EVBP

Q U A L I T Y I M P R O V E M E N T I N V I T E D A R T I C L E Trish M. Perl, Section Editor

Prevention of Central Line–Associated Bloodstream Infections Through Quality Improvement Interventions: A Systematic Review and Meta-analysis

Koen Blot,1 Jochen Bergs,3 Dirk Vogelaers,1,2 Stijn Blot,1,4 and Dominique Vandijck1,2,3

1Faculty of Medicine and Health Sciences, Ghent University, 2General Internal Medicine, Ghent University Hospital, Ghent, and 3Health Economics and Patient Safety, Hasselt University, Hasselt, Belgium; and 4Burns, Trauma and Critical Care Research Centre, The University of Queensland, Brisbane, Australia

This systematic review and meta-analysis examines the impact of quality improvement interventions on central line–associated bloodstream infections in adult intensive care units. Studies were identified through Medline and manual searches (1995–June 2012). Random-effects meta-analysis obtained pooled odds ratios (ORs) and 95% confidence intervals (CIs). Meta-regression assessed the impact of bundle/checklist interventions and high baseline rates on intervention effect. Forty-one before–after studies identified an infection rate de- crease (OR, 0.39 [95% CI, .33–.46]; P < .001). This effect was more pronounced for trials implementing a bundle or checklist approach (P = .03). Furthermore, meta-analysis of 6 interrupted time series studies revealed an in- fection rate reduction 3 months postintervention (OR, 0.30 [95% CI, .10–.88]; P = .03). There was no difference in infection rates between studies with low or high baseline rates (P = .18). These results suggest that quality improvement interventions contribute to the prevention of central line–associated bloodstream infections. Im- plementation of care bundles and checklists appears to yield stronger risk reductions.

Keywords. central line–associated bloodstream infection; catheter-related bloodstream infection; quality improvement intervention; meta-analysis.

Central venous catheters are indispensable devices in the intensive care unit (ICU), necessary for infusion of medication, fluid, or blood products; hemodialy- sis; blood withdrawal; or hemodynamic monitoring. However, these invasive devices predispose patients to preventable central line–associated bloodstream infec- tions (CLABSIs), defined as bloodstream infections in patients with a central line 48 hours before infection onset, not related to another site (Table 1). CLABSIs are associated with increased morbidity, leading to

increased length of hospitalization and resource use [3, 4], and might impact mortality and compromise patient prognosis [5–7].

Infection prevention measures during central line in- sertion or maintenance, such as hand hygiene, maximal sterile barriers during catheter insertion, chlorhexidine skin disinfection, optimal catheter site selection, and daily review of line necessity with prompt removal of un- necessary lines, are known to decrease CLABSI risk [8,9]. The Institute for Healthcare Improvement (IHI) recom- mends use of aforementioned items, in a central line care bundle, to decrease CLABSI occurrence. Despite the availability of evidence-based interventions summarized in guidelines [10, 11], CLABSI remains a substantial threat for hospitalized patients, with pooled estimated mean occurrence rates of 4.4 CLABSIs per 100 devices in- serted (95% confidence interval [CI], 4.1–4.9) and 2.7 CLABSIs per 1000 catheter-days (95% CI, 2.6–2.9) [12].

In recent years, it has become clear that the lim- iting factor to infection prevention resides in the

Received 4 February 2014; accepted 2 April 2014; electronically published 9 April 2014.

Presented in part: 26th Annual Congress of the European Society of Intensive Care Medicine, Paris, France, 5–9 October 2013.

Correspondence: Koen Blot, BSc, Ghent University, Faculty of Medicine and Health Sciences, 9000 Ghent, Belgium ([email protected]).

Clinical Infectious Diseases 2014;59(1):96–105 © The Author 2014. Published by Oxford University Press on behalf of the Infectious Diseases Society of America. All rights reserved. For Permissions, please e-mail: [email protected]. DOI: 10.1093/cid/ciu239

96 • CID 2014:59 (1 July) • QUALITY IMPROVEMENT

D ow

nloaded from https://academ

ic.oup.com /cid/article/59/1/96/404195 by guest on 13 M

ay 2021

implementation of published recommendations [13]. Introduc- ing prevention measures may be hampered by factors such as lack of problem awareness, poor familiarity or nonagreement with guidelines, low self-efficacy, inability to change practice, or lack of resources [14,15].Quality improvement interventions such as personnel education or catheter care bundles and check- lists aim to decrease CLABSIs by improving adherence to pre- vention measures [16]. However, efficacy of these interventions has not been fully assessed.

This study examined whether quality improvement interven- tions reduce CLABSI rates in adult ICUs. Subgroup analysis as- sessed whether bundle/checklist interventions, high study power, or high baseline CLABSI rates influenced the intervention effect.

METHODS

Search Strategy Medline was systematically searched (1995–June 2012) through a combination of search terms: catheter-related infections/pre- vention and control; catheterization, central venous/adverse ef- fects; catheters, indwelling/adverse effects; infection control/ methods; infection control/standards; intensive care units; qual- ity control; quality of healthcare; and bundle (Supplementary

Appendix 1). Extra studies were identified via reference lists, manually and through Ovid and ScienceDirect databases.

Study Selection Eligible studies used before–after, interrupted time series (ITS), controlled before–after, nonrandomized controlled trial, or ran- domized controlled trial study designs that complied with the Cochrane Effective Practice and Organisation of Care Group methodological criteria. ITS studies report at least 3 data points before and after a defined point in time in which the interven- tion is implemented. Participants consisted of adult ICU pa- tients with central line catheters. Trials implemented quality improvement interventions aimed at increasing professional ad- herence to evidence-based infection prevention processes. The primary outcome measure was the number of CLABSIs per catheter-days pre- and postintervention. Only English-language papers were included. Medline search results were screened by title and abstract. Selected papers underwent a full-text assess- ment, and eligibility issues were resolved between authors.

