Pharmacist-Led Chronic Disease Management: A Systematic Review of Effectiveness and Harms Compared With Usual Care

Integration of clinical pharmacists in patient care may offer increased access to health care and improved patient outcomes. Recently introduced legislation would establish pharmacists as health care providers and enable coverage of pharmacists' services through Medicare Part B in medically underserved communities (1, 2). Pharmaceutical care involves pharmacist collaboration with health team members to optimize therapeutic outcomes (3). Since 1995, the Department of Veterans Affairs has allowed clinical pharmacy specialists an expanded scope of practice, with independent prescribing privileges (4). The scope of practice for clinical pharmacy specialists includes pharmaceutical care or comprehensive medication management for patients with chronic diseases in addition to less complex services, such as counseling patients on medications or responding to questions about drug information. A 2014 systematic review evaluated outpatient medication therapy management interventions (5). Study inclusion criteria required that interventions have, at minimum, 3 elements: comprehensive medication review; patient-directed education and counseling; and coordination of care, including prescriber-directed interventions. The researchers concluded that evidence was insufficient for health outcomes, although interventions may reduce the frequency of some medication-related problems and decrease some health care use and costs. We determined the effectiveness and harms of pharmacist-led chronic disease management compared with usual care for community-dwelling adults. Chronic disease management is typically a multicomponent intervention (Figure 1), and we required that the pharmacist take a lead role in at least 1 component. Studies where the pharmacist was a collaborator on a Patient Aligned Care Team or similar care strategy were not included. This report summarizes findings of a Department of Veterans Affairs Evidence-based Synthesis Program review (available at www.hsrd.research.va.gov/publications/esp). Figure 1. Components of pharmacist-led chronic disease management. * Follow-up after prescription for medication effectiveness and safety and drug-related problems. Includes medication reconciliation. The pharmacist provides immunization, which was not an outcome of interest. Facilitate access to other health care professionals and provide education about disease, lifestyle changes, aspirin therapy, and tobacco cessation. Supplement. Study Protocol Methods We developed a protocol for the review with input from content experts and methodologists. The protocol was not registered but was available for comment during review development. Data Sources and Searches We searched MEDLINE (Ovid), CINAHL, the Cochrane Library, and International Pharmaceutical Abstracts from 1995 through February 2016 for English-language literature (Appendix Table 1). We obtained additional articles by hand-searching reference lists of systematic reviews and other articles and from peer-reviewers. Appendix Table 1. Search Strategy for MEDLINE (Ovid) Study Selection We focused on chronic disease management for outpatients in health care facilities, excluding retail pharmacies. We emphasized patient- or health system-centered outcomes but also addressed intermediate measures. MEDLINE abstracts were reviewed for eligibility by 2 researchers; abstracts from other databases were reviewed by a single researcher. Full-text articles of studies identified as potentially eligible on the basis of abstract review were obtained for further review. Each article was independently reviewed by 2 investigators or research associates. We included studies of any design that had a control group, evaluated outpatient adults with or at risk for a chronic disease, tested an intervention where the pharmacist was responsible for a distinct component of patient care, and were set in U.S. health care facilities. We excluded studies that did not involve interventions intended to manage or prevent 1 or more chronic diseases and studies of anticoagulation clinics because pharmacist management is considered standard care in those settings. The Appendix Figure shows the reasons for study exclusion at full-text review. Appendix Figure. Summary of evidence search and selection. CCT = controlled clinical trial; IPA = International Pharmaceutical Abstracts; RCT = randomized, controlled trial. Data Extraction and Quality Assessment Study data were extracted by one investigator and independently verified by a second. Our main outcomes of interest were clinical events, mortality, patient satisfaction, health care utilization (including urgent care or emergency department visits and hospitalizations), medication adherence, and goal attainment (Appendix Table 2) We also assessed access to care, quality of life, drug-related problems, and costs. We assessed risk of bias (low, medium, or high) for individual studies