Analyzing and Predicting Peritonitis in Nephrological Care: A Decision-Making Tool for Doctors at Sahloul Hospital

Abstract

The study aims to identify factors contributing to peritonitis in peritoneal dialysis patients and develop Machine Learning (ML) models to predict peritonitis precocity, marking the first application of such models in developing Countries. This retrospective cohort study involved 89 patients enrolled in peritoneal dialysis between 2004 and 2023. Episodes of peritonitis were recorded at the time of onset, and peritonitis was parameterized as a time-dependent variable for analysis. We utilized various Machine Learning models, including Decision Tree (DT), K-Nearest Neighbors (KNN), Random Forest (RF), and Support Vector Machine (SVM), to analyze the data and assess the likelihood of new patients developing peritonitis based on their characteristics. The models demonstrated varying levels of accuracy, with the Decision Tree model achieving the highest predictive performance. Machine learning models can effectively predict peritonitis in peritoneal dialysis patients, potentially improving patient outcomes through early intervention for the first time in Tunisia.