The Effectiveness of Medication Adherence Interventions Among Patients With Coronary Artery Disease: A Meta-analysis

Abstract

Background: Despite the known benefits of medication therapy for secondary prevention of coronary artery disease (CAD), many patients do not adhere to prescribed medication regimens. Medication nonadherence is associated with poor health outcomes and higher healthcare cost.

Objective: The purpose of this meta-analysis was to determine the overall effectiveness of interventions designed to improve medication adherence (MA) among adults with CAD. In addition, sample, study design, and intervention characteristics were explored as potential moderators to intervention effectiveness.

Methods: Comprehensive search strategies helped in facilitating the identification of 2-group, treatment-versus-control-design studies testing MA interventions among patients with CAD. Data were independently extracted by 2 trained research specialists. Standardized mean difference effect sizes were calculated for eligible primary studies, adjusted for bias, and then synthesized under a random-effects model. Homogeneity of variance was explored using a conventional heterogeneity statistic. Exploratory moderator analyses were conducted using meta-analytic analogs for analysis of variance and regression for dichotomous and continuous moderators, respectively.

Results: Twenty-four primary studies were included in this meta-analysis. The overall effect size of MA interventions, calculated from 18,839 participants, was 0.229 (P < .001). The most effective interventions used nurses as interventionists, initiated interventions in the inpatient setting, and informed providers of patients' MA behaviors. Medication adherence interventions tested among older patients were more effective than those among younger patients. The interventions were equally effective regardless of number of intervention sessions, targeting MA behavior alone or with other behaviors, and the use of written instructions only.

Conclusions: Interventions to increase MA among patients with CAD were modestly effective. Nurses can be instrumental in improving MA among these patients. Future research is needed to investigate nurse-delivered MA interventions across varied clinical settings. In addition, more research testing MA interventions among younger populations and more racially diverse groups is needed.

Figure 1. Forest Plot of Main Effects

Forest plot of meta-analysis of two-group posttest comparisons of medication adherence outcomes listed by year of publication. Effect sizes calculated using a random effects model. Study weight is proportional to the area of each square.