Introduction: Accurate documentation of neonatal morbidity and mortality is limited in many countries in sub-Saharan Africa. This project aimed to establish a surveillance system for neonatal conditions as an approach to improving the quality of neonatal care. Methods: A systematic data capture and surveillance system was established at Jinja Regional Referral Hospital, Uganda using a standardised neonatal medical record form which collected detailed individual patient level data. Additionally, training and mentorship were conducted and basic equipment was provided. Results: A total of 4178 neonates were hospitalised from July 2014 to December 2016. Median (IQR) age on admission was one day (1-3) and 48.0% (1851/3859) were male. Median (IQR) duration of hospitalisation was 17 days (IQR 10-40) and the longest duration of hospitalisation was 47 days (IQR 41-58). The majority were referrals from government health facilities (54.4%, 2012/3699), though 30.6% (1123/3669) presented as self-referrals. Septicaemia (44.9%, 1962/4371), prematurity (21.0%, 917/4371) and birth asphyxia (19.1%, 833/4371) were the most common diagnoses. The overall mortality was 13.8% (577/4178) and the commonest causes of death included septicaemia (26.9%, 155/577), prematurity (24.3%, 140/577), birth asphyxia (21.0%, 121/577), hypothermia (9.9%, 57/577) and respiratory distress (8.0%, 46/577). The majority of deaths (51.5%, 297/577) occurred within the first 24 h of hospitalisation although a significant proportion of deaths also occurred after 7 days of hospitalisation (24.1%, 139/577). A modest decrease in mortality and improvement in clinical outcome were observed. Conclusion: Improvement in neonatal data capture and quality of care was observed following establishment of an enhanced surveillance system, training and mentorship.
