A comprehensive and bias-free machine learning approach for risk prediction of preeclampsia with severe features in a nulliparous study cohort

Abstract

Preeclampsia is one of the leading causes of maternal morbidity, with consequences during and after pregnancy. Because of its diverse clinical presentation, preeclampsia is an adverse pregnancy outcome that is uniquely challenging to predict and manage. In this paper, we developed racial bias-free machine learning models that predict the onset of preeclampsia with severe features or eclampsia at discrete time points in a nulliparous pregnant study cohort. To focus on those most at risk, we selected probands with severe PE (sPE). Those with mild preeclampsia, superimposed preeclampsia, and new onset hypertension were excluded.The prospective study cohort to which we applied machine learning is the Nulliparous Pregnancy Outcomes Study: Monitoring Mothers-to-be (nuMoM2b) study, which contains information from eight clinical sites across the US. Maternal serum samples were collected for 1,857 individuals between the first and second trimesters. These patients with serum samples collected are selected as the final cohort.Our prediction models achieved an AUROC of 0.72 (95% CI, 0.69-0.76), 0.75 (95% CI, 0.71-0.79), and 0.77 (95% CI, 0.74-0.80), respectively, for the three visits. Our initial models were biased toward non-Hispanic black participants with a high predictive equality ratio of 1.31. We corrected this bias and reduced this ratio to 1.14. This lowers the rate of false positives in our predictive model for the non-Hispanic black participants. The exact cause of the bias is still under investigation, but previous studies have recognized PLGF as a potential bias-inducing factor. However, since our model includes various factors that exhibit a positive correlation with PLGF, such as blood pressure measurements and BMI, we have employed an algorithmic approach to disentangle this bias from the model.The top features of our built model stress the importance of using several tests, particularly for biomarkers (BMI and blood pressure measurements) and ultrasound measurements. Placental analytes (PLGF and Endoglin) were strong predictors for screening for the early onset of preeclampsia with severe features in the first two trimesters.

Keywords: Ensemble model; Fairness in machine learning; Machine learning; PlGF; Preeclampsia; Preeclampsia with severe features.

Fig. 1

Data process timeline, This figure shows the gestational weeks at each visit. For each visit, the number of features at that visit is listed and the category of new feature included is also shown

Fig. 2

Final study cohort selection process. Out of the participants from the placental analytes sub-study, we excluded participants with conditions such as new onset hypertension, mild preeclampsia, and missing label for preeclampsia to focus on the participants that are most at risk

Fig. 3

The training process of PEPrML pipeline. Samples were balanced for train and test sets. fivefold grid search cross-validation was used to select the hyperparameters for each trial. We repeated 100 trials and recorded the results

Fig. 4

sPE + E vs NPH and early sPE vs late sPE + E model performance. a Average AUC for 100 trials per visit for 4 classifiers. b RF classifier has best performance across visits for both comparisons. The ROC curve demonstrated the tradeoff between the true positive rate versus false positive rate. This summarizes the results for 100 trials

Fig. 5

Interpreting machine learning model for sPE + E vs NPH and early sPE vs late sPE + E. a V2 model features importance for sPE + E vs NPH model, b V2 model feature importance for early sPE vs late sPE + E, c—d PDP for BMI and PlGF based on model build for sPE + E vs NPH. V2 model includes both features from V1 and V2

Fig. 6

Fairness check for sPE + E vs NPH mode. The threshold set based on the four-fifth rule is 0.8 and 1.25. A Ceribus Paribus plot was used to adjust the prediction threshold for the Black population