbid older adults

PLOS ONE | https://doi.org/10.1371/journal.pone.0265507 April 11, 2022 11 / 17

https://doi.org/10.1371/journal.pone.0265507.t004

https://doi.org/10.1371/journal.pone.0265507

The OPERAM trial intervention was numerically dominant, as it resulted in cost savings of

about CHF 3’500 per patient and a gain of about 0.025 QALYs per patient over the one-year

trial observation period. However, these estimates of incremental costs and QALYs were not

statistically significant. Results of the analysis from a societal perspective suggested an even

higher potential cost saving per patient (CHF -4’214). Of note, the coefficient for Ireland dif-

fered substantially between the two analyses (CHF -299 vs CHF 11’507 when adding the

opportunity cost of informal care) due to a strong reliance on informal care in the country (see

also S8 Table in S1 File). Positive results of our analysis do not contradict the main outcome of

the OPERAM trial [10]. Although the trial did not find a statistically significant reduction in

DRAs resulting from the STRIP intervention, there was some evidence in terms of successfully

implemented, pharmacotherapy-related recommendations in 474 (51.7%) participants in the

intervention group. Also, the hazard ratios for specific secondary outcomes of the trial were

not statistically significant but quite consistently in the direction of favouring the intervention

(0.96 for the first fall; 0.90 for death). We speculate that the combination of these non-signifi-

cant but numerically positive results, suggesting intervention effectiveness, may have trans-

lated into our estimates of higher QALYs and lower health care utilisation costs.

Previous studies addressing effectiveness of medication-optimising interventions mainly

focused on medical outcomes and yielded very mixed evidence. Our results for costs can be

compared with two Cochrane reviews looking at interventions to optimise prescribing for

older adults in nursing homes that also looked at the effects on drug costs [5]. In one of the

reviews, 12 relevant studies were identified of which only 5 found that medication optimising

Fig 2. Cost-effectiveness plane for all countries. Note: The X-axis shows the difference in QALYs between the

OPERAM trial arms: a positive QALY difference means an increase in QALYs in the intervention arm and is

represented on the right-hand side. By contrast, a negative QALY difference is represented on the left-hand side. The

Y-axis shows the cost differential. If the intervention is associated with a cost reduction, the cost differential is negative

(lower part of the graph), while a positive cost differential (upper part of the graph) indicates an increase in costs due to

the intervention.

https://doi.org/10.1371/journal.pone.0265507.g002

PLOS ONE Cost-effectiveness analysis of a structured medication review approach for multimorbid older adults

PLOS ONE | https://doi.org/10.1371/journal.pone.0265507 April 11, 2022 12 / 17

https://doi.org/10.1371/journal.pone.0265507.g002

https://doi.org/10.1371/journal.pone.0265507

interventions were associated with a reduction in drug costs (and this reduction was small in

all 5 studies) [5]. In the other review which assessed pharmacist-directed medication manage-

ment interventions for older adults in nursing homes, non-significant cost reductions were

found to result from intervention [6], which is similar to the results of this study.

Results by country were heterogeneous. They went in the same direction for Switzerland,

Belgium and Ireland where intervention arm costs were lower. However, in the case of the

Netherlands, the intervention arm showed higher costs. We could not find a substance-matter

explanation for this deviation from the overall trend. It could to some extent be due to differ-

ences between countries in the way primary care physicians received and implemented the

STRIPA-based recommendations, generated in the inpatient setting. It is also possible that

some primary care physicians were more compliant, others less so, with STRIPA recommen-

dations. Observed heterogeneity between centres/countries, which may have resulted from

health system characteristics and resulting subtle differences in the implementation of the

intervention, represents a limitation of the trial. Therefore, the results of our CEA should be

interpreted with caution and consider local implementation. Subgroup analyses suggested

more favourable results for groups of persons at higher risk (i.e. having at least seven comor-

bidities or taking at least ten drugs). It should be noted, however, that patient numbers for the

country-level and subgroup analysis were quite limited, and we cannot exclude an increased

probability of chance effects.

We undertook a fully pre-planned health economic data collection and approach to analy-

sis. The methodological approaches pursued may inform the conduct of other CEAs of clus-

ter-randomized trials with similar design and validity issues. In our multiple imputation of

missing data, we explicitly accounted for the clustered nature of the data. We subsequently

used a regression-based approach (GSEM) to take into account the simultaneity and thus,

potential correlation, of costs and QALY results for each patient, and also the presence of clus-

ters given trial design. Robustness checks confirmed the validity of the GSEM approach.

We also detected between-country heterogeneity in the main outcomes. We included coun-

try fixed effects in the main GSEM analysis, where we were primarily interested in the overall

effect of the treatment. In addition, we ran separate CEAs for each country to further address

the heterogeneity issue (see S10 Table in S1 File). Other analysis methods reported in the liter-

ature recommend adding interaction terms between arm and country (as we did in a robust-

ness check), or consider countries as random effects [13]. We omitted the latter option given

the relatively small country-level sample sizes and small number of countries in the OPERAM

trial. Also, it is common in CEAs of multinational trials to apply the unit costs of one country

to all participating countries [20, 55, 56]. We alternatively applied local costs to each country

and made them comparable through the use of PPP. This approach has been reported to better

capture differences in mean, spread and skewness across centres [13, 31–33].

This study had some limitations. Firstly, the presence of missing data, despite being

addressed through the multiple imputation procedure, remains a potential source of bias for

our analyses. Secondly, there were some unclear data points in the trial database, especially in

the drug utilisation data. In about a quarter of cases, values representing dose, unit of dose, or

start or end date information were unclear and had to be treated as missing. This may have led

to data distortion to some extent. Thirdly, giving the absence of a Swiss EQ-5D valuation algo-

rithm, the main analysis used a German algorithm on the basis of geographic proximity [57].

Finally, the sample size of the trial was based on the clinical primary endpoint, i.e. drug-related

hospital admissions, and may have been insufficient to detect statistically significant changes

in economic outcomes, particularly QALYs.

There is a recognized lack of research on the economic impact of pharmacotherapy optimi-

sation, and inconsistency in the little available evidence on the effects of related interventions

PLOS ONE Cost-effectiveness analysis of a structured medication review approach for multimorbid older adults

PLOS ONE | https://doi.org/10.1371/journal.pone.0265507 April 11, 2022 13 / 17

https://doi.org/10.1371/journal.pone.0265507

for both costs and effectiveness. This study emphasizes the relevance of including CEA in clini-

cal trials and other studies aimed at optimizing pharmacotherapy. The OPERAM trial inter-

vention was numerically dominant, with a potential saving of about CHF 3’500 and a gain of

0.025 QALYs per patient over a one-year time horizon. However, results were not statistically

significant at the 5% level. Hence, more research is needed to support policy makers to address

resources in an efficient way. The methodological approaches of the present CEA could be

used to inform the planning of future CEAs of multi-country, cluster-randomized trials facing

similar challenges.

Supporting information

S1 File.

(DOCX)

Acknowledgments

We would like to thank Flavia Lazzeri for her availability in providing and interpreting Spitex

data, Mattia Minke from Canton Basel Stadt and the Insel Data Science Centre for providing

hospitalizations data, Helsana for providing unit cost data, Sven Trelle for his support in the

phases of planning and conducting the study, Matteo Quartagno and Günther Fink for their

valuable comments. The EQ-5D instrument is used by permission of the EuroQol Group.

Author Contributions

Conceptualization: Paola Salari, Nicolas Rodondi, Matthias Schwenkglenks.

Data curation: Paola Salari, Cian O’Mahony, Séverine Henrard, Paco Welsing, Marie Roumet,

Thomas Beck.

Formal analysis: Paola Salari.

Funding acquisition: Nicolas Rodondi, Matthias Schwenkglenks.

Investigation: Paola Salari.

Methodology: Paola Salari, Arjun Bhadhuri, Matthias Schwenkglenks.

Project administration: Paola Salari, Shanthi Beglinger, Katharina Tabea Jungo, Stefanie Hos-

smann, Matthias Schwenkglenks.

Resources: Matthias Schwenkglenks.

Software: Paola Salari.

Supervision: Nadine Schur, Denis O’Mahony, Matthias Schwenkglenks.

Validation: Paola Salari, Wilma Knol, Matthias Schwenkglenks.

Writing – original draft: Paola Salari.

Writing – review & editing: Paola Salari, Cian O’Mahony, Séverine Henrard, Paco Welsing,

Arjun Bhadhuri, Nadine Schur, Marie Roumet, Shanthi Beglinger, Thomas Beck, Katharina

Tabea Jungo, Stephen Byrne, Stefanie Hossmann, Wilma Knol, Denis O’Mahony, Anne

Spinewine, Nicolas Rodondi, Matthias Schwenkglenks.

PLOS ONE Cost-effectiveness analysis of a structured medication review approach for multimorbid older adults

PLOS ONE | https://doi.org/10.1371/journal.pone.0265507 April 11, 2022 14 / 17

http://www.plosone.org/article/fetchSingleRepresentation.action?uri=info:doi/10.1371/journal.pone.0265507.s001

https://doi.org/10.1371/journal.pone.0265507

References 1. Pirmohamed M, James S, Meakin S, Green C, Scott AK, Walley TJ, et al. Adverse drug reactions as

cause of admission to hospital: prospective analysis of 18 820 patients. BMJ. 2004; 329(7456):15–9.

https://doi.org/10.1136/bmj.329.7456.15 PMID: 15231615

2. Moriarty F, Cahir C, Bennett K, Fahey T. Economic impact of potentially inappropriate prescribing and

related adverse events in older people: a cost-utility analysis using Markov models. BMJ Open. 2019; 9

(1):e021832. https://doi.org/10.1136/bmjopen-2018-021832 PMID: 30705233

3. Hyttinen V, Jyrkkä J, Valtonen H. A Systematic Review of the Impact of Potentially Inappropriate Medi-

cation on Health Care Utilization and Costs Among Older Adults. Medical Care. 2016; 54(10). https://

doi.org/10.1097/MLR.0000000000000587 PMID: 27367864

4. Gillespie U, Alassaad A, Henrohn D, Garmo H, Hammarlund-Udenaes M, Toss H, et al. A comprehen-

sive pharmacist intervention to reduce morbidity in patients 80 years or older: a randomized controlled

trial. Arch Intern Med. 2009; 169(9):894–900. https://doi.org/10.1001/archinternmed.2009.71 PMID:

19433702

5. Alldred DP, Kennedy MC, Hughes C, Chen TF, Miller P. Interventions to optimise prescribing for older

people in care homes. Cochrane Database of Systematic Reviews. 2016(2). https://doi.org/10.1002/

14651858.CD009095.pub3 PMID: 26866421

6. Kwak A, Moon YJ, Song Y-K, Yun H-Y, Kim K. Economic Impact of Pharmacist-Participated Medication

Management for Elderly Patients in Nursing Homes: A Systematic Review. Int J Environ Res Public

Health. 2019; 16(16):2955. https://doi.org/10.3390/ijerph16162955 PMID: 31426382

7. Gillespie P, Clyne B, Raymakers A, Fahey T, Hughes CM, Smith SM. REDUCING POTENTIALLY

INAPPROPRIATE PRESCRIBING FOR OLDER PEOPLE IN PRIMARY CARE: COST-EFFECTIVE-

NESS OF THE OPTI-SCRIPT INTERVENTION. International Journal of Technology Assessment in

Health Care. 2017; 33(4):494–503. https://doi.org/10.1017/S0266462317000782 PMID: 29019297

8. Crowley EK, Sallevelt BTGM, Huibers CJA, Murphy KD, Spruit M, Shen Z, et al. Intervention protocol:

OPtimising thERapy to prevent avoidable hospital Admission in the Multi-morbid elderly (OPERAM): a

structured medication review with support of a computerised decision support system. BMC Health Ser-

vices Research. 2020; 20(1):220. https://doi.org/10.1186/s12913-020-5056-3 PMID: 32183810

9. Adam L, Moutzouri E, Baumgartner C, Loewe AL, Feller M, M’Rabet-Bensalah K, et al. Rationale and

design of OPtimising thERapy to prevent Avoidable hospital admissions in Multimorbid older people

(OPERAM): a cluster randomised controlled trial. BMJ Open. 2019; 9(6):e026769. https://doi.org/10.

1136/bmjopen-2018-026769 PMID: 31164366

10. Blum MR, Sallevelt BTGM, Spinewine A, O’Mahony D, Moutzouri E, Feller M, et al. Optimizing Therapy

to Prevent Avoidable Hospital Admissions in Multimorbid Older Adults (OPERAM): cluster randomised

controlled trial. BMJ. 2021; 374:n1585. https://doi.org/10.1136/bmj.n1585 PMID: 34257088

11. Glick H, Doshi JA, Sonnad SS, Polsky D. Economic evaluation in clinical trials. 2nd ed. Oxford: Oxford

University Press; 2015.

12. Grieve R, Nixon R, Thompson SG. Bayesian hierarchical models for cost-effectiveness analyses that

use data from cluster randomized trials. Med Decis Making. 2010; 30(2):163–75. https://doi.org/10.

1177/0272989X09341752 PMID: 19675321

13. Grieve R, Nixon R, Thompson SG, Cairns J. Multilevel models for estimating incremental net benefits in

multinational studies. Health Econ. 2007; 16(8):815–26. https://doi.org/10.1002/hec.1198 PMID:

17191271

14. O’Mahony D, O’Sullivan D, Byrne S, O’Connor MN, Ryan C, Gallagher P. STOPP/START criteria for

potentially inappropriate prescribing in older people: version 2. Age and Ageing. 2014; 44(2):213–8.

https://doi.org/10.1093/ageing/afu145 PMID: 25324330

15. Drenth-van Maanen AC, Leendertse AJ, Jansen PAF, Knol W, Keijsers CJPW, Meulendijk MC, et al.

The Systematic Tool to Reduce Inappropriate Prescribing (STRIP): Combining implicit and explicit pre-

scribing tools to improve appropriate prescribing. Journal of Evaluation in Clinical Practice. 2018; 24

(2):317–22. https://doi.org/10.1111/jep.12787 PMID: 28776873

16. Acock AC. Discovering Structural Equation Modeling Using Stata, Revised Edition: College Station,

TX. Stata Press; 2013.

17. Manca A, Willan AR. ’Lost in translation’: accounting for between-country differences in the analysis of

multinational cost-effectiveness data. Pharmacoeconomics. 2006; 24(11):1101–19. https://doi.org/10.

2165/00019053-200624110-00007 PMID: 17067195

18. Manca A, Rice N, Sculpher MJ, Briggs AH. Assessing generalisability by location in trial-based cost-

effectiveness analysis: the use of multilevel models. Health Econ. 2005; 14(5):471–85. https://doi.org/

10.1002/hec.914 PMID: 15386662

PLOS ONE Cost-effectiveness analysis of a structured medication review approach for multimorbid older adults

PLOS ONE | https://doi.org/10.1371/journal.pone.0265507 April 11, 2022 15 / 17

https://doi.org/10.1136/bmj.329.7456.15

http://www.ncbi.nlm.nih.gov/pubmed/15231615

https://doi.org/10.1136/bmjopen-2018-021832

http://www.ncbi.nlm.nih.gov/pubmed/30705233

https://doi.org/10.1097/MLR.0000000000000587

http://www.ncbi.nlm.nih.gov/pubmed/27367864

https://doi.org/10.1001/archinternmed.2009.71

http://www.ncbi.nlm.nih.gov/pubmed/19433702

https://doi.org/10.1002/14651858.CD009095.pub3

http://www.ncbi.nlm.nih.gov/pubmed/26866421

https://doi.org/10.3390/ijerph16162955

http://www.ncbi.nlm.nih.gov/pubmed/31426382

https://doi.org/10.1017/S0266462317000782

http://www.ncbi.nlm.nih.gov/pubmed/29019297

https://doi.org/10.1186/s12913-020-5056-3

http://www.ncbi.nlm.nih.gov/pubmed/32183810

https://doi.org/10.1136/bmjopen-2018-026769

http://www.ncbi.nlm.nih.gov/pubmed/31164366

https://doi.org/10.1136/bmj.n1585

http://www.ncbi.nlm.nih.gov/pubmed/34257088

https://doi.org/10.1177/0272989X09341752

http://www.ncbi.nlm.nih.gov/pubmed/19675321

https://doi.org/10.1002/hec.1198

http://www.ncbi.nlm.nih.gov/pubmed/17191271

https://doi.org/10.1093/ageing/afu145

http://www.ncbi.nlm.nih.gov/pubmed/25324330

https://doi.org/10.1111/jep.12787

http://www.ncbi.nlm.nih.gov/pubmed/28776873

https://doi.org/10.2165/00019053-200624110-00007

http://www.ncbi.nlm.nih.gov/pubmed/17067195

https://doi.org/10.1002/hec.914

http://www.ncbi.nlm.nih.gov/pubmed/15386662

https://doi.org/10.1371/journal.pone.0265507

19. Willke RJ, Glick HA, Polsky D, Schulman K. Estimating country-specific cost-effectiveness from multi-

national clinical trials. Health Economics. 1998; 7(6):481–93. https://doi.org/10.1002/(sici)1099-1050

(199809)7:6<481::aid-hec353>3.0.co;2-k PMID: 9809707

20. Cook JR, Drummond M, Glick H, Heyse JF. Assessing the appropriateness of combining economic

data from multinational clinical trials. Stat Med. 2003; 22(12):1955–76. https://doi.org/10.1002/sim.1389

PMID: 12802815

21. Ramsey SD, Willke RJ, Glick H, Reed SD, Augustovski F, Jonsson B, et al. Cost-effectiveness analysis

alongside clinical trials II-An ISPOR Good Research Practices Task Force report. Value Health. 2015;

18(2):161–72. https://doi.org/10.1016/j.jval.2015.02.001 PMID: 25773551

22. Agborsangaya CB, Lahtinen M, Cooke T, Johnson JA. Comparing the EQ-5D 3L and 5L: measurement

properties and association with chronic conditions and multimorbidity in the general population. Health

Qual Life Outcomes. 2014; 12:74. https://doi.org/10.1186/1477-7525-12-74 PMID: 24885017

23. Conner-Spady BL, Marshall DA, Bohm E, Dunbar MJ, Loucks L, Al Khudairy A, et al. Reliability and

validity of the EQ-5D-5L compared to the EQ-5D-3L in patients with osteoarthritis referred for hip and

knee replacement. Qual Life Res. 2015; 24(7):1775–84. https://doi.org/10.1007/s11136-014-0910-6

PMID: 25555837

24. Craig BM, Pickard AS, Lubetkin EI. Health problems are more common, but less severe when mea-

sured using newer EQ-5D versions. J Clin Epidemiol. 2014; 67(1):93–9. https://doi.org/10.1016/j.

jclinepi.2013.07.011 PMID: 24075597

25. Feng Y, Devlin N, Herdman M. Assessing the health of the general population in England: how do the

three- and five-level versions of EQ-5D compare? Health Qual Life Outcomes. 2015; 13:171. https://

doi.org/10.1186/s12955-015-0356-8 PMID: 26489956

26. Rabin R, de Charro F. EQ-5D: a measure of health status from the EuroQol Group. Ann Med. 2001; 33

(5):337–43. https://doi.org/10.3109/07853890109002087 PMID: 11491192

27. Bailey H, Kind P. Preliminary findings of an investigation into the relationship between national culture

and EQ-5D value sets. Qual Life Res. 2010; 19(8):1145–54. https://doi.org/10.1007/s11136-010-9678-

5 PMID: 20496167

28. Norman R, Cronin P, Viney R, King M, Street D, Ratcliffe J. International comparisons in valuing EQ-5D

health states: a review and analysis. Value Health. 2009; 12(8):1194–200. https://doi.org/10.1111/j.

1524-4733.2009.00581.x PMID: 19695009

29. Herdman M, Gudex C, Lloyd A, Janssen M, Kind P, Parkin D, et al. Development and preliminary testing

of the new five-level version of EQ-5D (EQ-5D-5L). Qual Life Res. 2011; 20(10):1727–36. https://doi.

org/10.1007/s11136-011-9903-x PMID: 21479777

30. Brazier J, Ara R, Azzabi I, Busschbach J, Chevrou-Séverac H, Crawford B, et al. Identification, Review,

and Use of Health State Utilities in Cost-Effectiveness Models: An ISPOR Good Practices for Outcomes

Research Task Force Report. Value in Health. 2019; 22(3):267–75. https://doi.org/10.1016/j.jval.2019.

01.004 PMID: 30832964

31. Jönsson B, Cook JR, Pedersen TR. The cost-effectiveness of lipid lowering in patients with diabetes:

results from the 4S trial. Diabetologia. 1999; 42(11):1293–301. https://doi.org/10.1007/s001250051441

PMID: 10550412

32. Briggs AH, Glick HA, Lozano-Ortega G, Spencer M, Calverley PM, Jones PW, et al. Is treatment with

ICS and LABA cost-effective for COPD? Multinational economic analysis of the TORCH study. Eur

Respir J. 2010; 35(3):532–9. https://doi.org/10.1183/09031936.00153108 PMID: 19717476

33. Thompson SG, Nixon RM, Grieve R. Addressing the issues that arise in analysing multicentre cost

data, with application to a multinational study. J Health Econ. 2006; 25(6):1015–28. https://doi.org/10.

1016/j.jhealeco.2006.02.001 PMID: 16540192

34. World Bank. PPP conversion factor, GDP, (LCU per international $). https://data.worldbank.org/

indicator/PA.NUS.PPP (accessed 16/4/2020). 2020.

35. Kanavos P, Mossialos E. International Comparisons of Health Care Expenditures: What We Know and

What We Do not Know. Journal of Health Services Research & Policy. 1999; 4(2):122–6.

36. Swiss Federal Office of Public Health. Spezialitätenliste. http://www.spezialitätenliste.ch/. Accessed

2020-10-12.

37. Department of Health and Children. Case-mix measurement in Irish hospitals: A broad outline of the

main features. Dublin, Ireland; 2003.

38. Open Drug Database (ODDB). Accessed 2021-01-15.

39. Kanters TA, Bouwmans CAM, van der Linden N, Tan SS, van Roijen LH. Update of the Dutch manual

for costing studies in health care. PLOS ONE. 2017; 12(11):1–11. https://doi.org/10.1371/journal.pone.

0187477 PMID: 29121647

PLOS ONE Cost-effectiveness analysis of a structured medication review approach for multimorbid older adults

PLOS ONE | https://doi.org/10.1371/journal.pone.0265507 April 11, 2022 16 / 17

https://doi.org/10.1002/%28sici%291099-1050%28199809%297%3A6%26lt%3B481%3A%3Aaid-hec353%26gt%3B3.0.co%3B2-k

http://www.ncbi.nlm.nih.gov/pubmed/9809707

https://doi.org/10.1002/sim.1389

http://www.ncbi.nlm.nih.gov/pubmed/12802815

https://doi.org/10.1016/j.jval.2015.02.001

http://www.ncbi.nlm.nih.gov/pubmed/25773551

https://doi.org/10.1186/1477-7525-12-74

http://www.ncbi.nlm.nih.gov/pubmed/24885017

https://doi.org/10.1007/s11136-014-0910-6

http://www.ncbi.nlm.nih.gov/pubmed/25555837

https://doi.org/10.1016/j.jclinepi.2013.07.011

http://www.ncbi.nlm.nih.gov/pubmed/24075597

https://doi.org/10.1186/s12955-015-0356-8

http://www.ncbi.nlm.nih.gov/pubmed/26489956

https://doi.org/10.3109/07853890109002087

http://www.ncbi.nlm.nih.gov/pubmed/11491192

https://doi.org/10.1007/s11136-010-9678-5

http://www.ncbi.nlm.nih.gov/pubmed/20496167

https://doi.org/10.1111/j.1524-4733.2009.00581.x

http://www.ncbi.nlm.nih.gov/pubmed/19695009

https://doi.org/10.1007/s11136-011-9903-x

http://www.ncbi.nlm.nih.gov/pubmed/21479777

https://doi.org/10.1016/j.jval.2019.01.004

http://www.ncbi.nlm.nih.gov/pubmed/30832964

https://doi.org/10.1007/s001250051441

http://www.ncbi.nlm.nih.gov/pubmed/10550412

https://doi.org/10.1183/09031936.00153108

http://www.ncbi.nlm.nih.gov/pubmed/19717476

https://doi.org/10.1016/j.jhealeco.2006.02.001

http://www.ncbi.nlm.nih.gov/pubmed/16540192

https://data.worldbank.org/indicator/PA.NUS.PPP

http://www.spezialittenliste.ch/

https://doi.org/10.1371/journal.pone.0187477

http://www.ncbi.nlm.nih.gov/pubmed/29121647

https://doi.org/10.1371/journal.pone.0265507

40. National Institute for Health and Disability Insurance (NIHDI). https://www.riziv.fgov.be/fr/themes/cout-

remboursement/par-mutualite/prestations-individuelles/prix/Pages/default.aspx. (Accessed 20 April

2020). 2020.

41. Park B, Ock M, Lee HA, Lee S, Han H, Jo M-W, et al. Multimorbidity and health-related quality of life in

Koreans aged 50 or older using KNHANES 2013–2014. Health and quality of life outcomes. 2018; 16

(1):186–. https://doi.org/10.1186/s12955-018-1016-6 PMID: 30219061

42. Healthcare Pricing Office. Admitted Price List 2020 http://www.hpo.ie/abf/

ABF2020AdmittedPatientPriceList.pdf. (Accessed 15 September 2019).

43. Federal Statistical Office. Health. Pocket Statistics 2018. https://www.bfs.admin.ch/asset/en/1543-

1800. (Accessed October 2019). 2018.

44. Meirmans J. Nursing Homes Support Scheme Trends and Figures. Dublin: Irish Government Eco-

nomic & Evaluation Service; 2017.

45. Swiss Federal Office of Public Health. Spezialitätenliste. http://www.spezialitätenliste.ch/. Accessed

2019-10-12. 2019.

