Quantitative Research
Potentially Suboptimal Prescribing for Older Veteran Nursing Home Patients with Dementia
Joseph T. Hanlon, PharmD, MS, Health Scientist with the Geriatric Research Education and Clinical Center and Center for Health Equity Research and Promotion at the VA Pittsburgh Health System Pittsburgh, PA and a Professor, Division of Geriatric Medicine, Department of Medicine, School of Medicine, University of Pittsburgh, Pittsburgh, PA
Sherrie L. Aspinall, PharmD, MS, Clinical Pharmacist Specialist with the Veterans Affairs Center for Medication Safety, Hines, IL and Associate Professor, Department of Pharmacy and Therapeutics, School of Pharmacy University of Pittsburgh, Pittsburgh, PA
Steven M. Handler, MD, PhD, Assistant Professor, Division of Geriatric Medicine, Department of Medicine, School of Medicine, University of Pittsburgh, Pittsburgh, PA
Walid F. Gellad, MD, MPH, Core investigator with the Center for Health Equity Research and Promotion at the VA Pittsburgh Healthcare System Pittsburgh, PA and an Assistant Professor, Division of General Internal Medicine, Department of Medicine, School of Medicine, University of Pittsburgh, Pittsburgh, PA
Roslyn A. Stone, PhD, Health Scientist with Center for Health Equity Research and Promotion at the VA Pittsburgh Healthcare System Pittsburgh, PA and an Associate Professor of Biostatistics, School of Public Health, University of Pittsburgh, Pittsburgh, PA
Todd P. Semla, PharmD, MS, Clinical Pharmacist Specialist with the Veterans Affairs Pharmacy Benefits Management Services, Hines, IL and an Associate Professor in the Departments of Medicine and Psychiatry & Behavioral Sciences, School of Medicine, Northwestern University, Evanston, IL
Mary Jo V. Pugh, PhD, and
CORRESPONDING AUTHOR AND REQUEST FOR REPRINTS: Dr. Hanlon at the Department of Medicine (Geriatrics), University of Pittsburgh, Kaufman Medical Building-Suite 514, 3471 5th Ave, Pittsburgh, PA 15213; Tel#: 412-864-2507; Fax#: 412-692-2370; jth14@pitt.edu.
CONFLICT OF INTEREST: Dr. Semla serves on the Omnicare, Inc. Pharmacy and Therapeutics Committee, and he is a member of AARP’s Caregiving Advisory Panel. He is an author and editor for LexiComp, Inc; his spouse is an employee of AbbVie and owns stock in AbbVie, Abbott Labs, and Hospira. The remaining authors have no conflicts to report.
PRESENTATION OF WORK This work was presented at the 27th International Conference on Pharmacoepidemiology & Therapeutic Risk Management, August 15, 2011, Chicago, Illinois.
NIH Public Access Author Manuscript Ann Pharmacother. Author manuscript; available in PMC 2015 January 01.
Published in final edited form as: Ann Pharmacother. 2015 January ; 49(1): 20–28. doi:10.1177/1060028014558484.
N IH
-P A
A u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t
Research Health Scientist at the South Texas Veterans Healthcare System, San Antonio TX and an Associate Professor in the Department of Epidemiology and Biostatistics in the School of Medicine, University of Texas Health Science Center, San Antonio, TX
Maurice W. Dysken, MD Staff Physician with the Geriatric Research Education and Clinical Center, Veterans Affairs Medical Center Minneapolis, MN and Professor, Department of Psychiatry, School of Medicine, University of Minnesota, Minneapolis, MN
Abstract
Background—Nursing home patients with dementia may be more likely to suffer adverse drug events from suboptimal prescribing. Previous studies have not had national samples nor have they
examined multiple types of suboptimal prescribing by dementia severity.
Objective—To examine the prevalence of, and factors associated with, potentially suboptimal prescribing in older Veteran nursing home patients with dementia.
Methods—This is a retrospective descriptive study of 1303 Veterans 65 years or older admitted between 1/1/04–6/3/05 with dementia for long stays (90+ days) to 133 Veterans Affairs
Community Living Centers. Dementia severity was determined by Cognitive Performance Scale
and functional status dependencies.
Results—Overall,70.2% with mild-moderate dementia (n = 1076) had underuse as they did not receive an acetylcholinesterase inhibitor (AChEI), and 27.2% had evidence of inappropriate use
due to a drug-disease or drug-drug-disease interaction. Of the 227 with severe dementia, 36.1%
had overuse by receiving an AChEI, lipid-lowering or other agents, and 25.1% had evidence of
inappropriate use due to a drug-disease or drug-drug interaction. Multinomial logistic regression
analyses among those with mild to moderate dementia identified that living in the South versus
other regions was the single factor associated with all three types of suboptimal prescribing. In
those with severe dementia, antipsychotic use was associated with all three suboptimal prescribing
types.
Conclusions—Potentially suboptimal prescribing was common in older Veteran nursing home patients with dementia. Clinicians should develop a heightened awareness of these problems.
Future studies should examine associations between potentially suboptimal prescribing and health
outcomes in patients with dementia.
