Associated factors and machine learning-based prediction model construction for positive reproductive tract group B streptococcus screening in late-pregnancy women
- 1Department of Obstetrics 5, The Fourth Hospital of Shijiazhuang, Tangu Campus, Shijiazhuang, China
- 2Department of Obsterics, The Fourth Hospital of Shijiazhuang, No.16, Tangu Beida Street, Chang’an District, Shijiazhuang, 050000 China
Abstract
Background: Group B Streptococcus (GBS) is a key pathogen in perinatal health. This study aimed to develop and internally validate a risk prediction model for positive reproductive tract GBS screening in late-pregnancy women using routinely collected clinical data.
Methods: This single-center retrospective study included 363 late-pregnancy women who underwent combined vaginal-rectal GBS screening from January 2023 to May 2024. Data were split 7:3 into training (n = 254) and validation (n = 109) sets using stratified sampling. Predictive variables were selected using least absolute shrinkage and selection operator (LASSO) regression combined with the Boruta algorithm. Seven machine learning models were developed in the training set. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, specificity, F1 score, Brier score, calibration curves, and decision curve analysis.
Results: The GBS positivity rate was 25.1% (91/363). Feature selection identified seven predictors: education level, platelet count, lymphocyte count, vaginitis type, abnormal vaginal discharge, gestational thyroid disease, and gestational weight gain. The random forest model showed the best performance in the validation set, achieving an AUC of 0.868 (95% CI: 0.811–0.922) and a Brier score of 0.153.
Conclusions: The random forest model, built from routinely available clinical data, demonstrated relatively favorable discrimination and calibration in internal validation for predicting positive GBS screening in late pregnancy. It may offer potential for risk stratification and screening adherence management, but it should not replace standard microbiological testing or clinical judgment. Multicenter prospective validation is required before clinical application.
Supplementary Information: The online version contains supplementary material available at https://doi.org/10.1186/s12884-026-09661-6.
Related
- Association of clinical and CT features with the prognosis of lung cancer associated with cystic airspaces and construction of a risk prediction model
- Navigating result disclosure in the context of reproductive genetic carrier screening: an interview study of genetic healthcare professionals in Belgium
- Construction and validation of a frailty screening model for older adults with cataract: a cross-sectional study
- Prediction of chemical miscarriage in assisted reproductive technology pregnancies using machine learning: A developmental study
- Clinical translation of liquid biopsy in early cancer screening: biofluid selection, model construction, and clinical validation
BibTeX
@article{liu2026associated,
title = {{Associated factors and machine learning-based prediction model construction for positive reproductive tract group B streptococcus screening in late-pregnancy women}},
author = {Fangfang Liu and Lixia Wu and Fang Li and Qianqian He and Xiaofen Duan and Jing Gao},
journal = {BMC Pregnancy and Childbirth},
year = {2026},
month = jul,
doi = {10.1186/s12884-026-09661-6},
url = {https://doi.org/10.1186/s12884-026-09661-6},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - JOUR
TI - Associated factors and machine learning-based prediction model construction for positive reproductive tract group B streptococcus screening in late-pregnancy women
AU - Fangfang Liu
AU - Lixia Wu
AU - Fang Li
AU - Qianqian He
AU - Xiaofen Duan
AU - Jing Gao
PY - 2026
DA - 2026/07/30
JO - BMC Pregnancy and Childbirth
DO - 10.1186/s12884-026-09661-6
UR - https://doi.org/10.1186/s12884-026-09661-6
AB - Background: Group B Streptococcus (GBS) is a key pathogen in perinatal health. This study aimed to develop and internally validate a risk prediction model for positive reproductive tract GBS screening in late-pregnancy women using routinely collected clinical data. Methods: This single-center retrospective study included 363 late-pregnancy women who underwent combined vaginal-rectal GBS screening from January 2023 to May 2024. Data were split 7:3 into training (n = 254) and validation (n = 109) sets using stratified sampling. Predictive variables were selected using least absolute shrinkage and selection operator (LASSO) regression combined with the Boruta algorithm. Seven machine learning models were developed in the training set. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, specificity, F1 score, Brier score, calibration curves, and decision curve analysis. Results: The GBS positivity rate was 25.1% (91/363). Feature selection identified seven predictors: education level, platelet count, lymphocyte count, vaginitis type, abnormal vaginal discharge, gestational thyroid disease, and gestational weight gain. The random forest model showed the best performance in the validation set, achieving an AUC of 0.868 (95% CI: 0.811–0.922) and a Brier score of 0.153. Conclusions: The random forest model, built from routinely available clinical data, demonstrated relatively favorable discrimination and calibration in internal validation for predicting positive GBS screening in late pregnancy. It may offer potential for risk stratification and screening adherence management, but it should not replace standard microbiological testing or clinical judgment. Multicenter prospective validation is required before clinical application. Supplementary Information: The online version contains supplementary material available at https://doi.org/10.1186/s12884-026-09661-6.
SN - 1471-2393
ER -