Data Extraction Extracted data included author and year of publication, settings and study populations, study designs and periods, quality im- provement and preventive interventions implemented in the baseline and intervention periods, compliance measures, num- ber of CLABSI and catheter-days, and applied CLABSI defini- tions. Study authors were not contacted for additional data. To obtain effect sizes for ITS studies, infection rate data were ex- tracted from study figures using the program Plot Digitizer. Re- sults reported as a mix from both included and excluded study participants were included. Quality improvement interventions were classified under general headers (Table 2), and only pre- ventive interventions described by Centers for Disease Control and Prevention (CDC) guidelines [10] and applicable to the majority of ICU patients were noted.

Quality Assessment The Downs and Black checklist ascertained study methodolog- ical risk of bias [17]. It consists of 27 questions that evaluate the reporting, external validity, internal validity, and power of non- randomized studies of healthcare interventions. Studies were scored based on these item criteria, adapted for CLABSI preven- tion research.

Statistical Analysis A random-effects meta-analysis using the DerSimonian-Laird estimator obtained odds ratios (ORs) and 95% CIs for CLABSI rate reductions. The Higgins I2 test was predefined to quantify heterogeneity (I2 ≤ 25% for low, 25% < I2 < 50% for moderate, and I2 ≥ 50% for high), and funnel plots assessed publication bias. Subgroup analysis through meta-regression for before– after study designs compared studies with or without bundle/

Table 1. Centers for Disease Control and Prevention Definitions for Central Line–Associated Bloodstream Infection Terminology

Terminology Definition

CLABSI An LCBI where a central line was in place for >2 calendar days and a central line was in place on the date of event or the day before.

LCBI To be defined as LCBI, it must meet 1 of the following criteria:

(1) Patient has a recognized pathogen cultured from 1 or more blood cultures, and organism cultured from blood is not related to an infection at another site;

(2) Patient has at least 1 of the following signs or symptoms: fever (>38°C), chills, or hypotension, and positive laboratory results are not related to an infection at another site and the same common commensal is cultured from 2 or more blood cultures drawn on separate occasions.

Central line days A daily count of the number of patients with a central line in the patient care location during a time period. A patient with multiple central lines for a day only counts as 1 central line day.

Patient-days A daily count of the number of patients in the patient care location during a time period.

Device utilization ratio

Central line utilization ratio is calculated by dividing the number of central line days by the number of patient-days.

Adapted from the Centers for Disease Control and Prevention [1, 2].

Abbreviations: CLABSI, central line–associated bloodstream infection; LCBI, laboratory-confirmed bloodstream infection.

QUALITY IMPROVEMENT • CID 2014:59 (1 July) • 97

D ow

nloaded from https://academ

ic.oup.com /cid/article/59/1/96/404195 by guest on 13 M

ay 2021

checklist interventions, baseline rates above or below 4.0 CLAB- SIs per 1000 catheter-days, and power scores above or below 0.75. Univariate analysis calculated changes in device utilization rates. Sensitivity analysis identified heterogeneous studies that influenced the meta-analysis.

Monthly ITS data were standardized for meta-analysis by dividing the outcome and standard error (SE) by the standard deviation (SD) of the preintervention trend. One study reported annual data points, which were used for the 12- and 24-month follow-up analyses [18]. SPSS version 22 calculated the inter- vention effect using segmented time series regression analysis, adjusting for time trend and autocorrelation. A negative change

in level or slope indicated an infection rate reduction [19]. A P value <.05 was considered statistically significant.

RESULTS

The search algorithm identified 634 records (627 in PubMed and 7 in Ovid and ScienceDirect). Forty-three studies, pub- lished in English between January 1995 and June 2012 involving 584 ICUs, were included for meta-analysis (Figure 1). Two stud- ies [20, 21] continued their quality improvement initiatives and republished old data with new results [9,22].The older study by Coopersmith et al [20] was included for ITS analysis, and the

Table 2. Classification of Quality Improvement Interventions and Number of Studies

Quality Improvement Intervention (No. of Studies) Definition and Examples

Education (n = 33) Teaching lectures transmitting theoretical knowledge concerning CLABSI

• Monthly sessions or a single lecture concerning CLABSI epidemiology or preventive interventions • Educational modules with pre- and posttest

Training (n = 4) Training sessions for practical skills associated with CVC care and maintenance

• Personnel training for aseptic CVC insertion on patient simulators

Feedback (n = 20) Reporting of CLABSI or care item compliance rates to ICU personnel

• Monthly reporting of infection rates at staff meetings • Posters of improving prevention measure compliance or CLABSI rates

Clinical reminders (n = 15) Reminders of optimal clinical practice strategically placed to improve awareness or application of prevention measures

• Posters advertising proper hand hygiene or monthly CLABSI rates • A daily reminder per patient asking whether a CVC can be discontinued • Stickers placed on CVCs or wearable lapel buttons reminding personnel of the importance of care

processes • Procedure fact sheets, flowcharts, algorithms, or daily goals sheet

Bundle (n = 11) A short list of at least 2 IHI prevention measures to be used during CVC insertion and/or maintenance

• A bundle consists of 2–5 IHI care items • A bundle does not contain quality improvement interventions aiming at improving compliance with

prevention measures

Checklist (n = 18) Checklist of bundled care item prevention measures to increase adherence to evidence-based infection prevention practices

• A checklist of 2–5 IHI care items is used during CVC insertion to improve adherence to a bundle of care items

Empowerment to stop procedure (n = 10)

Nurses are empowered to halt and restart CVC insertion care or maintenance when a prevention measure is not implemented correctly to ensure optimal catheter care

Surveillance: compliance monitoring (n = 12)

Nurses intermittently or continually supervise CVC insertion or maintenance prevention measures, with/without use of a bundle/checklist

Leader designation (n = 11) A leader is designated to facilitate implementation of quality intervention processes by planning activities to improve awareness or introduction of bundled care items

Prepackaging of CVC materials (n = 16) Use of a CVC cart or kit stocked with all necessary supplies to insert or maintain a central line

Infrastructure changes (n = 2) Changes to hospital infrastructure to facilitate adherence to prevention measures

• Installation of hand-rub dispensers to lower the threshold for personnel to apply antiseptic hand hygiene

Organizational changes (n = 4) Organizational changes in personnel staffing or duties to improve adherence to prevention measures

• Staffing of extra personnel to decrease the patient-to-nurse ratio • Only trained personnel may insert CVCs • Assistance in central line insertion or trainee supervision

Abbreviations: CLABSI, central line–associated bloodstream infection; CVC, central venous catheter; ICU, intensive care unit; IHI, Institute for Healthcare Improvement.

