PRACTICE RESEARCH REPORT

Effect of medication reconciliation interventions on

outcomes: A systematic overview of systematic reviews
Supplementary material is available Purpose. To evaluate and summarize published evidence from systematic
with the full text of this article at

Downloaded from https://academic.oup.com/ajhp/article-abstract/76/24/2028/5649011 by Columbia University user on 06 December 2019

reviews examining medication reconciliation.
AJHP online.
Methods. MEDLINE, the Cochrane Database of Systematic Reviews, and
the Database of Abstracts of Reviews of Effects were searched for English-
Laura J. Anderson, Ph.D., M.P.H., language systematic reviews published from January 2004 to March 2019.
Department of Medicine, Cedars-Sinai Reviewers independently extracted information and scored review quality
Medical Center, Los Angeles, CA.
using the Assessment of Multiple Systematic Reviews (AMSTAR) tool.
Jeff L. Schnipper, M.D., M.P.H., For reviews with AMSTAR scores above 7, Grading of Recommendations
Department of Medicine, Brigham and
Women’s Hospital, Boston, MA. Assessment, Development and Evaluation (GRADE) methodology was
Teryl K. Nuckols, M.D., M.S.H.S., applied to assess evidence quality, with evidence summarized and
Department of Medicine, Cedars-Sinai conclusions compared across reviews.
Medical Center, Los Angeles, CA.
Rita Shane, Pharm.D., Department of Results. Eleven reviews met the inclusion criteria, 5 of which used meta-
Pharmacy, Cedars-Sinai Medical Center, analytic pooling. Most systematic reviews included primary studies of
Los Angeles, CA.
comprehensive bundled interventions that featured medication recon-
Michael M. Le, B.S., David Geffen ciliation as a central component. Reviews largely focused on transitions
School of Medicine, Los Angeles, CA.
into and out of hospital settings. Five reviews focused exclusively on
Karen Robbins, M.D., Department of
Medicine, Cedars-Sinai Medical Center,
pharmacist-led interventions. Of the 5 reviews that considered all types of
Los Angeles, CA. medication discrepancies, 3 reviews found very low-quality evidence that
Joshua M. Pevnick, M.D., M.S.H.S., interventions reduced medication discrepancies. Neither of the 2 reviews
Department of Medicine, Cedars-Sinai that examined clinically significant medication discrepancies found any
Medical Center, Los Angeles, CA. intervention effect. Of the 5 reviews that examined healthcare utilization
Members of the PHARM-DC group: outcomes, only 1 found any intervention effect, and that finding was based
Carmel Hughes, Ph.D., School of on low- to very low-quality evidence. Four reviews considered clinical out-
Pharmacy, Queen’s University Belfast,
comes, but none found any intervention effect.
Belfast, UK.
Cynthia A. Jackevicius, BScPhm, Conclusion. An overview of systematic reviews of medication reconcili-
Pharm.D., M.Sc., Department of
Pharmacy Practice and Administration, ation interventions found 9 high-quality systematic reviews. A minority of
Western University of Health Sciences, those reviews’ conclusions were consistent with medication reconciliation
Pomona, CA, and VA Greater Los alone having a measurable impact, and such conclusions were almost all
Angeles Healthcare System, Los Angeles,
CA. based on very low-quality evidence.
Denis O’Mahony, M.D., Department of
Medicine (Geriatrics), University College Keywords: intervention, medication reconciliation, medication errors,
Cork, Cork, Ireland. medication review, review, systematic review
Catherine Sarkisian, M.D., Division
of Geriatrics, David Geffen School of Am J Health-Syst Pharm. 2019; 76:2028-2040
Medicine at UCLA, Los Angeles, CA.

C linicians frequently have difficulty

obtaining a complete list of pa-
tients’ medications or understanding
Up to two-thirds of patients have at least
1 unintended medication discrepancy
upon hospital admission,2,3 and up to
medication regimens. This is especially 81% of patients discharged from the
true when patients are taking many hospital experience at least 1 medica-
Address correspondence to Dr. Anderson medications, have difficulty commu- tion discrepancy.4 One-third of unin-
(anderson.laurajane@gmail.com).
nicating or understanding medication tended discrepancies at admission have
regimens, or are undergoing care tran- the potential to cause severe discomfort
© American Society of Health-System sitions.1 Discrepancies in medication or clinical deterioration.3 Furthermore,
Pharmacists 2019. All rights reserved.
For permissions, please e-mail: journals. reconciliation have the potential to neg- higher rates of 30-day readmissions
permissions@oup.com. atively impact the quality and safety of have been observed among patients ex-
DOI 10.1093/ajhp/zxz236 patient care and healthcare utilization. periencing medication discrepancies.5

2028 AM J HEALTH-SYST PHARM | VOLUME 76 | NUMBER 24 | DECEMBER 15, 2019

 MEDICATION RECONCILIATION PRACTICE RESEARCH REPORT

Medication reconciliation, or pur- without meta-analyses) that exam-

posefully working to understand and KEY POINTS ined the effects of interventions that
document the medication regimen • Eleven high-quality systematic included medication reconciliation as
that a patient has been prescribed reviews addressing medication a core part of an intervention imple-
and is taking, frequently requires rec- reconciliation interventions, mented in adult or pediatric popula-
onciling medication records from dif- predominantly focused on tions. For the purposes of this overview,
ferent sources (e.g., patient, caregiver, actions at hospital admis- we considered a systematic review to

