Effect of a scaled-up neonatal resuscitation quality improvement package on intrapartum-related mortality in Nepal: A stepped-wedge cluster randomized controlled trial

Background Improving quality of intrapartum care will reduce intrapartum stillbirth and neonatal mortality, especially in resource-poor settings. Basic neonatal resuscitation can reduce intrapartum stillbirth and early neonatal mortality, if delivered in a high-quality health system, but there is a dearth of evidence on how to scale up such evidence-based interventions. We evaluated the scaling up of a quality improvement (QI) package for neonatal resuscitation on intrapartum-related mortality (intrapartum stillbirth and first day mortality) at hospitals in Nepal. Methods and findings We conducted a stepped-wedge cluster randomized controlled trial in 12 hospitals over a period of 18 months from April 14, 2017, to October 17, 2018. The hospitals were assigned to one of four wedges through random allocation. The QI package was implemented in a stepped-wedge manner with a delay of three months for each step. The QI package included improving hospital leadership on intrapartum care, building health workers' competency on neonatal resuscitation, and continuous facilitated QI processes in clinical units. An independent data collection system was set up at each hospital to gather data on mortality through patient case note review and demographic characteristics of women using semi-structured exit interviews. The generalized linear mixed model (GLMM) and multivariate logistic regression were used for analyses. During this study period, a total of 89,014 women-infant pairs were enrolled. The mean age of the mother in the study period was 24.0 +/- 4.3 years, with 54.9% from disadvantaged ethnic groups and 4.0% of them illiterate. Of the total birth cohort, 54.4% were boys, 16.7% had gestational age less than 37 weeks, and 17.1% had birth weight less than 2,500 grams. The incidence of intrapartum-related mortality was 11.0 per 1,000 births during the control period and 8.0 per 1,000 births during the intervention period (adjusted odds ratio [aOR], 0.79; 95% CI, 0.69-0.92; p = 0.002; intra-cluster correlation coefficient [ICC], 0.0286). The incidence of early neonatal mortality was 12.7 per 1,000 live births during the control period and 10.1 per 1,000 live births during the intervention period (aOR, 0.89; 95% CI, 0.78-1.02; p = 0.09; ICC, 0.1538). The use of bag-and-mask ventilation for babies with low Apgar score (<7 at 1 minute) increased from 3.2% in the control period to 4.0% in the intervention period (aOR, 1.52; 95% CI, 1.32-1.77, p = 0.003). There were two major limitations to the study; although a large sample of women-infant pairs were enrolled in the study, the clustering reduced the power of the study. Secondly, the study was not sufficiently powered to detect reduction in early neonatal mortality with the number of clusters provided. Conclusion These results suggest scaled-up implementation of a QI package for neonatal resuscitation can reduce intrapartum-related mortality and improve clinical care. The QI intervention package is likely to be effective in similar settings. More implementation research is required to assess the sustainability of QI interventions and quality of care.