Keywords
dementia; nursing home; Veterans; drug utilization; quality
INTRODUCTION
Suboptimal prescribing can be defined as potential under-, over-, or inappropriate
medication use.1 Suboptimal prescribing can lead to considerable morbidity, especially in
older patients with dementia who may be more vulnerable to adverse drug events.2–4 This is
important as more than one half of nursing home patients have some form of dementia.5
Hanlon et al. Page 2
Ann Pharmacother. Author manuscript; available in PMC 2015 January 01.
N IH
-P A
A u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t
In patients with mild to moderate dementia, potentially suboptimal prescribing can include
the underuse of acetylcholinesterase inhibitors (AChEIs) because omission of these agents
may prevent an improvement in cognition. 6–8 In patients with severe dementia and limited
life expectancy, medications (e.g., lipid lowering agents) that may harm or not improve
quality of life can be overused.4,6,9,10 Drug-drug (e.g., AChEI and an anticholinergic) and
drug-disease (e.g., benzodiazepine use that can worsen cognition) interactions are two
important types of inappropriate prescribing for nursing home residents with both mild-
moderate and severe dementia.6,11
To the best of our knowledge, no published studies have stratified dementia patients in
nursing homes by severity and simultaneously examined potential under-, over-, or
inappropriate medication use. Most previous studies examined only one type of suboptimal
drug use.10 The objective of this study is to examine the prevalence of, and factors
associated with, potentially suboptimal prescribing in older Veteran nursing home patients
with and mild-moderate or severe dementia.
METHODS
Study Design, Setting, Data Sources and Sample
This was a retrospective descriptive study of 3,692 long-stay (90 days or more) patients age
65 or older admitted to any one of the 133 Veteran Affairs (VA)Community Living Centers
(CLCs) located in the U.S. between January 1, 2004 and June 30, 2005. Residents admitted
for respite or hospice care were excluded.
An analytic database was created from three data sources: 1) Minimum Data Set (MDS 2.0);
2) medication dispensing information from the VA Pharmacy Benefits Management
Services (PBM); and 3) Medical SAS® files.12 The MDS was completed at nursing home
admission and contains information about the functional, psychological and health status
needs of residents.13 The VA PBM data includes the following information for each drug
dispensed during the first 90 days following admission: 1) start date; 2) drug name; 3)
strength; 4) dosage form; 5) directions for use; 6) VA therapeutic class; and 7) amount
dispensed.12,14 International Classification of Diseases-9 (ICD-9) Clinical Modification
codes for inpatient and outpatient diagnoses in the year prior to admission were extracted
from the Medical SAS® files.15
The study sample (n = 1303) included those with dementia identified during a VA
hospitalization or outpatient visit in the year prior to CLC admission (ICD-9 codes:
Alzheimer’s [AD]- 331.0; 290.0; Other 046.1, 331.82, 294.8, 290.1x, 331.19; Vascular
[VaD]-290.4x. These codes have been used by to identify dementia patients in previous
national studies of Veterans.16,17 In this sample, 392 Veterans were coded as having AD,
180 had VaD, and the remaining 731 were coded as other dementias (including not
otherwise specified). The relatively large other dementia category is consistent with that
reported by a recent VA study and is partly due to the fact that the code of dementia, not
otherwise specified (ICD-9 code 294.8), is offered in the VA electronic medical record as
one of three top choices for clinicians to select.18 In a subsequent chart review by two
clinicians of a sample of veteran patients coded as other dementia, over half appeared to
Hanlon et al. Page 3
Ann Pharmacother. Author manuscript; available in PMC 2015 January 01.
N IH
-P A
A u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t
have a specific type of dementia (most commonly AD or VaD).18 To determine dementia
severity, we used the admission MDS assessment to determine residents’ cognitive function
using the Cognitive Performance Scale [CPS], which has scores ranging from 0 to 6 (with
higher scores representing worse cognition).19 Sensitivity and specificity for the CPS in
detecting residents with cognitive impairment is similar to that of the Mini Mental State
Exam.19 We also determined admission activities of daily living (ADL) dependencies; the
ADL identifies the amount of assistance needed from staff for five activities (bathing,
dressing, grooming, toileting and eating) and has scores ranging from 0 to 20 points (with
higher scores representing more dependencies).20 Following the study recommendations of
van der Steen et al., severe dementia was defined as a CPS score ≥5 and ADL score ≥10.21
All others were categorized as having mild-moderate dementia. The Pittsburgh VA Research
and Human Subjects Committees approved the study.
