Effect of medication reconciliation interventions on outcomes: A systematic overview of systematic reviews

Abstract

Purpose

To evaluate and summarize published evidence from systematic reviews examining medication reconciliation.

Methods

MEDLINE, the Cochrane Database of Systematic Reviews, and the Database of Abstracts of Reviews of Effects were searched for English-language systematic reviews published from January 2004 to March 2019. Reviewers independently extracted information and scored review quality using the Assessment of Multiple Systematic Reviews (AMSTAR) tool. For reviews with AMSTAR scores above 7, Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology was applied to assess evidence quality, with evidence summarized and conclusions compared across reviews.

Results

Eleven reviews met the inclusion criteria, 5 of which used meta-analytic pooling. Most systematic reviews included primary studies of comprehensive bundled interventions that featured medication reconciliation as a central component. Reviews largely focused on transitions into and out of hospital settings. Five reviews focused exclusively on pharmacist-led interventions. Of the 5 reviews that considered all types of medication discrepancies, 3 reviews found very low-quality evidence that interventions reduced medication discrepancies. Neither of the 2 reviews that examined clinically significant medication discrepancies found any intervention effect. Of the 5 reviews that examined healthcare utilization outcomes, only 1 found any intervention effect, and that finding was based on low- to very low-quality evidence. Four reviews considered clinical outcomes, but none found any intervention effect.

Conclusion

An overview of systematic reviews of medication reconciliation interventions found 9 high-quality systematic reviews. A minority of those reviews’ conclusions were consistent with medication reconciliation alone having a measurable impact, and such conclusions were almost all based on very low-quality evidence.

KEY POINTS

• Eleven high-quality systematic reviews addressing medication reconciliation interventions, predominantly focused on actions at hospital admission and/or discharge, were identified.

• Two of 6 systematic reviews concluded that medication reconciliation interventions reduced medication discrepancies; however, these conclusions were based on very low-quality evidence, and there was little evidence that medication reconciliation interventions improved clinical outcomes or healthcare utilization.

• We recommend that medication reconciliation be targeted to care settings where its face validity is highest and that quality assessment efforts measure benefits achievable with medication reconciliation rather than the act of medication reconciliation itself.

Clinicians frequently have difficulty obtaining a complete list of patients’ medications or understanding medication regimens. This is especially true when patients are taking many medications, have difficulty communicating or understanding medication regimens, or are undergoing care transitions.¹ Discrepancies in medication reconciliation have the potential to negatively impact the quality and safety of patient care and healthcare utilization. Up to two-thirds of patients have at least 1 unintended medication discrepancy upon hospital admission,^2,3 and up to 81% of patients discharged from the hospital experience at least 1 medication discrepancy.⁴ One-third of unintended discrepancies at admission have the potential to cause severe discomfort or clinical deterioration.³ Furthermore, higher rates of 30-day readmissions have been observed among patients experiencing medication discrepancies.⁵

Medication reconciliation, or purposefully working to understand and document the medication regimen that a patient has been prescribed and is taking, frequently requires reconciling medication records from different sources (e.g., patient, caregiver, and family interviews; medical records; insurance claims). This strategy has been endorsed by several patient safety organizations, including the Joint Commission,⁶ the World Health Organization,⁷ and other international entities.^8,9 These endorsements and the resulting widespread dissemination of recommendations on medication reconciliation have led to extensive evaluations. Hundreds of primary studies and dozens of systematic reviews have assessed the impact of medication reconciliation on both clinical and utilization outcomes.

Aggregating the central findings of this literature can be unwieldy, not only due to its volume but also because findings may differ with differences in outcome measures, study populations, clinical settings, methods of analysis, and the quality of primary studies and reviews. Accordingly, this overview of reviews aims to appraise and summarize evidence from high-quality systematic reviews examining the impact of interventions involving medication reconciliation. Specifically, we sought to understand whether medication reconciliation has been shown to improve medication discrepancy rates and downstream patient-centered outcomes, including measures representing morbidity, mortality, and utilization.