98 • CID 2014:59 (1 July) • QUALITY IMPROVEMENT

D ow

nloaded from https://academ

ic.oup.com /cid/article/59/1/96/404195 by guest on 13 M

ay 2021

article by Pronovost et al [21]was accessed to supplement infor- mation. One trial was not included for subgroup analysis be- cause, although pre- and postintervention initiatives were qualitatively different, no new intervention types were imple- mented [23]. Another study included multiple data sets, of which the set with the longest follow-up period was chosen [24]. Eleven studies could not be included for ITS analysis be- cause they implemented interventions in a stepwise manner [22, 23, 25–33].

The 43 studies involved primarily medical-surgical ICUs, implemented quality improvement interventions without simultaneously introducing novel prevention measures, and ap- plied CDC methods and definitions for CLABSI diagnosis (Supplementary Appendix 2).

The 584 included ICUs consisted of 564 adult, 11 pediatric [24, 34, 35], and 9 neonatal units [24]. Four studies reported the number of adult ICUs studied, but did not specify the ICU type (n = 270) [35–38]. The remaining 294 adult ICUs in- volved medical-surgical (n = 135), medical (n = 51), and surgi- cal (n = 61).

The meta-analysis consisted of 35 before–after [8, 22, 24, 26– 28, 33–61], 7 ITS [18, 20, 40, 62, 63, 64, 65], and 1 controlled be- fore–after study [66]. Five ITS studies were included in the meta-analysis of before–after study designs [18, 40, 63–65].

Duration of study periods ranged from 9 months [58] to 180 months [18], with a mean length of 26.75 months.

Up to 14 different types of interventions were reported. Stud- ies introduced multiple quality improvement interventions in different combinations, usually implementing 1–5 interven- tions (n = 34). Four studies implemented initiatives through im- provement systems such as plan-do-study-act, Six Sigma, and root cause analysis [36, 38, 42, 52].

Quality improvement interventions, details of their descrip- tion, methods used to apply them, and compliance measure reporting varied. Educational interventions consisted of single, monthly, quarterly, or yearly sessions. Feedback reporting of infection or compliance rates occurred at monthly or quarterly intervals. Surveillance of compliance with preventive interven- tions was implemented daily, periodically, or at random inter- vals. Likewise, studies reported compliance with different items or only during the intervention period.

Twenty-eight studies reported before–after device utilization rates (n = 10) [8, 27, 28, 34, 38, 44, 45, 54, 56, 57], catheterization duration (n = 11) [22, 24, 27, 39, 43, 51, 53, 57, 58, 61, 63], or pre- vention measure compliance (n = 18) [22, 24, 27, 33, 35, 36, 45, 46, 49, 53, 55–60, 63, 66]. Some studies reduced [24, 51] or in- creased duration of catheterization [27, 58], yet most improved compliance (n = 10) [24, 27, 35, 36, 45, 46, 49, 56, 60, 66]. Anal- ysis of 7 studies revealed device utilization rate increases [38,45, 57] and decreases (Supplementary Appendix 3) [8, 54, 56].

Half of trials implemented bundles or checklists (n = 20). Tri- als either introduced bundles without checklists (n = 2) [8, 45], only checklists because bundles were used during baseline (n = 9) [38, 44, 52, 53, 55, 57, 59, 62, 63], or both bundle and checklist interventions (n = 9) [27, 28, 33, 35, 37, 41, 43, 47, 54].

Differing amounts of preventive care items were grouped to- gether to form a bundle or checklist. Two trials [52, 59] did not report which items their bundle comprised, and 1 trial used a checklist a sole item [53]. Other trials used all 5 (n = 7) [8, 27, 37, 38, 43, 54, 63], 4 (n = 5) [28, 33, 41, 47, 62], 3 (n = 3) [35, 44, 55], or 2 (n = 2) [45, 57] IHI items in their bundle or checklist. The items “optimal catheter site selection” and “daily review of line necessity” were included least (Figure 2).

Four studies targeted other healthcare-associated infections such as ventilator-associated pneumonia (VAP) [36], both VAP and catheter-associated urinary tract infections [28, 59], or VAP and surgical site infections [34]. Eight studies initiated new prevention measures alongside quality improvement inter- ventions [26, 39, 40, 42, 43, 48, 52, 53].

The baseline CLABSI incidence varied; rates ranged from 2.1 [34] to 46.3 CLABSIs per 1000 catheter-days [46]. Trials report- ed baseline rates <5 [26, 27, 34–37, 43, 51–53, 57–59, 61, 62, 64] and >15 CLABSIs per 1000 catheter-days [18, 24, 41, 46, 49, 60].

Downs and Black quality assessment scores ranged from 15 [59] to 26 [22, 24, 49], with a mean of 21.2 (Supplementary

Figure 1. Study selection flow diagram. Abbreviations: CLABSI, central line–associated bloodstream infection; QI, quality improvement.