46. Soley-Bori M, Ashworth M, Bisquera A, Dodhia H, Lynch R, Wang Y, et al. Impact of multimorbidity on

healthcare costs and utilisation: a systematic review of the UK literature. Br J Gen Pract. 2021; 71(702):

e39–e46. https://doi.org/10.3399/bjgp20X713897 PMID: 33257463

47. Kanavos P, Vandoros S, Irwin R, Nicod E, Casson M, Medical Technology Research Group (LSE

Health), et al. DIfferences in costs of and access to pharmaceutical products in the EU. Brussels: Euro-

pean Parliament; 2011.

48. Michalowsky B, Hoffmann W, Kennedy K, Xie F. Is the whole larger than the sum of its parts? Impact of

missing data imputation in economic evaluation conducted alongside randomized controlled trials. Eur

J Health Econ. 2020; 21(5):717–28. https://doi.org/10.1007/s10198-020-01166-z PMID: 32108274

49. Carpenter J, Kenward M. Missing Data in Randomised Controlled Trial: A Practical Guide. 2007.

50. Rubin DB. Multiple Imputation for Nonresponse in Surveys. New York: John Wiley and Sons; 1987.

51. Leurent B, Gomes M, Faria R, Morris S, Grieve R, Carpenter JR. Sensitivity Analysis for Not-at-Random

Missing Data in Trial-Based Cost-Effectiveness Analysis: A Tutorial. PharmacoEconomics. 2018; 36

(8):889–901. https://doi.org/10.1007/s40273-018-0650-5 PMID: 29679317

52. Wooldridge JM. Econometric Analysis of Cross Section and Panel Data. Cambridge, Massachusetts &

London, England: The MIT Press.

53. Lord J, Asante MA. Estimating uncertainty ranges for costs by the bootstrap procedure combined with

probabilistic sensitivity analysis. Health Econ. 1999; 8(4):323–33. https://doi.org/10.1002/(sici)1099-

1050(199906)8:4<323::aid-hec431>3.0.co;2-0 PMID: 10398525

54. Quartagno Matteo, Carpenter J. jomo: A package for Multilevel Joint Modelling Multiple Imputation.

URL https://CRAN.R-project.org/package=jomo. 2020.

55. Scott WG, Cooper BC, Scott HM. Pharmacoeconomic evaluation of roxithromycin versus amoxycillin/

clavulanic acid in a community-acquired lower respiratory tract infection study. Infection. 1995; 23(1):

S21–S4.

56. Dasbach EJ, Rich MW, Segal R, Gerth WC, Carides GW, Cook JR, et al. The Cost-Effectiveness of

Losartan versus Captopril in Patients with Symptomatic Heart Failure. Cardiology. 1999; 91(3):189–94.

https://doi.org/10.1159/000006908 PMID: 10516413

57. Joelson A, Wildeman P, Sigmundsson FG, Rolfson O, Karlsson J. Properties of the EQ-5D-5L when

prospective longitudinal data from 28,902 total hip arthroplasty procedures are applied to different Euro-

pean EQ-5D-5L value sets. The Lancet Regional Health-Europe. 2021; 8:100165. https://doi.org/10.

1016/j.lanepe.2021.100165 PMID: 34557854

PLOS ONE Cost-effectiveness analysis of a structured medication review approach for multimorbid older adults

PLOS ONE | https://doi.org/10.1371/journal.pone.0265507 April 11, 2022 17 / 17

https://www.riziv.fgov.be/fr/themes/cout-remboursement/par-mutualite/prestations-individuelles/prix/Pages/default.aspx

https://doi.org/10.1186/s12955-018-1016-6

http://www.ncbi.nlm.nih.gov/pubmed/30219061

http://www.hpo.ie/abf/ABF2020AdmittedPatientPriceList.pdf

https://www.bfs.admin.ch/asset/en/1543-1800

http://www.spezialittenliste.ch/

https://doi.org/10.3399/bjgp20X713897

http://www.ncbi.nlm.nih.gov/pubmed/33257463

https://doi.org/10.1007/s10198-020-01166-z

http://www.ncbi.nlm.nih.gov/pubmed/32108274

https://doi.org/10.1007/s40273-018-0650-5

http://www.ncbi.nlm.nih.gov/pubmed/29679317

https://doi.org/10.1002/%28sici%291099-1050%28199906%298%3A4%26lt%3B323%3A%3Aaid-hec431%26gt%3B3.0.co%3B2-0

http://www.ncbi.nlm.nih.gov/pubmed/10398525

https://CRAN.R-project.org/package=jomo

https://doi.org/10.1159/000006908

http://www.ncbi.nlm.nih.gov/pubmed/10516413

https://doi.org/10.1016/j.lanepe.2021.100165

http://www.ncbi.nlm.nih.gov/pubmed/34557854

https://doi.org/10.1371/journal.pone.0265507

image15.emf

336 www.eymj.org

Effect of Pharmacist-Led Intervention in Elderly Patients through a Comprehensive Medication Reconciliation: A Randomized Clinical Trial

Sunmin Lee1,2,3, Yun Mi Yu1,2, Euna Han1,2, Min Soo Park1,4,5, Jung-Hwan Lee6, and Min Jung Chang1,2

1Department of Pharmaceutical Medicines and Regulatory Science, Colleges of Medicine and Pharmacy, Yonsei University, Incheon; 2Department of Pharmacy and Yonsei Institute of Pharmaceutical Sciences, College of Pharmacy, Yonsei University, Incheon; 3Department of Pharmacy, Inha University Hospital, Incheon; 4Department of Clinical Pharmacology, Severance Hospital, Yonsei University College of Medicine, Seoul; 5Department of Pediatrics, Yonsei University College of Medicine, Seoul; 6Department of Hospital Medicine, Inha University Hospital, Inha University School of Medicine, Incheon, Korea.

Purpose: Polypharmacy can cause drug-related problems, such as potentially inappropriate medication (PIM) use and med- ication regimen complexity in the elderly. This study aimed to investigate the feasibility and effectiveness of a collaborative medication review and comprehensive medication reconciliation intervention by a pharmacist and hospitalist for older pa- tients. Materials and Methods: This comprehensive medication reconciliation study was designed as a prospective, open-label, ran- domized clinical trial with patients aged 65 years or older from July to December 2020. Comprehensive medication reconcilia- tion comprised medication reviews based on the PIM criteria. The discharge of medication was simplified to reduce regimen complexity. The primary outcome was the difference in adverse drug events (ADEs) throughout hospitalization and 30 days after discharge. Changes in regimen complexity were evaluated using the Korean version of the medication regimen complex- ity index (MRCI-K). Results: Of the 32 patients, 34.4% (n=11/32) reported ADEs before discharge, and 19.2% (n=5/26) ADEs were reported at the 30- day phone call. No ADEs were reported in the intervention group, whereas five events were reported in the control group (p=0.039) on the 30-day phone call. The mean acceptance rate of medication reconciliation was 83%. The mean decreases of MRCI-K be- tween at the admission and the discharge were 6.2 vs. 2.4, although it was not significant (p=0.159). Conclusion: As a result, we identified the effect of pharmacist-led interventions using comprehensive medication reconciliation, including the criteria of the PIMs and the MRCI-K, and the differences in ADEs between the intervention and control groups at the 30-day follow-up after discharge in elderly patients. Trial Registration: (Clinical trial number: KCT0005994)

Key Words: Medication reconciliation, elderly, adverse drug event, potentially inappropriate medication, medication regimen complexity

Original Article

pISSN: 0513-5796 · eISSN: 1976-2437

Received: February 6, 2023 Revised: March 19, 2023 Accepted: March 28, 2023 Published online: April 20, 2023 Co-corresponding authors: Jung-Hwan Lee, MD, Department of Hospital Medicine, Inha University Hospital, 27 Inhang-ro, Jung-gu, Incheon 22332, Korea. E-mail: [email protected] and Min Jung Chang, PhD, Department of Pharmacy and Yonsei Institute of Pharmaceutical Sciences, College of Pharmacy, Yonsei University, 85 Songdogwahak-ro, Yeonsu-gu, Incheon 21983, Korea. E-mail: [email protected]

•The authors have no potential conflicts of interest to disclose.

© Copyright: Yonsei University College of Medicine 2023 This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (https://creativecommons.org/licenses/by-nc/4.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Yonsei Med J 2023 May;64(5):336-343 https://doi.org/10.3349/ymj.2022.0620

http://crossmark.crossref.org/dialog/?doi=10.3349/ymj.2022.0620&domain=pdf&date_stamp=2023-04-20

337

Sunmin Lee, et al.

https://doi.org/10.3349/ymj.2022.0620

INTRODUCTION

Polypharmacy, usually defined as the use of five or more med- ications, is common in the elderly population.1 Polypharmacy can cause potentially inappropriate medication (PIM) use in the elderly, which may lead to drug-related problems.2 A study found that 75% of discharged patients were prescribed a PIM, and the number of PIMs was thought to be associated with an increased risk of adverse drug events (ADEs) and all cause ad- verse events.3 PIMs have a significant effect on healthcare ser- vice use, especially hospitalization, among elderly patients.4

Polypharmacy can affect the complexity of the medication regimen, which may have a negative effect on clinical out- comes. A study indicated that complex regimens can lead to not only a higher likelihood of medication non-adherence but also hospitalization in older people.5 Furthermore, hospitaliza- tion itself was associated with increased medication regimen complexity in older patients during the inpatient period. Com- plex medication regimens at hospital admission were predic- tive of rehospitalizations for ADEs.6

In addition, polypharmacy results in medication discrepancy and complexity.7 Pharmacist-led activities to correct an inac- curate medication list at all transition points are called medi- cation reconciliation, and they are conducted to promote pa- tient safety in care transitions. There are divisions about the role of the pharmacist in the process, which varies depending on the hospital.8 However, pharmacists could contribute to resolving medication discrepancies after hospital discharge as well as during hospitalization, but data on effectiveness in terms of the clinical relevance of resolving discrepancies and health- care utilization is not clear.9

Pharmacist-led medication reviews have been proposed as an important part of the solution to medication-related prob- lems in terms of PIM, regimen complexity, and medication discrepancies. Studies have shown that inappropriate prescrib- ing in older patients has significantly decreased with collabora- tion between pharmacists and physicians in multidisciplinary teams onward across sectors.10 Pharmaceutical care has im- proved the appropriate use of medicines not only during a hos- pital stay but also after discharge.11 Clinical pharmacist medica- tion reviews could reduce the effect of hospitalization on the complexity of medication regimens of older patients.12 In par- ticular, participation of clinical pharmacists in hospital discharge transitions of care, including medication reconciliation, review, counseling, and post-discharge follow-up, had a positive effect on reduction in post-discharge hospital visits.13 Studies have found that medication-related interventions, such as medica- tion reconciliation, patient education, professional education, and transitional care, are more effective in preventing hospital readmission in older people.14,15

To date, randomized control studies of pharmacist interven- tion in elderly patients with polypharmacy, including medica- tion review and reconciliation, are lacking, although compre-

hensive approaches have been recommended to resolve drug- related problems among older adults with multimorbidity.16

Therefore, this study aimed to investigate the feasibility and effectiveness of a collaborative medication review and com- prehensive medication reconciliation intervention by a phar- macist and hospitalist for older patients.

MATERIALS AND METHODS

Study design The study was an open-label, randomized controlled trial conducted in a tertiary-level hospital in South Korea (Clinical trial number: KCT0005994). The study was performed in ac- cordance with the Declaration of Helsinki, and all participants provided written informed consent prior to inclusion in the study. The protocol was approved by the Institutional Review Board (IRB) of Inha Hospital (IRB# 2020-06-029).

Participants Patients aged 65 years or older admitted to the Department of Hospital Medicine at Inha University Hospital from July to De- cember 2020, who were taking at least five medications, were included. Written informed consent was obtained from pa- tients or caregivers prior to inclusion. Patients who were dis- charged within 24 h and those with a life expectancy of less than 3 months were excluded. Patients were randomly assigned to the intervention or control group using randomly generated blocks prior to patient enrollment.

Data collection Demographic information and laboratory data, including he- moglobin, sodium, potassium, albumin, alanine aminotrans- ferase, alkaline phosphatase, and creatinine clearance, were calculated using the Cockcroft-Gault equation and recorded. Clinical data, including the International Classification of Dis- eases 10th edition-Clinical Modification, length of stay, and destination after discharge were obtained from the hospital’s electronic medical record system. A face-to-face interview was conducted for all patients within 24 h of inclusion. The in- terview involved questions about the swallowing ability, pa- tient-reported adverse events, use of over-the-counter drugs, and complementary and alternative medicines. In addition, medical history, including syncope, delirium, dementia, cogni- tive impairment, gastric ulcer, constipation, and falls or frac- tures, were collected to identify PIMs that were not recom- mended for use due to drug-disease interactions, according to the BEER 2019.17 Antibiotic use and duration were recorded from medication history during hospitalization. The regimen complexity at admission was evaluated using the Korean ver- sion of the medication regimen complexity index (MRCI-K).18

The identified ADEs of the patients were recorded in the ADE reporting system at Inha University Hospital. Reports of ad-

338

Effect of a Comprehensive Medication Reconciliation

https://doi.org/10.3349/ymj.2022.0620

verse events completed during discharge were recorded. ADEs and serious adverse events (SAEs) were detected by patients, pharmacists, or physicians. Once they were reported, the phar- macist as an investigator evaluated them first, and the phy- sician as a principal investigator confirmed them. The phar- macist contacted all participants by telephone 30 days after discharge to follow-up patient-reported ADEs. The pharmacist contacted the next of kin or the caregiver of patients who were unable to communicate coherently. In addition, the degree of regimen complexity measured as MRCI-K was compared with the difference in scores between the intervention and control groups. After discharge, a follow-up telephone call to patients 30 days after discharge was conducted for reconfirmed medi- cation use and patient-reported ADEs.

Comprehensive medication reconciliation activity Comprehensive medication reconciliation is a series of clini- cal pharmacist activities, including medication reconciliation through a face-to-face interview at admission and daily medi- cation review, reducing regimen complexity and daily review of adverse events. The intervention group received a clinical pharmacy service from Monday to Friday, with a full-time phar- macist as an investigator, and the contents of the intervention were all recorded in case report forms, whereas the control group received the usual care during the inpatient period. The clinical pharmacy service provided to patients in the interven- tion group was as follows: within 24 h of admission, a compre- hensive list of current drugs was checked during a face-to-face interview with the pharmacist to identify medication discrep- ancies with self-medications and to evaluate the PIMs based on PIM criteria. Once assigned to the intervention group, omis- sions, duplication, and dosage errors were corrected through a complete understanding of all medications of patients. The analysis of PIMs was checked when initially performed at ad- mission and then checked again at the end of hospitalization.

During the hospital stay, medication reviews were conducted based on the BEER 2019,17 screening tool of older persons’ pre- scriptions (STOPP), and screening tool to alert to right treat- ment (START) 2015;19 and recommendations were given to physicians on drug selection, monitoring requirements, renal dose adjustment according to renal function (CrCl), and drug- drug interactions based on Lexicomp®, with the final decision made by the physician in charge. In addition, in order to reduce regimen complexity, the instructions were simplified during patient education about discharge medications, which result- ed in reduced dosing frequency by matching the administra- tion times of patients’ medications. All ADEs confirmed by the investigators were also reported in the ADE reporting system.

Study outcomes The primary outcome of this study was the difference in the number of ADEs reported during hospitalization and at the 30-day follow-up after discharge between the intervention and

control groups. All discharged patients were subject to drug side effect monitoring. Basically, an interview was conducted with the patient, while information on the complaints about ADEs were collected from the main caregiver in case contact with the patient was not possible. The secondary outcomes were the differences in the change of medication regimen com- plexity index, the number of PIMs, and the number of drugs be- tween the intervention and control groups. The PIMs were eval- uated based on the BEER criteria,17 STOPP, and START 2015.19 The number of PIMS was checked according to the criteria, ex- cluding overlapping drugs. The complexity of the medication regimen between admission and discharge was evaluated us- ing the validated Korean MRCI-K.18

Statistical analyses For power calculation, we estimated the ADE incidence to be 3.4% in the intervention group and 40% in the control group, based on previous randomized control studies that had a study design similar to the current study.20,21 Based on an 80% power of detection, there was a significant difference between the in- tervention and control groups at the 95% confidence limit. A targeted sample size of 20 patients was calculated with the ex- pected 10% rate of loss to follow-up (α=0.05, 1–β=0.80).

Patient characteristics were presented as median, interquar- tile range, and percentile (%). The chi-square test and Mann- Whitney U test were used to assess the differences between the intervention and control groups. Changes in the number of medications, MRCI-K, and PIMs between groups were analyzed using the Mann–Whitney U test. The Fischer exact and chi- squared tests were used to measure differences in the preva- lence of ADE reports between the two groups at discharge and 30-day follow-up phone calls. Analyses were performed using IBM SPSS Statistics for Windows (version 24.0; IBM Corp. Ar- monk, NY, USA).

RESULTS

Patient characteristics Forty patients were screened for participation in this study. Among them, eight patients were excluded due to readmis- sion to the intensive care unit (n=4), decline to enroll (n=3), and death before the intervention was finalized (n=1).

A total of 32 patients completed the study process during admission, and 26 patients completed the follow-up 30 days after discharge. Fig. 1 shows the flow of patients throughout the study. Patient characteristics (n=32) are shown in Table 1. The number of medications and the number of PIMs were similar between the groups. In the process of reconciliation, self-medication discrepancies in primary non-prescription drugs, such as health supplements and vitamins, were observed in both groups. There was no statistically significant difference between the groups in all values of MRCI-K, Charlson Comor-

339

Sunmin Lee, et al.

https://doi.org/10.3349/ymj.2022.0620

bidity Index, and length of stay.

Effects of pharmacist intervention in the medication reconciliation process In total, 41 suggested actions were identified in the 14 interven- tion groups during the study period. The most frequent sugges- tions were changes in drug therapy (n=13), reduction in dosage (n=8), and initiation of drug therapy (n=6). Suggested actions were implemented in 83% of the cases corresponding to 2.4 per person. A summary of the recommendations and acceptance is described in Table 2.

The changes from admission to discharge in the number of medications, MRCI-K, and PIMs are shown in Table 3. A total of 31 patient prescriptions were analyzed, excluding one pa- tient with incomplete prescription information. The interven- tion group had higher score changes at discharge than at ad-

Fig. 1. Patient flow diagram. ICU, intensive care unit.

Assessed for eligibility (n=40)

Declined (n=3)

Particepated to intervention (n=14)

Particepated to control (n=18)

Completed the 30 days follow-up (n=13)

Allocated to intervention (n=18) - Readmission to ICU (n=3) - Died before discharge (n=1)

Allocated to control (n=19) - Readmission to ICU (n=1)

Table 1. Baseline Characteristics (n=32)

Intervention group (n=14)

Control group (n=18)

p value

Sex, female 8 (57.1) 13 (72.2) 0.555 Age, yr 83 (71–87) 84.5 (72–94) 0.235 Medication management 0.777

Assisted 8 (57.1) 11 (61.1) Tube-feeding 5 (35.7) 7 (38.9) Self-administered 1 (7.2) 0 (0.0)

Number of medications 8.5 (4–14) 8.5 (3–15) 0.866 Self-medication discrepancy 0.457

Nonprescription drugs 2 (14.2) 5 (27.8) Prescription drugs 1 (7.1) 0�

Main diagnosis at admission 0.589 Pneumonia 3 (21.4) 7 (38.8) Sepsis 2 (14.2) 2 (11.1) Diabetes mellitus with hypoglycemia 0� 2 (11.1) Acute cholecystitis 1 (7.2) 1 (5.5) Acute kidney injury 1 (7.2) 1 (5.5) Others 7 (50) 4 (28)

PIMs criteria medication 1.5 (0–6) 3 (0–5) 0.253 Medical condition related to PIM criteria 0.723

Dementia 7 (50) 9 (50) Delirium 5 (35.7) 7 (38.9) Falls history 6 (42.9) 10 (55.5) Diabetes mellitus 9 (64.2) 11 (61.1)

CCI 5 (3–6) 4.5 (3–7) 0.301 Antibiotic use 12 (85.7) 17 (94.4) 0.199 Antibiotic duration 8 (3–41) 8 (3–39) Length of stay 7 (3–51) 8 (2–39) 0.837 Destination after discharge 0.446

Home 9 (64.2) 8 (44.4) Nursing home 3 (21.4) 5 (27.8) Transferred to another hospital 2 (14.4) 5 (27.8)

PIM, potentially inappropriate medication; CCI, Charlson Comorbidity Index. Data are presented as n (%) or median (Q1–Q3). Data were analyzed using the Fisher exact test, chi-square test, and Mann-Whitney U test.

Table 2. Number of Recommendations and Accepted Recommendations (n=14) Recommendation Number identified Number accepted Reference Acceptance rate (%)

Dosage adjustment Dosage low 1 1 Lexicomp® 100 Dosage too high 8 8 100

Need for additional therapy 6 2 PIMs criteria (‡START) 33.3 Change drug therapy 13 11 PIMs criteria (*BEER and †STOPP) 84.6 Drug-drug interactions 2 1 Lexicomp® 50 Drug duplication 1 1 Lexicomp® 100 Self-medication discrepancy 1 1 NA 100 Medication regimen simplification

Dose time 5 5 NA 100 Instruction modification 4 4 100

Total 41 34 NA 83 Number per person (/person) 2.9 2.4 NA PIM, potentially inappropriate medication. *BEER Criteria 2019; †STOPP 2015: Screening Tool of Older People’s Potentially Inappropriate Prescriptions; ‡START 2015 (Screening Tool to Alert to Right Treatment).

340

Effect of a Comprehensive Medication Reconciliation

https://doi.org/10.3349/ymj.2022.0620

mission, whereas the control group had lower score changes at discharge than at admission in all indices. There is a trend of de- creasing in MRCI-K after intervention between two groups (6.2 vs. 2.4), although it was not statistically significant (p=0.159).

ADEs in the medication reconciliation process A comparison of the ADEs and SAEs is shown in Table 4. ADEs confirmed by physicians and pharmacists were identified dur- ing the study period. Of the 32 patients, 34.4% (n=11) reported drug-related adverse events before the end of the discharge period. More ADEs were reported in the control group (44.4%, n=8) than in the intervention group (21.4%, n=3) during the overall study period, but the difference was not statistically sig- nificant (p=0.266). Among them, three patients in the control group had SAEs. No SAEs were reported in the intervention group compared with the control group with SAEs due to hypo- glycemia and drug-induced hepatitis.

Of the 26 follow-up patients, 19.2% (n=5) of ADEs were re- ported at the 30-day phone call. Among them, no ADEs were reported in the intervention group, whereas five events were reported in the control group (p=0.039), with a significant sta- tistical difference in the number of adverse events on the 30- day phone call. Of the five patients who reported adverse events at the 30-day call, three patients in the control group were the same patients who had adverse events during hospitalization. Patient 4 reported adverse events from the same drug, glimepiri- de (Table 5).

DISCUSSION

Studies have demonstrated that interventions by clinical phar- macists can improve drug-related problems and affect positive

clinical outcomes in both inpatient and outpatient care facili- ties.22,23 Pharmacy-led interventions via medication reconcili- ation are essential for reducing the occurrence of medication discrepancies that may lead to ADEs in the care transition pro- cesses.24 In particular, it is necessary to target interventions for high-risk patient populations, such as the elderly.25 Pharmacy- led medication reconciliation interventions have a greater im- pact when conducted at either admission or discharge.26 How- ever, studies have found that not only medication discrepancies, but also comprehensive approaches, such as structured medi- cation review and multidisciplinary cooperation, are required to resolve drug-related problems.27,28 To more clearly investi- gate the outcomes of pharmacist interventions, a comprehen- sive reconciliation process is required to evaluate the out- comes based on drug-related measurements, such as regimen complexity, inappropriate prescribing, and adherence. We con- ducted a randomized clinical study to investigate the effective- ness of pharmacist interventions for old age using tools, such as MRCI-K and PIMs, during hospital stays and at discharge.

The intervention group had fewer ADEs and no SAEs com- pared to the control group, although the differences between the groups were not significant. Furthermore, there was a sig- nificant difference in ADEs reported by patients between the intervention and control groups at the 30-day follow-up phone calls. This result was consistent with previous findings, which showed that intervention comprising medication reconcilia- tion by a pharmacist reduced the rate of preventable ADEs 30 days post-discharge.29 Among the patients who had ADEs in the control group, three, as reported before discharge, repeat- edly reported ADEs after 30 days. In addition, two patients in the control group newly reported ADEs after discharge. Patient 4, whose main diagnosis was pneumonia at admission in the control group, complained of adverse drug reactions due to the same drug, glimepiride. Our results could explain why clinical pharmacy services could lead to the reduction of ADEs after discharge, although there was no further clinical pharmacy ser- vice. This means that clinical pharmacy services during hospi- talization could affect the safety of the drug even after discharge.

In addition to medication review, risk assessment criteria tar- geting the elderly were introduced in the intervention group. To evaluate whether a pharmacist-led medication review is effec- tive during medication reconciliation, the number of PIMs us- ing the BEER 2019, STOPP 2015, and START 2015 were identi- fied. Both medication numbers and PIMs in the intervention

Table 4. Comparisons of ADEs Reporting between the Medication Rec- onciliation Group and Control Group

Intervention group

Control group

p value

ADE reported during hospitalization* 3 (21.4) 8 (44.4) 0.266 SAE reported during the study period* 0 3 (16.7) 0.529 ADE reported at 30-day phone call† 0 5 (38.5) 0.039 ADE, adverse drug event; SAE, serious adverse event. Data were analyzed using the Fisher’s exact test. *Intervention group (n=14), control group (n=18); †Intervention group (n=13), control group (n=13).

