To assess the impact and feasibility of artemether-lumefantrine deployment at community level, combined with phased introduction of rapid diagnostic tests (RDTs), on malaria transmission, morbidity, and mortality and health service use in a remote area of Ethiopia. Two-year pilot study in two districts: artemether-lumefantrine was prescribed after parasitological confirmation of malaria in health facilities in both districts. In the intervention district, artemether-lumefantrine was also made available through 33 community health workers (CHWs); of these, 50% were equipped with RDTs in the second year. At health facilities; 54 774 patients in the intervention and 100 535 patients in the control district were treated for malaria. In the intervention district, 75 654 patients were treated for malaria by community health workers. Use of RDTs in Year 2 excluded non-Plasmodium falciparumin 89.7% of suspected cases. During the peak of malaria transmission in 2005, the crude parasite prevalence was 7.4% (95% CI: 6.1-8.9%) in the intervention district and 20.8% (95% CI: 18.7-23.0%) in the control district. Multivariate modelling indicated no significant difference in risk of all-cause mortality between the intervention and the control districts [adjusted incidence rate ratio (aIRR) 1.03, 95%CI 0.87-1.21, P = 0.751], but risk of malaria-specific mortality was lower in the intervention district (aIRR 0.60, 95%CI 0.40-0.90, P = 0.013). Artemether-lumefantrine deployment through a community-based service in a remote rural population reduced malaria transmission, lowered the malaria case burden for health facilities and reduced malaria morbidity and mortality during a 2-year period which included a major malaria epidemic.