Main Outcome Measures
Three types of potential suboptimal medication use anytime during the first 90 days after
admission to a VA CLC was determined in patients with mild-moderate and severe
dementia. First, potential underuse was operationally defined as no pharmacy dispensing of
an ACHEI (i.e., donepezil, tacrine, galantamine, rivastigmine)to those in the mild-moderate
dementia group. The rationale is that many experts recommend AChEI treatment for those
with mild-moderate dementia (i.e., AD, VaD, or Lewy body dementia) as it may improve
cognitive function.6–8 Second, potential overuse was operationally defined in the severe
dementia group as medications/classes for which patients are unlikely to derive benefit
given their shortened life expectancy and risk of adverse drug events.6,9,10 Holmes et al.
convened 12 geriatricians and reached consensus on seven medications/classes examined in
this study that should never be used in these patients (i.e., ACHEIs, antiplatelet agents [e.g.
clopidogrel]), lipid lowering agents [e.g., statins]; antineoplastics [e.g., cyclophosphamide],
immune modulators [e.g., etanercept], leukotriene receptor antagonists [e.g., montelukast]
and sex hormones [e.g., testosterone]).9,10 It is important to note that in 2006 (one year after
the Holmes et al. panel was convened and the current study data were collected) the Food
and Drug Administration approved the indication for donepezil use in those with severe
dementia. The third outcome measure was potentially inappropriate use and was
operationally defined in both the mild-moderate and severe dementia groups as evidence of
either clinically important drug-disease interactions and/or clinically important drug-drug
interactions.6,11 Clinically important drug-disease interactions included the following
concomitant medications that can worsen cognitive function: 1) anticholinergics; 2)
barbiturates, and 3) benzodiazepine receptor agonists.6,11 Highly anticholinergic drugs were
those included in the Beers criteria (e.g., first generation antihistamines, tertiary tricyclic
antidepressants, gastrointestinal antispasmodics).11 Clinically important drug-drug
interactions included: 1) highly anticholinergic drugs that can block the pharmacodynamic
effects of ACHEIs; 2) cimetidine, ketoconazole, paroxetine or erythromycin that can inhibit
the hepatic metabolism of galantamine, and 3) ketoconazole or quinidine that can inhibit the
hepatic metabolism of donepezil.6
For the analysis in those with mild-moderate dementia, three groups were created: 1)
underuse only; 2) inappropriate use only, or 3) both under and inappropriate drug use. In
Hanlon et al. Page 4
Ann Pharmacother. Author manuscript; available in PMC 2015 January 01.
N IH
-P A
A u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t
those with severe dementia, three groups were created: 1) overuse only; 2) inappropriate use
only, or 3) both over and inappropriate use.
Independent Variables
Based on previous literature, our independent variables included demographic
characteristics, health status factors, and site level indicators.10,16, 22–24 Demographics
derived from the admission MDS assessment included categorical variables for age (65–74,
75–84, 85+), race (black, white, or other), gender (male or female) and educational level
(less than high school, high school, above high school).
Health status factors were derived from the admission MDS assessment, Medical SAS®
files, and VA PBM data. These factors included a continuous ADL dependencies measure
from the admission MDS.20 Using ICD-9 codes from the Medical SAS® files, we created a
continuous variable for the Charlson Comorbidity Index based on the methods of Deyo et
al., which creates a score (range 0–34) based on the presence of 17 chronic conditions
(excluding dementia).25,26 From VA PBM data, we created dichotomous medication use
variables (i.e., hypnotics, antidepressants, antipsychotics, mood stabilizers, and memantine)
and a categorical variable for polypharmacy involving the remaining medications (i.e., 0–1,
2–4, 5+ prescribed drugs at admission). Dichotomous variables for psychiatric/neurological
problems (i.e., aggressive behavior, post-traumatic stress disorder [PTSD], other anxiety
disorder, depression [Depression Rating Scale-DRS scores>3], seizure disorder, Parkinson’s
disease, neuropathic pain, bipolar disease, schizophrenia, hypertension) were created using
either ICD-9 codes from Medical SAS® files or MDS admission assessments.24,27
Categorical variables for site level indicators were created for urban/rural status, census
region (i.e., Northeast, Midwest, South, West), facility size (i.e., small [<60 beds], medium
[60–120 beds], and large [>120 beds]), and whether the patient resided in an Alzheimer’s/
Dementia Special Care Unit.16,24
Statistical Analyses
Descriptive statistics summarized independent variables and main outcome measures. To
include the approximately 1% of patients with missing data on education, we created a
dummy variable for a “missing” category. By dementia severity group status, we
summarized the number of patients with evidence of potential overuse, underuse, and
inappropriate use. Multinomial logistic regression analysis using a backward selection
approach (alpha = 0.10) identified those health status factors and site level indicators to be
included with the demographic characteristics in the final models. Among those with mild-
moderate dementia, we used multinomial logistic regression to identify demographic, health
status and site level factors associated with underuse only, inappropriate use only, or both.28
We also conducted logistic multinomial regression analysis in those with severe dementia to
identify demographic, health status and site level factors associated with potential overuse
only, potential inappropriate use only, or both.28 We report adjusted relative risk ratios
(RRRs) and 95% confidence intervals (CIs); robust standard errors adjusted for clustering by
CLC. Multi-parameter Wald tests quantified the association of each outcome with
categorical variables with more than two levels. Statistical analyses were performed using
SAS® (version 9; Cary, NC) and Stata® (College Station, TX) software.
Hanlon et al. Page 5
Ann Pharmacother. Author manuscript; available in PMC 2015 January 01.
