OBJECTIVE: To determine whether cost effectiveness evidence on adherence-enhancing interventions (AEIs) was of sufficient quality to inform policy decision-making. METHODS: A computerised search of Embase, Medline, Cinahl, Econlit, NHSEED, Psychlit, ePIC and Cochrane databases from January, 1980 to April, 2004 was performed. English-language human subject articles were identified using an inclusive search strategy. Studies that appeared to assess the cost effectiveness of medication adherence-enhancing interventions were included. Methodologic rigor was assessed using 15 clinical, economic and adherence-specific quality criteria developed by the authors. RESULTS: We found 45 comparative studies in 43 publications. Asthma (14 studies) and psychiatric illness (12 studies) were most commonly investigated. In 33 studies, interventions were educational, 20 had multiple components and 23 did not appear to be linked to proven reasons for non-adherence. No studies assessed management of unintentional non-adherence. No study met all quality criteria. Study quality has not improved with time, as some better studies are over ten years old. Many studies used inadequate or unidentifiable adherence measures. Critically, many were too small or not randomised. All studies assumed that patients were prescribed appropriate therapy for their condition, and no assessment of treatment quality was made by any study. Reporting of adherence and outcome results was often unclear. Cost data were poorer quality than outcome data, using average or estimated costs and omitting some cost elements. Nine studies carried out incremental economic analysis. CONCLUSIONS: We were not able to make definitive conclusions about the cost-effectiveness of medication adherence enhancing interventions due to the heterogeneity of the studies found and incomplete reporting of results. Important policy decisions need to be made about non-adherence, however, they are currently being made in a vacuum of adequate information. Medication adherence-enhancing interventions must be based on reasons for non-adherence and be evaluated using accepted clinical and economic quality criteria.