on the basis of sequence generation, allocation concealment, confounding (for nonrandomized studies), blinding, and incomplete or selective outcome reportinga modification of the Cochrane approach (6) (Appendix Table 2). Appendix Table 2. Evidence From Included Studies Appendix Table 2Continued Appendix Table 2Continued Appendix Table 2Continued Data Synthesis and Analysis We grouped studies by disease state and described and qualitatively compared the characteristics and findings of included studies. We rated the overall strength of the body of evidence across disease states for clinical events, patient satisfaction, target goal attainment, urgent care or emergency department visits and hospitalizations, and medication adherence by using the method of Owens and colleagues (7). Two trained reviewers graded the strength of evidence regarding the outcomes as high, moderate, low, or insufficient on the basis of risk of bias, consistency, precision, and directness. We resolved discrepancies in risk of bias and strength of evidence ratings by discussion with final determination arrived through consensus that included the principal investigator. Most findings are summarized in narrative form, but we pooled results if studies were randomized (noncluster) trials and enrolled participants and outcome measures were similar within each disease condition. DerSimonianLaird random-effects models were used to calculate pooled risk ratios (RRs) and 95% CIs. We assessed statistical heterogeneity by using the I 2 statistic (75% indicates substantial heterogeneity) (8). Role of the Funding Source The Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, Quality Enhancement Research Initiative assigned the topic and reviewed the key questions but was not involved in data collection, analysis, or manuscript preparation or submission. Results Search results are shown in the Appendix Figure. We included 72 articles representing 63 studies (40 randomized, controlled trials [RCTs]) with 65 unique study populations (940-4172-7381). We grouped studies by disease state of the population and study characteristics (Table 1). Most studies were conducted in primary care settings (community-based or university-affiliated), but a few were in specialty clinics (cardiology, nephrology, endocrinology). Interventions were typically delivered by pharmacists in person or with a mix of in-person and remote visits (in 89% of studies) over multiple sessions (89%). However, interventions varied in composition, delivery mode, and intensity, making it difficult to identify important intervention characteristics. Fifty-two studies were judged as low or medium risk of bias. Table 1. Summary of Included Studies, by Disease State and Study Characteristics* Studies were generally short term (12 months) and had small samples. They mainly assessed intermediate outcomes, such as blood pressure, cholesterol, or glucose goal attainment, in patients with hypertension, cardiovascular disease, or diabetes (45 of 65 unique studies). Goal attainment was listed as the primary study outcome in 23 unique studies. Clinical events were described in 15 unique studies and were the primary outcome in 2 of those studies. Office visits, medication adherence or nonadherence, hospitalizations, patient satisfaction, and health-related quality of life were noted in 26, 27, 20, 19, and 19 studies, respectively, but they rarely were the primary outcome. Additional study details are presented in Appendix Table 2. Findings Across All Disease Conditions In 15 trials (12 RCTs) reporting clinical events (for example, hypotension or hypoglycemia requiring additional intervention), pharmacist-led care and usual care (typically continuing a prestudy visit schedule) were generally similar (Appendix Table 3) (12, 14, 1922, 3437, 59, 62, 64, 71, 75, 77) Eight studies reported all-cause mortality, of which 6 found similar mortality in the pharmacist-led care and control groups and 2 did not report significance (12, 13, 16, 18, 28, 30, 37, 71). One of these studies in patients with diabetes also reported no disease-related deaths (28). However, of studies reporting clinical events or mortality, only 2 were designed to assess clinical events (13, 14), outcomes were sporadically and inconsistently reported, and there were few events (low strength of evidence) (Table 2). Evidence was insufficient to evaluate the effect of pharmacist-led care on patient satisfaction (overall and for each disease condition) and limited reporting of quality-of-life outcomes. Compared with usual care, pharmacist-led care improved study-selected goal attainment (typically for blood pressure, lipid, and glycemic goals). Strength of evidence was moderate, with consistent findings across disease conditions. Table 2. Strength of Evidence, by Outcome Evaluated* Appendix Table 3. Summary of Direction of Evidence For resource use outcomes, pharmacist-led care was similar to usual care for incidences or rates of office visits (26 studies), urgent care or emergenc