N IH
-P A
A u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t
RESULTS
Table 1 compares the characteristics of CLC patients who had mild-moderate dementia (n =
1,076) to those with severe dementia (n = 227). Those with mil-moderate dementia were less
likely to have greater ADL dependencies, aggressive behavior, seizure disorder, or
Parkinson’s disease. Patients with mild-moderate dementia were more likely to have a
greater number of comorbidities (on average), other anxieties, depression, neuropathic pain,
and to use antidepressants.
In those with mild-moderate dementia, 70.2% were not receiving an ACHEI, indicating
potential underuse (Table 2). Among those with severe dementia, 36.1% had potential
overuse; the most common medication classes involved were ACHEIs (23.4%), lipid
lowering agents (12.8%), and antiplatelet agents (5.3%).
Table 2 also shows potentially inappropriate drug use among both groups of patients with
dementia. The overall prevalence of potentially inappropriate medication use did not differ
significantly by dementia severity (27.2% for those with mild-moderate dementia versus
25.1% for those with severe dementia; p = 0.51). The most common drug-disease
interactions were the use of anticholinergics (24.4% and 19.8% vs. respectively)and
benzodiazepines (4.5% and 6.6%, respectively), both of which could exacerbate dementia.
The most common drug-drug interaction for both dementia groups involved concomitant use
of an anticholinergic with an ACHEI (7.9% and 5.3%, respectively). There were few other
drug-drug interactions with ACHEIs.
Table 3 shows the results of the multinomial logistic regression analyses among those with
mild-moderate dementia. Blacks were less likely than whites to have inappropriate use only
or both under and inappropriate use (i.e., significance of the predictor across the 3
responses). Those prescribed 5+ medications, were diagnosed with bipolar disorder,
schizophrenia and/or took memantine were significantly more likely than those without
these characteristics to have potential underuse only or both under and inappropriate use.
Those taking an antidepressant were significantly more likely than those who did not to have
inappropriate use only or both inappropriate and underuse. Those living in the South were
less likely than those living elsewhere to have evidence of suboptimal prescribing.
Table 4 shows the results of the multinomial logistic regression analyses among those with
severe dementia. The probability of overuse only or both overuse and inappropriate use
increased significantly with the number of comorbidities. Only antipsychotic use was
associated with a higher risk of all three types of potential suboptimal drug use (i.e., overuse,
inappropriate use or both).
DISCUSSION
Our study shows that seven in ten older Veterans with mild-moderate dementia residing in
these CLCs did not receive an ACHEI (i.e., potential underuse). However, the overall the
proportion of those with mild-moderate dementia receiving an ACEI (28.7%) is comparable
to the 30.0% reported based on the 2004 National Nursing Home Study.23 That study
reported that the oldest old, those with worse functional status, user of antidepressants or
Hanlon et al. Page 6
Ann Pharmacother. Author manuscript; available in PMC 2015 January 01.
N IH
-P A
A u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t
antipsychotics, and residents of larger nursing homes were more likely to use an ACHEI. In
our study, only the South region was significantly associated with less underuse of ACHEIs
in those with mild-moderate dementia. Therefore, clinicians practicing in regions besides the
southern US should have heightened awareness and consider the use of ACHEIs in those
with mild-moderate dementia.
Regarding potential overuse, nearly one quarter of older Veterans with severe dementia
received ACHEI agents (i.e., potential over-use). This compares with a national study by
Tjia et al. of 5,406 non-VA nursing home residents with advanced dementia where 36.4%
took an ACHEI.10 Clinicians should assess patients with severe dementia and consider
whether discontinuation of AChEIis indicated. Besides ACHEIs in those with severe
dementia, up to nearly 13% of these patients took a lipid lowering agent. In contrast, the
same Tjia et al. study reported 22.4% received a lipid lowering agent.10 In the current study
among those with severe dementia, only antipsychotic use was associated with a higher risk
of overuse, inappropriate use or both. Therefore, clinicians evaluating the appropriateness of
antipsychotic use should be aware that overuse and inappropriate use are more likely in
those with severe dementia.
Potentially inappropriate prescribing due to drug-disease or drug-drug interactions
regardless of dementia severity was common. Anticholinergic use was the most common
medication class involved in drug-disease interactions. The most common drug-drug
interaction in nearly a quarter of ACHEI users (100/386)was the concomitant use of a highly
anticholinergic agent, which can reduce the pharmacological effectiveness. The proportion
with this specific drug-drug interaction is similar to the 27.1% rate reported in a study of
3,251 non-VA nursing home patients from Indiana.22 These findings have some clinical
implications. One issue is for health care professionals working with older nursing home
patients with dementia who receive an ACHEI is to be more vigilant in avoiding the use of
anticholinergic medications given their potential for worsening cognition, as well as
increasing the risk of constipation, urinary retention, and syncope.11 In most cases a suitable
alternative is available. For example in patients with seasonal allergies, nasal steroids are
preferable to highly anticholinergic and sedative first generation antihistamines.29 Another
emerging and important issue is for clinicians to reexamine the need for continuing certain
medications in patients with shortened life expectancy, such as those with severe
dementia.9,10 Several reviews have addressed the ethics of not prescribing certain
medications, as well as techniques for medication discontinuation and avoidance of adverse
drug withdrawal events.3,4,30
This study has a number of potential limitations. First, there is potential disease
misclassification because dementia and its severity were not determined by an independent
research neurologist using standardized diagnostic criteria. Second, the criteria for
suboptimal drug use were developed by consensus of an expert panel. Currently, to the best
of our knowledge, evidence-based predictive validity data is not available for these criteria.
