A Machine Learning Approach to Predict Endometriosis

Abstract

Abstarct: Endometriosis is a chronic gynecological condition affecting 5-10% of women of reproductive age. Patients with endometriosis often face delays in diagnosis due to non-specific nature of symptoms. This study proposes a predictive model leveraging machine learning to detect endometriosis patients using available electronic health record data from gynecology clinic visits at University of Cincinnati Medical Center. Preliminary results indicated that among seven machine learning algorithms, naïve bayes performed the best with an AUC of 0.79. Presented at: AMIA Annual Symposium 2025Location: Atlanta, Georgia, USADates: November 15 to 19, 2025 Authors: Parand Shams, MS; Mayur Sarangdhar, PhD; Judith W. Dexheimer, PhD Affiliations: Department of Biostatistics, Health Informatics and Data Science, University of CincinnatiDivision of Biomedical Informatics, Cincinnati Children’s Hospital Medical CenterDepartment of Pediatrics, College of Medicine, University of CincinnatiDivision of Oncology, Cincinnati Children’s Hospital Medical Center Original abstract published in the AMIA proceedings: https://amia.secure-platform.com/symposium/gallery/rounds/82021/details/19365

Ähnliche Arbeiten