Table 3. Scores at Admission and Discharge and Change from Admission

Intervention group (n=14) Control group (n=18) p value

Admission Discharge Change Admission Discharge Change Number of medications 9 (4–14) 8 (3–11) -1 (-5–3) 9 (3–15) 9 (4–12) 0 (-9–5) 0.566 MRCI-K 29.5 (16–60) 29.2 (7–45) -8 (-21–17) 30 (14–58) 31 (14–56) 2 (-30–19) 0.159 PIM 1.5 (0–6) 0 (0–3) -0.5 (-6–0) 3 (0–5) 1 (0–4) -1 (-4–1) 0.968 MRCI-K, Korean version of medication regimen complexity index; PIM, potentially inappropriate medication. Data are presented as median (Q1–Q3). Data were analyzed using the Mann-Whitney U test.

341

Sunmin Lee, et al.

https://doi.org/10.3349/ymj.2022.0620

group were reduced compared to the control group, although no significant difference was found when comparing the change in drug-related problems between the groups. Although there was no difference in diabetic patients between the two groups in our study, there were reports of adverse events from glimepiri- de, which PIMs to avoid (n=4, 36.3%), and the side effects of this drug were repeated in the control group. Our finding that the same patient in the control group reported side effects of glimepiride explained the need for pharmacist intervention in medication review. PIM criteria as decision support tools were effective in reducing polypharmacy in older adults. However, the effect of the intervention on clinical outcomes is unclear.30 Our results have demonstrated the effectiveness of pharma- cist intervention by evaluating the patient-reported adverse effects related to PIMs, and further studies are needed for evi- dence-based practice.

We conducted drug instruction modifications, such as ad- ministration time and food-related instructions, to simplify prescription complexity. The overall change in the score of regi- men complexity expressed using MRCI-K was reduced after sim- plification of regimen complexity, although there was no signif- icant difference between both groups in the score change. MRCI has been used as a tool for evaluating pharmacist service activities at the time of hospitalization, and discharge and in- tervention studies are being conducted to establish guidelines and confirm effectiveness.12,31,32 Modifying usage, such as ad- ministering drugs at the same time, was a priority to consolidate prescription regimens most efficiently, as this was how multiple doses were dispensed in the Korean pharmacy practice. Our study found that dosage simplification is a major means of re- ducing the administration time in hospital pharmacy practice. When conducting routine in-hospital medication reviews, the simplification strategy for regimen complexity can help mini- mize the impact of hospitalization on the complexity of dis- charge medication regimens.

Our study was conducted to evaluate comprehensive med- ication reconciliation and clinical outcomes in real clinical practice in the context of standardized clinical pharmacy ser-

vice, including the MRCI-K, BEER 2019, STOPP, and START. Polypharmacy is associated with negative drug-related prob- lems, such as increased medication regimen complexity and in- appropriate drug use, which increase the care transition process and require intervention.6,25 Further research is needed to evaluate drug-related problems using objective criteria at vari- ous stages and to standardize the collaborative clinical phar- macy to confirm clinical outcomes.33,34 Thus, there is a need for a comprehensive pharmacist intervention approach to solve these problems, and an evaluation process must be developed.

There were some limitations to the present study. First, this was a preliminary descriptive study with limitations in deriv- ing decisive results. We identified the feasibility of using the MRCI-K and PIM criteria as medication reconciliation tools in a randomized study. We expect that differences in the distri- bution of scores and adverse events between the two groups could be used as a basis for future research. However, with a small sample size, further work needs to be done to establish the effectiveness of intervention. In addition, this study was conducted at a single center and lacked information on disease severity at the 30-day follow-up. We expect these findings to be implicated in a pilot randomized control study for further in- vestigation based on pharmacy-led intervention in multicenter research. Furthermore, the patients reported that adverse events at the 30-day phone call could be subjective which could be affected by confounders, whereas adverse events during hospitalization were confirmed by the pharmacist and the phy- sician. However, all ADEs, including patient reporting, have been monitored and recorded in the ADE reporting system of Inha Hospital.

In the comprehensive medication reconciliation process conducted as a randomized study, we identified the feasibility of pharmacist-led interventions using medication review, in- cluding the PIM criteria and MRCI-K. As a result of pharmacist interventions, we found that the number of ADEs before dis- charge was lower in the intervention group than in the control group, and that there were differences in ADEs at the 30-day follow-up after discharge.

Table 5. Detailed Description of Adverse Events Reported

Patient number Group Related medication ADEs SAE ADEs at 30-day phone call 6 Intervention Indapamide Hypotension No No 5 Intervention Glimepiride Hypoglycemia No No 27 Intervention Tamsulosin Hypotension No No 37 Control Glimepiride Hypoglycemia Yes No 4 Control Glimepiride Hypoglycemia Yes Yes 7 Control Glimepiride Hypoglycemia No No 9 Control Ampicillin+sulbactam Thrombocytopenia No No 12 Control Piperacillin tazobactam Drug fever No No 15 Control Ciprofloxacin Diarrhea No Yes 36 Control Dexibuprofen Drug induced hepatitis Yes No 18 Control Cefotaxime Thrombocytopenia No Yes

ADE, adverse drug event; SAE, serious adverse event.

342

Effect of a Comprehensive Medication Reconciliation

https://doi.org/10.3349/ymj.2022.0620

ACKNOWLEDGEMENTS

This research was funded by the National Research Founda- tion of Korea (#NRF- 2021R1F1A106009911), Daewoong Pharmaceutical Company (DFY2115P) and the Yonsei Uni- versity Research Fund of 2019-22-0082.

AUTHOR CONTRIBUTIONS

Conceptualization: Sunmin Lee, Jung-Hwan Lee, and Min Jung Chang. Data curation: Sunmin Lee and Jung-Hwan Lee. Formal anal- ysis: Sunmin Lee. Funding acquisition: Jung-Hwan Lee and Min Jung Chang. Investigation: Sunmin Lee, Jung-Hwan Lee, and Min Jung Chang. Methodology: all authors. Project administration: Jung-Hwan Lee and Min Jung Chang. Resources: Sunmin Lee and Jung-Hwan Lee. Software: Sunmin Lee. Supervision: Jung-Hwan Lee and Min Jung Chang. Validation: all authors. Visualization: Sunmin Lee. Writ- ing—original draft: Sunmin Lee. Writing—review & editing: Jung- Hwan Lee and Min Jung Chang. Approval of final manuscript: all au- thors.

ORCID iDs

Sunmin Lee https://orcid.org/0000-0002-0528-5098 Yun Mi Yu https://orcid.org/0000-0002-8267-9453 Euna Han https://orcid.org/0000-0003-2656-7059 Min Soo Park https://orcid.org/0000-0002-4395-9938 Jung-Hwan Lee https://orcid.org/0000-0001-7567-0664 Min Jung Chang https://orcid.org/0000-0002-8408-5907

REFERENCES

1. Masnoon N, Shakib S, Kalisch-Ellett L, Caughey GE. What is poly- pharmacy? A systematic review of definitions. BMC Geriatr 2017; 17:230.

2. Jetha S. Polypharmacy, the elderly, and deprescribing. Consult Pharm 2015;30:527-32.

3. Weir DL, Lee TC, McDonald EG, Motulsky A, Abrahamowicz M, Morgan S, et al. Both new and chronic potentially inappropriate medications continued at hospital discharge are associated with increased risk of adverse events. J Am Geriatr Soc 2020;68:1184-92.

4. Hyttinen V, Jyrkkä J, Valtonen H. A systematic review of the impact of potentially inappropriate medication on health care utilization and costs among older adults. Med Care 2016;54:950-64.

5. Wimmer BC, Cross AJ, Jokanovic N, Wiese MD, George J, Johnell K, et al. Clinical outcomes associated with medication regimen complexity in older people: a systematic review. J Am Geriatr Soc 2017;65:747-53.

6. Willson MN, Greer CL, Weeks DL. Medication regimen complex- ity and hospital readmission for an adverse drug event. Ann Phar- macother 2014;48:26-32.

7. Patel CH, Zimmerman KM, Fonda JR, Linsky A. Medication com- plexity, medication number, and their relationships to medica- tion discrepancies. Ann Pharmacother 2016;50:534-40.

8. Fernandes BD, Almeida PHRF, Foppa AA, Sousa CT, Ayres LR, Chemello C. Pharmacist-led medication reconciliation at patient discharge: a scoping review. Res Social Adm Pharm 2020;16:605-13.

9. McNab D, Bowie P, Ross A, MacWalter G, Ryan M, Morrison J. Systematic review and meta-analysis of the effectiveness of phar- macist-led medication reconciliation in the community after hos-

pital discharge. BMJ Qual Saf 2018;27:308-20. 10. Gillespie U, Alassaad A, Hammarlund-Udenaes M, Mörlin C,

Henrohn D, Bertilsson M, et al. Effects of pharmacists’ interven- tions on appropriateness of prescribing and evaluation of the in- struments’ (MAI, STOPP and STARTs’) ability to predict hospital- ization--analyses from a randomized controlled trial. PLoS One 2013;8:e62401.

11. Spinewine A, Swine C, Dhillon S, Lambert P, Nachega JB, Wilmo- tte L, et al. Effect of a collaborative approach on the quality of pre- scribing for geriatric inpatients: a randomized, controlled trial. J Am Geriatr Soc 2007;55:658-65.

12. Elliott RA, O’Callaghan C, Paul E, George J. Impact of an interven- tion to reduce medication regimen complexity for older hospital inpatients. Int J Clin Pharm 2013;35:217-24.

13. Van der Linden L, Hias J, Walgraeve K, Flamaing J, Tournoy J, Spri- et I. Clinical pharmacy services in older inpatients: an evidence- based review. Drugs Aging 2020;37:161-74.

14. Dautzenberg L, Bretagne L, Koek HL, Tsokani S, Zevgiti S, Rodon- di N, et al. Medication review interventions to reduce hospital re- admissions in older people. J Am Geriatr Soc 2021;69:1646-58.

15. Daliri S, Boujarfi S, El Mokaddam A, Scholte Op Reimer WJM, Ter Riet G, den Haan C, et al. Medication-related interventions deliv- ered both in hospital and following discharge: a systematic review and meta-analysis. BMJ Qual Saf 2021;30:146-56.

16. Tang JY, Lun P, Teng PHJ, Ang W, Tan KT, Ding YY. Intervention elements and behavior change techniques to improve prescrib- ing for older adults with multimorbidity in Singapore: a modified Delphi study. Eur Geriatr Med 2022;13:531-9.

17. 2019 American Geriatrics Society Beers Criteria® Update Expert Panel. American Geriatrics Society 2019 updated AGS Beers Cri- teria® for potentially inappropriate medication use in older ddults. J Am Geriatr Soc 2019;67:674-94.

18. Lee S, Jang J, Yang S, Hahn J, Min KL, Jung EH, et al. Development and validation of the Korean version of the medication regimen complexity index. PLoS One 2019;14:e0216805.

19. O’Mahony D, O’Sullivan D, Byrne S, O’Connor MN, Ryan C, Gal- lagher P. STOPP/START criteria for potentially inappropriate pre- scribing in older people: version 2. Age Ageing 2015;44:213-8.

20. Hanlon JT, Weinberger M, Samsa GP, Schmader KE, Uttech KM, Lewis IK, et al. A randomized, controlled trial of a clinical phar- macist intervention to improve inappropriate prescribing in el- derly outpatients with polypharmacy. Am J Med 1996;100:428-37.

21. Nachtigall A, Heppner HJ, Thürmann PA. Influence of pharmacist intervention on drug safety of geriatric inpatients: a prospective, controlled trial. Ther Adv Drug Saf 2019;10:2042098619843365.

22. Gillespie U, Alassaad A, Henrohn D, Garmo H, Hammarlund- Udenaes M, Toss H, et al. A comprehensive pharmacist interven- tion to reduce morbidity in patients 80 years or older: a random- ized controlled trial. Arch Intern Med 2009;169:894-900.

23. Gustafsson M, Sjölander M, Pfister B, Jonsson J, Schneede J, Lövheim H. Pharmacist participation in hospital ward teams and hospital readmission rates among people with dementia: a ran- domized controlled trial. Eur J Clin Pharmacol 2017;73:827-35.

24. Schnipper JL, Kirwin JL, Cotugno MC, Wahlstrom SA, Brown BA, Tarvin E, et al. Role of pharmacist counseling in preventing ad- verse drug events after hospitalization. Arch Intern Med 2006;166: 565-71.

25. Mueller SK, Sponsler KC, Kripalani S, Schnipper JL. Hospital- based medication reconciliation practices: a systematic review. Arch Intern Med 2012;172:1057-69.

26. Mekonnen AB, McLachlan AJ, Brien JA. Pharmacy-led medica- tion reconciliation programmes at hospital transitions: a system- atic review and meta-analysis. J Clin Pharm Ther 2016;41:128-44.

343

Sunmin Lee, et al.

https://doi.org/10.3349/ymj.2022.0620

27. Lenander C, Elfsson B, Danielsson B, Midlöv P, Hasselström J. Ef- fects of a pharmacist-led structured medication review in primary care on drug-related problems and hospital admission rates: a ran- domized controlled trial. Scand J Prim Health Care 2014;32:180-6.

28. Phatak A, Prusi R, Ward B, Hansen LO, Williams MV, Vetter E, et al. Impact of pharmacist involvement in the transitional care of high-risk patients through medication reconciliation, medication education, and postdischarge call-backs (IPITCH study). J Hosp Med 2016;11:39-44.

29. Al-Hashar A, Al-Zakwani I, Eriksson T, Sarakbi A, Al-Zadjali B, Al Mubaihsi S, et al. Impact of medication reconciliation and review and counselling, on adverse drug events and healthcare resource use. Int J Clin Pharm 2018;40:1154-64.

30. Earl TR, Katapodis ND, Schneiderman SR, Shoemaker-Hunt SJ. Using deprescribing practices and the screening tool of older per- sons' potentially inappropriate prescriptions criteria to reduce harm and preventable adverse drug events in older adults. J Pa- tient Saf 2020;16(3S Suppl 1):S23-35.

31. Morello CM, Rotunno T, Khoan J, Hirsch JD. Improved glycemic control with minimal change in medication regimen complexity in a pharmacist-endocrinologist diabetes intense medical manage- ment (DIMM) “tune up” clinic. Ann Pharmacother 2018;52:1091-7.

32. Sluggett JK, Hopkins RE, Chen EY, Ilomäki J, Corlis M, Van Emden J, et al. Impact of medication regimen simplification on medica- tion administration times and health outcomes in residential aged care: 12 month follow up of the SIMPLER randomized controlled trial. J Clin Med 2020;9:1053.

33. Johansen JS, Havnes K, Halvorsen KH, Haustreis S, Skaue LW, Ka- mycheva E, et al. Interdisciplinary collaboration across secondary and primary care to improve medication safety in the elderly (IM- MENSE study): study protocol for a randomised controlled trial. BMJ Open 2018;8:e020106.

34. Leguelinel-Blache G, Castelli C, Roux-Marson C, Bouvet S, Andrieu S, Cestac P, et al. Impact of collaborative pharmaceutical care on in- patients’ medication safety: study protocol for a stepped wedge cluster randomized trial (MEDREV study). Trials 2018;19:19.

image1.emf

Vol.:(0123456789)1 3

European Journal of Clinical Pharmacology (2023) 79:1391–1400 https://doi.org/10.1007/s00228-023-03551-y

RESEARCH

Implementation of clinical medication review in a geriatric ward to reduce potentially inappropriate prescriptions among older adults

Alexandre Meurant1,2 · Pascale Lescure1 · Claire Lafont1 · Wilhelm Pommier1 · Claire Delmas1 · Pablo Descatoire1,3 · Marie Baudon2 · Alexandra Muzard2 · Cédric Villain1,3 · Jean‑Pierre Jourdan4,5

Received: 23 February 2023 / Accepted: 9 August 2023 / Published online: 19 August 2023 © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2023

Abstract Purpose The impact of several pharmaceutical interventions to reduce the use of potentially inappropriate medications (PIMs) and potentially omitted medications (POMs) has been recently studied. We aimed to determine whether clinical medication review (CMR) (i.e. a systematic and patient-centred clinical assessment of all medicines currently taken by a patient) performed by a geriatrician and a pharmacist added to standard pharmaceutical care (SPC) (i.e. medication rec- onciliation and regular prescription review by the pharmacist) resulted in more appropriate prescribing compared to SPC among older inpatients. Methods A retrospective observational single-centre study was conducted in a French geriatric ward. Six criteria for appropri- ate prescribing were chosen: the number of PIMs and POMs as defined by the STOPP/STARTv2 list, the total number of drugs prescribed, the number of administrations per day and the number of psychotropic and anticholinergic drugs. These criteria were compared between CMR and SPC group using linear and logistic regression models weighted on propensity scores. Results There were 137 patients included, 66 in the CMR group and 71 in the SPC group. The mean age was 87 years, the sex ratio was 0.65, the mean number of drugs prescribed was 9, the mean MMSE was 21 and at admission 242 POMs, and 363 PIMs were prescribed. Clinical medication review did not reduce the number of PIMs at discharge compared to SPC (beta = − 0.13 [− 0.84; 0.57], p = 0.71) nor did it reduce the number of drugs prescribed (p = 0.10), the number of psy- chotropic drugs (p = 0.17) or the anticholinergic load (p = 0.87). Clinical medication review resulted in more POMs being prescribed than in standard pharmaceutical care (beta = − 0.39 [− 0.72; − 0.06], p = 0.02). Cardiology POMs were more implemented in the medication review group (p = 0.03). Conclusion Clinical medication review did not reduce the number of PIMs but helped clinicians introduce underused drugs, especially cardiovascular drugs, which are known to be associated with morbidity and mortality risk reduction.

Keywords Clinical pharmacy · Prescriptions · Medication reconciliation · Inappropriate prescribing · Pharmacy service · Hospital

Introduction

Adverse drug reactions (ADRs) are more frequent, more dif- ficult to diagnose and more severe in older adults. ADRs are responsible for 20% of emergency hospitalizations among patients over the age of 75 and 25% among those aged over 85 years [1]. However, it is estimated that ADRs could be avoided in 30 to 60% of cases [2]. With this aim, the lists of inappropriate prescriptions have been developed, such as the STOPP-STARTv2 in Europe and the Beers criteria in the USA [3, 4], which include potentially inappropriate medica- tions (PIMs) and, for the STOPP-STARTv2 list, potentially omitted medications (POMs). These lists help clinicians to prescribe more appropriate drugs. Multimorbidity and

* Alexandre Meurant [email protected]

1 Department of Geriatrics, University Hospital of Caen Normandie, Caen, France

2 Department of Pharmacy, University Hospital of Caen Normandie, Caen, France

3 Normandie University, Unicaen INSERM U1075, COMETE, Caen, France

4 Department of Pharmacy, Vire Hospital, Vire, France 5 Normandie University, UNICAEN, CERMN (Centre

d’Etudes et de Recherche sur le Médicament de Normandie), F-14032 Caen, France

http://crossmark.crossref.org/dialog/?doi=10.1007/s00228-023-03551-y&domain=pdf

1392 European Journal of Clinical Pharmacology (2023) 79:1391–1400

1 3

polypharmacy burden are high among older inpatients and are associated with an increased risk of inappropriate pre- scribing [5, 6]. Inappropriate prescribing varies from 25 to 50% for older inpatients and has been found to be associated with drug-related hospitalization as well as ADRs, functional decline, falls and an increase in health care costs [7–9].

Clinical pharmacy interventions are a way to prevent ADRs in older patients. Clinical pharmacists are frequently involved in geriatric wards to secure drug management. They can provide medication reconciliations which is the process of comparing a patient’s medication orders to all of the medications that the patient has been taking. This decrease drug discrepancies and might reduce the numbers of medical visits, drug-related hospitalizations and hospital readmissions [10, 11]. They can also provide prescription reviews to detect drug interactions, ADRs or dose issues. Medication reconciliation at admission and discharge and prescription review during hospitalization are defined here as standard pharmaceutical care (SPC). However, these activities are only aimed to secure drug treatment; SPC are not a formalized strategy to reassess and optimize all of the patient drug prescriptions to reduce inappropriate prescrip- tions. To identify, resolve and prevent ADRs and to optimize drug efficacy, a multifaceted approach involving a clinical pharmacist and a geriatrician has been described; this is so- called clinical medication review (CMR) [1, 7]. The latter intervention consists of a systematic and patient-centred clinical assessment of all medicines currently taken by a patient [1]. Clinical medication review has been recently found to be associated with a reduction in drug-related hospitalization, all-cause hospitalizations and emergency department visits [12, 13]. CMR has already been associ- ated with improved prescribing appropriateness compared to a controlled group with no pharmacist involved [14, 15]. In a recent Cochrane review of interventions to reduce inap- propriate prescriptions, no evidence was found regarding whether pharmacist interventions to improve appropriate polypharmacy led to a clinically significant improvement in patients’ health. However, two biased studies suggested that they may be beneficial in terms of reducing POMs [16]. In a recent large randomized controlled trial that aimed to show the benefit of CMR performed by a multidisciplinary team, the authors did not find any clinical improvement [17]. However, they did not compare the reduction in PIMs or POMs in the control and intervention groups.

To the best of our knowledge, no study has compared CMR to SPC. Furthermore, no study has investigated the impact of CMR on other indicators of polypharmacy burden, such as the reduction in anticholinergic load, number of antipsychotic drugs and number of drug administrations. We thus aimed to assess the impact of CMR on 6 appropriate prescribing criteria when compared with SPC in an acute geriatric ward.

Methods

Study design and participants

A single-centre, retrospective cohort study comparing before and after the implementation of CMR was carried out in an acute geriatric ward in a university hospital. The local research ethics committee approved this study. Due to the retrospec- tive design of this study, consent was not necessary to include patients, in accordance with the current Jardé’s law.

A clinical pharmacist team performed clinical pharmacy interventions such as medication reconciliation at admission and at discharge and prescription review. To fully reassess patients’ prescriptions, CMR was implemented in April 2020 and involved at least one geriatrician and one clinical phar- macist for each CMR [12]. Two groups were then defined. The patients in the SPC group received a SPC including medication reconciliations at admission and at discharge and a prescription review. Patients in the CMR group received CMR in addition to SPC. The medication review process added a systematic check of overprescribing (absence of a valid indication, drug duplications), underprescribing and misprescribing (unfavourable benefit/risk ratio, questionable efficacy, unsuitable duration and/or dose, risk of exacerba- tion of certain chronic clinical conditions and drug–drug interactions) [12, 18]. Geriatrician and clinical pharmacist then discuss during about 20 min the therapeutic modifica- tion to implement. Patients included in the SPC group were hospitalized in an acute geriatric ward from 1 January 2020 to 4 April 2020. Patients included in the CMR group were hospitalized from 5 April 2020 to 1 March 2021. Patients were excluded if they had been transferred to another depart- ment, if they had fewer than 5 drugs prescribed before hos- pitalization and if no medication reconciliations had been performed at admission and/or at discharge because they did not benefit of the SPC. Patients could not be included in both groups. The most complex patients were selected to benefit from CMR.

Data collection

All variables were collected using health medical records. Admission drugs were collected using medication recon- ciliation at admission. Discharge drugs were collected using medication reconciliation at discharge. A clinical pharma- cist collected all information. Comorbidity burden was evaluated using the Charlson comorbidity index adjusted for age. Functional status and autonomy were evaluated using the activities of daily living from Katz (ADL) and the instrumental activities of daily living from Lawton (IADL) scales. The mini–mental state examination (MMSE) was used to evaluate the patients’ cognition. PIMs and POMs

1393European Journal of Clinical Pharmacology (2023) 79:1391–1400

1 3

were identified according to the STOPP-STARTv2 list by an experienced clinical pharmacist and based on the list of diseases identified by physicians for each participant.

Outcomes

The primary outcome was the variation in the number of PIMs prescribed according to the STOPPv2 list between the admis- sion prescription and the discharge prescription [19]. The sec- ondary outcomes were the variation between the admission prescription and the discharge prescription in: (i) the number of POMs according to the STARTv2 list, (ii) the total number of drugs, (iii) the number of drug administrations per day, (iv) the number of psychotropic drugs and (v) the anticholinergic load of the prescription using the ADS. The STOPP/STARTv2 criteria were used, except for the START I1 criterion “An annual influenza vaccination, in early autumn - [prevention of influenza-related hospitalizations and mortality]”, which was not assessed due to the seasonality of the vaccine. The psy- chotropic drugs included antidepressant drugs, antipsychotic drugs, benzodiazepines and Z drugs. The ADS was chosen due to its large screening of drugs [20].

Statistical analysis

Quantitative variables are presented as mean ± standard deviation in the case of a normal distribution or medians [1st to 3rd quartiles], and categorical variables are presented as sample sizes and percentages. The characteristics of the SPC group and the CMR group were compared using Stu- dent’s t-tests for quantitative variables (or Mann–Whitney U tests if the conditions required for the t-test were not met) and chi-square tests for categorical variables. To limit the confounding bias related to the observational nature of this study, a propensity score analysis using the inverse probabil- ity of treatment weighting (IPTW) method was performed. In estimating the propensity score, variables associated with endpoints and variables representing potential confounders between the variable of interest (i.e. CMR) and the outcome were determined. The variables used for the computation of the propensity score are available in the Supplementary Materials. Propensity scores were determined by logis- tic regression, and their distribution between the 2 groups was graphically evaluated. To create a weighted database or “pseudopopulation” in which the confounding factors were distributed equally between the two groups, the calcu- lation of stabilized weights was performed using the aver- age treatment effect (ATE) method. Finally, the balance of the weighted variables between each group was checked by comparing the standardized difference in means between these variables (< 10%). In case of poor balance of one of the variables by the propensity score, the final model was adjusted for this variable. The final models assessing the

association between CMR and the outcome measures were logistic regression or linear regression models (according to the type of outcomes) weighted on the propensity score. Ninety-five per cent confidence intervals (95% CIs) were computed using bootstrapping.