Methods

We performed a systematic overview in accordance with the Preferred Reporting Items for Systematic Reviews and Meta Analyses (PRISMA) statement (eAppendix A).¹⁰ We also referenced published literature explicitly focused on methodological practices for overviews of systematic reviews.^11,12 We did not register a protocol prior to performing this review.

Data sources and searches.

We searched the literature in March 2019 using MEDLINE, the Cochrane Database of Systematic Reviews, and the Database of Abstracts of Reviews of Effects (DARE). Two investigators developed a search strategy that predominantly consisted of identifying systematic reviews with the term medication reconciliation in the title or abstract (see eAppendix B for details). The searches were limited to English-language articles published from January 2004 to March 2019. One investigator also searched the reference list of each included systematic review to identify other published or unpublished sources.

Selection of systematic reviews

Systematic reviews were eligible for inclusion in the overview if they were systematic reviews (with or without meta-analyses) that examined the effects of interventions that included medication reconciliation as a core part of an intervention implemented in adult or pediatric populations. For the purposes of this overview, we considered a systematic review to be any attempt to collate empirical evidence that fit prespecified eligibility criteria in order to answer a specific research question using explicit, systematic methods.¹³ Systematic reviews of studies with any study design and outcome(s) were included. We excluded reviews focused exclusively on interventions implemented in low- to middle-income countries due to differences in care practices and healthcare infrastructure.

Two investigators independently screened titles and abstracts in accordance with the prespecified inclusion and exclusion criteria. Next, full-text publications were retrieved and examined by 2 reviewers to determine eligibility; disagreements at either the title and abstract or full-text screening level were resolved by consensus in group meetings of research team members.

Quality evaluation

We assessed the methodological quality of each relevant systematic review with the Assessment of Multiple Systematic Reviews (AMSTAR) instrument.¹⁴ The tool consists of 11 domains, and each systematic review may receive a score ranging from 0 (lowest quality) to 11 (highest quality). Two reviewers independently applied the instrument, and scoring discrepancies were reconciled through oral discussion. Systematic reviews with an AMSTAR score below 8, a commonly applied threshold, were excluded from data extraction and synthesis.¹²

Data extraction

Research team members independently extracted data related to key characteristics using a standardized data extraction tool. Extracted variables included the following: dates of literature search; number and design of included primary studies; intervention types; patient populations; patient care settings; outcome measures, including medication discrepancies and measures representing morbidity, mortality, and utilization; and, if present, meta-analytic techniques and pooled estimates. Additionally, information concerning reviews’ major conclusions regarding intervention effectiveness were extracted. Reviewers compared extracted data and reconciled discrepancies through oral discussion.

GRADE quality of evidence

We documented the quality of evidence for each major conclusion within each review by applying Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology.¹⁵ We used objective criteria to assess evidence quality for each outcome in the following GRADE domains: study design; study quality; consistency (of effects between studies); directness (i.e., applicability of participants, interventions, and outcomes to the study question); and other modifying factors, including data sparseness (i.e., sample size) and strength of effect estimates. By combining item scores for each of these domains, we determined the level of evidence, which was classified into 1 of 4 categories:

• High—Further research is very unlikely to change our confidence in the estimate of effect.

• Moderate—Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.

• Low—Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.

• Very low—Any estimate of effect is very uncertain.

We did not assess the quality of the individual studies evaluated in the reviews but reported the risk of study bias as documented in the reviews. If a systematic review provided GRADE levels of evidence for outcomes (in the case of Cochrane Database reviews) we reported these assessments. Further details regarding how each GRADE domain was operationalized may be found in eAppendix D. One author assessed GRADE level of evidence for each systematic review, while a second author reviewed and verified the assessments; discrepancies were resolved through oral discussion.

Data synthesis

Two research team members independently examined each systematic review’s major conclusions regarding the effectiveness of intervention strategies on the reported primary outcomes and classified authors’ conclusions into 1 of 4 distinct categories: (1) a positive association between intervention strategy and outcome, (2) a negative association between intervention strategy and outcome, (3) a null association between intervention strategy and outcome, and (4) an inability to draw conclusions due to limited or low-quality studies. It was also documented whether conclusions were based on quantitative or qualitative (i.e., meta-analytic) assessments. Research team members compared synthesis information and reconciled discrepancies through oral discussion.

