An Analysis of Mortality Prediction by SOFA and APACHE II in Critically Maternal Patients Using the GiViTi Calibration Belt

Abstract

Objective: to assess the mortality prediction performance of SOFA and APACHE II scores in an obstetric population using the GiViTI calibration belt. Methods: a retrospective cohort study was conducted to evaluate the external validation of APACHE II and SOFA using the GiViTI calibration belt. Final analysis performed using SPSS, including explanation of absolute numbers, percentages, median and interquartile ranges, as well as ROC curves of the same prognostic scores. Results: the observed mortality rate was 3.57%. Both APACHE II and SOFA showed good discrimination for maternal death, with SOFA performing better. Conclusion: regarding the calibration belt, there was no significant deviation from the bisector; however, both APACHE II and SOFA showed less-than-perfect calibration belts with wide confidence intervals when expected mortality was >20%.