Medications and Prescribing Patterns as Factors Associated with Hospitalizations from Long-Term Care Facilities: A Systematic Review

Background Residents of long-term care facilities (LTCFs) are at high risk of hospitalization. Medications are a potentially modifiable risk factor for hospitalizations. Objective Our objective was to systematically review the association between medications or prescribing patterns and hospitalizations from LTCFs. Methods We searched MEDLINE, Embase, Cumulative Index to Nursing and Allied Health Literature (CINAHL) and International Pharmaceutical Abstracts (IPA) from inception to August 2017 for longitudinal studies reporting associations between medications or prescribing patterns and hospitalizations. Two independent investigators completed the study selection, data extraction and quality assessment using the Joanna Briggs Institute Critical Appraisal Tools. Results Three randomized controlled trials (RCTs), 22 cohort studies, five case–control studies, one case-time-control study and one case-crossover study, investigating 13 different medication classes and two prescribing patterns were included. An RCT demonstrated that high-dose influenza vaccination reduced all-cause hospitalization compared with standard-dose vaccination (risk ratio [RR] 0.93; 95% confidence interval [CI] 0.88–0.98). Another RCT found no difference in hospitalization rates between oseltamivir as influenza treatment and oseltamivir as treatment plus prophylaxis (treatment = 4.7%, treatment and prophylaxis = 3.5%; p = 0.7). The third RCT found no difference between multivitamin/mineral supplementation and hospitalization (odds ratio [OR] 0.94; 95% CI 0.74–1.20) or emergency department visits (OR 1.05; 95% CI 0.76–1.47). Two cohort studies demonstrated influenza vaccination reduced hospitalization. Four studies suggested polypharmacy and potentially inappropriate medications (PIMs) increased all-cause hospitalization. However, associations between polypharmacy (two studies), PIMs (one study) and fall-related hospitalizations were inconsistent. Inconsistent associations were found between psychotropic medications with all-cause and cause-specific hospitalizations (11 studies). Warfarin, nonsteroidal anti-inflammatory drugs, pantoprazole and vinpocetine but not long-term acetylsalicylic acid (aspirin), statins, trimetazidine, digoxin or β -blockers were associated with all-cause or cause-specific hospitalizations in single studies of specific resident populations. Most cohort studies assessed prevalent rather than incident medication exposure, and no studies considered time-varying me