Downloaded from https://academic.oup.com/ajhp/article-abstract/76/24/2028/5649011 by Columbia University user on 06 December 2019

and family interviews; medical re- sion and/or discharge, were be any attempt to collate empirical ev-
cords; insurance claims). This strategy identified. idence that fit prespecified eligibility
has been endorsed by several patient criteria in order to answer a specific
• Two of 6 systematic reviews
safety organizations, including the research question using explicit, sys-
concluded that medication
Joint Commission,6 the World Health tematic methods.13 Systematic reviews
reconciliation interventions
Organization,7 and other international of studies with any study design and
reduced medication dis-
entities.8,9 These endorsements and the outcome(s) were included. We ex-
crepancies; however, these
resulting widespread dissemination of cluded reviews focused exclusively on
conclusions were based on
recommendations on medication rec- interventions implemented in low- to
very low-quality evidence,
onciliation have led to extensive evalu- middle-income countries due to differ-
and there was little evidence
ations. Hundreds of primary studies ences in care practices and healthcare
that medication reconciliation
and dozens of systematic reviews have infrastructure.
interventions improved clin-
assessed the impact of medication rec- Two investigators independently
ical outcomes or healthcare
onciliation on both clinical and utiliza- screened titles and abstracts in accord-
utilization.
tion outcomes. ance with the prespecified inclusion
Aggregating the central findings • We recommend that medica- and exclusion criteria. Next, full-text
of this literature can be unwieldy, not tion reconciliation be targeted publications were retrieved and exam-
only due to its volume but also because to care settings where its face ined by 2 reviewers to determine eligi-
findings may differ with differences in validity is highest and that qual- bility; disagreements at either the title
outcome measures, study populations, ity assessment efforts measure and abstract or full-text screening level
clinical settings, methods of analysis, benefits achievable with medi- were resolved by consensus in group
and the quality of primary studies and cation reconciliation rather than meetings of research team members.
reviews. Accordingly, this overview of the act of medication reconcilia- Quality evaluation. We as-
reviews aims to appraise and summa- tion itself. sessed the methodological quality of
rize evidence from high-quality sys- each relevant systematic review with
tematic reviews examining the impact the Assessment of Multiple Systematic
of interventions involving medication Reviews (AMSTAR) instrument.14 The
reconciliation. Specifically, we sought tool consists of 11 domains, and each
to understand whether medication 2019 using MEDLINE, the Cochrane systematic review may receive a score
reconciliation has been shown to im- Database of Systematic Reviews, and ranging from 0 (lowest quality) to 11
prove medication discrepancy rates the Database of Abstracts of Reviews (highest quality). Two reviewers inde-
and downstream patient-centered out- of Effects (DARE). Two investigators pendently applied the instrument, and
comes, including measures representing developed a search strategy that pre- scoring discrepancies were reconciled
morbidity, mortality, and utilization. dominantly consisted of identifying through oral discussion. Systematic re-
systematic reviews with the term med- views with an AMSTAR score below 8,
Methods ication reconciliation in the title or ab- a commonly applied threshold, were
We performed a systematic over- stract (see eAppendix B for details). excluded from data extraction and
view in accordance with the Preferred The searches were limited to English- synthesis.12
Reporting Items for Systematic Reviews language articles published from Data extraction. Research team
and Meta Analyses (PRISMA) state- January 2004 to March 2019. One in- members independently extracted
ment (eAppendix A).10 We also refer- vestigator also searched the reference data related to key characteristics using
enced published literature explicitly list of each included systematic review a standardized data extraction tool.
focused on methodological practices to identify other published or unpub- Extracted variables included the fol-
for overviews of systematic reviews.11,12 lished sources. lowing: dates of literature search; number
We did not register a protocol prior to Selection of systematic re- and design of included primary studies;
performing this review. views. Systematic reviews were eli- intervention types; patient populations;
Data sources and searches. gible for inclusion in the overview if patient care settings; outcome measures,
We searched the literature in March they were systematic reviews (with or including medication discrepancies and

AM J HEALTH-SYST PHARM | VOLUME 76 | NUMBER 24 | DECEMBER 15, 2019  2029

 PRACTICE RESEARCH REPORT MEDICATION RECONCILIATION

measures representing morbidity, mor- evidence for each systematic review, pediatric and adult populations,20,22,32,33
tality, and utilization; and, if present, while a second author reviewed and veri- 4 concentrated strictly on adults,19,24,26,31
meta-analytic techniques and pooled fied the assessments; discrepancies were 1 concentrated on community-dwelling
estimates. Additionally, information resolved through oral discussion. adults,16 1 concentrated on adults with
concerning reviews’ major conclusions Data synthesis. Two research community-acquired pneumonia,18
regarding intervention effectiveness team members independently exam- and 1 concentrated on older adults.17
were extracted. Reviewers compared ex- ined each systematic review’s major The majority of reviews (n = 8) fo-

Downloaded from https://academic.oup.com/ajhp/article-abstract/76/24/2028/5649011 by Columbia University user on 06 December 2019

