PurposeTo identify approaches, enablers, barriers and outcomes of facility stillbirth and neonatal death audit in low-income and middle-income countries (LMICs).Data sourcesWe searched MEDLINE, CINAHL Complete, Academic Search Index, Science Citation Index, Complementary index and Global health electronic databases.Study selectionStudies were considered eligible when reporting the approaches, enablers, barriers and outcomes of facility-based stillbirth and neonatal death audit in LMICs.Data extractionTwo authors independently performed the data extraction using predefined templates made before data extraction.Results of data synthesisA total of 10 articles from 7 countries were included in the final analysis. Facility or external multidisciplinary teams performed death audits on a weekly or monthly basis. A total of 1018 stillbirths and neonatal deaths were audited. Of 18 audit enablers identified, nine were at the health provider level while 18 of 23 barriers to audit that were identified occurred at the facility level. The facility-level barriers cited by more than one study included: failure to implement change; inadequate training; limited time; increased workload; too many cases and poor documentation. Six studies reported that death audits resulted in structural improvements in physical structure, training, service organisation, supplies and equipment in the wards. Five studies reported that death audits improved the standard of care, with one study showing a significant improvement in measured standards. One study reported a significant reduction in newborn mortality rate of 29.4% (95% CI 0.6% to 2.4%; p=0.0015) and one study a reduction in perinatal mortality of 4.9% (52.8% in 2007 to 47.9% in 2008) before and after perinatal audit implementation.ConclusionStillbirth and neonatal death audit improves facility structures, processes of care and health outcomes in neonatal care. There is a need to enhance enablers and address barriers identified at both health provider and facility levels to improve the audit process.