QUALITY IMPROVEMENT • CID 2014:59 (1 July) • 99

D ow

nloaded from https://academ

ic.oup.com /cid/article/59/1/96/404195 by guest on 13 M

ay 2021

Appendix 4). The checklist revealed that 2 studies did not de- scribe CLABSI definitions [40, 59], 9 did not sufficiently de- scribe their quality improvement interventions [24, 33, 43, 46, 53, 56,57, 59, 65],and 34 measured prevention measure compli- ance [8,22,24,27,28,33–41,43–46,48–51,53–61,63,66].Twen- ty-eight studies reported confounding factors such as device utilization rates, catheterization duration, patient characteris- tics, or injury severity [8, 20, 22, 24, 26–28, 34, 38, 39, 41, 43– 46, 49–54, 56–58, 61, 63, 66], which were comparable between baseline and postintervention in 18 trials [20, 22, 24, 27, 28, 39, 46, 49, 50, 52–54, 57, 58, 61, 63, 66]. Two trials corrected for these measured differences in patient characteristics [41, 51]. Studies tended to have either low (n = 25) or high (n = 14) power scores [8, 18, 20, 22, 24, 26, 34, 37, 38, 47, 49, 52, 56, 65].

Ten before–after trials did not demonstrate CLABSI rate de- creases [36, 39, 41,50, 51,55,57, 59,64, 66]. Two studies revealed nonsignificant results for neurosurgical, neurological, cardio- thoracic, and coronary care units, yet decreased their total CLABSI rate [46, 47]. ITS analysis demonstrated beneficial changes in infection rate slope [18] and levels at 3 [62], 6 [40], 12 [18, 40], and 24 months postintervention [18].

Meta-analysis was performed on 41 before–after and 7 ITS study designs to assess the impact of quality improvement inter- ventions on the occurrence of CLABSIs. Before–after trials showed reductions in the CLABSI rate (OR, 0.39 [95% CI, .33–.46]; P < .0001, Figure 3) with high statistical heterogeneity (I2 = 85.4%). Analysis of 6 ITS studies, involving 11 ICUs, iden- tified a change in level for the CLABSI rate at 3 months postin- tervention (OR, 0.30 [95% CI, .10–.88]; P = .028, Figure 4) with low heterogeneity (I2 = 24.5%). Changes in infection rate slope (OR, 0.81 [95% CI, .59–1.13]; P = .216) and levels at 6 (OR, 0.36

[95% CI, .11–1.19]; P = .094), 12 (OR, 0.17 [95% CI, .02–1.27]; P = .084), and 24 months postintervention (OR, 0.052 [95% CI, .003–1.02]; P = .051) trended toward reductions, yet were not significant (Supplementary Appendix 5).

Subgroup analysis of before–after trials revealed that the CLABSI risk reduction was significantly stronger (P = .026; Fig- ure 3) in trials with care bundles or checklists (OR, 0.34 [95% CI, .27–.41]) than in those without them (OR, 0.45 [95% CI, .36–.55]). Further analysis revealed that studies with baseline rates >4.0 CLABSIs per 1000 catheter-days (OR, 0.37 [95% CI, .33–.46]) did not demonstrate more pronounced risk reduc- tions (P = .18) compared with studies below this baseline infec- tion rate (OR, 0.49 [95% CI, .37–.66]). Low-power (OR, 0.33 [95% CI, .26–.42]) and high-power studies (OR, 0.44 [95% CI, .36–.54]) exhibited near-different rate reductions (P = .06).

Funnel plots displayed an asymmetrical pattern for before– after, but not ITS, study designs (Supplementary Appendix 6). The results of the sensitivity analysis of before-after study designs suggest that 2 studies contribute to residual heterogeneity; re- moving them from the meta-analysis would reduce variability be- tween studies [49, 52]. However, because this did not affect the results, these studies were retained (Supplementary Appendix 7).

DISCUSSION

This meta-analysis of 43 studies, involving 584 ICUs, provides evidence that quality improvement interventions reduce CLAB- SI rates in adult ICUs. The effect size of 41 studies was signifi- cant yet highly heterogeneous. This infection rate decrease was more pronounced in studies using bundles or checklists, suggest- ing that their implementation alongside other initiatives leads to stronger rate reductions. The change in infection rate level

Figure 2. Frequency of included care items in bundles and checklists. Eighteen trials implementing bundle or checklist interventions reported which care items were implemented during the intervention period. One study used povidone-iodine in their bundle instead of chlorhexidine [57]. Hand hygiene: per- sonnel practiced antiseptic hand hygiene before/after procedure; chlorhexidine skin antisepsis: skin disinfection before catheter insertion; maximal sterile barrier precautions: personnel wore sterile coat and gloves, mask, and hat during insertion, optimal catheter site selection: personnel strived to insert catheters in the subclavian vein, daily review of line necessity: catheter need was assessed daily with prompt removal of unnecessary central lines.

100 • CID 2014:59 (1 July) • QUALITY IMPROVEMENT

D ow

nloaded from https://academ

ic.oup.com /cid/article/59/1/96/404195 by guest on 13 M

ay 2021

for 6 studies at 3 months postintervention also demonstrates the beneficial impact of quality improvement interventions, with low heterogeneity. However, only 1 of these studies showed significant rate decreases [62], and the overall intervention ef- fect was not sustained over longer follow-up periods. These findings may reflect the presence of the Hawthorne effect and

need for CLABSI awareness promotion through continuous stepwise, multifaceted quality improvement interventions.

This study offers a broad look on the state of current research and applicable interventions, and applies a novel classification system to synthesize evidence for quality improvement initia- tives. The meta-analysis is the first to include before–after

Figure 3. Meta-analysis and subgroup analysis of before-after studies. Bundle/checklist interventions: studies implementing a bundle and/or checklist. Non–bundle/checklist interventions: studies implementing neither a bundle nor checklist. Baseline period: before intervention implementation. Post period: after start of intervention implementation. Events: total number of central line–associated bloodstream infections. Total: total number of central line days. W, weight assigned per study. Abbreviations: CI, confidence interval; OR, odds ration.