tracted data and reconciled discrepan- conclusions regarding the effective- cused on interventions implemented
cies through oral discussion. ness of intervention strategies on the during hospital transitions (i.e., ad-
GRADE quality of evidence. We reported primary outcomes and clas- mission and/or discharge).18–20,22,24,26,32,33
documented the quality of evidence for sified authors’ conclusions into 1 of 4 However, in 1 review interventions
each major conclusion within each review distinct categories: (1) a positive asso- were restricted to the ambulatory care
by applying Grading of Recommendations ciation between intervention strategy setting16; 1 review evaluated studies in
Assessment, Development and Evaluation and outcome, (2) a negative association long-term care facilities,17 and 1 evalu-
(GRADE) methodology.15 We used objec- between intervention strategy and out- ated interventions in postdischarge
tive criteria to assess evidence quality for come, (3) a null association between community settings.31
each outcome in the following GRADE intervention strategy and outcome, and Interventions. All systematic
domains: study design; study quality; con- (4) an inability to draw conclusions due reviews considered interventions
sistency (of effects between studies); di- to limited or low-quality studies. It was incorporating medication reconciliation
rectness (i.e., applicability of participants, also documented whether conclusions as a central component. Five reviews
interventions, and outcomes to the study were based on quantitative or qualita- discussed pharmacist-led medication
question); and other modifying factors, tive (i.e., meta-analytic) assessments. reconciliation interventions,19,20,24,26,31 4
including data sparseness (i.e., sample Research team members compared reviews discussed medication reconcil-
size) and strength of effect estimates. By synthesis information and reconciled iation interventions of any type,16,17,22,32 1
combining item scores for each of these discrepancies through oral discussion. review assessed electronic medication
domains, we determined the level of ev- reconciliation,33 and 1 review evaluated
idence, which was classified into 1 of 4 structured discharge interventions with
categories: Results medication reconciliation.18 Most of
Study selection. The literature the included primary studies assessed
search identified 119 articles. Upon interventions that included more than
• High—Further research is very un-
screening titles and abstracts, 28 cit- only medication reconciliation. These
likely to change our confidence in the
ations were selected for full-text re- more comprehensive and multifaceted
estimate of effect.
view. After review of the full text of interventions frequently incorporated
• Moderate—Further research is likely
these articles, 18 systematic reviews medication reconciliation in a bundle
to have an important impact on our
met the inclusion criteria16-33; 11 re- of medication management activities.
confidence in the estimate of effect
ceived an AMSTAR score of greater Common supplemental components
and may change the estimate.
than 8 and were thus included in our included communication of medication
• Low—Further research is very likely
review.16–20,22,24,26,31–33 The PRISMA flow lists to primary care providers or com-
to have an important impact on our
diagram may be found in Figure 1, munity pharmacies, patient medication
confidence in the estimate of effect
and AMSTAR scores may be found in counseling, and medication review.
and is likely to change the estimate.
eAppendix C. Henceforth, we refer to these inter-
• Very low—Any estimate of effect is
Study characteristics. The ventions as “medication reconciliation
very uncertain.
major characteristics of included sys- interventions.”
tematic reviews may be found in Quality of evidence. A sum-
We did not assess the quality of the indi- Table 1. All 11 reviews were published mary of the GRADE level of evidence
vidual studies evaluated in the reviews after 2009,16–20,22,24,26,31–33 with 7 pub- for each systematic review’s major con-
but reported the risk of study bias as docu- lished since 2014.19,20,24,26,31–33 All of the clusions regarding primary outcomes
mented in the reviews. If a systematic re- identified reviews were published in may be found in Table 2. Details re-
view provided GRADE levels of evidence peer-reviewed journals.16–20,22,24,26,31–33 garding downgrades and/or upgrades
for outcomes (in the case of Cochrane Six reviews were exclusively descrip- for each GRADE domain by outcome
Database reviews) we reported these as- tive in nature16–20,22; in 5 reviews, re- may be found in eAppendix E. Most
sessments. Further details regarding how sults were pooled using meta-analytic conclusions drawn by the reviews were
each GRADE domain was operation- techniques.24,26,31–33 based on evidence of low quality (5 of
alized may be found in eAppendix D. Four of the systematic reviews in- 37 conclusions) or very low quality (25
One author assessed GRADE level of cluded articles pertaining to both of 37 conclusions). Five of 37 reviews’

2030 AM J HEALTH-SYST PHARM | VOLUME 76 | NUMBER 24 | DECEMBER 15, 2019

 MEDICATION RECONCILIATION PRACTICE RESEARCH REPORT

Figure 1. PRISMA flow diagram of publication selection for inclusion in Regarding the effectiveness of
overview. medication reconciliation interven-
tions, we begin by discussing the most
proximal and tightly linked outcome
that should be impacted by successful
Titles identified from Titles identified from
database searches
medication reconciliation interven-
searching references
(n = 118) (n = 1) tions: medication discrepancies. Of 6

Downloaded from https://academic.oup.com/ajhp/article-abstract/76/24/2028/5649011 by Columbia University user on 06 December 2019

reviews assessing all such discrepan-
cies, 2 were meta-analyses by the same
authors and focused on electronic33
Abstracts identified for dual review and pharmacist-led interventions,26
(n = 119) both of which were found to be effec-
tive in comparison to nonintervention
for different outcomes: the proportion
Abstracts excluded (n = 91) of medications associated with unin-
Not systematic review (n = 46) tentional discrepancies (relative risk
[RR], 0.55; 95% confidence interval
Not on topic (n = 41) [CI], 0.51–0.58)33 and the likelihood of
No interventions (n = 4) 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 evi-
Articles selected for full text review
(n = 28)
dence. A more recent (2018) publica-
tion concluded that community-based
pharmacist-led mediation reconcili-
Full text articles excluded (n = 17)
ation could identify and resolve med-
Not systematic review (n = 3) ication discrepancies; however, this
finding was also based on very low-
Not on topic (n = 5)
quality evidence.31 In the remaining 3
No interventions (n = 1) reviews among the 6 assessing all med-
ication discrepancies, the researchers
Not of sufficient quality (AMSTAR) (n = 7)
were precluded from drawing conclu-
Not in English (n = 1) sions due to sparse and/or low-quality
primary studies.16,20,32
Systematic reviews contributing to data synthesis We turn next to clinically significant
(n = 11) unintentional medication discrepan-
cies. Both reviews assessing this out-
come focused on the hospital setting.
Neither found interventions to be im-
conclusions were based on moderate- interest. We excluded 1 systematic re- pactful.20,22 Several reviews examined
quality evidence, while 2 reviews did view from the table, as the investigators other clinical outcomes, including
not provide sufficient data to make an did not prespecify primary outcomes.15 mortality19,20,24 and ADEs,19,32 but no re-
assessment. Downgrades in evidence Study outcomes assessed by system- views found medication reconciliation
quality were most commonly attrib- atic reviews were varied and included to improve these outcomes.
uted to poor quality of underlying ev- medication discrepancies, clinically Last, we examine review conclu-
idence, nonrandomized study design, significant medication discrepancies, sions regarding utilization. Seven re-
and inconsistency of effect estimates mortality, adverse drug events (ADEs), views18–20,22,24 reported on some aspect
across studies. A summary of the up- and several measures of healthcare of healthcare utilization, including
grades and downgrades may be found utilization. Seven reviews drew con- emergency department visits,19,24,31 hos-
in eAppendix F. clusions regarding the impact of inter- pital readmissions,18,19,24,31,32 hospital
Major study conclusions. In ventions on 2 widely reported process readmissions and/or emergency de-
addition to GRADE level of evidence, measures: all types of medication partment visits,18–20,22,24,32 length of stay,20
Table 2 summarizes each system- discrepancies16,20,26,31–33 and clinically postdischarge clinic visits,18 and pri-
atic review’s major conclusions with significant unintentional medication mary care workload.31 Among the 3 re-
respect to the primary outcomes of discrepancies.20,22 views assessing emergency department

