Pharmacy-supported interventions at transitions of care: an umbrella review

Abstract

Background Medication discrepancies arising at care transitions are prevalent and are linked with adverse drug events and increased healthcare utilization. Evidence is lacking about which pharmacy-supported interventions at care transitions are most effective for both the patient and the healthcare system. Aim of the review To invesitigate the content and effect of pharmacy-supported interventions at transitions of care. Method The PubMed, Ovid/Medline and Cochrane Database of Systematic Reviews databases were used. The search was limited to systematic reviews and meta-analyses published in English up to May 2018. Included reviews investigated any intervention related to medication therapy performed by pharmacists or multiple healthcare professionals, including a pharmacist, at transition points in any healthcare setting. Reviews were excluded if interventions were not clearly defined or were not performed at care transitions or were not related to medications. A quality assessment was performed using the PRISMA guidelines. The data extracted included general characteristics, methodology, point of transition, pharmacy-supported interventions and outcomes. For systematic reviews, narrative conclusions were extracted. For meta analyses, reported relative risks or odds ratios were extracted along with the 95% confidence intervals. Results Nine systematic reviews and 5 meta-analyses reporting 162 studies were included. The interventions analysed included medication reconciliation (7 reviews) and composite interventions (7 reviews). Six studies reviewed interventions performed by pharmacists alone, while 8 studies explored interventions by different healthcare professionals, including a pharmacist. A positive effect on either medication discrepancies or (potential) ADEs was observed in all reviews. Mixed effects were observed for hospitalizations rates (9 reviews) and costs (4 reviews), regardless of the intervention applied. Mixed effects were also observed for both medication reconciliation and composite interventions on the number of emergency department visit. Interventions showed no significant effect on mortality (4 reviews). The quality of the reviews showed significant variability. Conclusion Pharmacy-supported interventions at transitions of care are heterogeneous and potentially improve medication safety, but show no significant effect on mortality. The effect on healthcare utilization and costs is inconclusive.

Appendices

Appendix 1:

Search profile:

Search: (“transition of care”[Title/Abstract] OR “transitions of care”[Title/Abstract] OR “seamless care”[Title/Abstract] OR “integrated care”[Title/Abstract] OR “continuity of care”[Title/Abstract] OR “continuity of patient care”[Title/Abstract] OR “delivery of health care, integrated”[Title/Abstract] OR “transmural care”[Title/Abstract] OR “coordinated care”[Title/Abstract] OR “continuum of care”[Title/Abstract] OR “transitional care”[Title/Abstract] OR “medication reconciliation”[Title/Abstract] or “comprehensive care”[Title/Abstract]) AND (“drug” OR “medicament” OR “medicine” OR “drug prescriptions” OR “electronic prescribing” OR “medication” OR “medical product” OR “drug related problems” OR “medication related problems” OR “medication use” OR “medication errors” OR “medication review” OR “pharmacy”) AND (“patient admission” OR “hospital admission” OR “patient discharge” OR “hospital discharge” OR “patient readmission” OR “patient transfer” OR “transition” OR “transitional”)

Filters: Meta-Analysis, Systematic Reviews

Appendix 2:

Inclusion and exclusion criteria of included systematic reviews and meta-analyses

