Meta-analysis of interventions to improve drug adherence in patients with hyperlipidemia

To examine the results of meta-analyses addressing the net effect of tools and methods to enhance drug adherence in patients with hyperlipidemia.

We searched MEDLINE, International Pharmaceutical Abstracts (IPA), PsychLit, Educational Resource Information Clearinghouse (ERIC), and Excerpta Medica Database (EMBASE) databases (1996-2000) for randomized controlled trials of interventions to improve patients' adherence to drug regimens for hyperlipidemia. All trials selected had at least 10 subjects/group. Effect sizes, represented as the correlation coefficient, were calculated from Cohen's d and d'. Analysis of variance and the Q test were performed to evaluate statistical significance and heterogeneity.

Of the 484 articles evaluated, only four met the criteria for inclusion. Multiple interventions or study samples were identified in two articles. Each intervention was counted as a separate study, yielding seven cohorts with a total of 3077 subjects. Behavioral interventions accounted for five cohorts (2915 subjects). No cohorts received solely educational interventions. Two cohorts, with a total of 162 subjects, received a combination of behavioral and educational interventions. Homogeneity of groupings and effect sizes were calculated for the overall grouping as well as for each type of intervention. Overall, the effect size for the data was 0.14 (95% confidence interval [CI] 0.10-0.17) and the data were homogenous (Q = 7.36, df = 4, p = 0.29). The effect size for the behavioral interventions was 0.14 (95% CI 0.11-0.18, Q = 5.25, df = 4, p = 0.26). For interventions that combined behavioral and educational components, the effect size was 0.03 (CI -0.13-0.18, Q = 0.204, df = 1, p = 0.65).

The interventions evaluated had little impact on drug adherence. More studies are needed to assess how to improve drug adherence in patients with hyperlipidemia.