QUALITY IMPROVEMENT • CID 2014:59 (1 July) • 101

D ow

nloaded from https://academ

ic.oup.com /cid/article/59/1/96/404195 by guest on 13 M

ay 2021

studies and identify an additive preventive effect associated with bundle and checklist interventions. Two previous systema- tic reviews were unable to conclude which quality improve- ment interventions should be recommended for widespread implementation [16, 67]. Another recommended the use of educational programs and multidisciplinary teams [68]. A meta-analysis of ITS studies likewise demonstrated effect sizes with broad confidence intervals; however, they used different population criteria and studies, calculated rate reductions per quarter-year, reported mixed effects with small effect sizes, and did not investigate compliance measures. Additionally, the exclusion of before–after study designs discards much ob- servational evidence, negatively impacting the external validity of the results [19]. Comparable points of criticism were the low quality of included studies due to high baseline infection rates, inadequate reporting of multiple CLABSI data points, compli- ance measurements, and intervention details.

Although interventions implemented in settings with higher baseline rates would appear more likely to be successful, no dif- ference (P = .18) was found between studies with baseline infec- tion rates above or below a suboptimal rate of 4.0 CLABSI per 1000 catheter-days. Furthermore, high-power studies demon- strated CLABSI rate decreases not significantly different from low-power studies (P = .06). Noteworthy is that the study with the lowest baseline rate (2.1 CLABSI per 1000 catheter-days) still achieved a significant rate reduction by providing feedback of biannual infection rates [34].

Strengths of this study include the comprehensive search strat- egy encompassing various quality improvement interventions, the methodological quality assessment of trials, and the ran- dom-effects model analysis with multiple studies and ITS study designs. It is, however, hampered by certain limitations: a lack of randomized or controlled study designs, inconsistent reporting of prevention measure compliance, and heterogeneity. Before–after studies run a higher risk of bias due to their liberal study design, as they hamper the ability to recognize phenomena that influence

the CLABSI rate such as virulent epidemic outbreaks or sponta- neous regression to the mean [16].There is some evidence to sug- gest that the effects of quality improvement interventions are overestimated when based on before–after studies. Time series designs limit this risk of bias by detecting whether an interven- tion had an effect significantly greater than the underlying base- line trend [69].However, because these designs require initiatives to begin at a well-defined point in time, 11 studies with multifac- eted stepwise intervention implementation had to be excluded. This limitation could lead to an underestimation of the effect, as there is evidence for the effectiveness of gradual intervention introduction [70].

There are several issues related to the meta-analysis of be- fore–after studies. All quality improvement interventions were considered to have an equal impact, yet this assumption may not be fair. Assuming interventions take months to implement, those introduced in a later study period could have less effect compared with earlier initiatives. Inclusion of studies from identical authors can lead to bias [24, 49, 56, 60, 61, 65, 66]. Two of these studies were performed in the same hospital, which could overestimate the intervention effect due to hospital experience in intervention implementation [60, 66]. The forest plot of before–after studies revealed a lack of smaller studies with less drastic infection rate decreases, suggesting publication bias; however, subgroup analysis of high-power studies revealed CLABSI decreases. Nevertheless, analysis of ITS studies aims to avert these barriers, and there was little evidence of publication bias among those studies.

Interventions to change risk exposure confound results. Al- though statistically equivalent, a catheter-day from days 1–2 contains less infection risk than days 14–15 due to microbial bi- ofilm development and accumulating gaps in prevention mea- sure adherence. Studies that reduce device utilization rates with increased average catheterization duration, reflecting a cohort of patients no longer managed with short-term central line usage, could underestimate intervention effects and vice versa [27].

Figure 4. Meta-analysis of interrupted time series studies: change in central line-associated bloodstream infection rate level at 3 months postinterven- tion. Abbreviations: CI, confidence interval; OR, odds ratio.

102 • CID 2014:59 (1 July) • QUALITY IMPROVEMENT

D ow

nloaded from https://academ

ic.oup.com /cid/article/59/1/96/404195 by guest on 13 M

ay 2021

This impact is unclear, as studies with significant changes in de- vice utilization rates and duration of catheterization reported mixed effects. Analysis of catheterization duration was not fea- sible because CLABSI definitions do not account for usage of multiple catheters per patient.

Clinical and methodological heterogeneity stemmed from the use of differing intervention strategies, study designs, popula- tion characteristics, and baseline standards of care. No distinc- tion was made between interventions applied as part of a general program or introduced to solve a specific recurring problem. For example, one study formed a team of nurses to evaluate care processes related to an infection rate increase. By applying a comparable yet distinct multifaceted quality im- provement strategy, they decreased their rate from 1.5 to 0 CLABSIs per 1000 catheter-days [23]. Differing standards of care hinder comparison through meta-analysis. The effect of implemented quality improvement interventions is dependent on the efficacy or amount of baseline prevention measures. Si- multaneous introduction of daily chlorhexidine bathing along- side a quality improvement initiative may have influenced one ITS study’s intervention effect [40]. Last, this review did not aim to identify strategies that lead to optimal uptake of quality im- provement initiatives.

In conclusion, the results of this meta-analysis provide evi- dence that quality improvement interventions reduce CLABSI in adult ICUs. Forty-one before–after studies demonstrated consistent, beneficial results, which appeared to be more pro- nounced among studies implementing bundle and checklist in- terventions. Quality improvement interventions appeared equally effective in studies with low and high power or baseline CLABSI rate settings. The CLABSI rate reduction appears to be confirmed by the methodologically more robust interrupted time series studies. Further research should assess requirements for successful adaptation of quality improvement interventions, for example, through improvement systems, over longer follow- up periods. Studies should report before–after compliance measures, device utilization rates, and catheterization duration. These latter 2 items are necessary to assess confounding factors, because increased catheter use for shorter durations leads to intervention effect overestimations. To properly address these issues, studies need to account for the number of catheters per patient. Finally, studies should apply ITS study designs and, when introducing stepwise initiatives, enough time should be spaced between interventions to facilitate ITS analysis.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online (http://cid.oxfordjournals.org). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the

sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.