Missing data were taken into account by performing mul- tiple imputations, resulting in the creation of 20 databases, whose results were pooled according to Rubin’s rules. A p-value of less than 5% was considered statistically signifi- cant. The statistical analysis was performed using R software (R Foundation, Vienna, Austria).

In order to better understand our results, we decided to describe and analyse PIMs and POMs variations using uni- variate analysis. These results are available in Table 4 and in Supplementary Materials but have to be interpreted with caution as they are subject to confounding bias.

Results

A total of 137 patients were included, 71 (51.8%) in the SPC group and 66 (48.2%) in the CMR group (Fig. 1).

The median age was 86.7 years, and 83 (60.6%) women were included. There were significantly more patients who took their drugs on their own and patients with heart failure in the CMR group (p = 0.049). The distribution of the differ- ent classes of antidepressants differed according to patient group (p = 0.004) as did the total number of antihypertensive drugs (p = 0.02) (Table 1).

At admission, patients took an average of 9 drugs and 2 PIMs; 122 patients (89.1%) were prescribed at least 1 PIM. At admission, there were more PIMs (p = 0.01; Table 2) and a higher ratio of “PIM/total number of drugs” (p = 0.02) in the CMR group than in the SPC group.

Univariate analysis of outcome measures

At discharge, a greater number of drugs (p = 0.03) and administrations (p < 0.001) were prescribed in the CMR group, but fewer POMs (STARTv2 list) (p = 0.01) were pre- scribed compared to the SPC group. Clinical medication review was associated with a significant decrease in the pre- scription of psychotropic drugs (p = 0.02), PIMs (p = 0.01) and POMs (p = 0.03) in univariate analysis, compared to SPC (Table 3). Medication review was not associated with a variation in the total number of drugs, administrations or anticholinergic load compared to SPC. This univariate analysis is not adjusted on all.

Propensity score analysis

Some variables were poorly balanced by propensity scores (mean absolute difference > 10%; see Supplementary

1394 European Journal of Clinical Pharmacology (2023) 79:1391–1400

1 3

Materials). We therefore adjusted the final models for these variables (Table 3).

Clinical medication review was not more associated with a decrease in PIMs (p = 0.71) after weighting the linear regression models on propensity scores compared to SPC. It was not associated with a decrease in the num- ber of drugs, psychotropic drugs nor anticholinergic load. However, it was associated with a significant decrease in POMs (beta = − 0.39 [− 0.72; − 0.06]) and with a signifi- cant increase of the number of drugs compared to SPC (beta = 4.23; [2.02; 6.45]).

PIM and POM descriptions

Among the 242 POMs at admission, 41% were pre- scribed at discharge (n = 100). Among the POMs found at admission were pneumococcal vaccines (n = 82), vita- min D alone (n = 22) and combined with calcium (n = 19) and angiotensin-converting enzyme inhibitors (ACEis) (n = 18) (Table 4).

Potentially omitted cardiovascular drugs decreased more in the CMR group (− 47%) than in the SPC group (− 21%) (p = 0.03), especially lipid-lowering therapy.

CMR reduced vitamin D underuse by 92%, while SPC reduced vitamin D underuse by 50% (p = 0.06). In the SPC group, 44 patients had an indication for pneumococcal vac- cination; in the CMR group, 38 patients had this indication. The reduction in these POMs was 25% and 34% in the SPC and CMR groups, respectively (p = 0.11).

Of the 363 PIMs at admission, 120 were still prescribed at discharge. The most represented PIMs at admission were drugs without an indication (n = 137), and 11 drugs without an indication were still prescribed at discharge (see Supple- mentary Materials). Neuroleptics decreased by 33% in the CMR group, while their prescription increased by 100% in the SPC group at discharge.

Discussion

We performed the first study to compare clinical medication review in addition to a standard pharmaceutical care and a standard pharmaceutical care alone to reduce potentially inappropriate prescribing at discharge. The systematic and patient-centred clinical assessment of all medicines cur- rently taken by a patient in collaboration clinical pharma- cist–geriatrician in addition to a SPC was associated with a greater decrease in POMs between hospital admission and discharge than SPCs alone. It was not associated with a greater decrease in PIMs, total number of drugs, psycho- tropic drugs or anticholinergic load. It was also associated with an increased number of drug administrations per day.

These results are consistent with those of the 2018 Cochrane Medication Review, which concluded that all different interventions published in the literature to reduce inappropriate prescribing among older people did not reduce the total number of PIMs but may reduce the number of POMs compared to the standard of care [16]. However, none

Fig. 1 Flow chart

1395European Journal of Clinical Pharmacology (2023) 79:1391–1400

1 3

Table 1 Characteristics of the study population at admission

Variable CMR SPC p Total Missing values (%)

Number of patients 66 (48.2) 71 (51.8) 137 (100.0) Age 85.9 ± 6.5 87.4 ± 5.2 0.16 86.7 ± 5.9 0 (0.0) Female 43 (65.2) 40 (56.3) 0.38 83 (60.6) 0 (0.0) Hospitalization for fall 22 (33.3) 27 (38.0) 0.69 49 (35.8) 0 (0.0) Hospitalization via emergency department 54 (81.8) 58 (81.7) 1.00 112 (81.8) 0 (0.0) Comorbidities Charlson score (/42) 6 [4–7] 5 [5–6] 0.65 5 [5–7] 0 (0.0) Hypertension 46 (69.7) 47 (66.2) 0.80 93 (67.9) 0 (0.0) Diabetes mellitus 15 (22.7) 14 (19.7) 0.82 29 (21.2) 0 (0.0) Dyslipidaemia 23 (34.8) 24 (33.8) 1.00 47 (34.3) 0 (0.0) Heart failure 8 (12.1) 2 (2.8) 0.049 10 (7.3) 0 (0.0) Ischaemic heart disease 19 (28.8) 15 (21.1) 0.40 34 (24.8) 0 (0.0) Atrial fibrillation 18 (27.3) 25 (35.2) 0.41 43 (31.4) 0 (0.0) Peripheral arteritis 8 (12.1) 5 (7.0) 0.47 13 (9.5) 0 (0.0) Cognitive impairment 18 (27.3) 23 (32.4) 0.64 41 (29.9) 0 (0.0) Mini–mental state examination (/30) 23 [14–26] 16 [13–26] 0.62 21 [13–26] 72 (52.6) Depression 22 (33.3) 14 (19.7) 0.08 36 (26.3) 0 (0.0) Parkinson’s disease 3 (4.5) 2 (2.8) 0.67 5 (3.6) 0 (0.0) Cerebrovascular accident 5 (7.6) 8 (11.3) 0.65 13 (9.5) 0 (0.0) Chronic obstructive pulmonary disease 12 (18.2) 8 (11.3) 0.37 20 (14.6) 0 (0.0) Cancer 16 (24.2) 15 (21.1) 0.82 31 (22.6) 0 (0.0) Active cancer 3 (4.5) 6 (8.5) 0.50 9 (6.6) 0 (0.0) History of fall 24 (36.4) 25 (35.2) 1.00 49 (35.8) 0 (0.0) Fracture 12 (18.2) 9 (12.7) 0.51 21 (15.3) 0 (0.0) Glomerular filtration rate using CKD EPI for-

mula (ml/min/1.73 m2) 67 [51–81] 64 [45–79] 0.20 66 [46–79] 0 (0.0)

Functional status Activities of daily living (/6) 5.0 [3.5–5.5] 4.5 [2.0–5.5] 0.50 4.5 [2.5–5.5] 49 (35.8) Instrumental activities of daily living (/8) 2 [1–4] 1 [1–4] 0.94 1 [1–4] 53 (38.7) Ability to take her or his own drugs 25 (37.9) 15 (21.1) 0.049 40 (29.2) 0 (0.0) Nursing home residents 8 (12.1) 6 (8.5) 14 (10.2) 0 (0.0)

Nutritional status Body weight (kg) 65 [57–76] 62 [53–71] 0.33 64 [54–74] 7 (5.1) Body mass index (kg/m2) 25.1 ± 5.2 24.0 ± 4.3 0.59 24.5 ± 4.8 68 (49.6)

Drugs at admission Cardiovascular drugs Antiplatelet agents 31 (47.0) 26 (36.6) 0.29 57 (41.6) 0 (0.0) Anti-vitamin K 6 (9.1) 7 (9.9) 1.00 13 (9.5) 0 (0.0) Direct oral anticoagulants 13 (19.7) 15 (21.1) 1.00 28 (20.4) 0 (0.0) Angiotensin-converting enzyme inhibitors 20 (30.3) 18 (25.4) 0.65 38 (27.7) 0 (0.0) Angiotensin receptor blocker 14 (21.2) 7 (9.9) 0.11 21 (15.3) 0 (0.0) Loop diuretics 23 (34.8) 19 (26.8) 0.40 42 (30.7) 0 (0.0) Calcium channel blockers 20 (30.3) 20 (28.2) 0.93 40 (29.2) 0 (0.0) Beta blockers 35 (53.0) 28 (39.4) 0.15 63 (46.0) 0 (0.0) Thiazide diuretics 9 (13.6) 6 (8.5) 0.49 15 (10.9) 0 (0.0) Antialdosterones 4 (6.1) 4 (5.6) 1.00 8 (5.8) 0 (0.0) Other antihypertensive agents 2 (3.0) 2 (2.8) 1.00 4 (2.9) 0 (0.0) Number of antihypertensive drugs 2 [1–3] 2 [0–2] 0.02 2 [1–2] 0 (0.0) Lipid-lowering therapy 26 (39.4) 23 (32.4) 0.50 49 (35.8) 0 (0.0) Amiodarone 7 (10.6) 7 (9.9) 1.00 14 (10.2) 0 (0.0)

1396 European Journal of Clinical Pharmacology (2023) 79:1391–1400

1 3

Quantitative variables are presented as mean ± standard deviation in the case of a normal distribution or medians [1st to 3rd quartiles], and cat- egorical variables are presented as sample sizes and percentages. The characteristics of the two groups were compared using Student’s t-tests for quantitative variables (or Mann–Whitney U tests if the conditions required for the t-test were not met) and chi-square tests for categorical vari- ables. CMR clinical medication review, SPC standard pharmaceutical care, SRIs serotonin reuptake inhibitors, SNRIs serotonin and norepineph- rine reuptake inhibitors *This is a comparison of the distribution of the different classes of antidepressants between the two groups

Table 1 (continued)

Variable CMR SPC p Total Missing values (%)

Psychotropic drugs 0 (0.0) Antidepressants 36 (54.5) 32 (45.1) 0.35 68 (49.6) 0 (0.0) SRIs SNRIs Tetracyclic Other SNRIs and tetracyclics

13 (19.7) 10 (15.2) 10 (15.2) 3 (4.5) 0 (0.0)

15 (21.1) 6 (8.5) 10 (14.1) 0 (0.0) 1 (1.4)

0.004* 28 (20.4) 16 (11.7) 20 (14.6) 3 (2.2) 1 (0.7)

Anxiolytics 20 (30.3) 24 (33.8) 0.80 44 (32.1) 0 (0.0) Hypnotics 10 (15.2) 6 (8.5) 0.34 16 (11.7) 0 (0.0) Antipsychotics 9 (13.6) 5 (7.0) 0.32 14 (10.2) 0 (0.0) Thymoregulators 2 (3.0) 1 (1.4) 0.61 3 (2.2) 0 (0.0) Other drugs Native vitamin D 30 (45.5) 32 (45.1) 1.00 62 (45.3) 0 (0.0) Proton pump inhibitors 30 (45.5) 34 (47.9) 0.91 64 (46.7) 0 (0.0)

Table 2 Drug-related outcomes at admission and on discharge

Variable CMR SPC wp Total

Number of patients 66 (48.2) 71 (51.8) 137 (100.0) Admission Number of PIMs 3 [2–4] 2 [1–3] 0.01 2 [1–4]

≥ 1 PIM 61 (92.4) 61 (85.9) 0.34 122 (89.1) Ratio PIM/number of drugs (%) 29 [18–40] 20 [14–33] 0.02 25 [14–38] Number of POMs 2 [1–2] 2 [1–3] 0.49 2 [1–3] Number of drugs 10 [8–13] 9 [7–11] 0.07 9 [7–12] Number of administrations 11 [9–18] 10 [8–15] 0.11 11 [8–16] Number of psychotropic drugs 1 [0–2] 1 [0–2] 0.48 1 [0–2] Anticholinergic load 1 [0–2] 1 [0–2] 0.93 1 [0–2]

Discharge Number of PIMs 1 [0–1] 1 [0–1] 0.64 1 [0–1] ≥ 1 PIM 41 (62.1) 47 (66.2) 0.72 88 (64.2) Ratio PIM/number of drugs (%) 8 [0–11] 11 [0–13] 0.16 9 [0–13] Number of POMs 1 [0–1] 1 [0–2] 0.01 1 [0–2] Number of drugs 10 [9–12] 9 [8–11] 0.03 10 [8–12] Number of administrations 14 [11–18] 11 [8–14] < 0.001 12 [9–16] Number of psychotropic drugs 1 [0–2] 1 [0–2] 0.16 1 [0–2] Anticholinergic load 1 [1–2] 1 [1–2] 0.21 1 [1–2]

1397European Journal of Clinical Pharmacology (2023) 79:1391–1400

1 3

of the interventions included were compared to standard pharmaceutical care, and the estimated effect on POMs was based on only two studies. The intervention of the first study was conducted by six nursing home physicians who had

received a 10-h educational program about good prescribing for older adults followed by on-demand support by phone. The intervention showed a mean decrease of 0.86 POMs and 0.36 PIM per patient, using the START-STOPPv2

Table 3 Univariate linear regression and weighted linear regression on propensity score

SPC standard pharmaceutical care, CMR clinical medication review

Criterion Univariate linear regression Weighted linear regression on propensity score

CMR SPC Beta* 95% CI p Beta* 95% CI p

Difference in PIMs at discharge and admission − 2 [− 3; − 1] − 1 [− 2; 0] − 0.84 − 1.45; − 0.24 0.01 − 0.13 − 0.84; 0.57 0.71 Difference in POMs at discharge and admission − 1 [− 1; 0] 0 [− 1; 0] − 0.33 − 0.62; − 0.04 0.03 − 0.39 − 0.72; − 0.06 0.02 Difference in drugs at discharge and admission 1 [− 2; 2] 1 [− 1; 2] − 0.12 − 1.15; 0.91 0.82 1.03 − 0.21; 2.26 0.10 Difference in administrations at discharge and

admission 1 [− 2; 5] 0 [− 3; 4] 1.88 − 0.07; 3.82 0.06 4.23 2.02; 6.45 < 0.001

Difference in psychotropic drugs at discharge and admission

0 [− 1; 0] 0 [0; 1] − 0.36 − 0.66; − 0.06 0.02 − 0.25 − 0.60; 0.11 0.17

Difference in anticholinergic load at discharge and admission

0 [0; 1] 0 [− 1; 1] − 0.20 − 0.58; 0.18 0.29 0.04 − 0.42; 0.49 0.87

Table 4 POM detailed results

ACEis angiotensin-converting enzyme inhibitors, SRIs serotonin reuptake inhibitors, SPC standard pharmaceutical care, CMR clinical medica- tion review

POM classes POMs in SPC group POMs in CMR group

Admission [n] Discharge [n (% reduction in POMs)]

Total 130 89 (− 32%) 112 53 (− 53%) 0.01 Cardiovascular drugs 43 34 (− 21%) 19 10 (− 47%) 0.03 Oral anticoagulant drugs 4 1 (− 75%) 3 0 (− 100%) 1 Antiplatelet agents 8 6 (− 25%) 5 4 (− 20%) 1 ACEis 14 13 (− 7%) 4 3 (− 15%) 0.41 Beta blockers 9 7 (− 22%) 2 1 (− 50%) 1 Antihypertensive agents 1 0 (− 100%) 1 0 (− 100%) 1 Lipid-lowering therapy 7 7 (− 0%) 4 2 (− 50%) 0.11 Bone loss drugs 24 14 (− 42%) 35 14 (− 60%) 0.17 Vitamin D 10 5 (− 50%) 12 1 (− 92%) 0.06 Vitamin D + calcium 7 3 (− 57%) 12 5 (− 58%) 1 Bone resorption or anabolic inhibitors 5 5 (− 0%) 7 6 (− 24%) 1 Vitamin D + calcium + bisphosphonates 2 1 (− 50%) 4 2 (− 50%) 1 Other SRIs 2 0 (− 100%) 1 0 (− 100%) 1 Antidepressants 1 0 (100%) 2 0 (− 100%) 1 Dopamine agonists 0 0 1 1 (− 0%) 1 Acetylcholinesterase inhibitors 6 4 (− 33%) 3 3 (− 0%) 0.5 B2 adrenergic agonists 3 2 (− 33%) 0 0 1 Anti-glaucoma drugs 0 0 1 0 (− 100%) 1 Opioids 1 0 (100%) 3 0 (− 100%) 1 Laxatives 6 2 (− 67%) 7 0 (− 100%) 0.19 Alpha-blockers 0 0 2 0 (− 100%) 1 Pneumococcal vaccines 44 33 (− 25%) 38 25 (− 34%) 0.11

1398 European Journal of Clinical Pharmacology (2023) 79:1391–1400

1 3

criteria, compared to the control group [21]. In our study, the decreases were smaller (0.39 POMs and 0.13 PIMs). The number of PIMs and POMs at baseline was not much superior in their population (POM control group = 183 vs. 112 in our study, PIM intervention group = 211 vs. 203 in our study). These discrepancies could be explained by the fact that all geriatricians in our hospital are trained to detect inappropriate prescribing and are used to doing so. Con- trary to their results, PIMs and POMs decreased in our SPC group, from 130 to 89 POMs and from 160 to 62 PIMs.

The second study intervention was similar to that described in our protocol, and it also took place in an acute geriatric ward. Patients who received the intervention were prescribed on average 0.71 fewer POMs (baseline = 96 POMs) than the control group after the intervention accord- ing to the 7 criteria of the “ACOVE” list that defines POMs [22]. Patients who received the intervention were prescribed on average 0.05 PIMs (baseline = 96) according to the STOPPv2 criteria than the control group, which was not significant and similar to our results.

Recently, the OPERAM cluster randomized controlled trial evaluated the impact of CMR on drug-related admis- sions [23]. The included population was not hospitalized, but there were a comparable number of participants with at least one PIM, 72.6% vs. 89.1% in our study. At discharge, 53% of POMs were prescribed in our study, while 23% were prescribed 2 months after the recommendations were imple- mented in the OPERAM study. They deprescribed approxi- mately 52% according to the STOPPs criteria after 2 months, whereas we deprescribed 71% at discharge. This can be explained by the fact that we could easily change medica- tions during hospitalization, while it is more difficult when patients are at home because of the absence of continuous surveillance. Moreover, there seems to be a recurrent dif- ference in the implementation of recommendations between multicentre and single-centre trials [23].

As we found that POMs decreased more in the CMR group, the difference was significant for cardiovascular drugs. This result was expected as we added to a SPC a reap- praisal of all prescriptions by focusing on underprescribing. This included the prescription of drugs for secondary pre- vention of myocardial infarction (beta-blocker, antiplatelet therapy, statins and ACE inhibitors), the prescription of anticoagulation for atrial fibrillation or an antiplatelet therapy for secondary stroke prevention. Prescribing these omitted drugs brings a potential clinical benefit, reducing cardioembolic risks, ischaemic events and cardiac morbid- ity and mortality as a secondary prevention [24–26]. The greater increase in administrations in the CMR group may be explained by the greater introduction of POMs in this group, although the total number of drugs did not change. We did not expect this result; we thought we could improve patients drug regimen with CMR. At discharge, there were still 53

out of 112 POMs in the CMR group. This can be explained by pneumococcal vaccination and bisphosphonate treatment. This first is not usually administrated in the hospital because of acute intercurrent events, and vaccination is often left to the family physician. Second, multiple pretreatment check- ups (dental panoramic ± dental care) are an impediment to immediate bisphosphonate prescriptions.

At discharge, there were still 58 out of 203 admission PIMs in the CMR group. This could be due to barriers to deprescrip- tion, such as the short length of stay in the hospital, which prevents certain drugs from being stopped or modified. The lack of scientific studies on how to deprescribe, the difficulty in coordinating all the stakeholders involved in therapeutic management and the attachment of patients to their drugs are identified barriers. We did not expect the lack of effect on PIM decrease; we thought we could improve PIM deprescrib- ing with CMR as we added to a SPC a reappraisal of all pre- scriptions by focusing on overprescribing and misprescribing. The lack of an effect on PIM decrease may be explained by the fact that this study took place in an acute geriatric ward with experienced geriatricians that are used to detecting and deprescribing PIMs. Moreover, PIMS are often progressively deprescribed (benzodiazepines, proton pomp inhibitor to avoid anxiety rebound or gastric acidity rebound), and as only the discharge prescription was studied, we could not see an impact of these deprescriptions. The lack of difference in the decrease in the anticholinergic load of the prescriptions can be explained by a low initial anticholinergic load of 1 [0–2] in both groups and the frequent prescription of low-dose oral tramadol solu- tion for pain relief at discharge, which has demonstrated anticholinergic potential (score = 1) according to the ADS [20].

This study is original and complementary to the literature because of its real-world setting, the evaluated outcomes and the use of propensity score analysis to diminish the risk of bias. We also used the STOPP/STARTv2 list to define PIMs and POMs, whose use in predicting clinical events has been well established [19, 27, 28] d.

However, we must acknowledge some limitations. This was a retrospective, single-centre study with a CMR group that was selected for patients with more PIMs at admission. The quality of prescriptions was only assessed using explicit prescribing criteria, the STOPP/STARTv2 list, which does not take into account the majority of drug interactions, dose adjustments or the effectiveness of drugs with respect to therapeutic targets, unlike implicit criteria such as the medication appropriate index [29]. Another limitation was the absence of long-term follow-up of pre- scriptions, which did not make it possible to assess the gradual discontinuation of PIMs initiated in the hospital or the continuation of POMs prescribed at discharge. Finally, even if all PIMs and POMS were assessed by an experi- enced clinical pharmacist using STOPP-STARTv2, a part of subjectivity can persist in the assessment.

1399European Journal of Clinical Pharmacology (2023) 79:1391–1400

1 3

Conclusion

Clinical medication review was effective in introducing POMs but not preventing PIMs compared to SPC in an acute geriatric ward. We still need to confirm that drug modifi- cations are implemented in the long term and to prove an association with strong clinical outcomes, such as a decrease in rehospitalization or mortality. These results will help to encourage the implementation of CMR in acute geriatric wards by increasing the presence of pharmacists [30].

Supplementary Information The online version contains supplemen- tary material available at https:// doi. org/ 10. 1007/ s00228- 023- 03551-y.

Acknowledgements The authors would like to thank Romain Leguillon for his kind support

Author contribution Alexandre Meurant, Cédric Villain and Jean- Pierre Jourdan wrote the main manuscript. All authors reviewed the manuscript and contributed to implementing clinical medical reviews in the acute geriatric ward.

Availability of data and materials Data are available upon reasonable request.

Declarations

Ethical approval The local research ethics committee approved this study.

Consent to participate Due to the retrospective design of this study, consent was not necessary to include patients, in accordance with the current Jardé’s Law.

Consent for publication All authors consented to the publication of this manuscript.

Competing interests The authors declare no competing interests.

