Financial reinforcers for improving medication adherence: findings from a meta-analysis

Abstract

Background: Increasingly, financial reinforcement interventions based on behavioral economic principles are being applied in health care settings, and this study examined the use of financial reinforcers for enhancing adherence to medications.

Methods: Electronic databases and bibliographies of relevant references were searched, and a meta-analysis of identified trials was conducted. The variability in effect size and the impact of potential moderators (study design, duration of intervention, magnitude of reinforcement, and frequency of reinforcement) on effect size were examined.

Results: Fifteen randomized studies and 6 nonrandomized studies examined the efficacy of financial reinforcement interventions for medication adherence. Financial reinforcers were applied for adherence to medications for tuberculosis, substance abuse, human immunodeficiency virus, hepatitis, schizophrenia, and stroke prevention. Reinforcement interventions significantly improved adherence relative to control conditions with an overall effect size of 0.77 (95% confidence interval, 0.70-0.84; P<.001). Nonrandomized studies had a larger average effect size than randomized studies, but the effect size of randomized studies remained significant at 0.44 (95% confidence interval, 0.35-0.53; P<.001). Interventions that were longer in duration, provided an average reinforcement of $50 or more per week, and reinforced patients at least weekly resulted in larger effect sizes than those that were shorter, provided lower reinforcers, and reinforced patients less frequently.

Conclusion: These results demonstrate the efficacy of medication adherence interventions and underscore principles that should be considered in designing future adherence interventions. Financial reinforcement interventions hold potential for improving medication adherence and may lead to benefits for both patients and society.

Figure 1

Effect sizes (d) and 95% confidence intervals (CI) by disease state and study. Studies are listed by the disease state for which adherence to treatment was reinforced. Morisky 1990 (a) refers to an active sample of tuberculosis patients, and Morisky 1990 (p) refers to a preventive sample of tuberculosis cases. The total effect size represents all 21 studies.