AM J HEALTH-SYST PHARM | VOLUME 76 | NUMBER 24 | DECEMBER 15, 2019  2031

 PRACTICE RESEARCH REPORT MEDICATION RECONCILIATION

visits, 1 examined pharmacist-led inter- heterogeneous, which contributed to a example, medication reconciliation
ventions during hospital transitions GRADE rating of very low for quality of initiatives might initially focus on pa-
and reported statistically significant underlying evidence. Taken together, tients taking complex medication regi-
pooled effect estimates (RR for revisit, based on the literature examined in mens and on care transitions. Second,
0.72; 95% CI, 0.57–0.92)24; the other 2 this systematic overview, we would quality assessment efforts should seek
reviews focused on any type of medica- advise provider organizations not to to measure benefits achievable with
tion reconciliation intervention during expect to impact patient-centric out- medication reconciliation rather than

Downloaded from https://academic.oup.com/ajhp/article-abstract/76/24/2028/5649011 by Columbia University user on 06 December 2019

hospital transitions19 and community- comes (e.g., morbidity, mortality, utili- the quality of medication reconciliation
based, pharmacist-led medication rec- zation) with medication reconciliation itself. For example, a National Quality
onciliation,31 respectively, and both interventions alone. Forum–endorsed measure that is also
found no effect. Of 5 reviews examining Our findings are consistent with included in the Leapfrog Hospital
the impact of interventions on hospital those of the only other similar overview Survey tracks the quality of medica-
readmissions,18,19,24,31,32 only 1 review24 of which we are aware. Holte et al.34 tion reconciliation on medication or-
reported a positive impact (RR, 0.81; also used systematic overview method- ders placed at hospital admission and
95% CI, 0.70–0.95), and this finding was ology to understand the effect of med- discharge.35 As such, this measure fo-
based on very low-quality evidence. ication reconciliation. Although only cuses on care transitions, where there
None of the 6 articles examining hos- the executive summary is available in is evidence of increased medication
pital readmissions and/or emergency English, it reports 2 main findings: (1) discrepancies2–4; it also emphasizes the
department visits reported significant based on the conclusions of 7 of 9 sys- importance of medication reconcili-
improvements with use of medication tematic reviews, “medication recon- ation as part of the prescribing of new
reconciliation interventions.16,19,20,22,24,32 ciliation probably reduces the number medications, which impact patient
One review using meta-analytic of medication discrepancies,” but (2) care more directly than medication
methods reported that pharmacist-led there were “methodological issues in histories.
medication reconciliation interven- the primary studies.” The first finding is Our overview had several limita-
tions during hospital transitions signif- consistent with ours, and the methodo- tions. Because of the broad definition of
icantly reduced ADE-related hospital logical weaknesses are consistent with what constitutes medication reconcilia-
revisits (RR, 0.33; 95% CI, 0.20–0.53).24 the poor GRADE ratings we assigned. tion, there were sometimes differences
Reviews that assessed the effect of In the context of these under- in the medication reconciliation inter-
medication reconciliation interven- whelming findings, a justifiable initial ventions evaluated in different studies.
tions on length of stay,20 postdischarge reaction is to question the aforemen- Some researchers have proposed tax-
clinic visits,18 and primary care work- tioned mandates for and resultant onomies to classify different medica-
load31 reported no positive results. investments in medication reconcilia- tion reconciliation components,36 but
tion. However, it is important to place no standard scheme is widely accepted.
Discussion our findings in the context of real-world To be sure, it is most appropriate for any
We sought to understand whether clinical care. As we noted earlier in this medication reconciliation intervention
medication reconciliation has been article, many studies document the to be tailored to the care setting where it
shown to decrease medication dis- prevalence of unintentional medica- is implemented. Nonetheless, detailed
crepancies and improve downstream tion discrepancies.2–4 In this error-rich labeling of exactly what interventions
patient-centered outcomes. Our sys- environment, the act of medication are undertaken, as well as increased
tematic overview methodology identi- reconciliation has strong face validity. use of a taxonomy, would assist future
fied 11 high-quality systematic reviews Without it, we are left to wonder how systematic reviewers in understanding
assessing medication reconciliation clinicians can prescribe medications which intervention components pro-
interventions. Two reviews found very without knowing what medications a vide the most benefit or are useful in
low-quality evidence that medication patient has already been prescribed. certain situations. Other limitations
reconciliation interventions reduced How would they avoid the harm of pre- include those inherent to overviews
medication discrepancies. However, scribing a duplicate medication? How of systematic reviews, the most prom-
even though 2 reviews considered clin- would they be able to intervene to dis- inent of which is that we were limited
ically significant discrepancies, and 4 continue a hazardous medication? to analyzing evidence that had already
reviews considered clinical outcomes, Our recommendation for recon- been captured in a systematic review.
they found no evidence of any benefit. ciling this strong face validity with un- Although our findings are not easy
Finally, only 1 of 7 reviews examining certain evidence is to focus medication to reconcile with the strong face va-
healthcare utilization detected any im- reconciliation efforts in 2 ways. First, lidity of medication reconciliation, our
pact of medication reconciliation inter- efforts should concentrate on the care own experience indicates that med-
ventions; however, this review noted settings where the face validity for med- ication reconciliation is a necessary
that the results of primary studies were ication reconciliation is strongest. For part of good clinical care. We believe

