摘要：Background Rwanda has made significant reductions in child mortality; however, reductions in neonatal deaths have been slower. Despite near universal facility-based delivery, about half of neonatal deaths occur within 48 h of birth in health facilities. All Babies Count is an evidence-based 18-month change acceleration process that provides neonatal equipment and supplies, neonatal training and mentoring, and district-wide quarterly learning collaborative sessions to promote peer-to-peer learning and continuous quality improvement in interprofessional teams. The Rwanda Ministry of Health and Partners In Health are scaling-up All Babies Count to facilities in seven hospital catchment areas to improve quality of care and reduce neonatal mortality. This study describes the first year of implementation. Methods A quality-improvement adviser provided support for All Babies Count implementation in each hospital catchment area through mentorship, training, and quality-improvement coaching. Four catchment areas launched in June and July, 2017 (phase 1) and three in October and November, 2017 (phase 2), covering 76 rural health facilities (seven hospitals and 69 health centres) located in the northern, southern, and western provinces of Rwanda. We used data from Rwanda health management information systems to monitor indicators of antenatal, intrapartum, postnatal, and inpatient neonatal care; process data were gathered from activity logs kept by quality-improvement advisers and surveys were completed at learning collaborative sessions. We used a χ2 test to measure performance differences between baseline (phase 1, April to June, 2017; phase 2, July to September, 2017) and the most recent quarter of All Babies Count implementation (April to June, 2018). Findings The percentage of women who had their first antenatal care visit in the first trimester of pregnancy increased from 48% (4414/9141) to 60% (6296/10 486) (p<0·0001) and the proportion of women receiving first postnatal care consultation within 24 h increased from 85% (2862/3356) to 93% (3060/3282) (p<0·0001). The proportion of participants in the learning collaborative sessions who reported being “very or extremely confident” working in quality improvement rose from 54% before the sessions to 95% after the sessions (p<0·0001), with 89 quality improvement projects (69 in antenatal care, 10 intrapartum, three in postnatal care, and seven in neonatology) initiated after two learning collaborative sessions in phase 1 hospital catchment areas and one session in phase 2 catchment areas. Interpretation Improved performance in measures targeted by All Babies Count was seen halfway through programme implementation, along with active quality-improvement activities and increased confidence in working with quality-improvement measures. These preliminary results show promise for the All Babies Count programme to improve care delivery and reduce neonatal mortality in diverse geographic areas of Rwanda. Funding All Babies Count is funded by Saving Lives at Birth.