Third, potential overuse of ACHEIs in those with severe may have been overestimated in
part because interpretation of the VA criteria for use at the time of the study may have
varied at each CLC.10 Fourth, overuse would be overestimated if a discontinuation was
attempted, but because of patient deterioration, the AChEI was restarted. Despite this, it is
Hanlon et al. Page 7
Ann Pharmacother. Author manuscript; available in PMC 2015 January 01.
N IH
-P A
A u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t
important to keep in mind that the benefit in dementia patients is modest at best (only one
third of those with mild to moderate Alzheimer’s improved 4 points or more on a 70 point
ADAS-Cog scale).6,7 Fifth, potential underuse of ACHEIs may also be overestimated in
patients in whom it was started and stopped due to adverse drug events/perceived lack of
efficacy. Sixth, our power was limited, particularly in the multinomial modeling, due to the
relatively small sample of those with severe dementia, those who were female, and those
with specific psychiatric conditions. Finally, our findings in a primarily male VA nursing
home population may not generalize to non-VA nursing home settings where a majority of
dementia patients is older and female.
CONCLUSIONS
We conclude that potentially suboptimal prescribing was common in older Veteran nursing
home patients with dementia. Clinicians should develop a heightened awareness for these
problems. Future studies should examine the impact that potentially inappropriate
prescribing has on dementia patient health outcomes. Further study is also needed to better
understand dementia patient, family, and prescriber preferences for intensifying,
discontinuing or not initiating certain medications and the impact that these prescribing
actions have on health outcomes in dementia patients.
Acknowledgments
FUNDING
This study was not supported or funded by a specific grant.
We would like to thank Holly M. Holmes, MD for her consultation regarding specific medications that might represent potential overuse in those with severe dementia, Xiaoqiang Wang, MS for his help with the statistical analyses and Beth Thomas for her work on the Tables.
Dr. Hanlon is supported by National Institute on Aging grants (P30AG024827, T32 AG021885, K07AG033174, R01AG027017) and a VA Health Services Research grant (IIR12379).
References
1. Spinewine A, Schmader KE, Barber N, et al. Appropriate prescribing in elderly people: How can it be measured and optimized? Lancet. 2007; 370:173–184. [PubMed: 17630041]
2. Lau DS, Mercaldo ND, Shega JW, et al. Functional decline associated with polyphamacy and potentially inappropriate medication use in community –dwelling older adults with dementia. Am J Alz Dis OthDemen. 2011; 26:606–615.
3. O’Mahony D, O’Connor M. Pharmacotherapy at end of life. Age Ageing. 2011; 40:419–422. [PubMed: 21622981]
4. Parsons C, Hughes CM, Passmore AP, Lapane KL. Withholding, discontinuing and withdrawing medications in dementia patients at the end of life. Drugs Aging. 2010; 27:435–449. [PubMed: 20524704]
5. Seitz D, Purandare N, Conn D. Prevalence of psychiatric disorders among older adults in long-term care homes: A systematic review. Int Psychogeriatr. 2010; 22:1025–1039. [PubMed: 20522279]
6. VA PBM and Medical Advisory Panel. Cholinesterase inhibitor criteria for use. May. 2003
7. Wild R, Pettit T, Burns A. Cholinesterase inhibitors for dementia with Lewy bodies. Cochrane Database Syst Rev. 2003; (3):CD003672. [PubMed: 12917981]
8. Malouf R, Birks J. Evidence of efficacy of donepezil with mild or moderate vascular cognitive impairment. Cochrane Database Syst Rev. 2004; (1):CD004395. [PubMed: 14974068]
Hanlon et al. Page 8
Ann Pharmacother. Author manuscript; available in PMC 2015 January 01.
N IH
-P A
A u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t
9. Holmes HM, Sachs GA, Shega JW, et al. Integrating palliative medicine into the care of persons with advance dementia: identifying appropriate medication use. J Am Geriatr Soc. 2008; 56:1306– 11. [PubMed: 18482301]
10. Tjia J, Briesacher BA, Peterson D, Liu Q, Andrade SE, Mitchell SL. Use of medications of questionable benefit in advanced dementia. JAMA Intern Med. Published online September 08, 2014. doi:10.1001.