2032 AM J HEALTH-SYST PHARM | VOLUME 76 | NUMBER 24 | DECEMBER 15, 2019

 Table 1. Major Characteristics of All Systematic Reviews Included in Overview (n = 11)a
Literature Setting;
Researchers Search Number and Design Intervention Primary Outcome Pooled Effect
(Yr Published) Dates of Evaluated Studies Population Intervention Type(s) Timing Measure(s) Estimate(s) Researchers’ Major Conclusion(s)

Bayoumi et al.16 Through 4 studies (3 before- Community- Medication Ambulatory care; Medication NR There is no good-quality evidence
(2009) Mar and-after studies, dwelling reconciliation during care or discrepancies demonstrating the effectiveness of med-
2008 1 RCT) adults hospital dis- ication reconciliation in the primary care
charge setting.

Chhabra et al.17 Through 7 studies (2 RCTs, 4 Older adults Medication Long-term care Varied; no prespecified NR Improvement in outcomes examined was
(2012) Aug quasi-experimental reconciliation facilities; ad- outcomes shown in all studies; however, there were
2010 studies, 1 observa- mission and/or study design flaws. There is a need for
MEDICATION RECONCILIATION

tional study) discharge well-designed studies demonstrating the

effectiveness of medication reconciliation in
long-term care settings.

Domingo et al.18 NR 3 studies (2 RCTs, 1 Adults with Structured discharge Hospital; dis- Hospital readmissions, NR Structured discharge with medication recon-
(2012) quasi-RCT) community- with medication charge ED visits, and un- ciliation did not have a significant impact on
acquired reconciliation scheduled visits to hospital readmissions.
pneumonia healthcare provider
after discharge

Ensing et al.19 Inception– 30 RCTs Adults Pharmacists with a Hospital; admis- Mortality, hospital re- NR Multifaceted programs should combine med-
(2015) Nov proactive role in inter- sion and/or admissions, ED visits, ication reconciliation with active patient
2014 ventions during care discharge ADEs counseling and a clinical medication review;

AM J HEALTH-SYST PHARM
transitions, including performing medication reconciliation alone

|
medication reconcil- is insufficient.
iation

Hammad et al.20 Inception– 13 studies (3 RCTs; Children and Pharmacy-led medica- Hospital; ad- Medication discrep- NR Lack of evidence prevented drawing conclu-
(2017) Dec 6 prospective, un- adults tion reconciliation mission to ancies, clinically sions regarding effectiveness.
2015 controlled studies; discharge significant medica-

VOLUME 76
3 before-and- tion discrepancies,

|
after studies; 1 readmissions and/or
non-randomized, ED visits, length of
prospective obser- stay, mortality
vational study)

NUMBER 24
Kwan et al.22 Inception– 18 studies (5 RCTs, 1 Children or Medication Hospital; admis- Clinically significant NR Medication reconciliation alone likely does not

|
(2013) Jul quasi-experimental adults reconciliation sion and/or unintentional discrep- reduce postdischarge hospital utilization but
2012 study, 3 before- discharge ancies, ED visits and may do so when bundled with interventions
and-after studies, readmissions within aimed at improving care transitions.
9 postintervention 30 days
studies)

NcNab et al.31 Inception– 14 studies (5 RCTs, 6 Adults Community-based phar- Community; Medication discrepan- Proportion of Pharmacists can identify and resolve dis-
(2018) Sep cohort studies, 3 macist–led medication postdischarge cies, clinically relevant patients with crepancies when completing medication
2017 pre–post studies) reconciliation medication discrep- readmission for reconciliation after hospital discharge.
ancies, hospital re- any cause: RR, Improvements in ED visit and care workload
admissions, ED visits, 0.91 (95% CI, outcomes were not consistently seen. Meta-
primary care workload 0.66–1.25) analysis revealed a nonsignificant decrease
in hospital readmissions.
PRACTICE RESEARCH REPORT

DECEMBER 15, 2019  2033

Continued on next page

Downloaded from https://academic.oup.com/ajhp/article-abstract/76/24/2028/5649011 by Columbia University user on 06 December 2019

2034
Continued from previous page

Table 1. Major Characteristics of All Systematic Reviews Included in Overview (n = 11)a

Literature Setting;
Researchers Search Number and Design Intervention Primary Outcome Pooled Effect
(Yr Published) Dates of Evaluated Studies Population Intervention Type(s) Timing Measure(s) Estimate(s) Researchers’ Major Conclusion(s)

Mekonnen Inception– 10 studies (8 before- Children and Electronic medication Hospital; admis- Proportion of medica- Proportion of medi- Medication reconciliation supported by an

AM J HEALTH-SYST PHARM
et al.33 (2016) Nov and-after studies, adults reconciliation sion and/or tions associated with cations associated electronic tool can minimize the occurrence

|
2015 1 RCT, 1 NRCT) discharge unintentional discrep- with unintentional of unintended discrepancies, mainly drug
ancies, proportion of discrepancies: omissions; these interventions did not sig-
patients with medi- RR, 0.55 (95% CI, nificantly impact the proportion of patients
cation 0.51–0.58) with discrepancies or the mean number of
PRACTICE RESEARCH REPORT

discrepancies, mean Proportion of discrepancies per patient.

