BACKGROUND: WHO and UNICEF launched the Integrated Management of Childhood Illness (IMCI) strategy in the mid-1990s to reduce deaths from diarrhoea, pneumonia, malaria, measles, and malnutrition in children younger than 5 years. We assessed the effect of IMCI on health and nutrition of children younger than 5 years in Bangladesh. METHODS: In this cluster randomised trial, 20 first-level government health facilities in the Matlab subdistrict of Bangladesh and their catchment areas (total population about 350 000) were paired and randomly assigned to either IMCI (intervention; ten clusters) or usual services (comparison; ten clusters). All three components of IMCI-health-worker training, health-systems improvements, and family and community activities-were implemented beginning in February, 2002. Assessment included household and health facility surveys tracking intermediate outputs and outcomes, and nutrition and mortality changes in intervention and comparison areas. Primary endpoint was mortality in children aged between 7 days and 59 months. Analysis was by intention to treat. This study is registered, number ISRCTN52793850. FINDINGS: The yearly rate of mortality reduction in children younger than 5 years (excluding deaths in first week of life) was similar in IMCI and comparison areas (8.6%vs 7.8%). In the last 2 years of the study, the mortality rate was 13.4% lower in IMCI than in comparison areas (95% CI -14.2 to 34.3), corresponding to 4.2 fewer deaths per 1000 livebirths (95% CI -4.1 to 12.4; p=0.30). Implementation of IMCI led to improved health-worker skills, health-system support, and family and community practices, translating into increased care-seeking for illnesses. In IMCI areas, more children younger than 6 months were exclusively breastfed (76%vs 65%, difference of differences 10.1%, 95% CI 2.65-17.62), and prevalence of stunting in children aged 24-59 months decreased more rapidly (difference of differences -7.33, 95% CI -13.83 to -0.83) than in comparison areas. INTERPRETATION: IMCI was associated with positive changes in all input, output, and outcome indicators, including increased exclusive breastfeeding and decreased stunting. However, IMCI implementation had no effect on mortality within the timeframe of the assessment. FUNDING: Bill & Melinda Gates Foundation, WHO's Department of Child and Adolescent Health and Development, and US Agency for International Development.