BACKGROUND: Malaria infections are a major public health problem in Africa and prompt treatment is one way of controlling the disease and saving lives. METHODS: This cluster-randomised controlled community intervention conducted in 2003-2005 aimed at improving early malaria case management in under five children. Health workers were trained to train community-based women groups in recognizing malaria symptoms, providing first-line treatment for uncomplicated malaria and referring severe cases. Evaluation was through a pre- (2004) and a post-intervention survey (2005). Anaemia prevalence was the primary outcome. RESULTS: 1715 children aged 6-59 months were included in the pre-intervention survey and 2169 in the post-intervention survey. The prevalence of anaemia decreased significantly from 37% [95% CI 34.7-39.3] to 0.5% [95% CI 0.2-0.7] after the intervention (p<0.001); slightly more in the intervention (from 43.9% to 0.8%) than in the control (30.8% to 0.17%) group (p=0.038). Fever and reported fever decreased significantly and the mean body weight of the children increased significantly over the study period in both control and intervention groups. CONCLUSION: The decrease in anaemia was significantly associated with the intervention, whereas the fever and body weight trends might be explained by other malaria control activities or seasonal/climate effects in the area. The community intervention was shown to be feasible in the study context.