11. The American Geriatrics Society 2012 Beers Criteria Update Expert Panel. AGS Updated Beers Criteria for potentially inappropriate medication use in older adults. J Am Geriatr Soc. 2012; 60:616–631. [PubMed: 22376048]
12. Hanlon JT, Wang X, Good CB, et al. Racial differences in medication use among older long stay Veteran nursing home care unit patients. Consult Pharm. 2009; 24:439–46. [PubMed: 19555154]
13. Mor V. A comprehensive clinical assessment tool to inform policy and practice: applications of the Minimum Data Set. Med Care. 2004; 42(4 Suppl):III50–59. [PubMed: 15026672]
14. VHA Pharmacy Prescription Data: VIReC Research User Guide. U.S. Dept. of Veterans Affairs, Health Services Research and Development Service, VA Information Resource Center; Sep. 2003
15. VHA Medical SAS® Outpatient Datasets and Inpatient Encounters Dataset. U.S. Dept. of Veterans Affairs, Health Services Research and Development Service, VA Information Resource Center; Nov. 2010
16. Gellad WF, Aspinall SL, Handler SM, et al. Use of antipsychotics among older residents in Veterans Administration nursing homes. Med Care. 2012; 50:954–60. [PubMed: 23047785]
17. Krishnan LL, Petersen NJ, Snow AL, et al. Prevalence of dementia among Veterans Affairs medical care system users. Dement Geriatr Cogn Disord. 2005; 20:245–53. [PubMed: 16088141]
18. Butler D, Kowall NW, Lawler E, et al. Underuse of diagnostic codes for specific dementias in the Veterans Affairs New England Healthcare System. J Am Geriatr Soc. 2012; 60:910–915. [PubMed: 22587853]
19. Hartmaier SL, Sloane PD, Guess HA, et al. Validation of the Minimum Data Set Cognitive Performance Scale: agreement with the Mini-Mental State Examination. J Gerontol. 1995; 50:M128–133.
20. McConnell ES, Pieper CF, Sloane RJ, Branch LG. Effects of cognitive performance on change in physical function in long-stay nursing home residents. J Gerontol. 2002; 57:M778–M784.
21. van der Steen JT, Volicer L, Gerritsen DL, et al. Defining severe dementia with the Minimum Data Set. Int J Ger Psych. 2006; 21:1099–1106.
22. Modi A, Weiner M, Craig BA, et al. Concomitant use of anticholinergics with acetylcholinesterase inhibitors in medicaid recipients with dementia and residing in nursing homes. J Am Geriatr Soc. 2009; 57:1238–44. [PubMed: 19467148]
23. Seitz DP, Gruneir A, Conn DK, Rochon PA. Cholinesterase inhibitor use in U.S. nursing homes: results from the national nursing home survey. J Am Geriatr Soc. 2009; 57:2269–74. [PubMed: 19874411]
24. Hanlon JT, Wang X, Castle NG, et al. Potential underuse, overuse and inappropriate use of antidepressants in older Veteran nursing home patients. J Am Geriatr Soc. 2011; 59:1412–20. [PubMed: 21824120]
25. Charlson ME, Pompei P, Ales KL, Mackensie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987; 40:373– 383. [PubMed: 3558716]
26. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992; 45:613–619. [PubMed: 1607900]
27. Burrows AB, Morris JN, Simon SE, et al. Development of a minimum data set-based depression rating scale for use in nursing homes. Age Ageing. 2000; 29:165–172. [PubMed: 10791452]
28. Goldstein, H. Multilevel Statistical Models. 3. London, UK: Edward Arnold; 2003.
29. Slavin RG. Special considerations in treatment of allergic rhinitis in the elderly: role of intranasal corticosteroids. Allergy Asthma Proc. 2010; 31:179–84. [PubMed: 20615319]
Hanlon et al. Page 9
Ann Pharmacother. Author manuscript; available in PMC 2015 January 01.
N IH
-P A
A u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t
30. Bain KT, Holmes HM, Beers MH, et al. Discontinuing medications: a novel approach for revising the prescribing stage of the medication-use process. J Am Geriatr Soc. 2008; 56:1946–52. [PubMed: 18771457]
Hanlon et al. Page 10
Ann Pharmacother. Author manuscript; available in PMC 2015 January 01.
N IH
-P A
A u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t
N IH
-P A
A u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t
Hanlon et al. Page 11
Table 1
Characteristics of Older Veteran Nursing Home Patients by Dementia Severity
Factors Mild to Moderate Dementia (n = 1076) Severe Dementia (n = 227) p-value
n (%) n (%)
Demographics
Age
65–74 261 (24.3) 48 (21.1)
75–84 621 (57.7) 130 (57.3)
85+ 194 (18.0) 49 (21.6)
Race
White 889 (82.6) 166 (73.1)
Black 133 (12.4) 33 (14.5)
Other 54 (5.0) 28 (12.3)
Female gender 30 (2.8) 7 (3.1) 0.81
Education 0.29
Below High School 341 (31.7) 74 (32.6)
High School 475 (44.1) 111 (48.9)
Above High School 250 (23.2) 41 (18.1)
Missing 10 (0.9) 1 (0.4)
ADL Dependence (mean ± SD) 8.69 (6.13) 16.8 (3.49) <0.01
Comorbidity Index excluding dementia (mean ± SD) 2.10 (1.93) 1.48 (1.47) <0.01
Polypharmacy 0.05
0–1 416 (38.7) 74 (32.6)
2–4 392 (36.4) 79 (34.8)
5+ 268 (24.9) 74 (32.6)
Aggressive behavior 149 (13.8) 46 (20.3) 0.01
PTSD 106 (9.9) 14 (6.2) 0.08
Other anxiety 118 (11.0) 12 (5.3) 0.01
Depression 354 (32.9) 41 (18.1) <0.01
Seizure disorder 59 (5.5) 23 (10.1) 0.01
Parkinson’s disease 82 (7.6) 31 (13.7) <0.01
Neuropathic pain 242 (22.5) 24 (10.6) <0.01
Bipolar disease 54 (5.0) 6 (2.6) 0.12
Schizophrenia 121 (11.2) 24 (10.6) 0.77
Hypertension 705 (65.5) 135 (59.5) 0.08
Use of hypnotic 5 (0.5) 0 (0.0) 0.30
Use of antidepressant 571 (53.1) 86 (37.9) <0.01
Use of antipsychotic 445 (41.4) 95 (41.9) 0.89
Use of mood stabilizer 183 (17.0) 33 (14.5) 0.36
Use of memantine 73 (6.8) 16 (7.0) 0.89
Site Level Indicators
Rural 301 (28.0) 54 (23.8) 0.20
Ann Pharmacother. Author manuscript; available in PMC 2015 January 01.