Results

Study selection

The literature search identified 119 articles. Upon screening titles and abstracts, 28 citations were selected for full-text review. After review of the full text of these articles, 18 systematic reviews met the inclusion criteria^16-33; 11 received an AMSTAR score of greater than 8 and were thus included in our review.^{16–20,22,24,26,31–33} The PRISMA flow diagram may be found in Figure 1, and AMSTAR scores may be found in eAppendix C.

Figure 1.

PRISMA flow diagram of publication selection for inclusion in overview.

Study characteristics

The major characteristics of included systematic reviews may be found in Table 1. All 11 reviews were published after 2009,^{16–20,22,24,26,31–33} with 7 published since 2014.^{19,20,24,26,31–33} All of the identified reviews were published in peer-reviewed journals.^{16–20,22,24,26,31–33} Six reviews were exclusively descriptive in nature^16–20,22; in 5 reviews, results were pooled using meta-analytic techniques.^{24,26,31–33}

Table 1.

Major Characteristics of All Systematic Reviews Included in Overview (n = 11)^a

Researchers (Yr Published)	Literature Search Dates	Number and Design of Evaluated Studies	Population	Intervention Type(s)	Setting; Intervention Timing	Primary Outcome Measure(s)	Pooled Effect Estimate(s)	Researchers’ Major Conclusion(s)
Bayoumi et al.16 (2009)	Through Mar 2008	4 studies (3 before-and-after studies, 1 RCT)	Community-dwelling adults	Medication reconciliation	Ambulatory care; during care or hospital discharge	Medication discrepancies	NR	There is no good-quality evidence demonstrating the effectiveness of medication reconciliation in the primary care setting.
Chhabra et al.17 (2012)	Through Aug 2010	7 studies (2 RCTs, 4 quasi-experimental studies, 1 observational study)	Older adults	Medication reconciliation	Long-term care facilities; admission and/or discharge	Varied; no prespecified outcomes	NR	Improvement in outcomes examined was shown in all studies; however, there were study design flaws. There is a need for well-designed studies demonstrating the effectiveness of medication reconciliation in long-term care settings.
Domingo et al.18 (2012)	NR	3 studies (2 RCTs, 1 quasi-RCT)	Adults with community-acquired pneumonia	Structured discharge with medication reconciliation	Hospital; discharge	Hospital readmissions, ED visits, and unscheduled visits to healthcare provider after discharge	NR	Structured discharge with medication reconciliation did not have a significant impact on hospital readmissions.
Ensing et al.19 (2015)	Inception–Nov 2014	30 RCTs	Adults	Pharmacists with a proactive role in interventions during care transitions, including medication reconciliation	Hospital; admission and/or discharge	Mortality, hospital readmissions, ED visits, ADEs	NR	Multifaceted programs should combine medication reconciliation with active patient counseling and a clinical medication review; performing medication reconciliation alone is insufficient.
Hammad et al.20 (2017)	Inception–Dec 2015	13 studies (3 RCTs; 6 prospective, uncontrolled studies; 3 before-and-after studies; 1 non-randomized, prospective observational study)	Children and adults	Pharmacy-led medication reconciliation	Hospital; admission to discharge	Medication discrepancies, clinically significant medication discrepancies, readmissions and/or ED visits, length of stay, mortality	NR	Lack of evidence prevented drawing conclusions regarding effectiveness.
Kwan et al.22 (2013)	Inception–Jul 2012	18 studies (5 RCTs, 1 quasi-experimental study, 3 before-and-after studies, 9 postintervention studies)	Children or adults	Medication reconciliation	Hospital; admission and/or discharge	Clinically significant unintentional discrepancies, ED visits and readmissions within 30 days	NR	Medication reconciliation alone likely does not reduce postdischarge hospital utilization but may do so when bundled with interventions aimed at improving care transitions.
