Adherence to artemisinin-based combination therapies (ACTs) for malaria is important for effective treatment. This thesis compares adherence to ACTs obtained in the public and private retail sectors, describes an intervention to improve dispenser knowledge and patient adherence, and addresses challenges of measuring patient adherence to ACTs. A cluster randomised trial of a text message intervention targeted at dispensers in Accredited Drug Dispensing Outlets (ADDOs) was conducted in Tanzania to improve provision of advice on artemether-lumefantrine (AL) and as a result patient adherence. An observational study was also conducted among patients obtaining AL from public health facilities. In a third study, smart blister packs that recorded when pills were removed were used to assess the validity of self-report. Adherence was measured as completion of all doses (“completed treatment”) and completion of each dose at the correct time (“timely completion”). The intervention improved dispenser knowledge, but had no effect on patient completion of treatment (intervention 68.3%, control 69.8%, p [adjusted] = 0.6), or timely completion (intervention 33.1%, control 32.6%, p [adjusted] = 0.9). ADDO patients were wealthier, more educated, older, sought care later in the day, and were less likely to test positive for malaria than health facility patients. The adjusted odds of completed treatment and of timely completion for ADDO patients were 0.65 (95% CI: 0.43, 1.00) and 0.69 (95% CI: 0.47, 1.01) times that of health facility patients. Timely completion, but not completed treatment, was lower by smart blister packs than by self-report (37% vs. 24%, p<0.0001). Adherence to AL in both sectors was suboptimal. As the private sector continues to be important for malaria treatment, better understanding is needed of which aspects of patient care are most important to maximise adherence and how methods of assessing adherence can be improved.