N IH
-P A
A u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t
Hanlon et al. Page 12
Factors Mild to Moderate Dementia (n = 1076) Severe Dementia (n = 227) p-value
n (%) n (%)
Region 0.23
North East 264 (24.5) 60 (26.4)
Midwest 179 (16.6) 39 (17.2)
South 472 (43.9) 106 (46.7)
West 161 (15.0) 22 (9.7)
Bed Size 0.73
Small (<60) 92 (8.6) 23 (10.1)
Medium (60–120) 519 (48.2) 109 (48.0)
Large (>120) 465 (43.2) 95 (41.9)
Alzheimer’s/Dementia 0.36
Special Care Unit 231 (21.5) 55 (24.2)
Abbreviations: ADL = Activities of Daily Living; PTSD = Post Traumatic Stress Disorder
Ann Pharmacother. Author manuscript; available in PMC 2015 January 01.
N IH
-P A
A u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t
Hanlon et al. Page 13
Table 2
Underuse, Potential Inappropriate Prescribing, and Overuse in Older Veteran Nursing Home Patients with
Dementia
Mild to Moderate Dementia (n = 1076) Severe Dementia (n = 227)
n (%) n (%)
Underuse of ACHEI 755 (70.2)
Potentially Inappropriate Prescribing (Overall) 293 (27.2) 57 (25.1)
Drug/Disease Interaction
Anticholinergics 262 (24.4) 45 (19.8)
Barbiturates 2 (0.2) 0 (0.0)
Benzodiazepine 48 (4.5) 15 (6.6)
Drug-Drug Interaction*
Anticholinergics /ACHEI 85 (7.9) 12 (5.3)
Overuse (Overall) 82 (36.1)
ACHEI 65 (23.4)
Lipid lowering agent 29 (12.8)
Antiplatelet agent 12 (5.3)
Other classes 8 (3.5)
Abbreviations: ACHEI = Acetylcholinesterase Inhibitor;
* One person with severe dementia had evidence of paroxetine/galantamine drug-drug interaction; 6 persons with mild-moderate dementia had a
drug interaction with galantamine, and another 6 had one with donepezil.
Ann Pharmacother. Author manuscript; available in PMC 2015 January 01.
N IH
-P A
A u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t
Hanlon et al. Page 14
Table 3
Factors Associated with Potential Underuse Only, Inappropriate Use Only and Both in those with Mild to
Moderate Dementia (n = 1076)
Factors Underuse Only (n = 556) Inappropriate Only (n = 94) Both (n = 199)
Adj. RRR (95% CI) Adj. RRR (95% CI) Adj. RRR (95% CI)
Demographics
Age
65–74 reference reference reference
75–84 0.56 (0.34–0.93) 0.56 (0.26–1.21) 0.37 (0.19–0.73)
85+ 0.70 (0.36–1.35) 1.12 (0.41–3.01) 0.53 (0.24–1.16)
Race*
White reference reference reference
Black 0.85 (0.52–1.38) 0.41 (0.19–0.88) 0.30 (0.15–0.61)
Other 1.47 (0.69–3.11) 0.22 (0.04–1.12) 1.31 (0.55–3.14)
Female gender 2.98 (0.69–12.9) 4.08 (0.76–21.99) 6.70 (1.41–31.71)
Education
Below high school reference reference reference
High school 0.86 (0.61–1.20) 0.94 (0.54–1.64) 0.96 (0.60–1.53)
Above high school 0.96 (0.58–1.58) 0.75 (0.40–1.40) 0.69 (0.38–1.26)
Health Status
Comorbidity Index (excluding dementia) 1.08 (0.99–1.18) 0.98 (0.86–1.11) 1.12 (1.01–1.24)
Other Medications*
0–1 reference reference reference
2–4 1.01 (0.72–1.42) 1.01 (0.61–1.68) 0.74 (0.45–1.20)
5+ 2.72 (1.76–4.21) 1.12 (0.53–2.38) 2.02 (1.09–3.71)
Bipolar* 4.43 (1.31–15.08) 3.42 (0.70–16.64) 5.91 (1.65–21.13)
Schizophrenia* 2.45 (1.26–4.79) 0.89 (0.32–2.45) 3.64 (1.62–8.15)
Hypnotic use 0.28 (0.28–2.86) too few to calculate 1.52 (0.27–8.66)
Antidepressant use* 0.83 (0.61–1.13) 2.12 (1.27–3.52) 1.59 (1.02–2.47)
Antipsychotic use 0.96 (0.91–1.01) 1.08 (1.03–1.12) 1.04 (0.99–1.08)
Memantine use* 0.32 (0.17–0.60) 0.58 (0.22–1.48) 0.09 (0.18–0.41)
Site Level Indicators
Small Bed Size (<60) reference reference reference
Medium Bed Size (60–120) 1.04 (0.58–1.87) 1.79 (0.67–4.77) 1.17 (0.57–2.41)
Large Bed Size (>120) 0.70 (0.36–1.35) 1.53 (0.57–4.12) 0.89 (0.42–1.89)
Rural 1.26 (0.73–2.13) 1.22 (0.72–2.07) 1.39 (0.74–2.62)
North East Region* reference reference reference
Midwest Region 0.76 (0.42–1.38) 0.47 (0.25–0.87) 0.76 (0.41–1.40)
South Region 0.60 (0.37–0.97) 0.38 (0.21–0.70) 0.48 (0.27–0.86)
West Region 1.25 (0.69–2.23) 0.57 (0.18–1.74) 0.86 (0.42–1.77)
Abbreviations: Adj. = adjusted; CI = confidence interval; RRR = relative rate ratio
Ann Pharmacother. Author manuscript; available in PMC 2015 January 01.