NcNab et al.31 (2018)	Inception–Sep 2017	14 studies (5 RCTs, 6 cohort studies, 3 pre–post studies)	Adults	Community-based pharmacist–led medication reconciliation	Community; postdischarge	Medication discrepancies, clinically relevant medication discrepancies, hospital readmissions, ED visits, primary care workload	Proportion of patients with readmission for any cause: RR, 0.91 (95% CI, 0.66–1.25)	Pharmacists can identify and resolve discrepancies when completing medication reconciliation after hospital discharge. Improvements in ED visit and care workload outcomes were not consistently seen. Meta-analysis revealed a nonsignificant decrease in hospital readmissions.
Mekonnen et al.33 (2016)	Inception– Nov 2015	10 studies (8 before-and-after studies, 1 RCT, 1 NRCT)	Children and adults	Electronic medication reconciliation	Hospital; admission and/or discharge	Proportion of medications associated with unintentional discrepancies, proportion of patients with medication discrepancies, mean number of discrepancies per patient	Proportion of medications associated with unintentional discrepancies: RR, 0.55 (95% CI, 0.51–0.58) Proportion of patients with medication discrepancies: RR, 0.37 (95% CI, 0.08–1.70) Mean number of discrepancies per patient: mean difference, –0.18 (95% CI, –0.45 to 0.09)	Medication reconciliation supported by an electronic tool can minimize the occurrence of unintended discrepancies, mainly drug omissions; these interventions did not significantly impact the proportion of patients with discrepancies or the mean number of discrepancies per patient.
Mekonnen et al.26 (2016)	Inception– Dec 2014	19 studies (11 RCTs, 8 non-RCTs)	Adults	Pharmacy-led medication reconciliation	Hospital; admission and/or discharge	Medication discrepancies	Single-transition interventions (at either admission or discharge): RR, 0.34 (95% CI, 0.23–0.50) Multiple-transition interventions (both admission and discharge): RR, 0.88 (95% CI, 0.77–1.02)	Pharmacy-led medication reconciliation interventions were found to be an effective strategy to reduce medication discrepancies; interventions had a greater impact when conducted at either admission or discharge but were less effective during multiple transitions in care.
Mekonnen et al.24 (2016)	Inception– Dec 2014	17 studies (8 RCTs, 6 before-and-after studies, 3 NRCTs)	Adults	Pharmacist-led medication reconciliation	Hospital; admission and/or discharge	All-cause readmissions, ED visits, composite rate of readmission and/or ED visits, mortality, ADE-related hospital visits	All-cause readmissions: RR, 0.81 (95% CI, 0.70–0.95) ED visits: RR, 0.72 (95% CI, 0.57–0.92) Composite of readmission and/or ED visits: RR, 0.95 (95% CI, 0.90–1.00); Mortality: RR, 1.05 (95% CI, 0.95–1.16) ADE-related hospital visits: RR, 0.33 (95% CI, 0.20–0.53)	Pharmacist-led medication reconciliation programs are effective at improving posthospitalization healthcare utilization.
Redmond et al.32 (2018)	Inception– Jan 2018	25 RCTs	Adults and children	Medication reconciliation	Hospital or postdischarge settings; admission through postdischarge period	Medication discrepancies	Proportion of patients with ≥1 medication discrepancy: RR, 0.53 (95% CI, 0.42–0.67) Number of reported medication discrepancies: mean difference, –1.18 (95% CI, –2.58 to 0.23) Number of medication discrepancies per medication: RR, 0.13 (95% CI, 0.01–1.29) Preventable ADEs: RR, 0.37 (95% CI, 0.09–1.57) ADEs: RR, 1.09 (95% CI, 0.91–1.30) Unplanned hospital readmissions: RR, 0.72 (95% CI, 0.44–1.18) Composite measure of hospital utilization and ED visits: RR, 0.78 (95% CI, 0.50–1.22)	The impact of medication reconciliation interventions on medication discrepancies is uncertain due to the certainty of the evidence quality being very low. There was no certainty of the effect of the interventions on the secondary clinical outcomes of ADEs, preventable ADEs, and healthcare utilization.

