Establishing a neonatal database in a tertiary hospital in Rwanda - an observational study.

Background: Monitoring and evaluation is vital in the quest to improve the quality of care and to reduce the morbidity and mortality of neonates in a resource-limited setting. Databases offer several advantages such as data on large cohorts of neonates and from multiple centres. Aim: To establish a minimal dataset neonatal database in Kigali, Rwanda and to assess the quality and timing of the data entry process. Secondary objectives were to describe survival rates and associated risk factors. Methods: A cross-sectional, observational study was undertaken at a tertiary hospital in Kigali, Rwanda. The Rwanda Neonatal Data Collection Form was designed specifically for the database, based on the Vermont-Oxford Network neonatal data-collection tool with locally relevant amendments. All admitted neonates were enrolled during the study period of 2011-2017 with ongoing data-collection. Infants were recruited and data collected prospectively and cross-checked retrospectively with the inclusion of basic data on neonates who were not initially recruited prospectively. Results: 3391 analysable cases were recruited: 1420 prospective and 1971 retrospective cases. Prospective data collection peaked at 90%. Data entry was not always complete with data-points left blank with only 21% having adequate data available (0-25% missing). All-cause mortality during the study period was 16% and annual mortality ranged from 12% to 24%. On multivariate analysis, place of birth (AOR 2.17), small-for-gestational-age (AOR 2.05) and gestational age were all positively associated with survival. Conclusions: An academic setting in a low- or middle-income country can create and maintain a neonatal database without funding and produce a wealth of actionable results. Throughout the process, there were considerable challenges which must be addressed if such a database is to be optimised, maintained and created in other clinical sites.