Notes

Author contributions. K. B. conceived of and designed the study; per- formed the search of published work, literature search, data acquisition, in- terpretation and synthesis, and statistical analysis; and wrote the paper. J. B. performed the statistical analysis, contributed to data interpretation, and re- vised the statistical portions of the report. D. Vo. substantially contributed to data analysis and interpretation and critically revised the final manuscript. S. B. designed the study; substantially contributed to the search of published work, data interpretation and synthesis; and critically revised the final manuscript. D. Va. conceived of and designed the study; substan- tially contributed to data interpretation and synthesis; and critically revised the final manuscript. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication. Financial support. S. B. holds a research mandate of the Specific Re-

search Fund at Ghent University. Potential conflicts of interest. D. Vo. has received an institutional grant

for work under consideration for publication from Pfizer, and has been a consultant for Astellas, Pfizer, and Tibotec. All other authors report no po- tential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential

Conflicts of Interest. Conflicts that the editors consider relevant to the con- tent of the manuscript have been disclosed.

References

1. Centers for Disease Control and Prevention. Central line–associated bloodstream infection (CLABSI) event protocol, 2014: 1–10. Available at: http://www.cdc.gov. Accessed 9 February 2014.

2. Centers for Disease Control and Prevention. NHSN key terms. Atlanta, GA: CDC, 2014: 1–16. Available at: http://www.cdc.gov. Accessed 22 March 2014.

3. Blot SI, Depuydt P, Annemans L, et al. Clinical and economic outcomes in critically ill patients with nosocomial catheter-related bloodstream infections. Clin Infect Dis 2005; 41:1591–8.

4. Umscheid CA, Mitchell MD, Doshi JA, Agarwal R, Williams K, Bren- nan PJ. Estimating the proportion of healthcare-associated infections that are reasonably preventable and the related mortality and costs. In- fect Control Hosp Epidemiol 2011; 32:101–14.

5. Rosenthal VD. Central line–associated bloodstream infections in limit- ed‐resource countries: a review of the literature. Clin Infect Dis 2009; 49:1899–907.

6. Olaechea PM, Palomar M, Álvarez-Lerma F, et al. Morbidity and mor- tality associated with primary and catheter-related bloodstream infec- tions in critically ill patients. Rev Esp Quimioter 2013; 26:21–9.

7. Januel JM, Harbarth S, Allard R, et al. Estimating attributable mortality due to nosocomial infections acquired in intensive care units. Infect Control Hosp Epidemiol 2010; 31:388–94.

8. Marra AR, Cal RG, Durao MS, et al. Impact of a program to prevent central line-associated bloodstream infection in the zero tolerance era. Am J Infect Control 2010; 38:434–9.

9. Pronovost PJ, Goeschel CA, Colantuoni E, et al. Sustaining reductions in catheter related bloodstream infections in Michigan intensive care units: observational study. BMJ 2010; 340:c309.

10. O’Grady NP, Alexander M, Burns LA, et al. Guidelines for the preven- tion of intravascular catheter-related infections. Clin Infect Dis 2011; 52:e162–93.

11. Yokoe DS, Mermel LA, Anderson DJ, et al. Executive summary: a compendium of strategies to prevent healthcare‐associated infec- tions in acute care hospitals. Infect Control Hosp Epidemiol 2008; 29:S12–21.

QUALITY IMPROVEMENT • CID 2014:59 (1 July) • 103

D ow

nloaded from https://academ

ic.oup.com /cid/article/59/1/96/404195 by guest on 13 M

ay 2021

12. Maki DG, Kluger DM, Crnich CJ. The risk of bloodstream infection in adults with different intravascular devices: a systematic review of 200 published prospective studies. Mayo Clin Proc 2006; 81:1159–71.

13. Blot S, Vandijck D, Vogelaers D, Labeau S. Bridging the gap between theory and practice. ICU Management 2011; 11:40–11.

14. Labeau SO, Vandijck DM, Rello J, et al. Centers for Disease Control and Prevention guidelines for preventing central venous catheter-related in- fection: results of a knowledge test among 3405 European intensive care nurses. Crit Care Med 2009; 37:320–23.

15. Cabana MD, Rand CS, Powe NR, et al. Why don’t physicians follow clinical practice guidelines? A framework for improvement. JAMA 1999; 282:1458–65.

16. Ranji SR, Shetty K, Posley KA, et al. Closing the quality gap: a critical analysis of quality improvement strategies. Vol 6: Prevention of health- care–associated infections. Rockville, MD: Agency for Healthcare Re- search and Quality, 2007.

17. Downs SH, Black N. The feasibility of creating a checklist for the assess- ment of the methodological quality both of randomised and non-rand- omised studies of health care interventions. J Epidemiol Community Health 1998; 52:377–84.

18. Fraher MH, Collins CJ, Bourke J, Phelan D, Lynch M. Cost-effectiveness of employing a total parenteral nutrition surveillance nurse for the pre- vention of catheter-related bloodstream infections. J Hosp Infect 2009; 73:129–34.

19. Flodgren G, Conterno LO, Mayhew A, Omar O, Pereira CR, Shepperd S. Interventions to improve professional adherence to guidelines for prevention of device-related infections. Cochrane Database Syst Rev 2013. doi:10.1002/14651858.CD006559.pub2.

20. Coopersmith CM, Rebmann TL, Zack JE, et al. Effect of an education program on decreasing catheter-related bloodstream infections in the surgical intensive care unit. Crit Care Med 2002; 30:59–64.