^aRCT = randomized controlled trial, NR = not reported, ED = emergency department, ADE = adverse drug event, RR = risk ratio, CI = confidence interval.

Four of the systematic reviews included articles pertaining to both pediatric and adult populations,^20,22,32,33 4 concentrated strictly on adults,^19,24,26,31 1 concentrated on community-dwelling adults,¹⁶ 1 concentrated on adults with community-acquired pneumonia,¹⁸ and 1 concentrated on older adults.¹⁷ The majority of reviews (n = 8) focused on interventions implemented during hospital transitions (i.e., admission and/or discharge).^{18–20,22,24,26,32,33} However, in 1 review interventions were restricted to the ambulatory care setting¹⁶; 1 review evaluated studies in long-term care facilities,¹⁷ and 1 evaluated interventions in postdischarge community settings.³¹

Interventions

All systematic reviews considered interventions incorporating medication reconciliation as a central component. Five reviews discussed pharmacist-led medication reconciliation interventions,^{19,20,24,26,31} 4 reviews discussed medication reconciliation interventions of any type,^16,17,22,32 1 review assessed electronic medication reconciliation,³³ and 1 review evaluated structured discharge interventions with medication reconciliation.¹⁸ Most of the included primary studies assessed interventions that included more than only medication reconciliation. These more comprehensive and multifaceted interventions frequently incorporated medication reconciliation in a bundle of medication management activities. Common supplemental components included communication of medication lists to primary care providers or community pharmacies, patient medication counseling, and medication review. Henceforth, we refer to these interventions as “medication reconciliation interventions.”

Quality of evidence

A summary of the GRADE level of evidence for each systematic review’s major conclusions regarding primary outcomes may be found in Table 2. Details regarding downgrades and/or upgrades for each GRADE domain by outcome may be found in eAppendix E. Most conclusions drawn by the reviews were based on evidence of low quality (5 of 37 conclusions) or very low quality (25 of 37 conclusions). Five of 37 reviews’ conclusions were based on moderate-quality evidence, while 2 reviews did not provide sufficient data to make an assessment. Downgrades in evidence quality were most commonly attributed to poor quality of underlying evidence, nonrandomized study design, and inconsistency of effect estimates across studies. A summary of the upgrades and downgrades may be found in eAppendix F.

Table 2.

Conclusions Regarding Primary Outcomes Reported in Systematic Reviews With Prespecified Outcomes (n = 10), with GRADE Rating of Level of Evidence^a

Authors (Yr Published)	Measured Outcomes of Medication Reconciliation Interventions
	Process Measures		Healthcare Utilization						Patient Outcomes
	Medication Discrepancies	Clinically Significant Discrepancies	Hospital Readmissions	ED Visits	Hospital Readmissions and/or ED Visits	Postdischarge Clinic Visits	Primary Care Workload	Length of Stay	Mortality	ADEs
Bayoumi et al.16 (2009)	Very low
Domingo et al.18 (2012)			Moderate		No data	No data
Ensing et al.19 (2015)			Low	Low	Moderate				Very low	Moderate
Hammad et al.20 (2017)	Very low	Very low			Very low			Very low	Low
Kwan et al.22 (2013)		Very low			Very low
McNab et al.31 (2018)	Very low		Very lowb	Very low			Moderate
Mekonnen et al.33 (2016)	% prescriptions Very lowb % patients, no. prescriptions Very lowb
Mekonnen et al.26 (2016)	Single transition Very lowb Multiple transitions Lowb
Mekonnen et al.24 (2016)			Any readmission Very lowb ADE-related readmissions Lowb	Very lowb	Lowb				Very lowb
Redmond et al.32 (2018)	% patients with ≥1 discrepancy Very lowb No. discrepancies per patient Very lowb No. discrepancies per medication Very low		Unplanned readmissions Moderateb		Composite measure Very lowb					Preventable ADEs Very lowb All ADEs Lowb

^aADE = adverse drug event, GRADE = Grading of Recommendations Assessment, Development and Evaluation. The plus and minus symbols denote improved and worsened outcomes, respectively; the equal sign denotes that outcome was assessed and investigators reported no effect; the question mark symbol denotes outcome was assessed and investigators were precluded from drawing conclusions due to limited or low-quality studies.