References

1. Haute Autorité de Santé (2014) Comment améliorer la qualité et la sécurité des prescriptions de médicaments chez la personne âgée ? Haute Autorité de Santé, Saint-Denis

2. Michel P, Minodier C, Lathelize M, et al (2010) Les événements indésirables graves associés aux soins dans les établissements de santé. Direction de la recherche des études de l’évaluation et des statistiques

3. O’Mahony D, Gallagher P, Ryan C et al (2010) STOPP & START criteria: a new approach to detecting potentially inappropriate pre- scribing in old age. Eur Geriatr Med 1:45–51. https:// doi. org/ 10. 1016/j. eurger. 2010. 01. 007

4. Fixen DR (2019) 2019 AGS Beers criteria for older adults. Phar- macy Today 25:42–54. https:// doi. org/ 10. 1016/j. ptdy. 2019. 10. 022

5. Davies LE, Spiers G, Kingston A et al (2020) Adverse outcomes of polypharmacy in older people: systematic review of reviews. J Am Med Dir Assoc 21:181–187. https:// doi. org/ 10. 1016/j. jamda. 2019. 10. 022

6. Fortin M, Hudon C, Haggerty J, et al (2010) Prevalence estimates of multimorbidity: a comparative study of two sources. BMC Health Serv Res 10. https:// doi. org/ 10. 1186/ 1472- 6963- 10- 111

7. Spinewine A, Schmader KE, Barber N et al (2007) Appropriate prescribing in elderly people: how well can it be measured and optimised? Lancet 370:173–184. https:// doi. org/ 10. 1016/ S0140- 6736(07) 61091-5

8. Counter D, Millar JWT, McLay JS (2018) Hospital readmissions, mortality and potentially inappropriate prescribing: a retrospective study of older adults discharged from hospital. Br J Clin Pharma- col 84. https:// doi. org/ 10. 1111/ bcp. 13607

9. Mekonnen AB, Redley B, de Courten B, Manias E (2021) Poten- tially inappropriate prescribing and its associations with health- related and system-related outcomes in hospitalised older adults: a systematic review and meta-analysis. Br J Clin Pharmacol 87

10. Mekonnen AB, McLachlan AJ, Brien JAE (2016) Effectiveness of pharmacist-led medication reconciliation programmes on clinical out- comes at hospital transitions: a systematic review and meta-analysis. BMJ Open 6. https:// doi. org/ 10. 1136/ bmjop en- 2015- 010003

11. Mekonnen AB, McLachlan AJ, Brien JAE (2016) Pharmacy-led medication reconciliation programmes at hospital transitions: a systematic review and meta-analysis. J Clin Pharm Ther 41:128– 144. https:// doi. org/ 10. 1111/ jcpt. 12364

12. Renaudin P, Boyer L, Esteve MA et al (2016) Do pharmacist-led medication reviews in hospitals help reduce hospital readmis- sions? A systematic review and meta-analysis. Br J Clin Pharma- col 82:1660–1673. https:// doi. org/ 10. 1111/ bcp. 13085

13. Mizokami F, Mizuno T, Kanamori K et al (2019) Clinical medication review type III of polypharmacy reduced unplanned hospitalizations in older adults: a meta-analysis of randomized clinical trials. Geriatr Gerontol Int 19:1275–1281. https:// doi. org/ 10. 1111/ ggi. 13796

14. Anthonywalsh K, O’riordan D, Kearney PM et al (2016) Improv- ing the appropriateness of prescribing in older patients: a sys- tematic review and meta-analysis of pharmacists’ interventions in secondary care. Age Ageing 45:201–209. https:// doi. org/ 10. 1093/ AGEING/ AFV190

15. Léguillon R, Varin R, Pressat-Laffouilhère T et al (2023) Clinical pharmacist intervention reduces potentially inappropriate prescrip- tions in a geriatric perioperative care unit dedicated to hip fracture. Gerontology 69:386–395. https:// doi. org/ 10. 1159/ 00052 6595

16. Rankin A, Cadogan CA, Patterson SM et al (2018) Interventions to improve the appropriate use of polypharmacy for older people. Cochrane Database of Systematic Reviews 2018

17. Adam L, Moutzouri E, Baumgartner C et al (2019) Rationale and design of optimising therapy to prevent avoidable hospital admissions in multimorbid older people (OPERAM): a cluster randomised controlled trial. BMJ Open 9. https:// doi. org/ 10. 1136/ bmjop en- 2018- 026769

18. Roux B, Berthou-Contreras J, Beuscart JB et al (2021) Review of potentially inappropriate medication pr[e]scribing in seniors (REMEDI[e]S): French implicit and explicit criteria. Eur J Clin Phar- macol 77:1713–1724. https:// doi. org/ 10. 1007/ S00228- 021- 03145-6

19. O’mahony D, O’sullivan D, Byrne S et al (2015) STOPP/START cri- teria for potentially inappropriate prescribing in older people: version 2. Age Ageing 44:213–218. https:// doi. org/ 10. 1093/ ageing/ afu145

20. Carnahan RM, Lund BC, Perry PJ et al (2006) The anticholinergic drug scale as a measure of drug-related anticholinergic burden: associations with serum anticholinergic activity. J Clin Pharmacol 46:1481–1486. https:// doi. org/ 10. 1177/ 00912 70006 292126

21. García-Gollarte F, Baleriola-Júlvez J, Ferrero-López I et al (2014) An educational intervention on drug use in nursing homes improves health outcomes resource utilization and reduces inap- propriate drug prescription. J Am Med Dir Assoc 15:885–891. https:// doi. org/ 10. 1016/j. jamda. 2014. 04. 010

22. Spinewine A, Swine C, Dhillon S et al (2007) Effect of a col- laborative approach on the quality of prescribing for geriatric inpatients: a randomized, controlled trial. J Am Geriatr Soc 55:658–665. https:// doi. org/ 10. 1111/j. 1532- 5415. 2007. 01132.x

https://doi.org/10.1007/s00228-023-03551-y

https://doi.org/10.1016/j.eurger.2010.01.007

https://doi.org/10.1016/j.ptdy.2019.10.022

https://doi.org/10.1016/j.jamda.2019.10.022

https://doi.org/10.1186/1472-6963-10-111

https://doi.org/10.1016/S0140-6736(07)61091-5

https://doi.org/10.1111/bcp.13607

https://doi.org/10.1136/bmjopen-2015-010003

https://doi.org/10.1111/jcpt.12364

https://doi.org/10.1111/bcp.13085

https://doi.org/10.1111/ggi.13796

https://doi.org/10.1093/AGEING/AFV190

https://doi.org/10.1159/000526595

https://doi.org/10.1136/bmjopen-2018-026769

https://doi.org/10.1007/S00228-021-03145-6

https://doi.org/10.1093/ageing/afu145

https://doi.org/10.1177/0091270006292126

https://doi.org/10.1016/j.jamda.2014.04.010

https://doi.org/10.1111/j.1532-5415.2007.01132.x

1400 European Journal of Clinical Pharmacology (2023) 79:1391–1400

1 3

23. Blum MR, Sallevelt BTGM, Spinewine A et al (2021) Optimizing therapy to prevent avoidable hospital admissions in multimorbid older adults (OPERAM): cluster randomised controlled trial. The BMJ 374. https:// doi. org/ 10. 1136/ bmj. n1585

24. Gibbons RJ, Abrams J, Chatterjee K et al (2003) ACC/AHA 2002 guideline update for the management of patients with chronic sta- ble angina - summary article: A report of the American College of Cardiology/American Heart Association Task Force on Prac- tice Guidelines (Committee on the Management of Patients With Chronic Stable Angina). Circulation 107

25. Hunt SA, Abraham WT, Chin MH et al (2005) ACC/AHA 2005 guideline update for the diagnosis and management of chronic heart failure in the adult: a report of the American College of Cardiol- ogy/American Heart Association Task Force on practice guidelines (writing committee to update the 2001 guidelines for the evaluation and management of heart failure). J Am Coll Cardiol 46

26. Amarenco P, Labreuche J, Lavallée P, Touboul PJ (2004) Statins in stroke prevention and carotid atherosclerosis: systematic review and up-to-date meta-analysis. Stroke 35

27. Hill-Taylor B, Walsh KA, Stewart SA et al (2016) Effectiveness of the STOPP/START (screening tool of older persons’ potentially inappropriate prescriptions/screening tool to alert doctors to the

right treatment) criteria: systematic review and meta-analysis of randomized controlled studies. J Clin Pharm Ther 41:158–169. https:// doi. org/ 10. 1111/ jcpt. 12372

28. Gallagher P, Ryan C, Byrne S et al (2008) STOPP (screening tool of older person’s prescriptions) and START (screening tool to alert doctors to right treatment). Consensus validation. Int J Clin Pharmacol Ther 46. https:// doi. org/ 10. 5414/ CPP46 072

29. Hanlon JT, Schmader KE, Samsa GP et al (1992) A method for assessing drug therapy appropriateness. J Clin Epidemiol 45. https:// doi. org/ 10. 1016/ 0895- 4356(92) 90144-C

30. Petrovic M, O’Mahony D, Cherubini A (2022) Inappropriate pre- scribing: hazards and solutions. Age Ageing 51

Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

https://doi.org/10.1136/bmj.n1585

https://doi.org/10.1111/jcpt.12372

https://doi.org/10.5414/CPP46072

https://doi.org/10.1016/0895-4356(92)90144-C

Implementation of clinical medication review in a geriatric ward to reduce potentially inappropriate prescriptions among older adults

Abstract

Purpose
Methods
Results
Conclusion

Introduction
Methods

Study design and participants
Data collection
Outcomes
Statistical analysis

Results

Univariate analysis of outcome measures
Propensity score analysis
PIM and POM descriptions

Discussion
Conclusion
Anchor 19
Acknowledgements
References

image2.emf

OR I G I N A L A R T I C L E

Effects of a comprehensive medication review intervention on health-related quality of life and other clinical outcomes in geriatric outpatients with polypharmacy: A pragmatic randomized clinical trial

Jonatan Kornholt1 | Shafika Tapia Feizi2 | Alexandra Storm Hansen1 |

Jannie Thaysen Laursen1 | Lene Ørskov Reuther1,3 |

Tonny Studsgaard Petersen1,3 | Eckart Pressel2 | Mikkel Bring Christensen1,3,4,5

1Department of Clinical Pharmacology, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Denmark

2Department of Geriatric and Palliative Medicine, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Denmark

3Department of Clinical Medicine, University of Copenhagen, Denmark

4Center for Clinical Metabolic Research, Copenhagen University Hospital, Herlev and Gentofte, Denmark

5Copenhagen Center for Translational Research, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Denmark

Correspondence

Mikkel Bring Christensen, Department of

Clinical Pharmacology, Copenhagen University

Hospital Bispebjerg, Bispebjerg Bakke

23, 2400 Copenhagen NV, Denmark.

Email: [email protected]

Funding information

The Capital Region of Denmark, Grant/Award

Number: E-12825-01-12-01; Velux

Foundation, Grant/Award Number: 00025835

Aim: To investigate the effects of a comprehensive medication review intervention

on health-related quality of life (HRQoL) and clinical outcomes in geriatric outpatients

exposed to polypharmacy.

Methods: Pragmatic, nonblinded, randomized clinical trial with follow-up after 4 and

13 months. Participants were geriatric outpatients taking ≥9 medicines. The interven-

tion was an additional consultation with a physician focusing on reviewing medica-

tion, informing patients about their medicines and increasing cross-sectoral

communication as supplement to and compared with usual care. The primary out-

come was change in HRQoL after 4 months measured with the EuroQoL

5-dimension 5-level (EQ-5D-5L) questionnaire. Secondary outcomes were HRQoL

after 13 months, mortality, admissions, falls and number of medicines after 4 and

13 months.

Results: Of 785 eligible patients, 408 were included (age: mean 80.6 [standard devia-

tion 7.22] years; number of medicines: median 12 [interquartile range 10–14];

females 71%). After 4 months, the adjusted between-group difference in EQ-5D-5L

index score was 0.066 in favour of the medication consultation (95% confidence

interval 0.01 to 0.12, P = .02). After 4 months, two (1%) participants had died in the

medication-consultation group and nine (4%) in the usual-care group (log-rank test,

P = .045). The medication consultation reduced the number of medicines by 2.0

The authors confirm that the Principal Investigators for this paper are Jonatan Kornholt and Mikkel Bring Christensen, and that they had direct clinical responsibility for patients.

Received: 19 November 2021 Revised: 26 January 2022 Accepted: 5 February 2022

DOI: 10.1111/bcp.15287

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any

medium, provided the original work is properly cited and is not used for commercial purposes.

3360 Br J Clin Pharmacol. 2022;88:3360–3369.wileyonlinelibrary.com/journal/bcp

https://orcid.org/0000-0003-2662-4519

https://orcid.org/0000-0002-9974-2738

https://orcid.org/0000-0002-8774-1797

mailto:[email protected]

https://doi.org/10.1111/bcp.15287

http://creativecommons.org/licenses/by-nc/4.0/

http://wileyonlinelibrary.com/journal/bcp

(15.8%) after 4 months and 1.3 (10.7%) after 13 months. There were no statistically

significant differences in mortality or HRQoL after 13 months, and no differences in

falls or admissions.

Conclusions: An additional consultation with medication review and increased com-

munication as supplement to usual geriatric outpatient care improved HRQoL and

reduced mortality after 4 months.

K E YWORD S

geriatrics, health-related quality of life, medication reviews, polypharmacy

1 | INTRODUCTION

The prevalence of polypharmacy is increasing,1–3 and it is particularly

prevalent in the multimorbid older population.2,4 Polypharmacy is

associated with inappropriate medicine use5 and numerous adverse

health outcomes, including adverse drug events and functional and

cognitive decline.6 Inappropriate medicine use occurs when the risk/

benefit profile of the treatment is no longer favourable or when the

treatment is no longer aligned with the patient's goal of care.7 One

tool to combat inappropriate polypharmacy is medication reviews,

where a pharmacist or physician revises the patient's medicine to

reduce/stop inappropriate medicines and initiate/increase treatment

with appropriate medicines.8 While medication reviews are often suc-

cessful in reducing the use of inappropriate medicines, a beneficial

effect on more important clinical outcomes has not been consistently

demonstrated in clinical trials.9 Reasons for the lack of clinical effects

following medication reviews include a low implementation rate of

proposed medicine changes10 and numerous barriers towards lasting

medicine changes, including deficient cross-sectoral communica-

tion.11,12 Other reasons for inconsistent and unfavourable results

include qualitative differences in the medication reviews, the qualifi-

cations of the person reviewing the medication (eg, physician or phar-

macist), the setting (eg, in-hospital, outpatients, nursing homes) and

the patient population (eg, number of medicines, comorbidities). While

health-related quality of life (HRQoL) is recognized as an important

patient-reported outcome, few trials of medication reviews have

investigated changes in HRQoL13 and most trials did not find an effect

on HRQoL.9,14–18

In this pragmatic trial, we investigated whether allocation of addi-

tional physician resources for medication reviews with direct imple-

mentation of the medicine changes based on patient wishes and prior

contact to the primary care physician could improve HRQoL in geriat-

ric outpatients exposed to polypharmacy.

2 | METHODS

A pragmatic, nonblinded, single-centre, randomized clinical trial with

follow-up at 4 months and 13 months after first visit. The trial reg-

istration is available at clinicaltrials.gov (NCT03911934).

2.1 | Participants

Participants were included from June 2017 to December 2019 at the

geriatric outpatient clinic, Copenhagen University Hospital, Frederiks-

berg, Denmark as part of routine outpatient care. Newly referred

patients treated with at least nine different medicines according to

the electronic prescription system were randomly allocated prior to

their first visit to receive either usual geriatric care or usual geriatric

care and an additional medication consultation with a physician focus-

ing on reviewing medication, aligning treatment with the patient's

wishes and ensuring cross-sectoral communication. Medicines were

defined as “different” based on the fifth level codes in the Anatomical

Therapeutic Chemical (ATC) Classification System,19 and the counted

What is already known about this subject

• Medication reviews can reduce inappropriate poly-

pharmacy, but the clinical impact is uncertain.

• Possible reasons for low clinical impact include low imple-

mentation rates following medication reviews and defi-

cient cross-sectoral communication.

• We investigated the effect of a consultation including

medication review, patient information and increased

cross-sectoral communication as supplement to geriatric

outpatient care.

What this study adds

• Medication review, patient information and increased

cross-sectoral communication reduced the number of

medicines and improved health-related quality of life and

survival after 4 months.

• Physician resources to review medication, inform patients

and increase cross-sectoral communication can improve

geriatric outpatients' health-related quality of life and sur-

vival in the short term.

KORNHOLT ET AL. 3361

http://clinicaltrials.gov

medicines included inhaled medicines, but excluded topical treatments

such as eye drops, ear drops and lotions, antibiotics with limited treat-

ment duration, multivitamins and protein drinks. The cut-off of nine

medicines was selected as a compromise between a steady patient

flow and a high risk of inappropriate medicines.5 Patients allocated to

the additional medication consultation had their medication reviewed

irrespective of their consent to participate in the follow-up. Follow-up

was conducted for patients who, during their visit, agreed to have

their administrative data collected and registered (by signing a written

consent), and had sufficient cognitive and linguistic abilities to com-

plete the EuroQoL 5-dimension 5-level (EQ-5D-5L) questionnaire.20

The study was classified as a quality improvement study by the

Regional Ethics Committee (correspondence number 17001679), as

are similar studies in Denmark.21 The project was approved by the

Danish Data Protection Agency (BFH-2017-031).

2.2 | Outcomes, blinding and data collection

The primary outcome was the between-group difference in change in

HRQoL from baseline to 4-month follow-up measured with EQ-5D-

5L index values.20 We amended the protocol to use the recently pub-

lished Danish utility weights22 to convert the EQ-5D-5L health states

to index values (ranges from �0.758 to 1 with 0 anchored at dead

and 1 anchored at full health). Robustness to choice of value set was

investigated in sensitivity analyses using the index values from the

EQ-5D Crosswalk Calculator.23 The questionnaire was completed dur-

ing the first visit in the outpatient clinic (baseline) and at follow-up

around 4 and 13 months after the first visit. Follow-up was by tele-

phone, letter with return envelope, email or visit in the patient's home

depending on the patient's hearing and sight, and the patient's wishes.

Blinding was not possible since the patients actively participated in

the medication reviews. Secondary outcomes pertaining to HRQoL

were (1) change in EQ-5D-5L index values from baseline to 4-month

follow-up excluding participants who died, (2) change in EQ-5D-5L

index values from baseline to 13-month follow-up including partici-

pants who died, (3) change in EQ-5D-5L index values from baseline to

13-month follow-up excluding participants who died and (4) change in

EQ-5D-5L Visual Analog Scale (VAS) ranging from 0 to 100 at the

same timepoints as EQ-5D-5L index values, both including and

excluding participants who died. Other clinical secondary outcomes

were (1) proportion of participants with at least one self-reported fall

within the last 3 months at 4- and 13-month follow-ups, (2) number

of admissions and admission days from baseline to 4-month follow-up

and from baseline to 13-month follow-up, (3) time to first admission

and (4) mortality at 4-month follow-up and 13-month follow-up. To

assess whether the additional medication consultation resulted in

medicine changes compared with usual care, we also collected medi-

cine data at baseline, after first visit and at 4- and 13-month follow-

ups. As part of patient and public involvement in research, we con-

ducted qualitative interviews to understand physicians'11 and patients'

opinions and interpretation of the intervention and the results.

2.3 | Usual care and the intervention

Usual care depended on the reason for referral to the outpatient

clinic. Patients referred due to falls were examined in an interdisci-

plinary team of nurses, geriatricians and physiotherapists, and rou-

tinely screened with biothesiometry, orthostatic blood pressure

measurements and often Holter monitor or cardiac event recorder.

Furthermore, the medication was by default reviewed, including

optimizing/minimizing use of analgesics, psychotropics and other

fall-inducing medicines, and osteoporosis was treated as necessary.

For patients referred to general geriatric assessment, the usual care

depended on whether the patient was a follow-up patient after

admission or a new referral. Follow-up after admissions were short

consultations with a geriatrician to follow-up on specific problems

identified during admission. New referrals had a complete geriatric

assessment by a geriatrician, including a standard array of blood

tests and screening of weight, height, food intake and more by a

nurse.

The intervention consisted of an additional medication consul-

tation with a physician from the Department of Clinical Pharmacol-

ogy focusing on reviewing the medication, ie ensuring the

appropriateness of all prescriptions, informing the patient about

their medicines, and ensuring enhanced cross-sectoral communica-

tion and collaboration. The clinical medication review consisted of

a thorough examination of the patient's medical chart to assess the

risk/benefit profiles of all the patient's prescribed medicines, an

interaction check using the Danish National Drug Interactions

Database,24 examination of prescription redemption patterns using

data from the electronic prescription system and screening of inap-

propriate medicines using various tools25–27 as per the physician's

discretion. A physician from the Department of Clinical Pharmacol-

ogy prepared and, when possible, discussed the medication review

by telephone with the patient's primary care physician prior to the

patient's first visit in the outpatient clinic. The potential changes in

medication were consulted with the usual-care geriatrician and

implemented in agreement with the patient's wishes. After the

consultation, the implemented changes and any medicine-related

considerations were sent to the patient's primary care physician

using the routine electronic correspondence module.

2.4 | Randomization

Randomization of patients was performed prior to obtaining consent

for data registration and follow-up. The medical secretaries at the out-

patient clinic screened newly referred patients' medicine in the elec-

tronic prescription system and randomized eligible patients using

REDCap.28 The randomization list was generated in R29 and concealed

for the secretaries. Patients were randomized 1:1 in blocks of four

stratified by age group (65-70, 71-80 or >80 years old), number of dif-

ferent medicines at randomization (9-11, 12-16 or >16 different medi-

cines) and sex (male or female).

3362 KORNHOLT ET AL.

2.5 | Sample size

Value sets for EQ-5D-5L differ between countries30 and the minimal

clinically important difference differs between patient populations.31

Geriatric outpatients are a heterogeneous population and it is

therefore difficult to establish an EQ-5D-5L minimal clinically

important difference. Based on data from Walters and Brazier,31 by

including 208 patients in each group this study was powered to

detect a treatment effect of 0.08 (SD = 0.29) with alpha = 0.05

and beta = 0.2, calculated using the formula for sample size for a

two-sample t-test with the pwr32 package in R.29

2.6 | Statistical methods

HRQoL outcomes were analysed with a baseline-constrained linear

mixed model33 with group, time and the three stratification variables

as covariates. Correlation of repeated measurements was modelled

with an unstructured variance-covariance matrix with correlation

between measurements per participant and unequal variance per time

point. The linear mixed model was fitted with the nlme package

version 3.1-15034 in R version 3.6.3.29

Differences in number of medicines were analysed with one

model per time point (due to differences in missing data at

F IGURE 1 Patient flow through the study. Only data collection of the primary outcome measure (EQ-5D-5L index value) is depicted

KORNHOLT ET AL. 3363

different time points) using a generalized linear model with a

quasi-Poisson distribution with the logarithm of number of differ-

ent medicines at baseline as offset and group as covariate. Differ-

ences in proportion of patients with falls were analysed using

Fisher's exact test with one test per follow-up. Differences in num-

ber of admissions and differences in number of days admitted

were analysed with nonparametric methods using bootstrapping for

calculation of confidence intervals and permutation testing for cal-

culation of P values with one analysis per endpoint per time point.

Mortality and time to first admission were analysed using Kaplan-

Meier curves with P values from log-rank tests and with estimates

and confidence intervals for the effect using Cox proportional haz-

ards models adjusted for the same covariates as the primary end-

point. All secondary outcomes are hypothesis generating and no

adjustments for multiple comparisons were performed.

Since the primary outcome required information from follow-up,

intention-to-treat analysis was not possible and available case analysis

was performed instead (missingness could be intermittent). Sensitivity

analyses were performed to explore the consequences of data missing

not at random using worst/best single imputation (mean ± 2 � SD for

the group for the time point).

3 | RESULTS

A total of 408 patients were included, with 196 in the medication-

consultation group and 212 in the usual-care group (see Figure 1 for

patient flow and Supporting Information Table S1 for completeness of

EQ-5D-5L index data). The baseline characteristics of the participants

are available in Table 1. The baseline characteristics for participants

with missing data are available in Supporting Information Table S2.

The median (IQR) number of visits with the physician from the

Department of Clinical Pharmacology was one (1 to 1) with no (0 to 0)

telephone follow-ups. The median (IQR) number of visits with the ger-

iatrician was two (1 to 2) in the medication-consultation group and

two (1 to 3) in the usual-care group with one (0 to 1) telephone

follow-up in the medication-consultation group and one (0 to 1) in the

usual-care group.

Regarding HRQoL, at baseline the mean (SD) EQ-5D-5L index

score was 0.615 (0.286) in the medication-consultation group and

0.581 (0.309) in the usual-care group (two-sample t-test for differ-

ence at baseline, P = .25). The analysis of the primary outcome

showed an adjusted between-group difference in EQ-5D-5L index

value of 0.066 (95% confidence interval [CI] 0.01 to 0.12, P = .02)

in favour of the medication consultation at 4-month follow-up

(Figure 2A). The mean (SD) EQ-5D-5L index score was 0.631

(0.302) in the medication-consultation group and 0.540 (0.337) in

the usual-care group at 4-month follow-up, and 0.571 (0.344) in

the medication-consultation group and 0.544 (0.340) in the usual-

care group at 13-month follow-up. The other analyses pertaining

to changes in HRQoL are presented in Figure 2A and Table 2. At

4-month follow-up, 2 (1%) participants had died in the medication-

consultation group compared with 9 (4%) in the usual-care group

(log-rank test, P = .045; Figures 2B and 2C). At 13-month follow-

up, 14 (7.1%) participants had died in the medication-consultation

group compared with 26 (12.2%) in the usual-care group (log-rank

test, P = .12; Figures 2B and 2C). Hazard ratios for the reduction

in mortality are presented in Figure 2B. There were no statistically

significant differences between groups at any time points regarding

number of admissions, days admitted or time to first admission and

no difference in proportion of patients with falls (Supporting Infor-

mation Table S3 and Supporting Information Figure S1).

TABLE 1 Baseline characteristics of the included patients

Usual care

(n = 212)

Usual care + medication

consultation (n = 196)

Age in years, mean (SD) 80.8 (7.3) 80.5 (7.2)

Females, n (%) 149 (70) 139 (71)

Number of medicines,

median (range)

12 (9, 24) 12 (9, 27)

Number of diagnoses,

median (range)

6 (1, 16) 6 (1, 15)

Charlson comorbidity

index, median (IQR)

5 (4, 6) 5 (4, 6)

FRAIL score,

median (IQR)

2 (2, 3) 3 (2, 3)

At least one fall in the

last 3 months, n (%)

128 (65) 136 (64)

Number of admissions

the last 3 months,

median (IQR)

1 (0, 2) 1 (1, 2)

Not motivated for

medicine changes,

n (%)

39 (18) 39 (20)

Home care, n (%)

None 58 (27) 61 (31)

Daily 52 (25) 63 (32)

Less than daily 77 (36) 57 (29)

Nursing home

resident

25 (12) 15 (7.7)

Medicine dispensed by, n (%)

The patient 88 (42) 91 (46)

Relative 14 (6.6) 14 (7.1)

Home nurse 82 (39) 73 (37)

Nursing home 25 (12) 15 (7.7)

Other 3 (1.4) 3 (1.5)

Referred from, n (%)

General practitioner 73 (34) 51 (26)

Geriatric department 56 (26) 59 (30)

Other departments 83 (39) 86 (44)

Referred to, n (%)

Geriatric assessment 111 (52) 108 (55)

Falls or hip fracture

clinic

101 (48) 88 (45)

3364 KORNHOLT ET AL.

During the outpatient clinic visit, there were 1180 changes to

the medicine in the medication-consultation group compared with

456 changes in the usual-care group. These changes were mostly

discontinuations (53% of the changes in the medication-

consultation group and 49% in the usual-care group), followed by

reduced dosage (17% and 18%), new prescriptions (16% and 21%),

increased dosage (4% and 6%) and other changes such as change

from as-needed to regular dosing (10% and 5%). These changes

resulted in a statistically significantly reduced number of prescribed

medicines in the medication-consultation group compared with the

usual-care group at all time points after baseline (Table 3).

