Introduction/Objective: To summarize the methodologies and reported impacts of interventions to improve use of medicines in primary care settings in developing countries, and to identify critical areas for further research.

Methods: Systematic review of published and unpublished intervention studies, categorized by type of intervention, by health problem addressed, by adequacy of research design, and by effect size.

Results: Fifty-nine studies (36 acceptable designs) met review criteria. Interventions included training (34%), community case management (CCM) (31%), group norm setting (10%), supervision or audit/feedback (14%), and essential drugs programs (12%). Most targeted ARI (44%), diarrhoea (36%), or malaria (10%) in children. Major prescribing issues addressed were antibiotic misuse (76%), polypharmacy (24%) and injections (20%). Authors reported at least one large impact (>25% improvement relative to controls) in 43% and moderate impacts (10-25% gain) in 36% of well-designed interventions; only 21% had no impact. The impact of training was higher in interventions with multiple training modalities; repeated sessions; focus on one clinical problem; training at the work site; and using opinion leaders or district-level staff as trainers. CCM interventions clearly reduced mortality, but their effects on drug use were not well studied. Dissemination of printed materials consistently had no impact. The administrative interventions tested achieved consistently moderate to large impacts on target practices; effects on a broader range of problems and the sustainability of improvements remain unexplored.

Discussion: Well-designed developing country drug use interventions are increasing in frequency, with consistent evidence for effectiveness of various strategies. Improving drug use in adults, for chronic conditions, in the private sector, and cost-effectiveness remain poorly studied.