^bConclusion based on meta-analytic data pooling.

Major study conclusions

In addition to GRADE level of evidence, Table 2 summarizes each systematic review’s major conclusions with respect to the primary outcomes of interest. We excluded 1 systematic review from the table, as the investigators did not prespecify primary outcomes.¹⁵ Study outcomes assessed by systematic reviews were varied and included medication discrepancies, clinically significant medication discrepancies, mortality, adverse drug events (ADEs), and several measures of healthcare utilization. Seven reviews drew conclusions regarding the impact of interventions on 2 widely reported process measures: all types of medication discrepancies^{16,20,26,31–33} and clinically significant unintentional medication discrepancies.^20,22

Regarding the effectiveness of medication reconciliation interventions, we begin by discussing the most proximal and tightly linked outcome that should be impacted by successful medication reconciliation interventions: medication discrepancies. Of 6 reviews assessing all such discrepancies, 2 were meta-analyses by the same authors and focused on electronic³³ and pharmacist-led interventions,²⁶ both of which were found to be effective in comparison to nonintervention for different outcomes: the proportion of medications associated with unintentional discrepancies (relative risk [RR], 0.55; 95% confidence interval [CI], 0.51–0.58)³³ and the likelihood of medication discrepancies with single-transition intervention (RR, 0.34; 95% CI, 0.23–0.50)²⁶; however, these results were based on very low-quality evidence. A more recent (2018) publication concluded that community-based pharmacist-led mediation reconciliation could identify and resolve medication discrepancies; however, this finding was also based on very low-quality evidence.³¹ In the remaining 3 reviews among the 6 assessing all medication discrepancies, the researchers were precluded from drawing conclusions due to sparse and/or low-quality primary studies.^16,20,32

We turn next to clinically significant unintentional medication discrepancies. Both reviews assessing this outcome focused on the hospital setting. Neither found interventions to be impactful.^20,22 Several reviews examined other clinical outcomes, including mortality^19,20,24 and ADEs,^19,32 but no reviews found medication reconciliation to improve these outcomes.

Last, we examine review conclusions regarding utilization. Seven reviews^{18–20,22,24} reported on some aspect of healthcare utilization, including emergency department visits,^19,24,31 hospital readmissions,^{18,19,24,31,32} hospital readmissions and/or emergency department visits,^{18–20,22,24,32} length of stay,²⁰ postdischarge clinic visits,¹⁸ and primary care workload.³¹ Among the 3 reviews assessing emergency department visits, 1 examined pharmacist-led interventions during hospital transitions and reported statistically significant pooled effect estimates (RR for revisit, 0.72; 95% CI, 0.57–0.92)²⁴; the other 2 reviews focused on any type of medication reconciliation intervention during hospital transitions¹⁹ and community-based, pharmacist-led medication reconciliation,³¹ respectively, and both found no effect. Of 5 reviews examining the impact of interventions on hospital readmissions,^{18,19,24,31,32} only 1 review²⁴ reported a positive impact (RR, 0.81; 95% CI, 0.70–0.95), and this finding was based on very low-quality evidence. None of the 6 articles examining hospital readmissions and/or emergency department visits reported significant improvements with use of medication reconciliation interventions.^{16,19,20,22,24,32} One review using meta-analytic methods reported that pharmacist-led medication reconciliation interventions during hospital transitions significantly reduced ADE-related hospital revisits (RR, 0.33; 95% CI, 0.20–0.53).²⁴ Reviews that assessed the effect of medication reconciliation interventions on length of stay,²⁰ postdischarge clinic visits,¹⁸ and primary care workload³¹ reported no positive results.

