Predictors and a scoring model for maternal near-miss and maternal death in Southern Thailand: a case-control study

To identify predictors and develop a scoring model to predict maternal near-miss (MNM) and maternal mortality.

A case–control study of 1,420 women delivered between 2014 and 2020 was conducted. Cases were women with MNM or maternal death, controls were women who had uneventful deliveries directly after women in the cases group. Antenatal characteristics and complications were reviewed. Multivariate logistic regression and Akaike information criterion were used to identify predictors and develop a risk score for MNM and maternal mortality.

Predictors for MNM and maternal mortality (aOR and score for predictive model) were advanced age (aOR 1.73, 95% CI 1.25–2.39, 1), obesity (aOR 2.03, 95% CI 1.22–3.39, 1), parity ≥ 3 (aOR 1.75, 95% CI 1.27–2.41, 1), history of uterine curettage (aOR 5.13, 95% CI 2.47–10.66, 3), history of postpartum hemorrhage (PPH) (aOR 13.55, 95% CI 1.40–130.99, 5), anemia (aOR 5.53, 95% CI 3.65–8.38, 3), pregestational diabetes (aOR 5.29, 95% CI 1.27–21.99, 3), heart disease (aOR 13.40, 95%CI 4.42–40.61, 5), multiple pregnancy (aOR 5.57, 95% CI 2.00–15.50, 3), placenta previa and/or placenta-accreta spectrum (aOR 48.19, 95% CI 22.75–102.09, 8), gestational hypertension/preeclampsia without severe features (aOR 5.95, 95% CI 2.64–13.45, 4), and with severe features (aOR 16.64, 95% CI 9.17–30.19, 6), preterm delivery <37 weeks (aOR 1.65, 95%CI 1.06–2.58, 1) and < 34 weeks (aOR 2.71, 95% CI 1.59–4.62, 2). A cut-off score of ≥4 gave the highest chance of correctly classified women into high risk group with 74.4% sensitivity and 90.4% specificity.

We identified predictors and proposed a scoring model to predict MNM and maternal mortality with acceptable predictive performance.