EC Paediatrics

Objective: The objective of this study was to develop and evaluate various machine learning algorithms for predicting preterm birth, aiming to overcome the limitations of conventional approaches in terms of reliability and predictive power.

Study Design: This study utilized health records from 5900 patients to train and assess different machine learning algorithms. The focus was on identifying patterns and predicting outcomes related to preterm birth, with an emphasis on the interpretability and generalizability of the models.

Result: Among the various algorithms evaluated, CatBoost emerged as the most effective, achieving an accuracy of 84%. This performance surpassed that of predictions based on manual reviews. Key factors contributing to the model's predictive ability were also identified, highlighting its potential in clinical applications.

Conclusion: The findings demonstrate the efficacy of machine learning, particularly the CatBoost algorithm, in predicting preterm birth. This suggests a significant advancement over traditional methods, offering a more reliable tool for healthcare professionals in anticipating and managing preterm birth complications.

 Keywords: Premature Birth; Machine Learning; Infant, Newborn; Pregnancy

1. “152 Million Babies Born Preterm in the Last Decade”. n.d.
2. Saigal S and Doyle LW. “An overview of mortality and sequelae of preterm birth from infancy to adulthood”. Lancet9608 (2008): 261-269.
3. Raju TNK., et al. “Long-term healthcare outcomes of preterm birth: An executive summary of a conference sponsored by the national institutes of health”. Journal of Pediatrics 181 (2017): 309-318.e1.
4. Crump C., et al. “Preterm birth and risk of type 1 and type 2 diabetes: a national cohort study”. Diabetologia 3 (2020): 508-518.
5. Frey HA and Klebanoff MA. “The epidemiology, etiology, and costs of preterm birth”. Seminars in Fetal and Neonatal Medicine2 (2016): 68-73.
6. Purisch SE and Gyamfi-Bannerman C. “Epidemiology of preterm birth”. Seminars in Perinatology 7 (2017): 387-391.
7. Oskovi Kaplan ZA and Ozgu-Erdinc AS. “Prediction of preterm birth: Maternal characteristics, ultrasound markers, and biomarkers: An updated overview”. Journal of Pregnancy (2018): 8367571.
8. Zierden HC., et al. “Next generation strategies for preventing preterm birth”. Advanced Drug Delivery Reviews 174 (2021): 190-209.
9. Auger N., et al. “Association between maternal comorbidity and preterm birth by severity and clinical subtype: retrospective cohort study”. BMC Pregnancy Childbirth 11 (2011): 67.
10. Koullali B., et al. “Risk assessment and management to prevent preterm birth”. Seminars in Fetal and Neonatal Medicine 2 (2016): 80-88.
11. Sharma V., et al. “Clinical risk prediction models: the canary in the coalmine for artificial intelligence in healthcare?” BMJ Health and Care Informatics 1 (2021): e100421.
12. Sharifi-Heris Z., et al. “Machine learning approach for preterm birth prediction using health records: Systematic review”. JMIR Medical Informatics 4 (2022): e33875.
13. Sun H., et al. “Machine learning-based prediction models for different clinical risks in different hospitals: Evaluation of live performance”. Journal of Medical Internet Research 6 (2022): e34295.
14. Huang J., et al. “Analysis of factors related to preterm birth: a retrospective study at Nanjing Maternity and Child Health Care Hospital in China”. Medicine (Baltimore)28 (2020): e21172.
15. Nieto-Del-Amor F., et al. “Combination of feature selection and resampling methods to predict preterm birth based on electrohysterographic signals from imbalance data”. Sensors (Basel)14 (2022): 5098.
16. Goldenberg RL., et al. “Biochemical markers for the prediction of preterm birth”. American Journal of Obstetrics and Gynecology 5 (2005): S36-S46.
17. Song Y-Y and Lu Y. “Decision tree methods: applications for classification and prediction”. Shanghai Archives of Psychiatry 2 (2015): 130-135.
18. Gonçalves DM., et al. “Predicting postoperative complications in cancer patients: A survey bridging classical and machine learning contributions to postsurgical risk analysis”. Cancers (Basel)13 (2021): 3217.
19. Chen T and Guestrin C. “XGBoost”. In: Proceedings of the 22^nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining ACM: New York, NY, USA (2016).
20. Breiman L. “Random Forests”. Machine Learning 45 (2001): 5-32.
21. Friedman JH. “Greedy function approximation: A gradient boosting machine”. Annals of Statistics 5 (2001): 1189-1132.
22. Freund Y and Schapire RE. “A decision-theoretic generalization of on-line learning and an application to boosting”. Journal of Computer and System Sciences 1 (1997): 119-139.
23. Dorogush AV., et al. “CatBoost: Gradient boosting with categorical features support”. arXiv [csLG] (2018).
24. Deo RC. “Machine learning in medicine”. Circulation 20 (2015): 1920-1930.
25. AlSaad R., et al. “PredictPTB: an interpretable preterm birth prediction model using attention-based recurrent neural networks”. BioData Mining 1 (2022): 6.
26. Zhang Y., et al. “Establishment of a model for predicting preterm birth based on the machine learning algorithm”. BMC Pregnancy Childbirth1 (2023): 779.
27. de Carvalho AAV., et al. “Association of midtrimester short femur and short humerus with fetal growth restriction”. Prenatal Diagnosis 2 (2013): 130-133.
28. Partap U., et al. “Fetal growth and the risk of spontaneous preterm birth in a prospective cohort study of nulliparous women”. American Journal of Epidemiology 2 (2016): 110-119.
29. Han Z., et al. “Maternal underweight and the risk of preterm birth and low birth weight: a systematic review and meta-analyses”. International Journal of Epidemiology 1 (2011): 65-71.
30. Liu K., et al. “Association of maternal obesity with preterm birth phenotype and mediation effects of gestational diabetes mellitus and preeclampsia: a prospective cohort study”. BMC Pregnancy Childbirth1 (2022).
31. van Zijl MD., et al. “The predictive capacity of uterine artery Doppler for preterm birth—A cohort study”. Acta Obstetricia et Gynecologica Scandinavica 4 (2020): 494-502.