SYSTEMATIC REVIEW BMJ Qual Saf: first published as 10.1136/bmjqs-2019-009552 on 7 August 2019. Downloaded from Effectiveness of double checking to reduce medication administration errors: a systematic review Alain K Koyama, 1 Claire-Sophie Sheridan Maddox,1 Ling Li,1 Tracey Bucknall,2,3 Johanna I Westbrook 1 ► Additional material is ABSTRACT INTRODUCTION published online only. To view Background Double checking medication Medication safety continues to present a please visit the journal online administration in hospitals is often standard practice, (http:// dx. doi. org/ 10. 1136/ serious challenge in hospitals. Processing bmjqs- 2019- 009552). particularly for high-risk drugs, yet its effectiveness medications involves multiple steps in reducing medication administration errors (MAEs) 1 and individuals. Medication errors can Centre for Health Systems and and improving patient outcomes remains unclear. We occur during different stages, with a Safety Research, Australian conducted a systematic review of studies evaluating high frequency occurring during admin- Institute of Health Innovation, evidence of the effectiveness of double checking to Faculty of Medicine and Health 1 2 reduce MAEs. istration. Medication administration Sciences, Macquarie University, Methods Five databases (PubMed, Embase, CINAHL, errors (MAEs) are reported to occur in Macquarie Park, New South 3 4 Wales, Australia Ovid@Journals, OpenGrey) were searched for studies 20% to 25% of dose administrations. 2School of Nursing and evaluating the use and effectiveness of double checking While prescribing and dispensing errors Midwifery, Centre for Quality on reducing medication administration errors in a can be intercepted as a medication order and Patient Safety Research, 5 hospital setting. Included studies were required to report proceeds towards patient administration, Faculty of Health, Deakin any of three outcome measures: an effect estimate University, Geelong, Victoria, interventions to reduce errors during Australia such as a risk ratio or risk difference representing the administration are especially critical as it 3Alfred Health, Melbourne, VIC, association between double checking and MAEs, or is the final step before a patient receives a Australia between double checking and patient harm; or a rate drug.5 Various strategies have been devel- representing adherence to the hospital’s double checking oped and implemented in clinical practice Correspondence to policy. Dr Alain K Koyama, Centre to minimise MAEs. Results Thirteen studies were identified, including http://qualitysafety.bmj.com/ for Health Systems and Safety 10 studies using an observational study design, two The process of double checking is Research, Australian Institute of adopted as standard safety practice in Health Innovation, Macquarie randomised controlled trials and one randomised trial University, Macquarie Park, in a simulated setting. Studies included both paediatric many hospitals, as well as in other high- NSW 2113, Australia; and adult inpatient populations and varied considerably hazard industries such as aviation and alain. koyama@ mq. edu. au in quality. Among three good quality studies, only nuclear power.6 one showed a significant association between double Received 12 March 2019 Double checking medication admin- checking and a reduction in MAEs, another showed no Revised 23 July 2019 istration involves two individuals veri- Accepted 24 July 2019 association, and the third study reported only adherence fying the same information, while single rates. No studies investigated changes in medication- checking involves a single individual veri- related harm associated with double checking. Reported on September 26, 2021 by guest. Protected copyright. fying the information. double checking adherence rates ranged from 52% to 97% of administrations. Only three studies reported if The potential safety benefits of double and how independent and primed double checking were checking rely on two key factors: two differentiated. separate individuals verifying key infor- Conclusion There is insufficient evidence that double mation and independent verification. Two versus single checking of medication administration is individuals should result in fewer errors associated with lower rates of MAEs or reduced harm. by minimising endogenous errors that © Author(s) (or their Most comparative studies fail to define or investigate employer(s)) 2019. Re-use arise from one individual and are there- permitted under CC BY-NC. No the level of adherence to independent double checking, fore independent from errors that may commercial re-use. See rights further limiting conclusions regarding effectiveness arise in another individual.7 Exogenous and permissions. Published by in error prevention. Higher-quality studies are needed errors that arise from external factors, BMJ. to determine if, and in what context (eg, drug type, such as illegible text, are potentially setting), double checking produces sufficient benefits To cite: Koyama AK, in patient safety to warrant the considerable resources reduced through independent double Maddox C-SS, Li L, et al. checking when verification is performed BMJ Qual Saf Epub ahead of required. print: [please include Day PROSPERO REGISTRATION NUMBER without one checker priming the other 7–9 Month Year]. doi:10.1136/ CRD42018103436. with information to be verified. Several bmjqs-2019-009552 studies have, however, demonstrated that Koyama AK, et al. BMJ Qual Saf 2019;0:1–9. doi:10.1136/bmjqs-2019-009552 1 Systematic review BMJ Qual Saf: first published as 10.1136/bmjqs-2019-009552 on 7 August 2019. Downloaded from organisational double checking policies often differ in report at least one of the following quantitative meas- their level of detail of how the double-checking process ures: (1) an effect estimate such as a risk ratio or risk should be conducted, contributing to variation in the difference representing the association between double application by nurses.10–12 Some organisations require checking (compared with single checking) and MAEs; double checking for all medications while others only (2) an analogous effect estimate for the outcome of for high-risk medications such as opioids, chemo- patient harm. As the effectiveness of double checking therapeutic agents and intravenous drugs. Significant can depend in part on the extent to which nurses resources are required, given the process requires two adhere to the double checking policy, studies were individuals instead of one. also included if they reported a rate of adherence to Evidence that the process is effective in reducing double checking. Studies using either an observational errors is central to ensuring that this ingrained policy study design or randomised controlled trial (RCT) is justified in terms of resource use and workflow design were included. Studies using an observational disruptions.13 14 Double checking has been imple- study design are non-experimental studies that do not mented in hospitals based on an assumption that it randomise an intervention, which in this instance is will result in fewer MAEs, and at times as a response the double-checking process. The observational study to incidents when single checking was assessed to design categorisation is distinguished from the method have contributed to a serious error.15–19 However, its of ‘direct observation’ which is the use of observers to effectiveness in reducing MAEs and improving patient collect information during a study. Studies involving outcomes remains unclear. Much previous literature administrations of all medication types or select groups on double checking involves qualitative studies. A of medications were included as were those involving previous systematic review of studies published prior adult and/or paediatric populations. Only English-lan- to October 2010 investigated double checking during guage studies published in peer-reviewed journals were medication dispensing and administration and found included; abstracts and case studies were excluded. only three quantitative studies which provided insuf- ficient evidence to assess the effectiveness of double Data extraction and quality assessment checking in error reduction.20 Serious review of Data extraction was performed by two reviewers this safety procedure is unable to proceed without a (AKK and C-SSM) using a standardised extraction sound evidence-base. Thus, we performed a system- form. Abstracted variables were used to characterise atic review to examine contemporary evidence of the studies and assess study quality. Variables included first effectiveness of double checking to reduce medication author name, year of publication, country of study, administration errors and associated harm to identify years of data collection, study design, patient popu- both the strength of that evidence and where future lation, sample size and types of medications studied. http://qualitysafety.bmj.com/ research needs to focus. If medication errors were assessed, the types of medi- cations and errors, as well as the method of assessing METHODS errors were recorded. Variables specific to the double- Search strategy checking process included the definition of double Two reviewers (AKK and C-SSM) independently checking used, how double checking was measured performed each step of the literature search. A (eg, through self-report or direct observation) and the Boolean search strategy (online supplementary eTables policy of double checking at the participating institu- 1 and 2) was used