# Multilevel Interventions to Improve Medication Adherence in Older Adults: A Systematic Review and Meta-Analysis of Cognitive, Digital, Behavioral, and Socioeconomic Strategies (2015–2025)

## TL;DR

This study reviews strategies to improve medication adherence in older adults, finding that caregiver support, digital tools, and simplified regimens can help, though results vary widely.

## Contribution

The paper provides a comprehensive analysis of multilevel interventions for medication adherence in older adults from 2015 to 2025.

## Key findings

- Caregiver involvement improved adherence in dementia and schizophrenia patients.
- Digital tools and machine learning models effectively predicted and improved adherence in chronic diseases.
- Simplified medication regimens and subsidies reduced nonadherence due to socioeconomic barriers.

## Abstract

Objectives: Medication adherence in elderly patients is shaped by cognitive, behavioral, systemic, and socioeconomic factors. This review aimed to identify determinants and effective strategies to improve adherence in older adults. Methods: A systematic search of PubMed, Scopus, and ScienceDirect (2015–2025) followed PRISMA 2020 guidelines. From 5116 records, 53 studies met inclusion criteria. Randomized controlled trials were meta-analyzed using standardized mean differences under a random-effects model. Risk of bias in the 10 pooled trials was assessed using the Cochrane RoB 2 tool, and certainty of evidence was evaluated using the GRADE framework. Results: Adherence ranged from 25.3% in institutionalized patients to 97.6% in pharmacist-led schizophrenia programs. Cognitive impairment and frailty reduced adherence (54.2%), while caregiver involvement improved rates, especially in dementia and schizophrenia (77.4–97.6%). Socioeconomic barriers, including medication cost, contributed to nonadherence but were mitigated by subsidies. Digital tools enhanced adherence in chronic disease, and machine learning models accurately predicted nonadherence (AUC up to 0.935). Effective interventions—caregiver support, digital platforms, and single-pill regimens—increased adherence by 25–59% and reduced cardiovascular events. The meta-analysis demonstrated a significant pooled effect (Standardized Mean Difference, SMD = 0.71, 95% CI: 0.11–1.54), although heterogeneity was high (I2 = 99%). The RoB 2 assessment of the 10 pooled trials identified 2 at low risk, 4 with some concerns, and 4 at high risk of bias; the GRADE certainty of evidence was rated Very Low. Conclusions: Multiple factors, including frailty, cognitive deficits, socioeconomic barriers, regimen complexity, and the level of caregiver support, appear to be consistently associated with medication adherence in older adults. Strategies such as caregiver engagement, digital health tools, regimen simplification, and mental health support may contribute to improved adherence, although effect sizes vary considerably across study contexts. Given the substantial heterogeneity, Very Low certainty of evidence (GRADE), and variable study quality, findings should be interpreted with caution. System-level reforms, financial assistance programs, and culturally tailored approaches may further support adherence, while the successful implementation of digital health solutions will require addressing literacy, accessibility, and integration challenges.