Objectives: To validate the WHO maternal near-miss criteria and develop a benchmark tool for severe maternal morbidity assessments.
Methods: In a multicenter cross-sectional study implemented in 27 referral maternity hospitals in Brazil, a one-year prospective surveillance on severe maternal morbidity and data collection was carried out. Diagnostic accuracy tests were used to assess the validity of the WHO maternal near-miss criteria. Binary logistic regression was used to model the death probability among women with severe maternal complications and benchmark the management of severe maternal morbidity.
Results: Of the 82,388 women having deliveries in the participating health facilities, 9,555 women presented pregnancy-related complications, including 140 maternal deaths and 770 maternal near misses. The WHO maternal near-miss criteria were found to be accurate and highly associated with maternal deaths (Positive likelihood ratio 106.8 (95% CI 99.56-114.6)). The maternal severity index (MSI) model was developed and found to able to describe the relationship between life-threatening conditions and mortality (Area under the ROC curve: 0.951 (95% CI 0.909-0.993)).
Conclusion: The identification of maternal near-miss cases using the WHO list of pregnancy-related life-threatening conditions was validated. The MSI model can be used as a tool for benchmarking the performance of health services managing women with severe maternal complications and provide case-mix adjustment.