To describe the implementation of bubble-CPAP in a referral hospital in a developing country and to investigate: the feasibility of nurses implementing bubble-CPAP and the impact of bubble-CPAP on need for mechanical ventilation and mortality. Retrospective evaluation of prospectively collected data from two time periods: 18 months before and 18 months after the introduction of bubble-CPAP. The introduction of bubble-CPAP was associated with a 50 per cent reduction in the need for mechanical ventilation; from 113 of 1106 (10.2 per cent) prior to bubble-CPAP to 70 of 1382 (5.1%) after introduction of CPAP (χ2, p<0.001). In the 18 months prior to bubble-CPAP there were 79 deaths (case fatality of 7.1 per cent). In the 18 months after bubble-CPAP there were 74 deaths (CF 5.4 per cent), relative risk: 0.75 (0.55–1.02, χ2, p=0.065). Nurses could safely apply bubble-CPAP after 1–2 months of on-the-job training. Equipment for Bubble-CPAP cost 15 per cent of the cost of the cheapest mechanical ventilator. The introduction of bubble-CPAP substantially reduced the need for mechanical ventilation, with no difference in mortality. In models of neonatal care for resource-limited countries, bubble-CPAP may be the first type of ventilatory support that is recommended. Its low cost and safety when administered by nurses makes it ideal for this purpose. Bubble-CPAP has the potential for being available at even lower cost than the current commercially available bubble systems used in this study.