Abstract

Background: The predictive models used for general population have been used to predict outcome of obstetric patients as well. However, obstetric patients have unique physiology and hence these models do not perform well. Recently, attempts have been made to devise models specific for obstetric population.Objectives: To evaluate the accuracy of 2 scores (Carle’s OEWS and obstetrically modified SOFA score) in predicting mortality and morbidity in obstetric patients admitted in CCU.Methods: 500 patients admitted to obstetric CCU, were included in the study. The patients were managed as per the standard departmental protocols of the given condition. Both scores were calculated. The patients were followed up till discharge/death.Results: Both models had significant discriminatory power to predict mortality as well as morbidity. Obstetric SOFA score was the better predictor of mortality at cut-off point of >6 with area under curve of 0.966 for correctly predicting mortality. Standardized maternal mortality for Carle’s OEWS score and obstetric SOFA score was 54.74% and 49.52%, respectively. Hence, it can be seen that both scores tend to over-predict the maternal mortality. Between the two models, Obstetric SOFA score was the better predictor of morbidity at cut off point of >3 with area under curve of 0.811 for correctly predicting morbidity.Conclusion: Both the obstetric specific scores tend to over-predict the maternal mortality.