OBJECTIVES:To further evaluate our earlier findings on the feasibility and effectiveness of home-based management of neonatal sepsis by analysing 7 years data (1996 to 2003) from the field trial in Gadchiroli, India.STUDY DESIGN:Neonates in 39 villages were monitored by trained village health workers (VHWs) from 1995 onwards. In 1996, we trained VHWs to diagnose sepsis by using a clinical algorithm and provide domiciliary treatment using intramuscular gentamicin and oral co-trimoxazole. Health records for all neonates were kept by the VHWs, checked by field supervisors, and computerized. Live births and neonatal deaths were recorded by an independent vital statistics collection system. We evaluated the feasibility and effectiveness of this approach.RESULTS:During September 1996 to March 2003, VHWs monitored 93% of all neonates in 39 villages (N=5268). As compared to 552 cases of sepsis diagnosed by computer algorithm, VHWs correctly diagnosed 492 cases (89%). Parents agreed to home-based treatment for the majority of infants (448, 91%), refused treatment in 31 (6.4%) cases, and hospitalized 13 infants (2.6%). VHWs treated 470 neonates with antibiotics, that is, 8.9% of all neonates in community. Of 552 cases diagnosed by computer, VHWs correctly treated 448 (81.2%) and gave unnecessary treatment to 22/470 (4.7%) of treated neonates. The case fatality (CF) was 6.9% in treated cases vs 22% in untreated or 16.6% in the pre-intervention period (p<0.001). Home-based treatment resulted in 67.2% reduction in %CF among preterm and a 72% reduction among LBW neonates.CONCLUSIONS:Home-based management of neonates with suspected sepsis is acceptable to most parents, safe, and effective in reducing sepsis case fatality by nearly 60%. With proper selection, training, and supervision of health workers, this method may be applicable in areas in developing countries where access to hospital care is limited.