Low birth weight (LBW) puts infants at greater risk of life-threatening respiratory and gastrointestinal conditions that account for much of the mortality in early infancy. Dramatic improvements in child survival in developing countries, there is limited success in reducing infant mortality. The Kangaroo Mother Care (KMC) method, whereby hospital-born stabilized LBW newborns are placed in an upright position in twenty-four hour skin-to-skin contract on the mother's breast has been shown to significantly reduce the incidence of life-threatening morbidity. Yet it has never been adapted for community-based implementation, where it could prove to be the best means of non-institutional stabilization and significantly reduce neonatal and infant mortality. This is critical in developing countries where fewer than 50% of women give birth in health facilities. We will adapt the KMC for community-based application and conduct a randomized controlled cluster trial of 3,200 infants to determine its effectiveness to reduce neonatal and infant mortality. The Population Council will conduct this study with the Bangladesh Rural Advancement Committee (BRAC) in Bangladesh where 95% of births occur at home and 50% of births nationally are LBW, in Sylhet where IMR and NMR are 13.8% and 8.5%, the highest in Bangladesh. BRAC has an extensive community development program in Sylhet, including nurses and community-based workers that can be trained to be KMC instructors, and has an ongoing periodic household data collection system in the area. Project area nurses will train BRAC community-based workers to instruct village women in the experimental study area in KMC. All consenting women expected to deliver within the first year of the project will be enrolled in the study. Birth weight will be measured within 48 hours. Information on subject characteristics will be obtained by interview at enrollment, and reported morbidity, breastfeeding, skin-to-skin contact, neonatal and infant mortality, and arm circumference measurement will be collected at 3- month intervals through each infants first birthday. Data will be directly recorded into handheld computers to speed data processing, preliminary analyses, and consistency checks for immediate resolution simultaneous to data collection. Bivariate comparisons and adjusted Cox proportional hazards analyses will be conducted by intent-to-treat and by receipt of and KMC compliance.