The sensitivity analysis of the primary endpoint showed that

results were robust to the choice of valuation set. Further sensitivity

analyses adjusting the HRQoL and mortality data for the effect of

nursing home status showed similar results as the primary analyses

(Supporting Information Table S4). For missing data considered not

missing at random, the worst/best sensitivity analysis showed a worst

effect of the intervention of �0.057 and a best effect of 0.194.

4 | DISCUSSION

In this study, we investigated whether an allocation of additional phy-

sician resources could improve health outcomes for geriatric outpa-

tients taking ≥9 medicines compared with usual care in a geriatric

outpatient clinic. The additional physician focused on improving the

F IGURE 2 (A) Results from the analysis of the primary outcome measure showing the difference in change from baseline in health-related quality of life between groups from baseline to follow-up after 4 months, including subjects who died during follow-up. The error bars are 95% confidence intervals for the estimated marginal means. The secondary outcome (change from baseline to follow-up after 13 months) is also depicted since both outcomes were analysed in the same constrained linear mixed model adjusted for the stratification variables: number of medicines at baseline (three levels), age at baseline (three levels) and sex (two levels). CI, confidence interval; EQ-5D-5L, EuroQoL 5-dimension 5- level. (B) Cumulative incidence curves showing the incidence of death in the control and intervention groups. Hazard ratios were calculated using an adjusted Cox proportional hazards model. CI, confidence interval; HR, hazard ratio. (C) Numbers at risk of dying during the study. Only four patients withdrew consent and were censored prior to the end of the study

KORNHOLT ET AL. 3365

care by performing a thorough medication review and carefully com-

municating with patients and primary care providers about the medi-

cation. We found that there were statistically significant

improvements in HRQoL and mortality after 4 months, but that these

effects were no longer statistically significant after 13 months.

The changes in HRQoL after 4 months were evident measured

both with EQ-5D-5L index values and EQ-5D-5L VAS values. Part of

the effect on HRQoL was due to a statistically significant difference in

mortality, but there was still an effect on HRQoL when excluding par-

ticipants who died (Table 2). After 13 months, there was no evidence

of an effect on HRQoL both including and excluding participants who

died during the study (Table 2), and the between group difference in

mortality was also no longer statistically significant even though the

effect estimate still favoured the intervention (Figures 2B and 2C).

This study was not powered to detect a difference in mortality, but

given the relatively large effect on mortality and that the survival cur-

ves for the groups were still clearly separated at the end of the study,

a potential effect on mortality should be investigated in a new trial.

The effect of the intervention on the EQ-5D-5L index value of

0.066 was lower than the value used in our sample size calculation

(0.080). This raises the question whether the statistically significant

HRQoL difference is also clinically relevant. The clinical relevance of

HRQoL changes is generally difficult to ascertain. Across various

populations the minimal clinically important differences in EQ-5D-5L

have been reported to range from 0.02835 to 0.4636 depending on

patient population and calculation method. Another method to

describe the minimal clinically important difference is to use Cohen's

d effect sizes, where 0.20-0.50 often is considered the minimal clini-

cally important difference.37 The Cohen's d effect in this study was

0.23 corresponding to a small effect. To involve patients and further

develop the outpatient offer, we conducted interviews with five poly-

pharmacy patients not included in the trial. All these patients would

attend an outpatient clinic one or more times to achieve the observed

effect. Also, the reduction in medicine was important for these

patients irrespective of any gain in HRQoL especially if the reduction

would lead to fewer daily administrations of medicine. Overall, we

believe the change in HRQoL after 4 months constitutes a small but

clinically relevant effect.

Our study resembles a recent trial by Romskaug et al,38 where

geriatric assessments with focus on medication reviews resulted in

increased HRQoL measured with 15D39 after 16 weeks in home-

dwelling older polypharmacy patients. In contrast to the trial by

Romskaug et al, in this study both groups were treated by a geriatri-

cian as part of usual care. Nevertheless, the Cohen's d effect size of

0.24 in the trial by Romskaug et al was nearly identical to the Cohen's

d effect size of 0.23 in this study corresponding to a small clinically

important difference.37 As such, these studies complement each other

and show that medication reviews can improve the HRQoL of geriat-

ric patients exposed to polypharmacy.

The 13-month follow-up revealed that the HRQoL in this popula-

tion is decreasing, and that more than two-thirds of the patients were

admitted at least once and approximately one in 10 died. The seem-

ingly diminishing effects over time of the intervention on number ofT A B L E 2

A na

ly se s o f th e se co

nd ar y o ut co

m e m ea

su re s pe

rt ai ni ng

to he

al th -r el at ed

qu al it y o f lif e

E Q -5 D -5 L m ea

su re m en

t G ro up

U na

dj us te d sc o re s, m ea

n (S D )

A dj us te d m ea

n d if fe re n ce

in ch

an ge

fr o m

b as el in e (9 5 %

C I)a

B as el in e

Fo llo

w -u p

4 m o nt hs

Fo llo

w -u p

1 3 m o nt hs

Fo llo

w -u p 4 m o n th s

P va

lu e

F o llo

w -u p 1 3 m o n th s

P va

lu e

In de

x va lu e ex

cl ud

in g th o se

w ho

di ed

U su al ca re

+ m ed

ic at io n

co ns ul ta ti o n

0 .6 2 1 (0 .2 6 8 )

0 .6 4 1 (0 .2 8 8 )

0 .6 3 5 (0 .2 9 2 )

0 .0 5 1 (�

0 .0 0 1 to

0 .1 0 0 )

.0 5 5

�0 .0 0 8 (�

0 .0 6 3 to

0 .0 4 7 )

.7 8

U su al ca re

0 .5 9 2 (0 .2 8 3 )

0 .5 7 6 (0 .3 0 3 )

0 .6 3 9 (0 .2 7 0 )

V A S va lu e in cl ud

in g th o se

w ho

di ed

U su al ca re

+ m ed

ic at io n

co ns ul ta ti o n

5 7 .3

(2 0 .4 )

5 9 .3

(1 9 .9 )

5 1 .1

(2 6 .1 )

7 .7

(3 .3

to 1 2 .1 )

< .0 0 1

4 .1

(� 1 .8

to 1 0 )

.1 8

U su al ca re

5 3 .5

(2 1 .4 )

5 0 .1

(2 4 .3 )

4 6 .4

(2 8 .4 )

V A S va lu e ex

cl ud

in g th o se

w ho

di ed

U su al ca re

+ m ed

ic at io n

co ns ul ta ti o n

5 7 .3

(2 0 .4 )

6 0 .1

(1 8 .9 )

5 5 .3

(2 1 .6 )

6 .2

(2 .1

to 1 0 .2 )

.0 0 3

0 .6

(� 4 .1

to 5 .3 )

.8 0

U su al ca re

5 3 .5

(2 1 .4 )

5 2 .7

(2 2 .0 )

5 7 .0

(2 0 .5 )

A bb

re vi at io ns :C

I, co

nf id en

ce in te rv al ;E

Q -5 D -5 L,

E ur o Q o L 5 -d im

en si o ns

5 -l ev

el ;S

D ,s ta nd

ar d de

vi at io n;

V A S,

vi su al an

al o gu

e sc al e.

a B as el in e- co

ns tr ai ne

d lin

ea r m ix ed

m o de

lw it h an

un st ru ct ur ed

va ri an

ce -c o va ri an

ce m at ri x fo r th e re si du

al s al lo w in g fo r co

rr el at io n be

tw ee

n re pe

at ed

m ea

su re m en

ts fo r p ar ti ci p an

ts an

d fo r u n eq

u al va ri an

ce p er

ti m e po

in t. T he

m o de

li nc

lu de

s da

ta fr o m

bo th

fo llo

w -u ps .T

he m o de

li s ad

ju st ed

fo r th e st ra ti fi ca ti o n va ri ab

le s: nu

m be

r o f m ed

ic in es

at ba

se lin

e (t h re e le ve

ls ), ag e at

b as el in e (t h re e le ve

ls )a

n d se x (t w o le ve

ls ).

C o nf id en

ce lim

it s an

d P va lu es

ar e no

t ad

ju st ed

fo r m ul ti pl e co

m pa

ri so ns .

3366 KORNHOLT ET AL.

medicines, HRQoL and mortality suggest that perhaps follow-up med-

ication reviews are needed to ensure a continuous alignment of treat-

ment goals and favourable risk/benefit profile.

We did not find any difference in admissions (including emer-

gency department visits) between the groups, which has been

reported in other trials of medication reviews performed in hospi-

tal.10,21 The setting may have impacted this finding as the modifiable

risk factors for admission and the baseline risk of admission may differ

between outpatients and patients acutely admitted to a hospital.

4.1 | Strengths and limitations

The major limitation to this trial is the unblinded design, which may

bias the patient-reported outcome measure in favour of the interven-

tion. However, it is not possible to blind the participants to the addi-

tional medication consultation and this bias was likely limited as the

additional consultation was provided as part of normal routine in the

outpatient clinic. The trial was conducted in a single centre with the

same geriatricians providing usual care for both groups, which may

have introduced contamination bias and reduced the effect of the

additional physician consultation. Likewise, there is a general focus in

the geriatric outpatient clinic on medication reviews and the signifi-

cant reductions of medicine in the usual care group are likely also to

reduce the effect of the additional medication consultation. Due to

the complex intervention, it is impossible to distinguish effects due to

the medication review and the accompanying reduction in medicine

from effects due to the increased communication with patient and pri-

mary care physicians.

The strengths of this trial include the long follow-up period, the

cross-sectoral alliance with the primary care physician and the prag-

matic nature, which means that, based on these results, a similar offer

could easily be established within other geriatric outpatient clinics.

5 | CONCLUSION

Allocation of additional physician resources focusing on the patient's

medication with thorough medication review and enhanced

communication with patients and primary care providers can persis-

tently reduce medicine use for geriatric outpatients exposed to poly-

pharmacy and increase HRQoL in the short term. Selecting

appropriate patients for additional resource use seems important for

combating the negative effects of polypharmacy.

ACKNOWLEDGEMENTS

We would like to thank Ditte-Marie Schärfe Kjærgaard and Linea

Fugmann Thamdrup for assistance with data collection. We would like

to thank all the physicians, nurses and medical secretaries at the geri-

atric outpatient clinic at Copenhagen University Hospital, Bispebjerg

and Frederiksberg for assistance with the conduct of this study. This

work was supported by the Velux Foundation [00025835] and the

Capital Region of Denmark [E-12825-01-12-01]. The funders had no

role in the design and conduct of the study, the collection, manage-

ment, analysis and interpretation of the data, the preparation, review

or approval of the manuscript or the decision to submit the manu-

script for publication.

COMPETING INTERESTS

No competing interests have been declared by the authors.

CONTRIBUTORS

J.K. conceptualization, methodology, formal analysis, investigation,

data curation, writing – original draft, writing – review and editing,

visualization, project administration. S.T.F. methodology, investigation,

writing – review and editing. A.S.H. methodology, investigation, writ-

ing – review and editing, project administration. J.T.L. methodology,

investigation, writing – review and editing, project administration. L.Ø.

R. conceptualization, methodology, writing – review and editing.

T.S.P. conceptualization, methodology, formal analysis, investigation,

writing – review and editing, supervision. E.P. conceptualization,

methodology, resources, writing – review and editing, supervision.

M.B.C. conceptualization, methodology, investigation, writing – origi-

nal draft, writing – review and editing, supervision, project administra-

tion, funding acquisition.

DATA AVAILABILITY STATEMENT

Data available on request from the authors.

TABLE 3 Analyses of the number of different medicines after the first visit in the outpatient clinic and during follow-up

Follow-up time

Proportion of medicines compared with baseline, a mean number at follow-up/mean number at baseline (%) Comparison between groupsb

Usual care + medication consultation Usual care Rate ratio (95% CI) P value

After first visit 10.1/12.4 (81.7) 11.6/12.2 (95.1) 0.859 (0.829 to 0.890) < .001

After 4 months 10.4/12.4 (84.2) 11.6/12.2 (95.3) 0.883 (0.848 to 0.921) < .001

After 13 months 11.0/12.3 (89.3) 11.8/12.1 (97.6) 0.915 (0.873 to 0.960) < .001

Abbreviations: ATC, Anatomical Therapeutic Chemical classification system; CI, confidence interval. aOnly different medicines are counted, where different entails unique ATC codes at the fifth level. Only patients alive at follow-up are included and

therefore the number of baseline medicines may differ between time points. bGeneralized linear model with a quasi-Poisson distribution with the logarithm of the number of different medicines at baseline as offset, number of

medicines at follow-up as dependent variable and group as independent variable. Presented results are exponentiated. Confidence limits and P values are

not adjusted for multiple comparisons.

KORNHOLT ET AL. 3367

ORCID

Jonatan Kornholt https://orcid.org/0000-0003-2662-4519

Tonny Studsgaard Petersen https://orcid.org/0000-0002-9974-

2738

Mikkel Bring Christensen https://orcid.org/0000-0002-8774-1797

REFERENCES

1. Hovstadius B, Hovstadius K, Astrand B, et al. Increasing polypharmacy

– an individual-based study of the Swedish population 2005-2008.

BMC Clin Pharmacol. 2010;10(1):10-16. doi:10.1186/1472-6904-

10-16

2. Qato DM, Wilder J, Schumm LP, Gillet V, Alexander GC. Changes in

prescription and over-the-counter medication and dietary supplement

use among older adults in the United States, 2005 vs 2011. JAMA

Intern Med. 2016;176(4):473-482. doi:10.1001/jamainternmed.2015.

8581

3. Masnoon N, Shakib S, Kalisch-Ellett L, Caughey GE. What is poly-

pharmacy? A systematic review of definitions. BMC Geriatr. 2017;17:

230. doi:10.1186/s12877-017-0621-2

4. Kornholt J, Christensen MB. Prevalence of polypharmacy in Denmark.

Dan Med J. 2020;67:A12190680.

5. Steinman MA, Landefeld CS, Rosenthal GE, et al. Polypharmacy and

prescribing quality in older people. J Am Geriatr Soc. 2006;54(10):

1516-1523. doi:10.1111/j.1532-5415.2006.00889.x

6. Maher RL, Hanlon JT, Hajjar ER. Clinical consequences of poly-

pharmacy in elderly. Expert Opin Drug Saf. 2014;13(1):57-65. doi:

10.1517/14740338.2013.827660

7. World Health Organization. Medication Safety in Polypharmacy:

Technical Report. Geneva: World Health Organization; 2019.

8. Blenkinsopp A, Bond C, Raynor DK. Medication reviews. Br J Clin

Pharmacol. 2012;74(4):573-580. doi:10.1111/j.1365-2125.2012.

04331.x

9. Huiskes VJB, Burger DM, van den Ende CHM, van den Bemt BJF.

Effectiveness of medication review: a systematic review and meta-

analysis of randomized controlled trials. BMC Fam Pract. 2017;18:5.

doi:10.1186/s12875-016-0577-x

10. Christensen M, Lundh A. Medication review in hospitalised patients

to reduce morbidity and mortality. Cochrane Database Syst Rev. 2016;

2:CD008986. doi:10.1002/14651858.CD008986.pub3

11. Laursen J, Kornholt J, Betzer C, Petersen TS, Christensen MB. Gen-

eral practitioners’ barriers toward medication reviews in

polymedicated multimorbid patients: How can a focus on the pharma-

cotherapy in an outpatient clinic support GPs? Health Serv Res Manag

Epidemiol. 2018;5:1–7. doi:10.1177/2333392818792169 12. Bokhof B, Junius-Walker U. Reducing polypharmacy from the

perspectives of general practitioners and older patients: A synthesis

of qualitative studies. Drugs Aging. 2016;33(4):249-266. doi:10.1007/

s40266-016-0354-5

13. Beuscart J-B, Pont LG, Thevelin S, et al. A systematic review of the

outcomes reported in trials of medication review in older patients:

the need for a core outcome set. Br J Clin Pharmacol. 2017;83(5):

942-952. doi:10.1111/bcp.13197

14. Renaudin P, Boyer L, Esteve M, et al. Do pharmacist-led medication

reviews in hospitals help reduce hospital readmissions? A systematic

review and meta-analysis. Br J Clin Pharmacol. 2016;82(6):1660-1673.

doi:10.1111/bcp.13085

15. Holland R, Desborough J, Goodyer L, Hall S, Wright D, Loke YK. Does

pharmacist-led medication review help to reduce hospital admissions

and deaths in older people? A systematic review and meta-analysis.

Br J Clin Pharmacol. 2008;65(3):303-316. doi:10.1111/j.1365-2125.

2007.03071.x

16. Almutairi H, Stafford A, Etherton-Beer C, Flicker L. Optimisation of

medications used in residential aged care facilities: a systematic

review and meta-analysis of randomised controlled trials. BMC

Geriatr. 2020;20:236. doi:10.1186/s12877-020-01634-4

17. Alldred DP, Raynor DK, Hughes C, et al. Interventions to optimise

prescribing for older people in care homes. Cochrane Database Syst

Rev. 2013;CD009095. doi:10.1002/14651858.CD009095.pub2

18. Abbott RA, Moore DA, Rogers M, Bethel A, Stein K, Coon JT. Effec-

tiveness of pharmacist home visits for individuals at risk of

medication-related problems: a systematic review and meta-analysis

of randomised controlled trials. BMC Health Serv Res. 2020;20:39. doi:

10.1186/s12913-019-4728-3

19. WHO Collaborating Centre for Drug Statistics Methodology. ATC

Classification Index with DDDs. Oslo, Norway: Oslo, Norway; 2020.

20. EuroQol Group. EuroQol – a new facility for the measurement of

health-related quality of life. Health Policy. 1990;16(3):199-208. doi:

10.1016/0168-8510(90)90421-9

21. Ravn-Nielsen LV, Duckert M-L, Lund ML, et al. Effect of an in-hospital

multifaceted clinical pharmacist intervention on the risk of

readmission. JAMA Intern Med. 2018;178(3):375-382. doi:10.1001/

jamainternmed.2017.8274

22. Jensen CE, Sørensen SS, Gudex C, Jensen MB, Pedersen KM,

Ehlers LH. The Danish EQ-5D-5L Value Set: A hybrid model using

cTTO and DCE data. Appl Health Econ Health Policy. 2021;19(4):

579-591. doi:10.1007/s40258-021-00639-3

23. van Hout B, Janssen MF, Feng Y-S, et al. Interim scoring for the EQ-

5D-5L: mapping the EQ-5D-5L to EQ-5D-3L value sets. Value Health.

2012;15(5):708-715. doi:10.1016/j.jval.2012.02.008

24. Aagaard L, Kristensen MB. The national drug interactions database.

Ugeskr Laeger. 2005;167(35):3283-3286.

25. O'Mahony D, O'Sullivan D, Byrne S, O'Connor MN, Ryan C,

Gallagher P. STOPP/START criteria for potentially inappropriate pre-

scribing in older people: version 2. Age Ageing. 2015;44(2):213-218.

doi:10.1093/ageing/afu145

26. Sumukadas D, McMurdo MET, Mangoni AA, et al. Temporal trends in

anticholinergic medication prescription in older people: repeated

cross-sectional analysis of population prescribing data. Age Ageing.

2014;43(4):515-521. doi:10.1093/ageing/aft199

27. Carnahan RM, Lund BC, Perry PJ, Pollock BG, Culp KR. The Anticho-

linergic Drug Scale as a measure of drug-related anticholinergic bur-

den: associations with serum anticholinergic activity. J Clin Pharmacol.

2006;46(12):1481-1486. doi:10.1177/0091270006292126

28. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG.

Research electronic data capture (REDCap) – A metadata-driven

methodology and workflow process for providing translational

research informatics support. J Biomed Inform. 2009;42(2):377-381.

doi:10.1016/j.jbi.2008.08.010

29. R Core Team. R: A Language and Environment for Statistical Computing.

Vienna, Austria; 2019.

30. Gerlinger C, Bamber L, Leverkus F, et al. Comparing the EQ-5D-5L

utility index based on value sets of different countries: impact on the

interpretation of clinical study results. BMC Res Notes. 2019;12(1):18.

doi:10.1186/s13104-019-4067-9

31. Walters SJ, Brazier JE. Comparison of the minimally important differ-

ence for two health state utility measures: EQ-5D and SF-6D. Qual

Life Res. 2005;14(6):1523-1532. doi:10.1007/s11136-004-7713-0

32. Champely S. pwr: Basic Functions for Power Analysis. R package

version 1.3-0; 2020. https://CRAN.R-project.org/package=pwr

33. Liu GF, Lu K, Mogg R, Mallick M, Mehrotra DV. Should baseline be a

covariate or dependent variable in analyses of change from baseline

in clinical trials? Stat Med. 2009;28(20):2509-2530. doi:10.1002/

sim.3639

34. Pinheiro J, Bates D, DebRoy S, Sarkar D, R Core Team. nlme: Linear

and Nonlinear Mixed Effects Models. R package version 3.1-150;

2020. https://CRAN.R-project.org/package=nlme

35. Bae E, Choi S-E, Lee H, Shin G, Kang D. Validity of EQ-5D utility

index and minimal clinically important difference estimation among

3368 KORNHOLT ET AL.

https://orcid.org/0000-0003-2662-4519

https://orcid.org/0000-0002-9974-2738

https://orcid.org/0000-0002-8774-1797

info:doi/10.1186/1472-6904-10-16

info:doi/10.1001/jamainternmed.2015.8581

info:doi/10.1186/s12877-017-0621-2

info:doi/10.1111/j.1532-5415.2006.00889.x

info:doi/10.1517/14740338.2013.827660

info:doi/10.1111/j.1365-2125.2012.04331.x

info:doi/10.1186/s12875-016-0577-x

info:doi/10.1002/14651858.CD008986.pub3

info:doi/10.1177/2333392818792169

info:doi/10.1007/s40266-016-0354-5

info:doi/10.1111/bcp.13197

info:doi/10.1111/bcp.13085

info:doi/10.1111/j.1365-2125.2007.03071.x

info:doi/10.1186/s12877-020-01634-4

info:doi/10.1002/14651858.CD009095.pub2

info:doi/10.1186/s12913-019-4728-3

info:doi/10.1016/0168-8510(90)90421-9

info:doi/10.1001/jamainternmed.2017.8274

info:doi/10.1007/s40258-021-00639-3

info:doi/10.1016/j.jval.2012.02.008

info:doi/10.1093/ageing/afu145

info:doi/10.1093/ageing/aft199

info:doi/10.1177/0091270006292126

info:doi/10.1016/j.jbi.2008.08.010

info:doi/10.1186/s13104-019-4067-9

info:doi/10.1007/s11136-004-7713-0

https://CRAN.R-project.org/package=pwr

info:doi/10.1002/sim.3639

https://CRAN.R-project.org/package=nlme

patients with chronic obstructive pulmonary disease. BMC Pulm Med.

2020;20:73. doi:10.1186/s12890-020-1116-z

36. Parker SL, Adogwa O, Paul AR, et al. Utility of minimum clinically

important difference in assessing pain, disability, and health state

after transforaminal lumbar interbody fusion for degenerative lumbar

spondylolisthesis: Clinical article. J Neurosurg Spine. 2011;14(5):

598-604. doi:10.3171/2010.12.SPINE10472

37. Samsa G, Edelman D, Rothman ML, Williams GR, Lipscomb J,

Matchar D. Determining clinically important differences in health

status measures. Pharmacoeconomics. 1999;15(2):141-155. doi:

10.2165/00019053-199915020-00003

38. Romskaug R, Skovlund E, Straand J, et al. Effect of clinical geriatric

assessments and collaborative medication reviews by geriatrician and

family physician for improving health-related quality of life in home-

dwelling older patients receiving polypharmacy. JAMA Intern Med.

2020;180(2):181-189. doi:10.1001/jamainternmed.2019.5096

39. Sintonen H. The 15D instrument of health-related quality of life:

properties and applications. Ann Med. 2001;33(5):328-336. doi:

10.3109/07853890109002086

SUPPORTING INFORMATION

Additional supporting information may be found in the online version

of the article at the publisher's website.

How to cite this article: Kornholt J, Feizi ST, Hansen AS, et al.

Effects of a comprehensive medication review intervention on

health-related quality of life and other clinical outcomes in

geriatric outpatients with polypharmacy: A pragmatic

randomized clinical trial. Br J Clin Pharmacol. 2022;88(7):

3360-3369. doi:10.1111/bcp.15287

KORNHOLT ET AL. 3369

info:doi/10.1186/s12890-020-1116-z

info:doi/10.3171/2010.12.SPINE10472

info:doi/10.2165/00019053-199915020-00003

info:doi/10.1001/jamainternmed.2019.5096

info:doi/10.3109/07853890109002086

info:doi/10.1111/bcp.15287

Effects of a comprehensive medication review intervention on health-related quality of life and other clinical outcomes in ...

1 INTRODUCTION
2 METHODS

2.1 Participants

What is already known about this subject
What this study adds

2.2 Outcomes, blinding and data collection
2.3 Usual care and the intervention
2.4 Randomization
2.5 Sample size
2.6 Statistical methods

3 RESULTS
4 DISCUSSION

4.1 Strengths and limitations

5 CONCLUSION
ACKNOWLEDGEMENTS
COMPETING INTERESTS
CONTRIBUTORS

DATA AVAILABILITY STATEMENT

REFERENCES

image3.emf

Age and Ageing 2026; 55: afag209 https://doi.org /10.1093/ageing /afag209

© The Author(s) 2026. Published by Oxford University Press on behalf of the British Geriatrics Society. This is an Open Access article distributed under the terms of the Creative Commons Attribution License

(https://creativecommons.org /licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.

