Development of a machine learning-based model for predicting adverse pregnancy outcomes in women with polycystic ovary syndrome (PCOS): a retrospective observational study protocol.

INTRODUCTION: Polycystic ovary syndrome (PCOS) is a prevalent endocrine condition in reproductive-aged women, which is associated with adverse maternal and neonatal outcomes in pregnancy. In the present study, pregnancy complications-like gestational diabetes, preterm birth, hypertension, low birth weight and neonatal intensive care unit admission-will be predicted with high accuracy by machine learning-based prediction models.

METHODS AND ANALYSIS: There are three substudies: (1) Developing a researcher-made questionnaire by literature review, data collection from existing medical records in maternal and neonatal systems of women with PCOS who have given birth, collected from hospitals and private clinics of East Azerbaijan Province and Tehran. Approximately 800-1000 women with PCOS will be included in the study and then data preprocessing will be performed. (2) Developing machine learning-based models for predicting adverse pregnancy outcomes in women with PCOS using decision trees, random forests, Extreme Gradient Boosting, support vector machines, k-nearest neighbours and neural network algorithms and (3) Developing a user-friendly application or interface that will operate on various devices. Based on model evaluation metrics, the model with the highest area under the receiver operating characteristic curve for predicting adverse pregnancy outcomes in women with PCOS will be used as the final model.

ETHICS AND DISSEMINATION: This study was approved by the Ethics Committee of Shahid Beheshti University of Medical Sciences, Tehran (Ethics Code: IR.SBMU.PHARMACY.REC.1404.017). Findings will be disseminated through peer-reviewed journals, conferences and social media and will also be shared with study participants.