Comparison of Multivariable Logistic Regression and Machine Learning Models for Predicting Bronchopulmonary Dysplasia or Death in Very Preterm Infants

Abstract

Bronchopulmonary dysplasia (BPD) is the most prevalent and clinically significant complication of prematurity. Accurate identification of at-risk infants would enable ongoing intervention to improve outcomes. Although postnatal exposures are known to affect an infant's likelihood of developing BPD, most existing BPD prediction models do not allow risk to be evaluated at different time points, and/or are not suitable for use in ethno-diverse populations. A comprehensive approach to developing clinical prediction models avoids assumptions as to which method will yield the optimal results by testing multiple algorithms/models. We compared the performance of machine learning and logistic regression models in predicting BPD/death. Our main cohort included infants <33 weeks' gestational age (GA) admitted to a Canadian Neonatal Network site from 2016 to 2018 (n = 9,006) with all analyses repeated for the <29 weeks' GA subcohort (n = 4,246). Models were developed to predict, on days 1, 7, and 14 of admission to neonatal intensive care, the composite outcome of BPD/death prior to discharge. Ten-fold cross-validation and a 20% hold-out sample were used to measure area under the curve (AUC). Calibration intercepts and slopes were estimated by regressing the outcome on the log-odds of the predicted probabilities. The model AUCs ranged from 0.811 to 0.886. Model discrimination was lower in the <29 weeks' GA subcohort (AUCs 0.699–0.790). Several machine learning models had a suboptimal calibration intercept and/or slope (k-nearest neighbor, random forest, artificial neural network, stacking neural network ensemble). The top-performing algorithms will be used to develop multinomial models and an online risk estimator for predicting BPD severity and death that does not require information on ethnicity.

The past, current, and future of neonatal intensive care units with artificial intelligence: a systematic review

“… - Small sample size - Retrospective - Gestational age, postnatal age, sepsis and culture Reed et al, 2021 143 Comparison least absolute shrinkage and selection operator (LASSO) and random forest (RF) to expert-opinion driven logistic regression modeling Prediction of 30-day unplanned rehospitalization of preterm babies 5567 live-born babies and 3841 were included to the study Data derived exclusively from The population-based prospective cohort study of French preterm babies, EPIPAGE 2. The logistic regression model comprised 10 predictors, selected by expert clinicians, while the LASSO and random forest included 75 predictors 65% (AUC) RF 59% (AUC) LASSO 57% (AUC) LR + The first comparison of different modeling methods for predicting early rehospitalization + Large cohort with data variation - No accurate evaluation of rehospitalization causes - Data collection after discharge based on survey filled by mothers - 9% of babies were rehospitalized Khursid et al, 2021 70 K-nearest neighbor, random forest, artificial neural network, stacking neural network ensemble To predict, on days 1, 7, and 14 of admission to neonatal intensive care, the composite outcome of BPD/death prior to discharge. <33 weeks GA cohort ( n = 9006) And < 29 weeks GA were included For each set of models (Days 1, 7, 14), stratified random sampling.…”

“…Despite the advancement in neonatal care, it is crucial that preterm infants remain highly susceptible to mortality due to immaturity of organ systems and increased susceptibility to early and late sepsis 69 . Addressing these permanent risks necessitates the utilization of ML to predict mortality 63 – 66 , 68 , 70 . Early studies employed ANN and fuzzy linguistic models and achieved an AUC of 85–95% and accuracy of 90% 62 , 68 .…”

“…While various biomarkers have been linked to the development of respiratory distress syndrome (RDS), no clinically relevant prognostic tests are available for BPD at birth 125 . There are ML studies aiming to predict BPD from birth 70 , 126 , gastric aspirate content 125 and genetic data 127 and it has been shown that BPD could be predicted with an accuracy of up to 86% in the best-case scenario 70 (Table 5 ), analysis of responsible genes with ML could predict BPD development with an AUC of 90% 127 (Table 3 ) and combination of gastric aspirate after birth and clinical information analysis with SVM predicted BPD development with a sensitivity of 88% 125 (Table 5 ).…”

Construction and Application of Early Stratification Dynamic Prediction Model for Bronchopulmonary Dysplasia in Extremely Premature / Very Low Birth Weight Infants

“…Studies employing machine learning algorithms for the prediction and management of BPD have shown substantial advancements [40] . Nevertheless, there persists contention regarding the performance of machine learning models in predicting BPD, where they haven't conclusively outperformed traditional LR models [11,41,42] . Our research, by constructing LR, RF, GBDT, and XGB models at various timepoints within the rst postnatal week, identi ed that the LR and XGBoost models in particular perform well on days 1, 3, and 7 for early strati ed prediction of infants at high risk for BPD.…”