VOLUME 76
number of patients with

|
discrepancies per medication dis-
patient crepancies: RR,
0.37 (95% CI,
0.08–1.70)
Mean number of

NUMBER 24
discrepancies

|
per patient: mean
difference, –0.18
(95% CI, –0.45 to
0.09)

Mekonnen Inception– 19 studies (11 RCTs, Adults Pharmacy-led medica- Hospital; admis- Medication discrepan- Single-transition Pharmacy-led medication reconciliation
et al.26 (2016) Dec 8 non-RCTs) tion reconciliation sion and/or cies interventions (at interventions were found to be an effective
2014 discharge either admission strategy to reduce medication discrepan-
or discharge): RR, cies; interventions had a greater impact

DECEMBER 15, 2019

0.34 (95% CI, when conducted at either admission or
0.23–0.50) discharge but were less effective during
Multiple-transition multiple transitions in care.
interventions (both
admission and
discharge): RR,
0.88 (95% CI,
0.77–1.02)

Continued on next page

MEDICATION RECONCILIATION

Downloaded from https://academic.oup.com/ajhp/article-abstract/76/24/2028/5649011 by Columbia University user on 06 December 2019

 Continued from previous page
MEDICATION RECONCILIATION

Table 1. Major Characteristics of All Systematic Reviews Included in Overview (n = 11)a

Literature Setting;
Researchers Search Number and Design Intervention Primary Outcome Pooled Effect
(Yr Published) Dates of Evaluated Studies Population Intervention Type(s) Timing Measure(s) Estimate(s) Researchers’ Major Conclusion(s)

Mekonnen Inception– 17 studies (8 RCTs, Adults Pharmacist-led Hospital; admis- All-cause readmissions, All-cause readmis- Pharmacist-led medication reconcilia-
et al.24 (2016) Dec 6 before-and-after medication sion and/or ED visits, composite sions: RR, tion programs are effective at improving
2014 studies, 3 NRCTs) reconciliation discharge rate of readmission 0.81 (95% CI, posthospitalization healthcare utilization.
and/or ED visits, mor- 0.70–0.95)
tality, ADE-related ED visits: RR,

AM J HEALTH-SYST PHARM
hospital visits 0.72 (95% CI,

|
0.57–0.92)
Composite of re-
admission and/
or ED visits: RR,
0.95 (95% CI,

VOLUME 76
0.90–1.00);

|
Mortality: RR,
1.05 (95% CI,
0.95–1.16)
ADE-related hos-
pital visits: RR,

NUMBER 24
0.33 (95% CI,

|
0.20–0.53)

Continued on next page

PRACTICE RESEARCH REPORT

DECEMBER 15, 2019  2035

Downloaded from https://academic.oup.com/ajhp/article-abstract/76/24/2028/5649011 by Columbia University user on 06 December 2019
2036
Continued from previous page

Table 1. Major Characteristics of All Systematic Reviews Included in Overview (n = 11)a

Literature Setting;
Researchers Search Number and Design Intervention Primary Outcome Pooled Effect
(Yr Published) Dates of Evaluated Studies Population Intervention Type(s) Timing Measure(s) Estimate(s) Researchers’ Major Conclusion(s)

Redmond et al.32 Inception– 25 RCTs Adults and chil- Medication recon­ Hospital or Medication discrepan- Proportion of The impact of medication reconciliation
(2018) Jan dren ciliation postdischarge cies patients with interventions on medication discrepan-

AM J HEALTH-SYST PHARM
2018 settings; ≥1 medication cies is uncertain due to the certainty of the

|
admission discrepancy: RR, evidence quality being very low. There was
through 0.53 (95% CI, no certainty of the effect of the interven-
postdischarge 0.42–0.67) tions on the secondary clinical outcomes
period Number of reported of ADEs, preventable ADEs, and healthcare
PRACTICE RESEARCH REPORT

medication dis- utilization.

VOLUME 76
crepancies: mean

|
difference, –1.18
(95% CI, –2.58 to
0.23)
Number of medica-
tion discrepancies

NUMBER 24
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:

DECEMBER 15, 2019

RR, 0.72 (95% CI,
0.44–1.18)
Composite measure
of hospital utiliza-
tion and ED visits:
RR, 0.78 (95% CI,
0.50–1.22)

a
RCT = randomized controlled trial, NR = not reported, ED = emergency department, ADE = adverse drug event, RR = risk ratio, CI = confidence interval.
MEDICATION RECONCILIATION

Downloaded from https://academic.oup.com/ajhp/article-abstract/76/24/2028/5649011 by Columbia University user on 06 December 2019

 Table 2. Conclusions Regarding Primary Outcomes Reported in Systematic Reviews With Prespecified Outcomes (n = 10), with GRADE Rating of Level of
Evidencea
Measured Outcomes of Medication Reconciliation Interventions

Process Measures Healthcare Utilization Patient Outcomes

Medication Discrepancies Clinically Hospital ED Visits Hospital Postdischarge Primary Length Mortality ADEs
Authors (Yr Significant Readmissions Readmissions Clinic Visits Care of Stay
Published) Discrepancies and/or ED Visits Workload
MEDICATION RECONCILIATION

Bayoumi et al.16
(2009)
Very low

Domingo
et al.18 (2012) No data
Moderate No data

Ensing et al.19
(2015)
Low Low Moderate Very low Moderate

AM J HEALTH-SYST PHARM
Hammad

|
et al.20 (2017)
Very low Very low Very low Very low Low

Kwan et al.22
(2013)

