OBJECTIVE: To validate the accuracy of four early warning scores for early identification of women at risk.
METHODS: This was a retrospective study of pregnant women admitted in obstetrics Critical Care Unit (ICU). Capacity of the Modified Obstetric Early Warning Score (MOEWS), ICNARC Obstetric Early Warning Score (OEWS), Maternal Early Obstetric Warning System (MEOWS chart), and Maternal Early Warning Trigger (MEWT) were compared in predicting severe maternal morbidity. Area under receiver operator characteristic (AUROC) curve was used to evaluate the predictive performance of scoring system.
RESULTS: A total of 352 pregnant women were enrolled and 290 were identified with severe maternal morbidity. MOEWS was more sensitive than MEOWS chart, ICNARC OEWS and MEWT (96.9 % vs. 83.4 %, 66.6 % and 44.8 %). MEWT had the highest specificity (98.4 %), followed by MOEWS (83.9 %), ICNARC OEWS (75.8 %) and MEOWS chart (48.4 %). AUROC of MOEWS, ICNARC OEWS, MEOWS chart, and MEWT for prediction of maternal mortality were 0.91 (95 % CI: 0.874-0.945), 0.765(95 % CI: 0.71-0.82), 0.657(95 % CI: 0.577-0.738), and 0.716 (95 % CI, 0.659-0.773) respectively. MOEWS had the highest AUCs in the discrimination of serious complications in hypertensive disorders, cardiovascular disease, obstetric hemorrhage and infection. For individual vital signs, maximum diastolic blood pressure (DBP), maximum systolic blood pressure (SBP), maximum respiratory rate (RR) and peripheral oxygen saturation (SPO2) demonstrated greater predictive ability.
CONCLUSIONS: MOEWS is more accurate than ICNARC OEWS, MEOWS chart, and MEWT in predicting the deterioration of women. The prediction ability of DBP, SBP, RR and SPO2 are more reliable.

BACKGROUND: Maternal mortality is still a major challenge for health systems, while severe maternal complications are the primary causes of maternal death. Our study aimed to determine whether severe maternal morbidity is effectively predicted by a newly proposed Modified Obstetric Early Warning Score (MOEWS) in the setting of an obstetric intensive care unit (ICU).
METHODS: A retrospective study of pregnant women admitted in the ICU from August 2019 to August 2020 was conducted. MOEWS was calculated 24 h before and 24 h after admission in the ICU, and the highest score was taken as the final value. For women directly admitted from the emergency department, the worst value before admission was collected. The aggregate performance of MOEWS in predicting critical illness in pregnant women was evaluated and finally compared with that of the Acute Physiology and Chronic Health Evaluation II (APACHE II) score.
RESULTS: A total of 352 pregnant women were enrolled; 290 women (82.4%) with severe maternal morbidity were identified and two of them died (0.6%). The MOEWSs of women with serious obstetric complications were significantly higher than those of women without serious obstetric complications [8(6, 10) vs. 4(2, 4.25), z = -10.347, P < 0.001]. MOEWSs of 24 h after ICU admission had higher sensitivity, specificity and AUROC than MOEWSs of 24 h before ICU admission. When combining the two MOEWSs, sensitivity of MOEWS was 99.3% (95% CI: 98-100), specificity 75.8% (95% CI: 63-86), positive predictive value (PPV) 95.1% (95% CI: 92-97) and negative predictive value (NPV) 95.9% (95% CI: 86-100). The areas under the receiver operator characteristic (ROC) curves of MOEWS were 0.92 (95% CI: 0.88-0.96) and 0.70 (95% CI: 0.63-0.76) of the APACHE II score.
CONCLUSIONS: The newly proposed MOEWS has an excellent ability to identify critically ill women early and is more effective than APACHE II. It will be a valuable tool for discriminating severe maternal morbidity and ultimately improve maternal health.