The association between medication non-adherence and adverse health outcomes in ageing populations: A systematic review and meta-analysis

Abstract

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.

Keywords: ageing population; hospitalisation; medication adherence; mortality.

Figure 1

Study selection process for systematic review of the association between medication (non‐)adherence and health outcomes

Figure 2

Forest plots of medication non‐adherence and association with all‐cause hospitalisations (A) and disease‐specific hospitalisations (B). OR = odds ratio; CI = confidence interval; PDC = proportion of days covered; MPR = medication possession ratio. Reference group: Adherent (PDC/MPR ≥80%). Intermediate adherence (PDC 40–79%) results are reported for Rasmussen et al. in (A). All‐cause hospitalisation (Figure 2A): Heterogeneity: χ² = 0.08 (d.f. = 2), P = .958, I² = 0.0% τ² = 0.0000. Test for overall effect: Z= 7.65, P < .0001. Disease‐specific hospitalisation (Figure 2B): Heterogeneity: χ² = 4.26 (d.f. = 2), P = .119, I² = 53.0%, τ²= 0.0035. Test for overall effect: Z= 1.47, P =.143

Figure 3

Forest plot of medication non‐adherence and association with emergency department visits. OR = odds ratio; CI = confidence interval; PDC = proportion of days covered; MPR = medication possession ratio. Reference group: Adherent (PDC/MPR ≥ 80%). Heterogeneity: χ² = 2.51 (d.f. = 1), P = .113, I² = 60.2%, τ²= 0.0084. Test for overall effect: Z= 0.57, P = .566

Figure 4

Forest plot of low medication adherence (A) and association with mortality (follow up≥1 year) and good medication adherence (B) and association with mortality (follow up≥1 year). HR = hazard ratio; CI = confidence interval; PDC = proportion of days covered. (Figure 2A) All studies measure adherence categorically. Reference group: Adherent (PDC ≥ 80%); *low medication adherence: PDC ≤ 60%; #low medication adherence: PDC < 40%; Heterogeneity: χ² = 0.01 (d.f. = 3), P = .927, I² = 0.0%, τ²= 0.0000. Test for overall effect: Z = 5.09, P < .001. (Figure 2B) All studies measured adherence dichotomously. Reference group for all studies: non‐adherent (PDC ≤ 80%). Heterogeneity: χ² = 130.72 (d.f. = 7), P < .001, I² = 94.6%, τ² = 0.0906. Test for overall effect: Z = 2.12, P = .034