Background: Adherence to combination antiretroviral therapy (ART) is a key predictor of the success of human immunodeficiency virus (HIV) treatment, and is potentially amenable to intervention. Insight into predictors or correlates of non-adherence to ART may help guide targets for the development of adherence-enhancing interventions. Our objective was to review evidence on predictors/correlates of adherence to ART, and to aggregate findings into quantitative estimates of their impact on adherence.
Methods: We searched PubMed for original English-language papers, published between 1996 and June 2014, and the reference lists of all relevant articles found. Studies reporting on predictors/correlates of adherence of adults prescribed ART for chronic HIV infection were included without restriction to adherence assessment method, study design or geographical location. Two researchers independently extracted the data from the same papers. Random effects models with inverse variance weights were used to aggregate findings into pooled effects estimates with 95% confidence intervals. The standardized mean difference (SMD) was used as the common effect size. The impact of study design features (adherence assessment method, study design, and the United Nations Human Development Index (HDI) of the country in which the study was set) was investigated using categorical mixed effects meta-regression.
Results: In total, 207 studies were included. The following predictors/correlates were most strongly associated with adherence: adherence self-efficacy (SMD = 0.603, P = 0.001), current substance use (SMD = -0.395, P = 0.001), concerns about ART (SMD = -0.388, P = 0.001), beliefs about the necessity/utility of ART (SMD = 0.357, P = 0.001), trust/satisfaction with the HIV care provider (SMD = 0.377, P = 0.001), depressive symptoms (SMD = -0.305, P = 0.001), stigma about HIV (SMD = -0.282, P = 0.001), and social support (SMD = 0.237, P = 0.001). Smaller but significant associations were observed for the following being prescribed a protease inhibitor-containing regimen (SMD = -0.196, P = 0.001), daily dosing frequency (SMD = -0.193, P = 0.001), financial constraints (SMD -0.187, P = 0.001) and pill burden (SMD = -0.124, P = 0.001). Higher trust/satisfaction with the HIV care provider, a lower daily dosing frequency, and fewer depressive symptoms were more strongly related with higher adherence in low and medium HDI countries than in high HDI countries.
Conclusions: These findings suggest that adherence-enhancing interventions should particularly target psychological factors such as self-efficacy and concerns/beliefs about the efficacy and safety of ART. Moreover, these findings suggest that simplification of regimens might have smaller but significant effects.