21. Pronovost P, Needham D, Berenholtz S, et al. An intervention to de- crease catheter-related bloodstream infections in the ICU. N Engl J Med 2006; 355:2725–32.

22. Coopersmith CM, Zack JE, Ward MR, et al. The impact of bedside be- havior on catheter-related bacteremia in the intensive care unit. Arch Surg 2004; 139:131–6.

23. Richardson J, Tjoelker R. Beyond the central line-associated blood- stream infection bundle: the value of the clinical nurse specialist in con- tinuing evidence-based practice changes. Clin Nurse Spec 2012; 26:205–11.

24. Rosenthal VD, Maki DG, Rodrigues C, et al. Impact of International Nosocomial Infection Control Consortium (INICC) strategy on central line–associated bloodstream infection rates in the intensive care units of 15 developing countries. Infect Control Hosp Epidemiol 2010; 31:1264–72.

25. Berenholtz SM, Pronovost PJ, Lipsett PA, et al. Eliminating catheter-re- lated bloodstream infections in the intensive care unit. Crit Care Med 2004; 32:2014–20.

26. Cherry-Bukowiec JR, Denchev K, Dickinson S, et al. Prevention of cath- eter-related blood stream infection: back to basics? Surg Infect (Larchmt) 2011; 12:27–32.

27. Longmate AG, Ellis KS, Boyle L, et al. Elimination of central-venous- catheter-related bloodstream infections from the intensive care unit. BMJ Qual Saf 2011; 20:174–80.

28. Miller RS, Norris PR, Jenkins JM, et al. Systems initiatives reduce healthcare-associated infections: a study of 22,928 device days in a sin- gle trauma unit. J Trauma 2010; 68:23–31.

29. Munoz-Price LS, Dezfulian C, Wyckoff M, et al. Effectiveness of step- wise interventions targeted to decrease central catheter-associated bloodstream infections. Crit Care Med 2012; 40:1464–9.

30. Ong A, Dysert K, Herbert C, et al. Trends in central line–associated bloodstream infections in a trauma-surgical intensive care unit. Arch Surg 2011; 146:302–7.

31. Ramos ER, Reitzel R, Jiang Y, et al. Clinical effectiveness and risk of emerging resistance associated with prolonged use of antibiotic-

impregnated catheters: more than 0.5 million catheter days and 7 years of clinical experience. Crit Care Med 2011; 39:245–51.

32. Royer T. Implementing a better bundle to achieve and sustain a zero central line-associated bloodstream infection rate. J Infus Nurs 2010; 33:398–406.

33. Seddon ME, Hocking CJ, Mead P, Simpson C. Aiming for zero: decreas- ing central line associated bacteraemia in the intensive care unit. N Z Med J 2011; 124:9–21.

34. Gastmeier P, Geffers C, Brandt C, et al. Effectiveness of a nationwide nosocomial infection surveillance system for reducing nosocomial in- fections. J Hosp Infect 2006; 64:16–22.

35. McLaws ML, Burrell AR. Zero risk for central line-associated blood- stream infection: are we there yet? Crit Care Med 2012; 40:388–93.

36. Bonello RS, Fletcher CE, Becker WK, et al. An intensive care unit qual- ity improvement collaborative in nine Department of Veterans Affairs hospitals: reducing ventilator-associated pneumonia and catheter-relat- ed bloodstream infection rates. Jt Comm J Qual Patient Saf 2008; 34:639–45.

37. Koll BS, Straub TA, Jalon HS, Block R, Heller KS, Ruiz RE. The CLABs collaborative: a regionwide effort to improve the quality of care in hos- pitals. Jt Comm J Qual Patient Saf 2008; 34:713–23.

38. Render ML, Hasselbeck R, Freyberg RW, Hofer TP, Sales AE, Almenoff PL. Reduction of central line infections in Veterans Administration intensive care units: an observational cohort using a central infrastruc- ture to support learning and improvement. BMJ Qual Saf 2011; 20: 725–32.

39. Bijma R, Girbes AR, Kleijer DJ, Zwaveling JH. Preventing central ve- nous catheter-related infection in a surgical intensive-care unit. Infect Control Hosp Epidemiol 1999; 20:618–20.

40. Dixon JM, Carver RL. Daily chlorohexidine gluconate bathing with im- pregnated cloths results in statistically significant reduction in central line-associated bloodstream infections. Am J Infect Control 2010; 38:817–21.

41. Duane TM, Brown H, Borchers CT, et al. A central venous line protocol decreases bloodstream infections and length of stay in a trauma inten- sive care unit population. Am Surg 2009; 75:1166–70.

42. Frankel HL, Crede WB, Topal JE, Roumanis SA, Devlin MW, Foley AB. Use of corporate Six Sigma performance-improvement strategies to re- duce incidence of catheter-related bloodstream infections in a surgical ICU. J Am Coll Surg 2005; 201:349–58.

43. Galpern D, Guerrero A, Tu A, Fahoum B, Wise L. Effectiveness of a cen- tral line bundle campaign on line-associated infections in the intensive care unit. Surgery 2008; 144:492–5.

44. Gozu A, Clay C, Younus F. Hospital-wide reduction in central line-as- sociated bloodstream infections: a tale of two small community hospi- tals. Infect Control Hosp Epidemiol 2011; 32:619–22.

45. Guerin K, Wagner J, Rains K, Bessesen M. Reduction in central line- associated bloodstream infections by implementation of a postinsertion care bundle. Am J Infect Control 2010; 38:430–3.

46. Higuera F, Rosenthal VD, Duarte P, Ruiz J, Franco G, Safdar N. The effect of process control on the incidence of central venous catheter- associated bloodstream infections and mortality in intensive care units in Mexico. Crit Care Med 2005; 33:2022–7.