Author	Inclusion criteria	Exclusion criteria
Chhabra et al. [22]	Studies that evaluated an intervention involving medication reconciliation in patients transferred to and/or from long-term care settings, such as nursing homes, skilled nursing facility, residential care facilities, assisted living facilities, homes for the aged, and hospice care	Studies involving transitions to or from home health care agencies were excluded because they typically serve patients for a short duration and are in contact with a patient’s health care provider, family members, and/or caregivers. Studies published before 2000, not written in English, that were not empirical, that did not involve medication reconciliation as an intervention, or that were not experimental or quasi-experimental also were excluded
De Oliveira et al. [19]	RCTs that compared a pharmacist intervention (i.e., reconciliation and/or patient education) with an inactive (“usual care”) control group in patients undergoing transitions of care from a hospital setting back to the community. Studies had to report at least on medication errors or healthcare utilization (hospital readmissions and/or emergency room visits) as outcomes. No minimum sample size was required for inclusion in the metaanalysis	Excluded were trials reporting on the effect of pharmacist interventions in non-transition of care settings or involving pediatric patients. Trials that evaluated the effect of the pharmacist intervention compared with an active comparator (e.g., nurse intervention, community health worker) were excluded to optimize clinical homogeneity. Studies containing a concurrent use of an alternative intervention (e.g., bundle interventions) were excluded if a direct comparison of pharmacist and control could not be established
Ensing et al. [21]	Study design (quasi) (cluster) randomized controlled trial published in English Population Adult participants admitted to a hospital and discharged home Setting Intervention conducted in hospital and/or community pharmacy and/or patients’ home Intervention The intervention involved a pharmacist, pre-graduate pharmacist, or pharmacy technician. The pharmacist had a proactive role, meaning for all interventions, patients received an active pharmacist intervention (e.g., excluding reactively responding to physician questions during ward rounds). Interventions were performed before, during, or up to 30 days after hospitalization. The post-discharge time limit was chosen to ensure connection to transitional care. The intervention was designed to improve transitional care and aimed at medication-related issues Comparison The intervention was compared with a control group that received usual care Clinical outcomes At least 1 of the following outcomes was measured: mortality, readmissions, emergency department visits, and adverse drug events	Participant Intervention conducted solely on pediatric patients or psychiatric patients due to their specific population characteristics Setting Interventions in a palliative care setting or in an intensive care ward due to their specific setting characteristics Intervention Interventions solely targeted at specific drugs (e.g., improving adherence of statins) Interventions not aimed at transitional care (e.g., interventions in outpatient clinics without transmission of relevant information from earlier health care provider encounters in the hospital or interventions on heart failure guideline adherence)
Hammad et al. [26]	Adults and children receiving pharmacy-led medication reconciliation within hospital inpatient settings	Studies evaluating non-pharmacy-led medication reconciliation at only one end of patient care or transfer were not included. Studies evaluating pharmacy-led medication reconciliation using a qualitative approach and studies evaluating enhanced interventions, including telephone helpline and post discharge follow-up calls, were excluded. Telephone helpline and follow-up calls were not considered part of medication reconciliation and suspected to influence readmissions and health care utilization
Kwan et al. [20]	Studies evaluating medication reconciliation with an eligible outcome reported	Studies in which the person conducting medication reconciliation provided the sole assessment of clinical significance for identified discrepancies were excluded. The authors also required that studies explicitly distinguish unintentional discrepancies from other (intentional) medication changes through direct communication with the medical team
Lehnbom et al. [27]	Studies evaluating medication reconciliation or medication review that also reported the impact of identifying and rectifying discrepancies and medication-related problems	Articles that only reported the number of identified discrepancies or medication-related problems but failed to identify the potential or actual impact of these discrepancies on patient outcomes were excluded. Articles were also excluded if the sample size was fewer than 50 patients because the impact on clinical outcomes could not be reliably evaluated with such a small sample size
McNab et al. [18]	Population Patients discharged from hospital to their permanent residence (home, residential unit or nursing home) Intervention of interest Medicines reconciliation completed by a pharmacist based in the community Comparator Usual care processes for medication reconciliation Outcome measure Discrepancy identification, Discrepancy categorisation, Healthcare usage (readmission, emergency department attendance, GP attendance), Workload/efficiency measures—time to complete medicines reconciliation, effect on number of primary and secondary care appointments needed, and economic outcomes Study design RCTs, cluster RCTs, quasi-RCTs, cluster quasi-RCTs, controlled pre–post intervention studies, interrupted-time-series, cohort studies (prospective or retrospective), case–control studies, uncontrolled pre–post intervention studies Language No limitation Publication date No limitation	Studies focused on medication review (e.g., recommendations to optimise medication regimens) were excluded
Mekonnen et al. [16]	Studies reporting medication reconciliation intervention primarily and that provide data on any of these clinical end points (all-cause readmission, emergency department visits, mortality, adverse drug events)	The following types of studies were excluded: other medication reconciliation practices (e.g., nurse-led) or practices as part of a multicomponent intervention (e.g., medication therapy management), case studies, systematic reviews, qualitative outcomes and non-research articles
Mekonnen et al. [17]	Studies comparing pharmacy-led medication reconciliation interventions with usual care and measured medication discrepancies as an outcome of interest	Interventions with medication reconciliation where physicians or nurses assessed medication discrepancies were excluded. Studies assessed discrepancies in medical histories, for example – documentation of allergy, were excluded
Mekonnen et al. [15]	Studies which reported data related to the effectiveness of electronic medication reconciliation intervention, and provided data on medication discrepancies or errors	Studies with a focus on other types of medication errors (e.g. prescribing errors) that were identified through the non-reconciliation process were excluded
Mueller et al. [28]	Medication reconciliation was the primary focus of the intervention. English language, hospital setting, intervention during hospitalization and/or transition, quantitative results	Not reported
Nazar et al. [23]	Patients who were identified as post-discharged from hospital, intervention performed by a community pharmacist, intervention focused on continuity of care, CT or RCT. Outcomes reported	Not reported
Simoens et al. [29]	Studies had to compare an intervention to improve seamless care focusing on medication with usual care	Not reported
Spinewine et al. [25]	Inclusion of patients admitted to and/or discharged from hospital	The following exclusion criteria were applied: (i) studies where the focus of the intervention was broader than medications and without specific measure to evaluate the effect of the intervention on the medication component; (ii) studies where the intervention focused on medications but with a scope that was broader than continuity of care (e.g. clinical pharmacists doing admission histories and discharge counselling along with interventions to improve prescribing during hospital stay) and without specific measure to evaluate the effect of the intervention on the continuity of care component; (iii) studies with no control group or ‘before-after studies’ with no control group (i.e. in which the control group is an historical group) and (iv) studies with <30 patients per group. Systematic reviews and meta-analyses on related topics were not included as such, but the list of articles included in the corresponding reviews was checked for eligibility