VOLUME 76
Very low

|
Very low
31
McNab et al.
(2018)
Very low Very lowb Very low Moderate

NUMBER 24
Mekonnen % prescriptions

|
et al.33 (2016)

Very lowb
% patients, no. prescriptions

Very lowb
Continued on next page
PRACTICE RESEARCH REPORT

DECEMBER 15, 2019  2037

Downloaded from https://academic.oup.com/ajhp/article-abstract/76/24/2028/5649011 by Columbia University user on 06 December 2019
 Continued from previous page

2038
Table 2. Conclusions Regarding Primary Outcomes Reported in Systematic Reviews With Prespecified Outcomes (n = 10), with GRADE Rating of Level of
Evidencea

Measured Outcomes of Medication Reconciliation Interventions

Process Measures Healthcare Utilization Patient Outcomes

Medication Discrepancies Clinically Hospital ED Visits Hospital Postdischarge Primary Length Mortality ADEs
Authors (Yr Significant Readmissions Readmissions Clinic Visits Care of Stay
Published) Discrepancies and/or ED Visits Workload

AM J HEALTH-SYST PHARM
|
Mekonnen Single transition
et al.26 (2016)

Very lowb
Multiple transitions
PRACTICE RESEARCH REPORT

VOLUME 76
|
Lowb

Mekonnen Any
et al.24 (2016) readmission

NUMBER 24
Very lowb Lowb Very lowb

|
Very lowb
ADE-related
readmissions

Lowb

DECEMBER 15, 2019

Redmond % patients with ≥1 discrepancy Unplanned re- Composite Prevent-
et al.32 (2018) admissions measure able
ADEs
Very lowb
No. discrepancies per patient Moderateb Very lowb
Very lowb
All ADEs
Very lowb
No. discrepancies per medi-
cation Lowb

Very low
MEDICATION RECONCILIATION

a
ADE = 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.
b
Conclusion based on meta-analytic data pooling.
Downloaded from https://academic.oup.com/ajhp/article-abstract/76/24/2028/5649011 by Columbia University user on 06 December 2019
 MEDICATION RECONCILIATION PRACTICE RESEARCH REPORT

that provider organizations and policy 6. Joint Comission. Hospital: 2017 18. Domingo GR, Reyes FC, Thompson FV
makers can integrate these findings National Patient Safety Goals (2016). et al. Effectiveness of structured dis-
https://www.jointcommission.org/ charge process in reducing hospital
into their guidelines and recommenda-
hap_2017_npsgs/ (accessed 2017 Sep readmission of adult patients with com-
tions by focusing less on documenting 14). munity acquired pneumonia: A sys-
the act of medication reconciliation it- 7. World Health Organization. Assuring tematic review. JBI Libr Syst Rev. 2012;
self and more on quantifying its down- medication accuracy at transition of 10(18):1086-121.
stream benefits (e.g., the avoidance of care: medication reconciliation (2014). 19. Ensing HT, Stuijt CC, van den Bemt BJ

Downloaded from https://academic.oup.com/ajhp/article-abstract/76/24/2028/5649011 by Columbia University user on 06 December 2019