Discussion

We sought to understand whether medication reconciliation has been shown to decrease medication discrepancies and improve downstream patient-centered outcomes. Our systematic overview methodology identified 11 high-quality systematic reviews assessing medication reconciliation interventions. Two reviews found very low-quality evidence that medication reconciliation interventions reduced medication discrepancies. However, even though 2 reviews considered clinically significant discrepancies, and 4 reviews considered clinical outcomes, they found no evidence of any benefit. Finally, only 1 of 7 reviews examining healthcare utilization detected any impact of medication reconciliation interventions; however, this review noted that the results of primary studies were heterogeneous, which contributed to a GRADE rating of very low for quality of underlying evidence. Taken together, based on the literature examined in this systematic overview, we would advise provider organizations not to expect to impact patient-centric outcomes (e.g., morbidity, mortality, utilization) with medication reconciliation interventions alone.

Our findings are consistent with those of the only other similar overview of which we are aware. Holte et al.³⁴ also used systematic overview methodology to understand the effect of medication reconciliation. Although only the executive summary is available in English, it reports 2 main findings: (1) based on the conclusions of 7 of 9 systematic reviews, “medication reconciliation probably reduces the number of medication discrepancies,” but (2) there were “methodological issues in the primary studies.” The first finding is consistent with ours, and the methodological weaknesses are consistent with the poor GRADE ratings we assigned.

In the context of these underwhelming findings, a justifiable initial reaction is to question the aforementioned mandates for and resultant investments in medication reconciliation. However, it is important to place our findings in the context of real-world clinical care. As we noted earlier in this article, many studies document the prevalence of unintentional medication discrepancies.^2–4 In this error-rich environment, the act of medication reconciliation has strong face validity. Without it, we are left to wonder how clinicians can prescribe medications without knowing what medications a patient has already been prescribed. How would they avoid the harm of prescribing a duplicate medication? How would they be able to intervene to discontinue a hazardous medication?

Our recommendation for reconciling this strong face validity with uncertain evidence is to focus medication reconciliation efforts in 2 ways. First, efforts should concentrate on the care settings where the face validity for medication reconciliation is strongest. For example, medication reconciliation initiatives might initially focus on patients taking complex medication regimens and on care transitions. Second, quality assessment efforts should seek to measure benefits achievable with medication reconciliation rather than the quality of medication reconciliation itself. For example, a National Quality Forum–endorsed measure that is also included in the Leapfrog Hospital Survey tracks the quality of medication reconciliation on medication orders placed at hospital admission and discharge.³⁵ As such, this measure focuses on care transitions, where there is evidence of increased medication discrepancies^2–4; it also emphasizes the importance of medication reconciliation as part of the prescribing of new medications, which impact patient care more directly than medication histories.

Our overview had several limitations. Because of the broad definition of what constitutes medication reconciliation, there were sometimes differences in the medication reconciliation interventions evaluated in different studies. Some researchers have proposed taxonomies to classify different medication reconciliation components,³⁶ but no standard scheme is widely accepted. To be sure, it is most appropriate for any medication reconciliation intervention to be tailored to the care setting where it is implemented. Nonetheless, detailed labeling of exactly what interventions are undertaken, as well as increased use of a taxonomy, would assist future systematic reviewers in understanding which intervention components provide the most benefit or are useful in certain situations. Other limitations include those inherent to overviews of systematic reviews, the most prominent of which is that we were limited to analyzing evidence that had already been captured in a systematic review.

Although our findings are not easy to reconcile with the strong face validity of medication reconciliation, our own experience indicates that medication reconciliation is a necessary part of good clinical care. We believe that provider organizations and policy makers can integrate these findings into their guidelines and recommendations by focusing less on documenting the act of medication reconciliation itself and more on quantifying its downstream benefits (e.g., the avoidance of medication order errors that are caused by medication reconciliation errors).

Conclusion

An overview of systematic reviews of medication reconciliation interventions found 9 high-quality systematic reviews. A minority of those reviews’ conclusions were consistent with medication reconciliation alone having a measurable impact, and such conclusions were almost all based on very low-quality evidence.

Funding

This research was supported by the American Society of Health-System Pharmacists (ASHP) Research and Education Foundation and the National Institute on Aging of the National Institutes of Health under awards K23AG049181 (JMP) and R01AG058911 (JMP).

Disclosures

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.