CLINICAL TRIAL

Deprescribing in older patients with hyperpolypharmacy: a cluster-randomised trial in primary care

GERT BAAS 1 , 2 , METTE HERINGA 1, SANNE VERDOORN 1, HENK-FRANS KWINT 1, EMAN BADAWY 1, JACOBIJN GUSSEKLOO 3 , JAKO BURGERS 4 , 5, PETRA DENIG 6, MARCEL L. BOUVY 2

1SIR Institute for Pharmacy Practice and Policy, Leiden, ZH, The Netherlands

2Department of Pharmaceutical Sciences, Utrecht University, Utrecht 125, The Netherlands

3Public Health and Primary Care, Leiden University Medical Center, Leiden, The Netherlands

4Department of Family Medicine, Maastricht University, Maastricht, LI, The Netherlands

5Dutch College of General Practitioners (NHG), Utrecht, The Netherlands

6Clinical Pharmacy & Pharmacology, University of Groningen, University Medical Center Groningen, PO Box 30001 EB71, Groningen 9700RB, The Netherlands

Address correspondence to: Gert Baas, SIR Institute for Pharmacy Practice and Policy, Leiden, ZH, The Netherlands. Email: [email protected]

Abstract

Background: Polypharmacy increases the risk of adverse drug events and potentially preventable hospital admissions. Depre- scribing may reduce medication-related harm, and a clinical medication review (CMR) provides an opportunity to support this. Objective: To investigate the effect of a deprescribing-focused CMR by trained pharmacists on the number of reduced and stopped medications among patients aged ≥75 years with hyperpolypharmacy using multidose-drug-dispensing (MDD)- systems. Methods: In this cluster-randomised controlled trial, pharmacists were trained to perform a deprescribing-focused CMR, while control pharmacies provided usual care. Eligible patients were aged ≥75 years, had hyperpolypharmacy and used MDD. Dispensing data were used to determine the number of reduced and stopped medications per patient after 6 months (primary outcome). Secondary outcomes were: total number of medications, health problems and quality-of-life (EQ-5D-5L, EQ-VAS). A generalised linear mixed model was used for the primary outcome. Results: A total of 318 patients from 58 pharmacies were enrolled (155 intervention, 163 control). At 6 months, patients in the intervention group had significantly more medications deprescribed than those in the control group (mean 2.5 vs. 1.7 per patient; mean difference 0.79, 95% CI 0.29–1.28; P = .002). There was no significant between-group difference in the total number of medications at 6 months (intervention − 0.30 vs. control +0.12; P = .108). No significant differences were observed in health problems or quality-of-life. Conclusion: Deprescribing-focused CMRs by trained pharmacists successfully increased reduction and stopping of medica- tions in older patients with hyperpolypharmacy using MDD-systems, while no changes in health problems or health-related quality of life could be observed over 6 months.

Keywords: deprescribing; (hyper)polypharmacy; older people; medication review; randomised controlled trial

https://creativecommons.org/licenses/by/4.0/

https://orcid.org/0000-0002-3159-509X

https://orcid.org/0000-0001-7186-8278

https://orcid.org/0000-0002-4596-0684

401 34311 a 401 34311 a

mailto:[email protected]

G. Baas et al.

Key Points • A pharmacist-led deprescribing intervention increased medication discontinuation and reduction in older people. • A deprescribing toolbox and training supports medication reviews to reduce and stop medication in older people. • Despite deprescribing, total medication counts did not differ between intervention and control groups after 6 months.

Introduction

More than half of older people live with multiple chronic conditions, leading to long-term use of multiple medications, often defined as polypharmacy (≥5 medications) or hyper- polypharmacy (≥10 medications) [ 1, 2]. While medications can be appropriate by improving health and preventing com- plications, polypharmacy is associated with increased hos- pital admissions, morbidity and mortality [ 3– 5]. The risk arises from age-related physiological changes, comorbidities and complex medication regimens [ 6], and is exacerbated by inappropriate prescribing. In primary care, potentially inap- propriate prescribing among older patients was identified in about one-third of patients [ 7– 9]. In the Netherlands, one in five patients aged 75 years and older use multidose drug dispensing (MDD) systems [ 10] to support medica- tion adherence, particularly among those with complex reg- imens and limited capacity for self-management [ 11]. MDD users represent a vulnerable group, as they generally use more medicines leading to hyperpolypharmacy [ 12, 13] and have a relatively high number of drug-related problems (DRPs) [ 9, 14]. Studies have also reported a poorer quality of drug treatment [ 12, 15– 17] and fewer changes in ongoing therapy among MDD users compared with patients receiving their medicines through standard dispensing, suggesting a reduced reconsideration of treatment in this population [ 13]. Subop- timal prescribing and overtreatment remain concerns, high- lighting the need for regular and collaborative medication review in this population [ 15, 16].

Given these risks, strategies to safely reduce overall medication use are needed for groups at highest risk. Deprescribing is the planned, health care provider (HCP)- supervised dose reduction or stopping of a medication [ 18]. Over the past decade, deprescribing has gained substantial international attention as a promising approach to address the risks of polypharmacy. Evidence shows that deprescribing is generally safe and can reduce medication use: a review of randomised trials on deprescribing interventions in older patients with polypharmacy reported that almost all interventions (13/14) reduced medication counts without adverse effects, with several also leading to improvements in health-related quality of life, reduced health care costs or fewer hospitalisations [ 19]. Furthermore, patient-centred deprescribing interventions achieve persistent reductions in drug use, with most discontinued medications remaining stopped after 12 months [ 20]. Among older people with frailty, studies suggest that deprescribing is safe and feasible, with reductions in the number of medications and potentially inappropriate medications [ 21]. So far, no studies have

assessed the impact of deprescribing initiatives among older patients with hyperpolypharmacy using MDD systems.

Deprescribing can be integrated into a CMR, a well- established intervention in primary care that brings together general practitioners (GP), community pharmacists (CP) and patients. Its design inherently supports deprescribing: prescribers and pharmacists jointly assess the appropriateness of current pharmacotherapies, beginning with a patient consultation to identify current health status, goals and preferences. Previous trials in the Netherlands demonstrated that CMRs can improve patient-reported outcomes in older patients with polypharmacy [ 22]. However, they did not explicitly focus on deprescribing, leaving it unclear whether a more targeted approach could be more effective. To explore this, we conducted a mixed-methods feasibility study of a deprescribing-focused CMR among older MDD users with hyperpolypharmacy [ 23]. The study showed that the intervention was acceptable to both patients and HCPs, and that deprescribing recommendations were implemented in most patients.

This study aimed to determine whether a deprescribing- focused CMR by trained pharmacists increases the number of reduced and stopped medications after 6 months, com- pared with usual care in older patients with hyperpolyphar- macy using MDD systems.

Methods

Study setting and design

A pragmatic cluster-randomised controlled trial was con- ducted in 58 community pharmacies in the Netherlands comparing a CMR focused on deprescribing with usual care. Pharmacies were recruited between March 2022 and Novem- ber 2022 via mailings and information letters distributed through an academic pharmacy network [ 24]. Before patient recruitment started, pharmacies were randomised in 2 arms (29 intervention, 29 control) by an independent researcher Adrianne Faber (AF), who was not otherwise involved in study conduct, using block randomisation with a block size of four to obtain equal numbers of pharmacies per group. Pharmacies within the same pharmacotherapy audit meeting (PTAM) local network were placed in the same block to avoid contamination. Because a PTAM network in the Netherlands rarely exceeds four pharmacies, the maximum block size was set at four. Due to the nature of the intervention, blinding of HCPs and patients was not possible. The random allocation sequence was generated and held exclusively by the AF and was not accessible to recruiting

Deprescribing in older patients with hyperpolypharmacy

personnel. Formal blinding of outcome assessors was also not applied. CPs in the intervention group received a one-day face-to-face training. CPs in the control group only attended a 1-hour online session solely focused on explaining study procedures and patient recruitment, ensuring consistency across study groups. Available pharmacy and pharmacist characteristics are presented in Supplementary Appendix I. Patient and public involvement informed the development of the intervention and study materials. Representatives from patient advocacy organisations (Zorgbelang Inclusief) and an older persons’ advisory panel (Ouderenberaad ZHN) provided feedback on study procedures and patient-facing materials, which were refined accordingly.

Study participants

Eligible patients were community-dwelling adults aged ≥75 years using an MDD system and ≥10 chronic med- ications (hyperpolypharmacy). Lists of potentially eligible patients were generated by CPs from their pharmacy information system and screened for exclusion criteria in collaboration with GPs. Patients were excluded by either CP or GP, using routinely documented information in the medical records, if they had: (i) an estimated life expectancy of ≤6 months, (ii) cognitive impairment (as the intervention relied on patient participation in consultations and shared deprescribing decision-making), (iii) residence in nursing homes, (iv) CMR within the past 12 months and (v) repeat prescriptions issued exclusively by a hospital specialist. Patients with a CMR within the past 12 months were excluded to avoid overlap with recently implemented medication optimisation interventions that could influence baseline medication use and reduce the ability to detect effects attributable to the study intervention.

Following pharmacy randomisation, eligible patients were recruited between October 2022 and March 2024. Patients were approached by telephone by CPs who had been instructed to contact patients without applying any fixed order (e.g. alphabetical or first-listed) or reviewing additional patient information to select patients. Those expressing interest received an information package, including an informed consent form, which they returned directly to the research team. Upon receipt, the researchers informed the relevant CPs of the patient’s inclusion. Participating patients were followed for 6 months.

Multilevel intervention

The intervention targeted both HCPs and patient level and consisted of complementary implementation and care com- ponents.

Deprescribing training and implementation support (HCP level)

The one-day face-to-face training for the intervention group was developed based on previous training programs [ 25] and the Dutch deprescribing guideline module to support

deprescribing-focused CMRs [ 26]. The training focused on: (i) the organisation of CMR focusing on deprescribing; (ii) communication skills to support deprescribing consul- tations; (iii) the application of recently developed drug- specific deprescribing factsheets in practice [ 26]. The training combined educational methods including interactive plenary teaching, case-based learning, role-play, peer discussion and feedback using practical polypharmacy cases relevant to older adults. During the training, the CPs received a toolbox with the following materials: (i) ten drug-specific deprescribing fact sheets; (ii) an information letter for GPs and (iii) a PowerPoint presentation to be used in a local PTAM, covering the same aspects as in the CPs’ training and (iv) consultation aid (A4) with tips and tricks for deprescribing consultations.

Deprescribing-focused clinical medication review (patient level)

The deprescribing-focused CMR was based on the Dutch multidisciplinary guideline on polypharmacy in older people (2019) and the deprescribing module (2020) as part of the guideline [ 26, 27]. The CMR followed a five-step process ( Figure 1): (i) patient interview identifying current health problems, goals and preferences, (ii) pharmacotherapeutic analysis using the guideline module, (iii) discussion and consensus with the GP, (iv) discussion and consensus with the patient about a pharmaceutical care plan and implementation of actions and (v) follow-up monitoring of actions and effects through patient interaction. Before the CMR took place, patients completed questionnaires on their health-related complaints and preferences.

Usual care

Patients in the control group received usual care, and no train- ing or deprescribing toolbox was provided to HCPs. Within the framework of usual care, HCPs retained the discretion to perform a CMR whenever deemed appropriate.

Outcome measures and data collection

The primary outcome was the number of medications stopped or reduced per patient at 6 months. Dispensing data over a 24-month period, covering a maximum of 18 months before and at least 6 months after inclusion for all study patients, were provided by Dutch Foundation for Pharmaceutical Statistics (Stichting Farmaceutische Kengetallen: SFK). SFK maintains a comprehensive national database of dispensing data provided by over 98% of community pharmacies and reflecting medication exposure in a population of approximately 16.5 million individuals (www.sfk.nl). For the analysis, deprescribing was defined as stopping a medication, reducing its dose or substituting it for an alternative with lower potency or reduced complexity, observed after 6 months using a stepwise approach applied for all medication classes. Dose reductions and substitutions to less potent or less complex alternatives were both categorised

https://academic.oup.com/ageing/article-lookup/doi/10.1093/ageing/afag209#supplementary-data

www.sfk.nl

G. Baas et al.

Figure 1. Five step approach of a CMR according to the Dutch multidisciplinary guideline ‘Polypharmacy in the Elderly’ [ 27] with special focus on deprescribing. The CMR begins with a consultation (step 1) between the patient and the CP, using input from the questionnaires to understand the patient’s pref- erences and health-related issues. This includes identifying any medications the patient prefers to continue or discontinue and any adverse effects they may be experiencing. This is followed by a pharmacotherapeutic analysis by the CP to identify potential DRPs and opportunities for deprescribing (step 2). The CP and GP then discuss the identified DRPs, patient preferences and possible actions, including deprescribing (step 3). If needed, the CP or GP consults other HCPs. Agreed deprescribing propos- als were subsequently discussed with the patient and incorpo- rated into a pharmaceutical care plan (step 4). Implementation took place through routine procedures. All deprescribing actions will be monitored through follow-up (step 5), which includes patient interaction by either the CP, nurse practitioners or GPs, and with interprofessional communication as needed, to adjust tapering based on the patient’s health or withdrawal symptoms.

as medication reduction. At the ATC-5 level, medications were checked for continuation. If present after 6 months, the daily dose was assessed to identify dose reductions. If not present, substitutions within the same therapeutic group (ATC-2 level) were identified. In such cases, equivalence and potency were evaluated according to the deprescribing factsheet [ 26] (or, when absent, primary care treatment guidelines), documenting lower potency as ‘reduced’. In the absence of guidelines, changes in medication complexity (e.g. number of active substances, number of devices and dosing schedules) were assessed to determine whether the substitution represented ‘reduced’, ‘similar’ or ‘increased’ medication use. If no substitution was identified, the medication was documented as ‘stopped’. Two researchers (E.B. and G.B.) independently reviewed dispensing records to identify deprescribing using predefined criteria applied on dispensing data obtained from the national SFK database. The criteria were based on the Dutch deprescribing fact

sheets and guidelines from the Dutch College of General Practitioners (Nederlands Huisartsen Genootschap: NHG).

Secondary outcomes included the total number of medi- cations in use, the number and type of health problems with an impact on daily life [ 22, 28], health-related quality of life measured with the EQ-5D-5L and the EQ Visual Analogue Scale (VAS), the number and type of DRPs identified during CMR, and the implementation rate of deprescribing-related recommendations. Health problems and health-related quality of life were assessed through patient questionnaires at baseline and 6 months. Demographics, the Integrated Sys- tematic Care for Older People (ISCOPE) questionnaire [ 29] and Self-Perceived Burden Scale Questionnaire questionnaire [ 30] were completed at baseline to evaluate the presence of complex problems, and inadequate health literacy. Question- naires were completed by telephone interviews, on paper or via a digital link sent by email, depending on patient’s pref- erence. All responses were entered into CASTOR Electronic Data Capture (CASTOR EDC) by the researchers.

CMR data were documented by CPs in CASTOR using Hepler and Strand’s classification of DRPs [ 31], including proposed actions and their implementation status. Two researchers (E.B. and G.B.) independently checked CMR data for completeness and consistency, resolving disagreements with CPs and/or a third researcher.

Sample size

The sample size calculation was based on detecting a difference of one deprescribed medication per patient after 6 months, assuming a standard deviation of 2.5, a two-sided α of 0.05, power of 80%, an intra-cluster correlation coefficient (ICC) of 0.05 and an average cluster size of seven patients. This required 38 clusters (266 patients). To allow for an expected 30% patient dropout, the target was increased to 10 patients per pharmacy. After adjusting for a 20% potential pharmacy dropout, the total target was 46 pharmacies.

Statistical analyses

Baseline characteristics were compared between the interven- tion and control group. Categorical variables were analysed using Pearson’s chi-square test. Continuous variables were assessed for normality using visual inspection of histograms and the Shapiro–Wilk test; as all continuous baseline variables were non-normally distributed, between-group differences were evaluated using the Mann–Whitney U test, and results are presented as medians with interquartile ranges (IQR).

Analyses were based on intention-to-treat, including all randomised patients with at least 4 months of medication dispensing data after baseline; patients with less dispensing data were excluded. A per-protocol analysis was additionally performed, excluding patients who, based on dispensing data verification, were found not to meet the predefined inclusion criterion for hyperpolypharmacy (≥10 chronic medications at baseline) and, within the intervention group, those who did not receive the intervention. The primary outcome (the number of medications stopped or reduced per patient at

Deprescribing in older patients with hyperpolypharmacy

6 months) was estimated using a Generalised Linear Mixed Model (GLMM) with a linear distribution, including inter- vention group as a fixed effect and a random intercept for cluster (pharmacy) to account for intra-cluster correlation. No covariates were included in the primary model. To assess whether the intervention effect differed across patient char- acteristics, we evaluated effect modification by sequentially adding prespecified covariates (sex, age, baseline medication count and ISCOPE complexity score) and their interaction terms with the intervention to the GLMM. Residuals at both patient and cluster level were assessed for normality using the Shapiro–Wilk test and visual inspection of quantile–quantile (QQ) plots and histograms.

Secondary outcomes were analysed using GLMMs. For questionnaire-based outcomes assessed at baseline and 6 months (number of health problems, EQ-5D-5L and EQ- VAS), models included fixed effects for intervention, time and their interaction, with baseline values as covariates, following an analysis of covariance (ANCOVA) approach. Time (in months) was included to adjust for potential temporal trends over the 6-month follow-up. Outcomes derived from the CMR process or dispensing data (DRPs, implementation of deprescribing-related recommendations and medication changes) were analysed using GLMMs including intervention as a fixed effect and cluster as a random effect, without inclusion of time.

When a statistically significant interaction with interven- tion was identified, pairwise contrasts were performed to estimate the intervention effect to obtain the corresponding P -values.

All statistical analyses were performed in IBM SPSS Statistics version 29.0.1.0 (171) (IBM Corporation, Armonk, NY, USA).

Ethics and confidentiality

The Medical Research Ethics Committee NedMec deter- mined that the study (21/672) was not subject to the Medical Research Involving Human Subjects Acts. The study protocol UPF2211 was approved by the Institutional Review Board of UPPER. All patients gave written informed consent. To ensure patient privacy, all data were anonymised. Data collection was conducted using CASTOR, a secure online data management platform. The trial was registered at ClinicalTrials.gov (ID: NCT05609981), where the study protocol and statistical analysis plan can be accessed. A completed CONSORT extension checklist for cluster randomised trials is provided in Supplementary Appendix II.

Results

A total of 58 community pharmacies were enrolled in the study, with 29 randomised to the intervention group and 29 to the control group ( Figure 2). In the intervention arm, 28 CPs completed the one-day deprescribing training, while 26 CPs in the control arm received online instructions.

After training the participating CPs, 244 patients in the intervention group and 287 patients in the control

group received an information package by post after being contacted by telephone by their CP. Of these, 89 patients in the intervention group and 123 in the control group declined participation. This resulted in 155 patients in the intervention group and 163 in the control group providing written informed consent (range per pharmacy: 1–18). Patient baseline characteristics for both groups are presented in Table 1. Over the 6-month follow-up, ten patients in the intervention group (nine deceased, one moved) and four in the control group (three deceased, one hospitalised) were lost to follow-up ( Figure 2), resulting in dropout rates of 6.5% and 2.5%, respectively. Most deaths in the intervention group occurred in patients who had not yet had any medications deprescribed or had not even undergone a CMR ( Supplementary Appendix III), making a causal relationship unlikely.

A total of 304 patients completed the study, providing data for the intention-to-treat analysis. The per-protocol anal- ysis was based on 240 patients, after excluding 64 patients. Patients using fewer than 10 medications at baseline were excluded (n = 54; 21 in the intervention group and 33 in the control group). In the intervention group, 10 additional patients were excluded because the planned CMR was not performed.

Primary outcome

At 6 months, the mean difference in the number of depre- scribed medications was 0.79 (95% confidence interval (CI) 0.29–1.28) in favour of the intervention (P = .002; Table 2). This reflects a mean of 2.5 medications stopped or reduced per patient in the intervention group (20.7% of baseline medications) compared with 1.7 medications (14.4%) in the control group ( Table 2). Per-protocol analysis, restricted to patients who fully met inclusion criteria and, within the intervention group, those who did receive the intervention (n = 240), yielded a similar difference of 0.76 medications (P = .008). The ICC for the primary outcome was 0.12.

In effect modification analyses, a higher baseline medi- cation count was associated with more deprescribing; each additional baseline medication corresponded to an average increase of 0.23 deprescribed medications after 6 months (95% CI 0.17 –0.29, P < .001). The interaction between base- line medication count and intervention was not statistically significant (β = 0.06, 95% CI − 0.07 to 0.18, P = .37), indicat- ing that this association was consistent across both groups.

Secondary outcomes

There was no significant difference in the change in total number of medications between groups (intervention vs control; − 0.30 vs +0.12 medications; P = .108; Table 2). In the intervention group, an average of 1.6 medications per patient was stopped compared with 1.3 in the control group, while newly initiated medications occurred to a similar extent in both groups (1.3 vs 1.4; Table 2). Dose reduction accounted for almost one medication per patient in the intervention group versus less than half in the control group (0.9 vs 0.4), without affecting the total number

ClinicalTrials.gov

https://academic.oup.com/ageing/article-lookup/doi/10.1093/ageing/afag209#supplementary-data

G. Baas et al.

Figure 2. Flow chart study population.

of medications. Consequently, the net effect on the total number of medications remained small in both arms. Of the medications deprescribed in the intervention group,

approximately half were chronic medications dispensed in the MDD system, compared with one-third in the control group ( Table 2). In 127 of the 139 CMRs (91%) in the

Deprescribing in older patients with hyperpolypharmacy

Table 1. Characteristics of patients in the intervention and control group.

Characteristic Intervention

(n = 155)

Control

(n = 163)

P -value

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Sociodemographic

Sex, female, n (%) a 89 (57) 85 (52) .345

Age, years, median (IQR) b 82 (78–86) 82 (79–86) .188 Age, category, n (%) 75–79 57 (37) 48 (29) .370

80–84 52 (34) 63 (39)

>85 46 (30) 52 (32)

Migration background (yes) c, n (%) 23 (16) 16 (11) .187

Complex health problems (ISCOPE >3) c, n (%) 65 (45) 76 (51) .315

Low health literacy (SBS-Q < 2) d, n (%) 42 (29) 42 (28) .823

Medication-related

Number of medications, median (IQR) b 11.0 (10.0–14.0) 12.0 (10.0–14.0) .472

Number of medications in MDD, median (IQR) b 9.0 (7.0–10.0) 9.0 (7.0–10.0) .547

Top 10 medication classes (ATC2 level), n (%)

A02 Drugs for acid-related disorders 147 (95) 151 (93)

B01 Antithrombotic agents 134 (86) 146 (90)

C10 Lipid-modifying agents 124 (80) 117 (72)

C07 Beta-blocking agents 111 (72) 112 (69)

C09 Agents acting on the renin-angiotensin

system

110 (71) 103 (63)

C03 Diuretics 98 (63) 111 (68)

A10 Drugs used in diabetes 84 (54) 81 (50)

A11 Vitamins 68 (44) 50 (31)

C08 Calcium channel blockers 66 (43) 68 (42)

C01 Cardiac therapy 62 (40) 67 (41)

a Analysed using Pearson’s chi-square test. b Analysed using the Mann–Whitney U test. c 25 missing (14 control, 11 intervention). d 26 missing (14 control, 12

intervention).

Table 2. Medication use and effects on reduced and stopped medications (n = 304).

Outcome measure Intervention Control

Mean (SD) % Mean (SD) % . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Number of medications

Baseline (t 0 ) 12.1 (2.6) n/a 11.8 (2.8) n/a

Follow-up (t 6 ) 11.8 (2.6) 11.9 (3.0)

� (t 6 − t 0 ) a − 0.30 (1.6) 0.12 (2.0)

Medications reduced and/or stopped (% relative to baseline)

Reduced & stopped b 2.5 (1.7) 20.7 1.7 (1.6) 14.4

Reduced c 0.9 (1.0) 7.4 0.4 (0.7) 3.4

Stopped 1.6 (1.3) 13.2 1.3 (1.4) 11.0

Number of (non)-MDD medications deprescribed (% of total ‘reduced & stopped’)

MDD medications 1.3 (1.2) 52.0 0.6 (1.0) 35.3

Non-MDD medications 1.2 (1.3) 48.0 1.1 (1.1) 64.7

Number of medications started

Started 1.3 (1.2) n/a 1.4 (1.5) n/a

a Between-group difference in change scores ( � t

6 − t 0 ) analysed using a GLMM with cluster (pharmacy) as random intercept, adjusted for baseline number of

medications (β = − 0.35, 95% CI − 0.78 to 0.08, P = .108). b Between-group difference in the number of medications reduced and/or stopped analysed using a

GLMM with cluster (pharmacy) as random intercept (β = 0.79, 95% CI 0.29 to 1.28, P = .002). c Medication reduction included dose reductions and lower-burden

substitutions. Such substitutions occurred in broadly similar proportions in the intervention and control groups (0.15 vs. 0.11 changes per patient; 22/145 vs. 18/159

patients).

intervention group, at least one medication was reduced or stopped. Deprescribing was observed across all intervention pharmacies, with a median cluster mean of 2.4 deprescribed medications per patient (IQR 2.0–3.0; range 1.0–3.75).

The number of health problems affecting daily life did not differ between groups at baseline or 6 months ( Table 3). The

most frequently reported problems were mobility issues (65% control, 71% intervention at baseline), fatigue (61% and 60%) and pain (49% and 56%). At 6 months, these issues remained the most prevalent, with only minor and comparable changes between groups. The change in EQ-VAS scores tended to be higher in the intervention group (mean change +4.1 vs +0.9;

G. Baas et al.

Table 3. Mean number of health problems with impact on daily life at baseline and follow-up.