47. Kim JS, Holtom P, Vigen C. Reduction of catheter-related bloodstream infections through the use of a central venous line bundle: epidemiolog- ic and economic consequences. Am J Infect Control 2011; 39:640–6.

48. Lopez AC. A quality improvement program combining maximal barrier precaution compliance monitoring and daily chlorhexidine gluconate baths resulting in decreased central line bloodstream infections. Dimens Crit Care Nurs 2011; 30:293–8.

49. Rosenthal VD, Guzman S, Pezzotto SM, Crnich CJ. Effect of an infec- tion control program using education and performance feedback on rates of intravascular device-associated bloodstream infections in inten- sive care units in Argentina. Am J Infect Control 2003; 31:405–9.

50. Santana SL, Furtado GH, Wey SB, Medeiros EA. Impact of an education program on the incidence of central line-associated bloodstream

104 • CID 2014:59 (1 July) • QUALITY IMPROVEMENT

D ow

nloaded from https://academ

ic.oup.com /cid/article/59/1/96/404195 by guest on 13 M

ay 2021

infection in 2 medical-surgical intensive care units in Brazil. Infect Con- trol Hosp Epidemiol 2008; 29:1171–3.

51. Seguin P, Laviolle B, Isslame S, Coue A, Malledant Y. Effectiveness of simple daily sensitization of physicians to the duration of central venous and urinary tract catheterization. Intensive Care Med 2010; 36:1202–6.

52. Shannon RP, Frndak D, Grunden N, et al. Using real-time problem solving to eliminate central line infections. Jt Comm J Qual Patient Saf 2006; 32:479–87.

53. Tsuchida T, Makimoto K, Toki M, Sakai K, Onaka E, Otani Y. The ef- fectiveness of a nurse-initiated intervention to reduce catheter-associat- ed bloodstream infections in an urban acute hospital: an intervention study with before and after comparison. Int J Nurs Stud 2007; 44:1324–33.

54. Venkatram S, Rachmale S, Kanna B. Study of device use adjusted rates in health care-associated infections after implementation of ‘bundles’ in a closed-model medical intensive care unit. J Crit Care 2010; 25:174. e11–18.

55. Wall RJ, Ely EW, Elasy TA, et al. Using real time process measurements to reduce catheter related bloodstream infections in the intensive care unit. Qual Saf Health Care 2005; 14:295–302.

56. Warren DK, Cosgrove SE, Diekema DJ, et al. A multicenter intervention to prevent catheter-associated bloodstream infections. Infect Control Hosp Epidemiol 2006; 27:662–9.

57. Yoo S, Ha M, Choi D, Pai H. Effectiveness of surveillance of central catheter-related bloodstream infection in an ICU in korea. Infect Con- trol Hosp Epidemiol 2001; 22:433–6.

58. Zingg W, Imhof A, Maggiorini M, Stocker R, Keller E, Ruef C. Impact of a prevention strategy targeting hand hygiene and catheter care on the incidence of catheter-related bloodstream infections. Crit Care Med 2009; 37:2167–80.

59. Chua C, Wisniewski T, Ramos A, Schlepp M, Fildes JJ, Kuhls DA. Mul- tidisciplinary trauma intensive care unit checklist: impact on infection rates. J Trauma Nurs 2010; 17:163–6.

60. Lobo RD, Levin AS, Gomes LM, et al. Impact of an educational program and policy changes on decreasing catheter-associated bloodstream

infections in a medical intensive care unit in Brazil. Am J Infect Control 2005; 33:83–7.

61. Warren DK, Zack JE, Cox MJ, Cohen MM, Fraser VJ. An educational intervention to prevent catheter-associated bloodstream infections in a nonteaching, community medical center. Crit Care Med 2003; 31:1959–63.

62. Berriel-Cass D, Adkins FW, Jones P, Fakih MG. Eliminating nosocomial infections at Ascension Health. Jt Comm J Qual Patient Saf 2006; 32:612–20.

63. Peredo R, Sabatier C, Villagra A, et al. Reduction in catheter-related bloodstream infections in critically ill patients through a multiple sys- tem intervention. Eur J Clin Microbiol Infect Dis 2010; 29:1173–7.

64. Perez Parra A, Cruz Menarguez M, Perez Granda MJ, Tomey MJ, Padi- lla B, Bouza E. A simple educational intervention to decrease incidence of central line-associated bloodstream infection (CLABSI) in intensive care units with low baseline incidence of CLABSI. Infect Control Hosp Epidemiol 2010; 31:964–7.

65. Warren DK, Zack JE, Mayfield JL, et al. The effect of an education pro- gram on the incidence of central venous catheter-associated blood- stream infection in a medical ICU. Chest 2004; 126:1612–8.

66. Lobo RD, Levin AS, Oliveira MS, et al. Evaluation of interventions to reduce catheter-associated bloodstream infection: continuous tailored education versus one basic lecture. Am J Infect Control 2010; 38:440–8.

67. Safdar N, Abad C. Educational interventions for prevention of health- care-associated infection: a systematic review. Crit Care Med 2008; 36:933–40.

68. Aboelela SW, Stone PW, Larson EL. Effectiveness of bundled behaviou- ral interventions to control healthcare-associated infections: a system- atic review of the literature. J Hosp Infect 2007; 66:101–8.

69. Eccles M, Grimshaw J, Campbell M, Ramsay C. Research designs for studies evaluating the effectiveness of change and improvement strate- gies. Qual Saf Health Care 2003; 12:47–52.

70. Prior M, Guerin M, Grimmer-Somers K. The effectiveness of clinical guideline implementation strategies—a synthesis of systematic review findings. J Eval Clin Pract 2008; 14:888–97.

QUALITY IMPROVEMENT • CID 2014:59 (1 July) • 105

D ow

nloaded from https://academ

ic.oup.com /cid/article/59/1/96/404195 by guest on 13 M

ay 2021