Prediction of preterm birth in nulliparous women using logistic regression and machine learning

Abstract

Objective To predict preterm birth in nulliparous women using logistic regression and machine learning. Design Population-based retrospective cohort. Participants Nulliparous women (N = 112,963) with a singleton gestation who gave birth between 20–42 weeks gestation in Ontario hospitals from April 1, 2012 to March 31, 2014. Methods We used data during the first and second trimesters to build logistic regression and machine learning models in a “training” sample to predict overall and spontaneous preterm birth. We assessed model performance using various measures of accuracy including sensitivity, specificity, positive predictive value, negative predictive value, and area under the receiver operating characteristic curve (AUC) in an independent “validation” sample. Results During the first trimester, logistic regression identified 13 variables associated with preterm birth, of which the strongest predictors were diabetes (Type I: adjusted odds ratio (AOR): 4.21; 95% confidence interval (CI): 3.23–5.42; Type II: AOR: 2.68; 95% CI: 2.05–3.46) and abnormal pregnancy-associated plasma protein A concentration (AOR: 2.04; 95% CI: 1.80–2.30). During the first trimester, the maximum AUC was 60% (95% CI: 58–62%) with artificial neural networks in the validation sample. During the second trimester, 17 variables were significantly associated with preterm birth, among which complications during pregnancy had the highest AOR (13.03; 95% CI: 12.21–13.90). During the second trimester, the AUC increased to 65% (95% CI: 63–66%) with artificial neural networks in the validation sample. Including complications during the pregnancy yielded an AUC of 80% (95% CI: 79–81%) with artificial neural networks. All models yielded 94–97% negative predictive values for spontaneous PTB during the first and second trimesters. Conclusion Although artificial neural networks provided slightly higher AUC than logistic regression, prediction of preterm birth in the first trimester remained elusive. However, including data from the second trimester improved prediction to a moderate level by both logistic regression and machine learning approaches.

Risk factors for spontaneous preterm birth among healthy nulliparous pregnant women in the Netherlands, a prospective cohort study

“…31,32 Some reported associations were absent in our study, including maternal age, BMI, maternal height, smoking during pregnancy, alcohol use, maternal medical history, ART, previous miscarriages or terminations of pregnancy, short cervical length. 7,10,30,33,34 Unfortunately, our study was underpowered to detect these associations. This is most likely for the associations with small ORs found in other studies (e.g., maternal age and BMI).…”

“…Previous studies among nulliparous women identified several of the associated factors with sPTB found in our cohort, such as socioeconomic deprivation, ethnicity, marital status, vaginal bleeding during early pregnancy. 7,10,[29][30][31] A Danish study showed that vaginal bleeding during the first pregnancy also increased the risk of PTB in a second pregnancy, but this finding was not confirmed in a smaller study executed in the United States of America. 31,32 Some reported associations were absent in our study, including maternal age, BMI, maternal height, smoking during pregnancy, alcohol use, maternal medical history, ART, previous miscarriages or terminations of pregnancy, short cervical length.…”

“…Various prediction models to identify women at risk for sPTB have been published. 7 , 8 , 9 , 10 A systematic review shows that published prediction models for sPTB failed to adequately identify women at risk, using only clinical risk factors such as previous PTB, ethnicity, and maternal age. 11 A subgroup analysis of 1284 nulliparous women revealed poor discriminative performance, with discrimination C‐statistic ranging from 0.51 (95% CI: 0.45–0.57) to 0.55 (95% CI: 0.49–0.60).…”

“…Various prediction models to identify women at risk for sPTB have been published 7–10 . A systematic review shows that published prediction models for sPTB failed to adequately identify women at risk, using only clinical risk factors such as previous PTB, ethnicity, and maternal age 11 .…”

Prediction of low Apgar score at five minutes following labor induction intervention in vaginal deliveries: machine learning approach for imbalanced data at a tertiary hospital in North Tanzania

“…In other words, probability is directly related to the cost function and thus the algorithm produces unbiased probability estimates [ 62 ]. This means Lreg model had returned a well calibrated predictions as it directly optimizes the “Log loss” (also known as cross-entropy loss) [ 63 , 64 ]. In other words, the tested models returned somewhat a biased probabilities compared to that shown by logistic regression.…”