https://www.who.int/patientsafety/ et al. Identifying the optimal role for
medication order errors that are caused
implementation/solutions/high5s/h5s- pharmacists in care transitions: A sys-
by medication reconciliation errors). sop.pdf (accessed 2017 Sep 14). tematic review. J Manag Care Spec
8. National Institute for Health and Pharm. 2015; 21(8):614-36.
Conclusion Care Excellence. Medicines recon- 20. Hammad EA, Bale A, Wright DJ,
An overview of systematic reviews ciliation (2009). https://www.nice. Bhattacharya D. Pharmacy led
org.uk/sharedlearning/medicines- medicine reconciliation at hospital:
of medication reconciliation interven-
reconciliation (accessed 2017 Sep 14). A systematic review of effects and
tions found 9 high-quality systematic 9. Accreditation Canada, Canadian costs. Res Social Adm Pharm. 2017;
reviews. A minority of those reviews’ Institute for Health Information, 13(2):300-12.
conclusions were consistent with med- Canadian Patient Safety Institute, and 21. Kaboli PJ, Hoth AB, McClimon BJ,
ication reconciliation alone having a Institute for Safe Medication Practices Schnipper JL. Clinical pharmacists
Canada. Medication reconciliation in and inpatient medical care: a system-
measurable impact, and such conclu-
Canada: raising the bar—progress to atic review. Arch Intern Med. 2006;
sions were almost all based on very date and the course ahead. https:// 166(9):955-64.
low-quality evidence. www.ismp-canada.org/download/ 22. Kwan JL, Lo L, Sampson M,
MedRec/20121101MedRecCanadaENG. Shojania KG. Medication reconcili-
Funding pdf (accessed 2017 Sep 14). ation during transitions of care as a
10. Moher D, Liberati A, Tetzlaff J, patient safety strategy: a systematic
This research was supported by the American
Altman DG; PRISMA Group. Preferred review. Ann Intern Med. 2013; 158(5 Pt
Society of Health-System Pharmacists (ASHP)
reporting items for systematic reviews 2):397-403.
Research and Education Foundation and
and meta-analyses: the PRISMA 23. Lehnbom EC, Stewart MJ, Manias E,
the National Institute on Aging of the
statement. J Clin Epidemiol. 2009; Westbrook JI. Impact of medication
National Institutes of Health under awards
62(10):1006-12. reconciliation and review on clinical
K23AG049181 (JMP) and R01AG058911 (JMP).
11. Smith V, Devane D, Begley CM, outcomes. Ann Pharmacother. 2014;
Clarke M. Methodology in conducting a 48(10):1298-312.
Disclosures systematic review of systematic reviews 24. Mekonnen AB, McLachlan AJ, Brien JA.
The content is solely the responsibility of the of healthcare interventions. BMC Med Effectiveness of pharmacist-led med-
authors and does not necessarily represent Res Methodol. 2011; 11(1):15. ication reconciliation programmes
the official views of the National Institutes of 12. Whitlock EP, Lin JS, Chou R et al. Using on clinical outcomes at hospital
Health. existing systematic reviews in complex transitions: a systematic review and
systematic reviews. Ann Intern Med. meta-analysis. BMJ Open. 2016;
2008; 148(10):776-82. 6(2):e010003.
References 13. Higgins JP, Green S, eds. Cochrane hand- 25. Mueller SK, Sponsler KC, Kripalani S,
1. Varkey P, Cunningham J, Bisping DS. book for systematic reviews of inter- Schnipper JL. Hospital-based medi-
Improving medication reconciliation in ventions version 5.1.0 (updated March cation reconciliation practices: a sys-
the outpatient setting. Jt Comm J Qual 2011). Available from http://handbook. tematic review. Arch Intern Med. 2012;
Patient Saf. 2007; 33(5):286-92. cochrane.org (accessed 2017 Sep 14). 172(14):1057-69.
2. Tam VC, Knowles SR, Cornish PL et al. 14. Shea BJ, Grimshaw JM, Wells GA et al. 26. Mekonnen AB, McLachlan AJ, Brien JA.
Frequency, type and clinical impor- Development of AMSTAR: a measure- Pharmacy-led medication recon-
tance of medication history errors at ad- ment tool to assess the methodological ciliation programmes at hospital
mission to hospital: a systematic review. quality of systematic reviews. BMC Med transitions: a systematic review and
CMAJ. 2005; 173(5):510-5. Res Methodol. 2007; 7:10. meta-analysis. J Clin Pharm Ther. 2016;
3. Cornish PL, Knowles SR, Marchesano R 15. Atkins D, Best D, Briss PA et al. Grading 41(2):128-44.
et al. Unintended medication discrep- quality of evidence and strength 27. Champion HM, Loosen JA,
ancies at the time of hospital admission. of recommendations. BMJ. 2004; Kennelty KA. Pharmacy students and
Arch Intern Med. 2005; 165(4):424-9. 328(7454):1490. pharmacy technicians in medication
4. Armor BL, Wight AJ, Carter SM. 16. Bayoumi I, Howard M, Holbrook AM, reconciliation: a review of the cur-
Evaluation of adverse drug events and Schabort I. Interventions to improve rent literature. J Pharm Pract. 2019;
medication discrepancies in transitions medication reconciliation in pri- 32(2):207-18.
of care between hospital discharge and mary care. Ann Pharmacother. 2009; 28. Cheema E, Alhomoud FK, Kinsara ASA
primary care follow-up. J Pharm Pract. 43(10):1667-75. et al. The impact of pharmacists-led
2016; 29(2):132-7. 17. Chhabra PT, Rattinger GB, Dutcher SK medicines reconciliation on health-
5. Coleman EA, Smith JD, Raha D, Min SJ. et al. Medication reconciliation during care outcomes in secondary care: a
Posthospital medication discrepancies: the transition to and from long-term systematic review and meta-analysis of
prevalence and contributing factors. care settings: a systematic review. Res randomized controlled trials. PLoS One.
Arch Intern Med. 2005; 165(16):1842-7. Social Adm Pharm. 2012; 8(1):60-75. 2018; 13(3):e0193510.

AM J HEALTH-SYST PHARM | VOLUME 76 | NUMBER 24 | DECEMBER 15, 2019  2039

 PRACTICE RESEARCH REPORT MEDICATION RECONCILIATION

29. DeAntonio JH, Nguyen T, Chenault G 32. Redmond P, Grimes TC, McDonnell R Norway: Norwegian Institute of Public
et al. Medications and patient safety in et al. Impact of medication reconcilia- Health; 2015.
the trauma setting: a systematic review. tion for improving transitions of care. 35. National Quality Forum. Measure 2456:
World J Emerg Surg. 2019; 14:5. Cochrane Database Syst Rev. 2018; medication reconciliation: number of
30. Kee KW, Char CWT, Yip AYF. A review on 8:Cd010791. unintentional medication discrepan-
interventions to reduce medication dis- 33. Mekonnen AB, Abebe TB, cies per patient (2014). http://www.
crepancies or errors in primary or ambu- McLachlan AJ, Brien JA. Impact of qualityforum.org/qps/2456 (accessed
latory care setting during care transition electronic medication reconciliation 2017 Sep 14).

Downloaded from https://academic.oup.com/ajhp/article-abstract/76/24/2028/5649011 by Columbia University user on 06 December 2019

from hospital to primary care. J Family interventions on medication discrepan- 36. Lo L, Kwan J, Fernandes OA,
Med Prim Care. 2018; 7(3):501-6. cies at hospital transitions: a systematic Shojania KG. Medication reconcilia-
31. McNab D, Bowie P, Ross A et al. review and meta-analysis. BMC Med tion supported by clinical pharmacists
Systematic review and meta-analysis Inform Decis Mak. 2016; 16:112. (NEW). In: Making health care safer
of the effectiveness of pharmacist-led 34. Holte HH, Hafstad E, Vist GE. NIPH II: an updated critical analysis of the
medication reconciliation in the com- systematic reviews: executive sum- evidence for patient safety practices. US
munity after hospital discharge. BMJ maries. Overview of reviews on Rockville, MD: Agency for Healthcare
Qual Saf. 2018; 27(4):308-20. effect of medical reconciliation. Oslo, Research and Quality; 2013.

2040 AM J HEALTH-SYST PHARM | VOLUME 76 | NUMBER 24 | DECEMBER 15, 2019