N IH
-P A
A u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t
Hanlon et al. Page 15
* P<0.05 for overall multiparameter Wald test for the factor
Ann Pharmacother. Author manuscript; available in PMC 2015 January 01.
N IH
-P A
A u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t N
IH -P
A A
u th
o r M
a n u scrip
t
Hanlon et al. Page 16
Table 4
Factors Associated with Potential Overuse Only, Inappropriate Use Only and Both in those with Severe
Dementia (n = 227)
Factors Overuse Only (n = 55) Inappropriate Only (N = 30) Both (n = 27)
Adj. RRR (95% CI) Adj. RRR (95% CI) Adj. RRR (95% CI)
Demographics
Age
65–74 reference reference reference
75–84 0.89 (0.32–2.46) 0.42 (0.14–1.28) 2.54 (0.58–11.13)
85+ 1.21 (0.37–3.98) 0.44 (0.11–1.67) 1.87 (0.30–11.55)
Race
White reference reference reference
Black 0.50 (0.14–1.73) 0.17 (0.02–1.21) 0.03 (0.01–0.35)
Other 2.84 (0.71–11.35) 1.77 (0.55–5.67) 1.18 (0.17–7.93)
Female gender 4.01 (0.55–35.21) too few to calculate too few to calculate
Education
Below high school reference reference reference
High school 1.13 (0.41–3.09) 1.42 (0.41–4.82) 1.26 (0.33–4.79)
Above high school 0.88 (0.31–2.45) 0.80 (0.18–3.53) 0.34 (0.02–6.68)
Health Status
Comorbidity index* (excluding dementia) 1.35 (1.05–1.72) 1.39 (0.97–2.00) 2.08 (1.47–2.94)
0–1 medications reference reference reference
2–4 medications 1.54 (0.64–3.70) 0.24 (0.08–0.72) 0.72 (0.20–2.54)
5+ medications 0.67 (0.26–1.70) 0.29 (0.08–0.72) 1.15 (0.29–4.54)
Neuropathic pain 2.35 (0.66–8.45) 4.05 (0.91–17.93) 24.71 (6.79–89.89)
Bipolar too few to calculate 0.12 (0.00–275.63) 0.64 (0.03–12.50)
Mood stabilizers 2.01 (0.40–9.96) 2.72 (0.37–19.84) 27.71 (4.17–184.02)
Antipsychotic use* 1.19 (1.04–1.36) 1.32 (1.13–1.54) 1.36 (1.20–1.55)
Memantine use 5.89 (0.92–37.59) too few to calculate 4.78 (0.34–66.62)
Site Level Indicators
Small Bed Size (<60) reference reference reference
Medium Bed Size (60–120) 0.34 (0.08–1.38) 2.75 (0.36–20.82) 0.15 (0.02–1.16)
Large Bed Size (>120) 1.08 (0.28–4.05) 4.48 (0.45–44.85) 0.21 (0.03–1.24)
Rural 1.71 (0.66–4.41) 4.56 (1.37–15.14) 0.60 (0.12–2.99)
North East Region reference reference reference
Midwest Region 0.53 (0.14–1.99) 0.79 (0.15–4.16) 0.20 (0.03–1.13)
South Region 0.84 (0.30–2.28) 1.37 (0.26–7.12) 0.28 (0.05–1.610
West Region 0.65 (0.11–3.74) 2.84 (0.56–14.42) 0.05 (0.01–0.41)
Abbreviations: Adj. = adjusted; CI = confidence interval; RRR = relative rate ratio
* P<0.05 for overall multiparameter Wald test for the factor
Ann Pharmacother. Author manuscript; available in PMC 2015 January 01.