RESUMEN
AIMS: The aim of this systematic review and meta-analysis was to synthesise the evidence relating to medication non-adherence and its association with health outcomes in people aged ≥50 years. METHODS: Seven databases were searched up to February 2019 for observational studies that measured medication (non-)adherence as a predictor of the following health outcomes in adults aged ≥50 years: healthcare utilisation (hospitalisation, emergency department visits, outpatient visits and general practitioner visits), mortality, adverse clinical events and quality of life. Screening and quality assessment using validated criteria were completed by 2 reviewers independently. Random effects models were used to generate pooled estimates of association using adjusted study results. The full methodological approach was published on PROSPERO (ID: CRD42017077264). RESULTS: Sixty-six studies were identified for qualitative synthesis, with 11 of these studies eligible for meta-analyses. A meta-analysis including 3 studies measuring medication non-adherence in adults aged ≥55 years showed a significant association with all-cause hospitalisation (adjusted odds ratio 1.17, 95% confidence interval [CI] 1.12, 1.21). A meta-analysis including 2 studies showed that medication non-adherence was not significantly associated with an emergency department visit (adjusted odds ratio 1.05, 95% CI 0.90, 1.22). Good adherence was associated with a 21% reduction in long-term mortality risk in comparison to medication non-adherence (adjusted hazard ratio 0.79, 95% CI 0.63, 0.98). CONCLUSION: Medication non-adherence may be significantly associated with all-cause hospitalisation and mortality in older people. Medication adherence should be monitored and addressed in this cohort to minimise hospitalisation, improve clinical outcomes and reduce healthcare costs.