BACKGROUND: We aimed to assess risk factors for neonatal mortality, quality of neonatal resuscitation (NR) on videos and identify potential areas for improvement.

METHODS: This prospective cohort study included women in childbirth and their newborns at four district hospitals in Pemba, Tanzania. Videos were analysed for quality-of-care. Questionnaires on quality-of-care indicators were answered by health workers (HW) and women. Risk factors for neonatal mortality were analysed in a binomial logistic regression model.

RESULTS: 1440 newborns were enrolled. 34 newborns died within the neonatal period (23.6 per 1000 live births). Ninety neonatal resuscitations were performed, 20 cases on video. Positive pressure ventilation (PPV) was inadequate in 15 cases (75%). Half (10/20) did not have PPV initiated within the first minute, and in one case (5.0%), no PPV was performed. PPV was not sustained in 16/20 (80%) newborns. Of the 20 videos analysed, death occurred in 10 newborns: 8 after resuscitation attempts and two within the first 24 h. Most of HW 49/56 (87.5%) had received training in NR.

CONCLUSIONS: Video analysis of NR revealed significant deviations from guidelines despite 87.5% of HW being trained in NR. Videos provided direct evidence of gaps in the quality of care and areas for future education, particularly effective PPV.

IMPACT: Neonatal mortality in Pemba is 23.6 per 1000 livebirths, with more than 90% occurring in the first 24 h of life. Video assessment of neonatal resuscitation revealed deviations from guidelines and can add to understanding challenges and aid intervention design. The present study using video assessment of neonatal resuscitation is the first one performed at secondary-level hospitals where many of the world's births are conducted. Almost 90% of the health workers had received training in neonatal resuscitation, and the paper can aid intervention design by understanding the actual challenges in neonatal resuscitation.