Number of health problems with impact Intervention Control P -value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Baseline (t 0 ), mean (SD) 4.6 (2.9) 4.5 (2.9)

Follow-up (t 6 ), mean (SD) 4.4 (3.2) 4.5 (3.0)

� (t 6 − t 0 ), mean (SD) − 0.14 (2.3) − 0.04 (2.1) .229 a

Baseline: 25 missing (11 intervention, 14 control); follow-up: 62 missing (37 intervention, 25 control). Analyses were based on all available data. a P -value for the

pairwise contrast of the estimated means at follow-up derived from the GLMM (mean difference − 0.47, 95% CI − 1.23 to 0.30); model including cluster as a random

effect and fixed effects for intervention, time (months), baseline number of health problems, and the time × intervention interaction.

Table 4. Effect of the intervention on change in quality of life (EQ-VAS, EQ-5D-5L).

Quality of life measure Intervention Control P -value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

EQ-VAS

Baseline (t 0 ), mean (SD) 63.1 (17.4) 62.9 (15.7)

Follow-up (t 6 ), mean (SD) 67.5 (16.1) 63.6 (15.0)

� (t 6 − t 0 ), mean (SD) 4.1 (19.0) 0.9 (14.0) .097 a

EQ-5D-5L

Baseline (t 0 ), mean (SD) 0.669 (0.22) 0.698 (0.19)

Follow-up (t 6 ), mean (SD) 0.688 (0.23) 0.677 (0.20)

� (t 6 − t 0 ), mean (SD) 0.014 (0.16) − 0.017 (0.15) .071 a

Baseline: 25 missing (11 intervention, 14 control); follow-up: 62 missing (37 intervention, 25 control). Analyses were based on all available data. a P -value for the

pairwise contrast of the estimated means at follow-up derived from the GLMM (EQ-VAS: 3.58, 95% CI − 0.65 to 7.80; EQ-5D-5L: 0.038, 95% CI − 0.003 to 0.079);

model including cluster as a random effect and fixed effects for intervention, time (months), baseline value of the outcome, and the intervention × time interaction.

P = .097; Table 4) but the difference was not statistically sig- nificant. EQ-5D utility scores showed no significant change in either group.

During 139 CMRs in the intervention group, CPs registered 515 DRPs ( Supplementary Appendix IV), on average 3.7 per patient. Overtreatment accounted for 34% (n = 174), potential adverse effects for 21% (n = 108), undertreatment for 13% (n = 69) and incorrect dosing for 12% (n = 63) of total DRPs. A total of 515 action proposals were made during the CMRs ( Table 5), of which 314 (61%) were related to deprescribing. The majority concerned stopping medication (205 proposals, 65% of deprescribing- related actions) or dose reduction (99 proposals, 32%). Implementation rates were 59% and 65%, respectively. Other deprescribing-related action proposals, such as medication substitution, were less common (<3%). In total, 189 depre- scribing proposals (60%) were implemented. In 99 CMRs (71%), at least one deprescribing-related action proposal was implemented.

Deprescribing was observed across the 10 most frequently used medication classes (ATC level 2) within the MDD system and the corresponding prevalence of deprescribing ( Table 6). Among these, deprescribing was most common for drugs for acid-related disorders (A02), drugs used in diabetes (A10) and lipid-modifying agents (C10) in the intervention group. In these classes, between 11% and 28% of users had at least one medication discontinued or reduced, compared with 4%–8% in the control group. Supplementary Appendix V provides a complete overview of all deprescribed MDD medications and shows that deprescribing was largely concentrated in medication classes covered by the deprescribing factsheets [ 26].

Discussion

This study demonstrated that a deprescribing-focused CMR in older patients with hyperpolypharmacy led to an increased number of medications being discontinued or reduced at 6 months, compared with usual care. On average, patients in the intervention group had 2.5 medications deprescribed, compared with 1.7 in the control group. However, the total number of medications in use did not change. No signifi- cant changes in health problems or health-related quality of life were observed alongside the increase in deprescribing. CPs identified overtreatment and potential adverse effects as the most common DRPs, and two-thirds of deprescribing recommendations were successfully implemented.

In our trial, patients in the intervention group depre- scribed an average of 2.5 medications, of which 1.6 were completely stopped. This aligns with the broader evidence base showing that medication review interventions can slow or reverse the growth in overall medication use. A systematic review of 12 high-quality randomised trials including 1972 patients found that medication review consistently led to a greater decrease or smaller increase in the number of drugs used compared with usual care [ 32]. In contrast, a previous randomised trial reported a slight increase in medication count after 6 months, our study showed a small, non-significant mean decrease ( − 0.30 medications in the intervention group vs. +0.12 in controls) [ 22]. A considerable proportion of the deprescribing actions consisted of dose reductions, which are not reflected in the total medication count. As newly initiated prescriptions occurred to a similar extent in both groups, the small net change in medication count should therefore be interpreted

https://academic.oup.com/ageing/article-lookup/doi/10.1093/ageing/afag209#supplementary-data

Deprescribing in older patients with hyperpolypharmacy

Table 5. Number, type and implementation rate of action proposals.

Action proposal Number

n (%)

Implemented

n (%) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Stop medication 205 (40) 120 (59)

Reduce dosage 99 (19) 64 (65)

Substitute medication, e.g. lower potency or reduced treatment complexity 50 (10) 26 (52) Start medication 46 (9) 28 (61)

Referral to another healthcare provider 22 (4) 21 (96)

Other 22 (4) 16 (73)

Provide information/advice 20 (4) 19 (95)

Increase dosage 19 (4) 16 (84)

Additional monitoring of laboratory values 19 (4) 17 (90)

Discuss use/adherence 13 (3) 13 (100)

Total 515 (100) 340 (66)

Percentages in the ‘Number’ column represent the proportion of total action proposals (n = 515). Percentages in the ‘Implemented’ column represent the proportion

of implemented proposals within each category.

Table 6. Top 10 most frequently used medication classes (MDD medication only), ranked by frequency of deprescribing in the intervention group, and corresponding prevalence of deprescribing (n = 304).

ATC2 Medication class Intervention (n = 145) Control (n = 159)

Total Deprescribing Total Deprescribing

n n (%) n n (%) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

A02 Drugs for acid-related disorders 129 37 (29) 145 6 (4)

A10 Drugs used in diabetes 112 16 (14) 103 8 (8)

C10 Lipid-modifying agents 124 14 (11) 125 6 (5)

C03 Diuretics 109 14 (13) 121 8 (7)

C07 Beta-blocking agents 101 13 (13) 108 8 (7)

C01 Cardiac therapy 46 10 (22) 59 4 (7) A11 Vitamins 55 10 (18) 44 3 (7)

B01 Antithrombotic agents 115 8 (7) 126 9 (7)

C09 Agents acting on the renin–angiotensin system 102 8 (8) 102 12 (12)

C08 Calcium channel blockers 63 3 (5) 68 4 (6)

in light of these uncounted dose reductions. This pattern likely reflects both the natural progression of multimorbidity in older adults and the broader objective of the CMR to optimise pharmacotherapy rather than solely deintensify treatment.

The natural progression of multimorbidity may have lim- ited changes within groups over time, thereby reducing the magnitude of any observable effects on health problems or health-related quality of life. In addition, no significant differ- ences between groups were observed for these outcomes. Our study was not designed to detect (small) changes in health problems or quality of life, as these were secondary outcomes mainly aimed at monitoring safety and potential negative outcomes. Importantly, the purpose of deprescribing is not only to solve current problems, but also to limit drug-related risks and decrease unnecessary medication use. In that light, when medication can be deprescribed without an increase in health problems or a decrease in quality of life, this is still a favourable outcome.

Furthermore, our findings on DRPs are broadly consistent with previous national and trial-based studies, which gener- ally report an average of three to five DRPs per patient [ 22,

32– 34]. Variation across studies can largely be explained by contextual factors such as differences in patient populations, the scope and maturity of CMR practice, and the method- ological focus of the intervention. For example, studies on national CMR data [ 33] show similar overall DRP counts but lower proportions of overtreatment, whereas goal-oriented interventions such as the DREAMeR trial [ 22] identify more DRPs but relatively fewer cases of overtreatment. Against this background, the higher proportion of overtreatment in our study likely reflects the explicit focus of our intervention on deprescribing. This pattern underscores that both the total number of DRPs and the distribution across DRP categories are sensitive to contextual and intervention-specific charac- teristics.

The deprescribing process involves sequential steps: identifying candidates for dose reduction or discontinua- tion, implementing deprescribing actions and maintaining persistence over time. Randomised trials have shown variable effectiveness of deprescribing interventions, particularly in primary care, reflecting barriers at each stage. In our trial, 40% of all action proposals involved stopping a medication, whereas in routine Dutch practice fewer than 10% of

G. Baas et al.

documented interventions during CMRs concern med- ication discontinuation [ 34]. In addition, deprescribing occurred in 71% of CMRs in our study, compared with approximately one-third in routine Dutch practice [ 34]. This suggests that explicit deprescribing guidance and focus enhance both identification and deprescribing actions.

Strengths and limitations

A strength of this study is that it was conducted in rou- tine primary care and included a large sample of more than 300 older patients. By targeting patients using MDD sys- tems, deprescribing actions focused on long-term medication use could be objectively monitored through dispensing data and facilitated adherence to tapering schedules by the short dispensing intervals inherent to MDD, which ensured that scheduled dose reductions or discontinuations were imple- mented and became visible in subsequent dispensing records. Assessing deprescribing after 6 months of follow-up, rather than immediately after the intervention, provided valuable insight into sustained deprescribing. Randomisation at the cluster level minimised contamination between groups, and baseline characteristics were well balanced.

Several limitations should be acknowledged. First, recruit- ment may have been affected by selection bias, as partici- pating HCPs and patients were likely more motivated or open to deprescribing than average. However, randomisation took place after HCPs participation was confirmed, reducing the likelihood of differential motivation between groups. Second, post-randomisation recruitment by non-blinded clinicians may have introduced recruitment bias. However, all pharmacists were instructed to approach patients without applying a fixed order or reviewing additional patient information to guide invitation. Furthermore, baseline characteristics between groups were well balanced, suggesting no major imbalance resulting from recruitment. Also, findings from our preceding feasibility study indicated that, after applying eligibility and exclusion criteria, pharmacists generally had to approach the large majority of eligible patients in order to achieve the intended inclusion target, limiting the opportunity for selective recruitment [ 23]. Third, cluster randomisation at the pharmacy level may have introduced between-practice variability. Although clustering was accounted for statistically using mixed models with a ran- dom intercept for pharmacy, residual confounding at cluster level cannot be fully excluded. The ICC for the primary out- come (0.12) indicated relevant clustering at pharmacy level, supporting the use of a cluster-adjusted analytical approach. However, the study included a relatively large number of pharmacies with modest cluster sizes, and available pharmacy- level characteristics appeared broadly comparable between groups, which indicates that practice variation related to these characteristic was evenly distributed. Furthermore, although the primary outcome was based on dispensing data, secondary outcomes based on questionnaires had missing data. However, the extent was limited and consistent with expectations for this population, likely reflecting the data

collection burden and the frailty of the study population. Finally, although deprescribing guidelines were available to control group participants and deprescribing occurs as part of routine care, they were not actively supported by training. This may have led to some contamination and attenuation of the observed intervention effect, resulting in a more conservative estimate of its true impact. At the same time, findings likely underscores the importance of combining structured training at HCP level with a structured patient- level intervention when implementing such materials in clinical practice.

Implications for practice and research

Our findings demonstrate that deprescribing can be inte- grated into CMRs in primary care and effectively stop and reduce medications in an older population on MDD by trained and motivated HCPs. Earlier Dutch research showed that when CMRs include an explicit focus on deprescribing, substantially more actions are implemented and significantly more dose reductions or discontinuations occurred, but this study primarily targeted cardiometabolic medication [ 25]. This trial shows that providing training and practical tools for HCPs enhances the implementation of deprescribing and increases the proportion of related actions on a range of medications in patients with hyperpolypharmacy. This is supported by the overlap between the medication classes most frequently deprescribed ( Supplementary Appendix V) and those addressed in the training materials, suggesting that structured guidance and practical tools are essential for achieving systematic deprescribing across therapeutic domains and should be incorporated into routine CMR prac- tice. Furthermore, patients with a higher baseline medication count experienced more deprescribing overall, suggesting that those with the greatest medication burden may be partic- ularly relevant candidates for deprescribing-focused CMRs.

In this trial, health problems and health-related quality of life remained rather similar over 6 months. A lack of improve- ment on such outcomes aligns with evidence from other interventions among patients with polypharmacy and mul- timorbidity, in which measurable improvements in patient- reported outcomes were often difficult to demonstrate over relatively short follow-up periods despite meaningful changes in care processes or medication use [ 35, 36]. As deprescribing in frail older adults is often gradual and iterative, a single CMR with 6 months of follow-up may not fully capture longer-term effects [ 35]. In our study, these outcomes were included primarily as indicators of potential negative effects of deprescribing but the trial was not powered to detect differ- ences. Nonetheless, no evidence of short-term deterioration was observed indicative of adverse drug withdrawal events. Future research should evaluate the long-term outcomes of deprescribing within CMRs, including effects on clinical endpoints, healthcare utilisation and costs. In addition, process and implementation evaluations are needed to better understand how deprescribing interventions are delivered, adopted and sustained in routine care, while economic

https://academic.oup.com/ageing/article-lookup/doi/10.1093/ageing/afag209#supplementary-data

Deprescribing in older patients with hyperpolypharmacy

evaluations may help determine whether deprescribing- focused CMRs provide sufficient value relative to their time investment. Studies are also needed to assess the appropriateness of deprescribing decisions in daily practice and to explore strategies for integrating deprescribing into other pharmacist- or physician-led interventions, to support broader and more sustainable implementation in routine care. Additionally, further work could examine how training and experience influence the efficiency of deprescribing- focused CMRs, with the aim of reducing time investment while maintaining effectiveness.

Conclusion

This pragmatic cluster-randomised trial demonstrates that a deprescribing-focused CMR by trained pharmacists was effective in reducing or stopping medications in older patients with hyperpolypharmacy using MDD systems at 6 months, without a reduction in the total number of medications. No changes in patient-reported health problems or health-related quality of life could be observed over 6 months.

Acknowledgements: The authors would like to thank all participating community pharmacists, general practitioners and patients for their contribution to this study. We also acknowledge the support of students Martha Wouda and Denise Sloekers for their assistance in data collection and processing. In addition, we thank Svetlana Belitser for her statistical advice and support.

Supplementary Data: Supplementary data is available at Age and Ageing online.

Declaration of Conflicts of Interest: None declared.

Declaration of Sources of Funding: This work was supported by The Netherlands Organisation for Health Research and Development (ZonMw) as part of the pro- gramme ‘Goed Gebruik Geneesmiddelen’ [project number 10140021910504]. The funding was awarded to the study team. The funder had no role in the design of the study, data collection, analysis, interpretation of the data or writing of the manuscript.

Research Data Transparency and Availability: De- identified participant data (including the data dictionary), statistical code and supporting study materials will be made publicly available in a research repository following publication. Details regarding repository access will be provided through the ClinicalTrials.gov registration.

References

1. Salive ME. Multimorbidity in older adults. Epidemiol Rev 2013;35:75–83. https://doi.org/10.1093/epirev/mxs009.

2. Masnoon N, Shakib S, Kalisch-Ellett L et al. What is polyphar- macy? A systematic review of definitions. BMC Geriatr 2017;17:230. https://doi.org/10.1186/s12877- 017- 0621- 2.

3. Gnjidic D, Hilmer SN, Blyth FM et al. Polypharmacy cutoff and outcomes: five or more medicines were used to identify community-dwelling older men at risk of different adverse outcomes. J Clin Epidemiol 2012;65:989–95. https://doi.o rg/10.1016/j.jclinepi.2012.02.018.

4. van den Bemt PMLA, de Smet PAGM, Hek K et al. Vervolgonderzoek medicatieveiligheid: eindrapport . Utrecht, The Netherlands: NIVEL (Netherlands Institute for Health Services Research); 2017. Report. https://www.nivel.nl/site s/default/files/bestanden/Vervolgonderzoek_Medicatieveili gheid_Eindrapport.pdf (18 March 2026, date last accessed).

5. Leendertse AJ, Egberts ACG, Egberts AC et al. Frequency of and risk factors for preventable medication-related hospital admissions in the Netherlands. JAMA Intern Med 2008;144:1890–6. https://doi.org/10.1001/archinternmed. 2008.3.

6. Mangoni AA, Jackson SHD. Age-related changes in pharma- cokinetics and pharmacodynamics: basic principles and prac- tical applications. Br J Clin Pharmacol 2004;57:6–14. https:// doi.org/10.1046/j.1365- 2125.2003.02007.x.

7. Bruin-Huisman L, Abu-Hanna A, Van Weert HCPM et al. Potentially inappropriate prescribing to older patients in pri- mary care in the Netherlands: a retrospective longitudinal study. Age Ageing 2017;46:614–9. https://doi.org/10.1093/a geing/afw243.

8. Liew TM, Lee CS, Goh SKL et al. The prevalence and impact of potentially inappropriate prescribing among older persons in primary care settings: multilevel meta-analysis. Age Ageing 2020;49:570–9. https://doi.org/10.1093/ageing/afaa057.

9. Josendal AV, Bergmo TS, Granas AG. Potentially inappropriate prescribing to older patients receiving multidose drug dispens- ing. BMC Geriatr 2020;20:272. https://doi.org/10.1186/ s12877- 020- 01665- x.

10. Stichting Farmaceutische Kengetallen (SFK) . Farmacie in cijfers. 3% van de bevolking gebruikt een GDV. Pharm Weekbl 2023;50:8.

11. Mertens BJ, Kwint HF, van Marum RJ et al. Are multidose drug dispensing systems initiated for the appropriate patients? Eur J Clin Pharmacol 2018;74:1159–64. https://doi.org/10.1007/ s00228- 018- 2478- 5.

12. Belfrage B, Koldestam A, Sjöberg C et al. Prevalence of suboptimal drug treatment in patients with and without multidose drug dispensing —a cross-sectional study. Eur J Clin Pharmacol 2014;70:867–72. https://doi.org/10.1007/ s00228- 014- 1683- 0 PubMed PMID: 24801148.

13. Sjöberg C, Ohlsson H, Wallerstedt SM. Association between multi-dose drug dispensing and drug treatment changes. Eur J Clin Pharmacol 2012;68:1095–101. https://doi.o rg/10.1007/s00228- 012- 1230- 9.

14. Kwint HF, Faber A, Gussekloo J et al. Effects of medica- tion review on drug-related problems in patients using auto- mated drug-dispensing systems: a pragmatic randomized con- trolled study. Drugs Aging 2011;28:305–14. https://doi.o rg/10.2165/11586850- 000000000- 00000.

15. Johnell K, Fastbom J. Multi-dose drug dispensing and inap- propriate drug use: a nationwide register-based study of over 700 000 elderly. Scand J Prim Health Care 2008;26:86–91. https://doi.org/10.1080/02813430802022196.

https://academic.oup.com/ageing/article-lookup/doi/10.1093/ageing/afag209#supplementary-data

ClinicalTrials.gov

https://doi.org/10.1093/epirev/mxs009

https://doi.org/10.1186/s12877-017-0621-2

https://doi.org/10.1016/j.jclinepi.2012.02.018

https://www.nivel.nl/sites/default/files/bestanden/Vervolgonderzoek_Medicatieveiligheid_Eindrapport.pdf

https://doi.org/10.1001/archinternmed.2008.3

https://doi.org/10.1046/j.1365-2125.2003.02007.x

https://doi.org/10.1093/ageing/afw243

https://doi.org/10.1093/ageing/afaa057

https://doi.org/10.1186/s12877-020-01665-x

https://doi.org/10.1007/s00228-018-2478-5

https://doi.org/10.1007/s00228-014-1683-0

https://doi.org/10.1007/s00228-012-1230-9

https://doi.org/10.2165/11586850-000000000-00000

https://doi.org/10.1080/02813430802022196

G. Baas et al.

16. Sjöberg C, Edward C, Fastbom J et al. Association between multi-dose drug dispensing and quality of drug treatment—a register-based study. PloS One 2011;6:e26574. https://doi.org/10.1371/journal.pone.0026574.

17. Wallerstedt SM, Fastbom J, Johnell K et al. Drug treat- ment in older people before and after the transition to a multi-dose drug dispensing system-a longitudinal analysis. PloS One 2013;8:e67088. https://doi.org/10.1371/journal.po ne.0067088.

18. Reeve E, Gnjidic D, Long J et al. A systematic review of the emerging definition of “deprescribing” with network analy- sis: implications for future research and clinical practice. Br J Clin Pharmacol 2015;80:1254–68. https://doi.org/10.1111/ bcp.12732.

19. Omuya H, Nickel C, Wilson P et al. A systematic review of randomised-controlled trials on deprescribing outcomes in older adults with polypharmacy. Int J Pharm Pract 2023:349–68. https://doi.org/10.1093/ijpp/riad025.

20. Zechmann S, Senn O, Valeri F et al. Effect of a patient- centred deprescribing procedure in older multimorbid patients in Swiss primary care—a cluster-randomised clinical trial. BMC Geriatr 2020;20:471. https://doi.org/10.1186/ s12877- 020- 01870- 8.

21. Ibrahim K, Cox NJ, Stevenson JM et al. A systematic review of the evidence for deprescribing interventions among older people living with frailty. BMC Geriatr 2021;21:258. https:// doi.org/10.1186/s12877- 021- 02208- 8.

22. Verdoorn S, Kwint HF, Blom JW et al. Effects of a clini- cal medication review focused on personal goals, quality of life, and health problems in older persons with polyphar- macy: a randomised controlled trial (DREAMER-study). PLoS Med 2019;16:e1002798. https://doi.org/10.1371/jou rnal.pmed.1002798.

23. Baas G, Heringa M, Verdoorn S et al. A clinical medi- cation review focused on deprescribing in older patients with hyperpolypharmacy: a mixed-methods feasibility study. Basic Clin Pharmacol Toxicol 2026;138:e70184. https://doi.o rg/10.1111/bcpt.70184.

24. Koster ES, Blom L, Philbert D et al. The Utrecht pharmacy practice network for education and research: a network of community and hospital pharmacies in the Netherlands. Int J Clin Pharmacol 2014;36:669–74. https://doi.org/10.1007/ s11096- 014- 9954- 5.

25. Crutzen S, Baas G, Denig P et al. Pharmacist-led intervention aimed at deprescribing and appropriate use of cardiometabolic medication among people with type 2 diabetes. Res Soc Adm Pharm 2023;19:783–92. https://doi.org/10.1016/j.sa pharm.2022.11.009.

26. Module Minderen en stoppen van medicatie Onderdeel van de multidisciplinaire richtlijn Polyfarmacie bij ouderen. Utrecht, The Netherlands: Nederlands Huisartsen Genootschap (NHG); 2020. https://richtlijnen.nhg.org/files/2020- 11/Fina l_Module%20Minderen%20en%20stoppen%20van%20medi catie.pdf (31 May 2025, date last accessed).

27. Module Medicatiebeoordeling Onderdeel van de multidisciplinaire richtlijn Polyfarmacie bij ouderen. Utrecht, The Netherlands: Nederlands Huisartsen Genootschap (NHG); 2019. https://richtlijnen.nhg.org/files/2020- 05/fi nal_module_medicatiebeoordeling_2019.pdf (26 May 2025, date last accessed).

28. Van Blijswijk SCE, Chan OYA, Van Houwelingen AH et al. Self-reported hindering health complaints of community- dwelling older persons: a cross-sectional study. PloS One 2015;10:e0142416. https://doi.org/10.1371/journal.pone.01 42416.

29. Blom J, Den EW, Houwelingen AHV et al. Effectiveness and cost-effectiveness of a proactive, goal-oriented, integrated care model in general practice for older people. A cluster randomised controlled trial: integrated systematic care for older people-the ISCOPE study. Age Ageing 2016;45:30–41. https://doi.org/10.1093/ageing/afv174.

30. Chew L, Bradley K, Boyko E. Brief questions to iden- tify patients with inadequate health literacy. Fam Med 2004;36:588–94.

31. Hepler CD, Strand LM. Opportunities and responsibilities in pharmaceutical care. Am J Hosp Pharm 1990;47:533–43.

32. Huiskes VJB, Burger DM, Van Den Ende CHM et al. Effec- tiveness of medication review: a systematic review and meta- analysis of randomized controlled trials. BMC Fam Pract 2017;18:5. https://doi.org/10.1186/s12875- 016- 0577- x.

33. Kempen TGH, Van De Steeg-Van Gompel CHPA, Hoogland P et al. Large scale implementation of clinical medication reviews in Dutch community pharmacies: drug-related problems and interventions. Int J Clin Pharmacol 2014;36:630–5. https:// doi.org/10.1007/s11096- 014- 9947- 4.

34. Heringa M, Vervloet M, Bakker-Verdoorn S et al. Optimaliseren van medicatie-gebruik bij ouderen: Een evaluatie in opdracht van het ministerie van VWS. Leiden/Utrecht, The Netherlands: SIR Institute for Pharmacy Practice and Policy and Nivel; 2024. Report. https://sirste venshof.nl/wp- content/uploads/2024/05/Eindrapport- opti maliseren- van- medicatiegebruik- bij- ouderen.pdf (18 March 2026, date last accessed).

35. Salisbury C, Man MS, Bower P et al. Management of multimor- bidity using a patient-centred care model: a pragmatic cluster- randomised trial of the 3D approach. Lancet. 2018;392:41–50. https://doi.org/10.1016/S0140- 6736(18)31308- 4.

36. McCarthy C, Clyne B, Boland F et al. GP-delivered med- ication review of polypharmacy, deprescribing, and patient priorities in older people with multimorbidity in Irish pri- mary care (SPPiRE Study): a cluster randomised controlled trial. PLoS Med 2022;19:e1003862. https://doi.org/10.1371/ journal.pmed.1003862.

Received 18 March 2026 ; accepted 11 June 2026

https://doi.org/10.1371/journal.pone.0026574