Early Preeclampsia Detection Using XGBoost-Cox Proportional Hazard Model

Various prognostic models based on survival analysis methods have beenproposed to predict the risk of preeclampsia (PE). To develop a more accurate yetinterpretable prediction approach, we utilized clinical data from pregnant womencollected at a hospital in Jakarta and applied the XGBoost-Cox ProportionalHazard Model (XGB-Cox). This model integrates the predictive power of theXGBoost machine learning algorithm with the Cox Proportional Hazard (Cox-PH)model, which estimates the effect of covariates on event time. Our results showthat the XGB-Cox model outperforms the traditional Cox-PH model based on fourevaluation metrics: log-likelihood, log-rank test, concordance index (C-index), andBrier score. The XGB-Cox model achieved a higher C-index of 0.8908 comparedto 0.7548 for Cox-PH, indicating improved risk discrimination.Kaplan-Meier curves suggest that XGB-Cox provides better separation across risk quartiles.While XGB-Cox generally yields lower Brier Scores, its performance declines atlater gestational weeks. The Cox-PH model remains superior in interpretability,offering clear hazard ratios, while XGB-Cox enhances model fitness and stillprovides meaningful insights into feature importance. Additionally, sensitivity analysis underscores the need to carefully determine the proportion of censored data, as excessive censoring affects model stability. These findings suggest that XGB-Cox provides a robust predictive framework for early PE risk assessment, supporting its potential application in clinical decision-making for maternalhealthcare.