BACKGROUND: Insecticide-treated nets have proven efficacy as a malaria-control tool in Africa. However, the transition from efficacy to effectiveness cannot be taken for granted. We assessed coverage and the effect on child survival of a large-scale social marketing programme for insecticide-treated nets in two rural districts of southern Tanzania with high perennial malaria transmission. METHODS: Socially marketed insecticide-treated nets were introduced step-wise over a 2-year period from May, 1997, in a population of 480000 people. Cross-sectional coverage surveys were done at baseline and after 1, 2, and 3 years. A demographic surveillance system (DSS) was set up in an area of 60000 people to record population, births, and deaths. Within the DSS area, the effect of insecticide-treated nets on child survival was assessed by a case-control approach. Cases were deaths in children aged between 1 month and 4 years. Four controls for each case were chosen from the DSS database. Use of insecticide-treated nets and potential confounding factors were assessed by questionnaire. Individual effectiveness estimates from the case-control study were combined with coverage to estimate community effectiveness. FINDINGS: Insecticide-treated net coverage of infants in the DSS area rose from less than 10% at baseline to more than 50% 3 years later. Insecticide-treated nets were associated with a 27% increase in survival in children aged 1 month to 4 years (95% CI 3-45). Coverage in such children was higher in areas with longer access to the programme. The modest average coverage achieved by 1999 in the two districts (18% in children younger than 5 years) suggests that insecticide-treated nets prevented 1 in 20 child deaths at that time. INTERPRETATION: Social marketing of insecticide-treated nets has great potential for effective malaria control in rural African settings.