Aim: This paper is a report of a synthesis of evidence on implementation of interventions to improve adherence to antiretroviral therapy.
Background: Evidence on efficacy must be supplemented with evidence on how interventions were implemented in practice and on how that implementation varied across populations and settings.
Data sources: Sixty-one reports were reviewed of studies conducted in the United States of America in the period 2001 to December 2008. Fifty-two reports were included in the final analysis: 37 reporting the effects of interventions and 15 reporting intervention feasibility, acceptability, or fidelity.
Review methods: An adaptation of Pawson's realist synthesis method was used, whereby a provisional explanatory model and associated list of propositions are developed from an initial review of literature. This model is successively refined to the point at which it best explains empirical findings from the reports reviewed.
Results: The final explanatory model suggests that individuals with HIV will be more likely to enroll in interventions that protect their confidentiality, to attend when scheduling is responsive to their needs, and both to attend and continue with an intervention when they develop a strong, one-to-one relationship with the intervener. Participants who have limited prior experience with antiretroviral therapy will be more likely to continue with an intervention than those who are more experienced. Dropout rates are likely to be higher when interventions are integrated into existing delivery systems than when offered as stand-alone interventions.
Conclusion: The explanatory model developed in this study is intended to provide guidance to clinicians and researchers on the points in the implementation chain that require strengthening.