Care during labor and birth for the prevention of intrapartum-related neonatal deaths: a systematic review and Delphi estimation of mortality effect

Background: Our objective was to estimate the effect of various childbirth care packages on neonatal mortality due to intrapartum-related events (“birth asphyxia”) in term babies for use in the Lives Saved Tool (LiST). Methods: We conducted a systematic literature review to identify studies or reviews of childbirth care packages as defined by United Nations norms (basic and comprehensive emergency obstetric care, skilled care at birth). We also reviewed Traditional Birth Attendant (TBA) training. Data were abstracted into standard tables and quality assessed by adapted GRADE criteria. For interventions with low quality evidence, but strong GRADE recommendation for implementation, an expert Delphi consensus process was conducted to estimate causespecific mortality effects. Results: We identified evidence for the effect on perinatal/neonatal mortality of emergency obstetric care packages: 9 studies (8 observational, 1 quasi-experimental), and for skilled childbirth care: 10 studies (8 observational, 2 quasi-experimental). Studies were of low quality, but the GRADE recommendation for implementation is strong. Our Delphi process included 21 experts representing all WHO regions and achieved consensus on the reduction of intrapartum-related neonatal deaths by comprehensive emergency obstetric care (85%), basic emergency obstetric care (40%), and skilled birth care (25%). For TBA training we identified 2 metaanalyses and 9 studies reporting mortality effects (3 cRCT, 1 quasi-experimental, 5 observational). There was substantial between-study heterogeneity and the overall quality of evidence was low. Because the GRADE recommendation for TBA training is conditional on the context and region, the effect was not estimated through a Delphi or included in the LiST tool. Conclusion: Evidence quality is rated low, partly because of challenges in undertaking RCTs for obstetric interventions, which are considered standard of care. Additional challenges for evidence interpretation include varying definitions of obstetric packages and inconsistent measurement of mortality outcomes. Thus, the LiST effect estimates for skilled birth and emergency obstetric care were based on expert opinion. Using LiST modelling, universal coverage of comprehensive obstetric care could avert 591,000 intrapartum-related neonatal deaths each year. Investment in childbirth care packages should be a priority and accompanied by implementation research and further evaluation of intervention impact and cost.