Abstract
IMPORTANCE: The World Health Organization (WHO) Safe Childbirth Checklist (SCC) has been adapted and implemented in at least 35 countries. Consistently, the SCC has shown increased adherence to practices, but there are mixed results regarding its association with health outcomes in different settings. OBJECTIVE: To examine the association of SCC implementation with mortality, accounting for variations in evidence-based practices (EBP) adherence. DESIGN, SETTING, AND PARTICIPANTS: In this meta-analysis, data were pooled from 3 cluster randomized trials of the SCC (January 1, 2014, to December 31, 2017). Intention to treat (ITT) and a complier average causal effect analysis (CACE) on EBPs and perinatal mortality were estimated via a generalized linear model. The primary facilities were in Uttar Pradesh, India; basic emergency obstetric facilities were in Aceh, Indonesia; and primary and secondary health centers were in Khyber Pakhtunkhwa, Pakistan. INTERVENTIONS: In India, the 8-month SCC intervention involved facility engagement, a launch event, and 8 months of tapered coaching. In Indonesia, the 6-month SCC intervention included 11 coaching visits. In Pakistan, the 12-month SCC intervention included light touch external monitoring, skills training, and supplies assessment. MAIN OUTCOMES AND MEASURES: Primary outcomes were stillbirth and perinatal and early neonatal mortality. Secondary outcomes were adherence to 15 EBPs, facility supply availability, and safety culture perceptions. RESULTS: Pooled data included 169 511 births, supply assessments from 163 facilities, and 6298 observed deliveries for EBPs and health workers' perceptions on safety culture. Mortality did not differ in the full sample; however, during months when EBP observations were conducted, stillbirth rates in the intervention facilities were lower by 9.8 per 1000 births (95% CI, -18.5 to -1.1; P = .03; q = .05) in the ITT analysis and 14.5 per 1000 births (95% CI, -27.2 to -1.7; P = .03; q = .05) in the CACE analysis compared with control facilities. EBP adherence was higher by 3.6 practices (95% CI, 3.3 to 4.1; P < .001; q = .001) in the ITT analysis and 6.0 practices (95% CI, 5.3 to 6.8; P < .001; q = .001) in the CACE analysis in intervention facilities. CONCLUSIONS AND RELEVANCE: In this meta-analysis, SCC use in lower-middle-income settings was associated with increased EBP adherence and lower rates of stillbirths when EBPs were directly observed. Further research is needed to identify additional factors to optimize SCC's potential impact on maternal and newborn safety outcomes.