Application of Machine Learning to Predict Symptomatic Recurrence Events for Patients With Kidney Stone.

Area of Science:

• Nephrology
• Medical Informatics
• Data Science

Background:

• Kidney stone recurrence poses a significant clinical challenge.
• Adherence to American Urological Association (AUA) guidelines can reduce recurrence.
• Predictive tools are needed to identify at-risk individuals.

Purpose of the Study:

• To assess the efficacy of machine learning (ML) in identifying patients prone to symptomatic kidney stone recurrence.
• To develop a predictive model for 1-year symptomatic recurrence risk.

Main Methods:

• Retrospective review of electronic health records (EHRs) from January 2008 to December 2023.
• Supervised machine learning model utilizing 952 features from EHR data.
• Model trained on 123,900 patients and tested on 30,976 patients.

Main Results:

• The ML model achieved an area under the receiver operating characteristic curve (AUC) of 0.727 in predicting 1-year symptomatic recurrence.
• The model demonstrated effectiveness in discriminating between high and low-risk patients.
• Analysis included over 154,000 patients and 1.4 million stone encounters.

Conclusions:

• Machine learning models can effectively predict the risk of symptomatic urinary stone recurrence.
• This predictive capability aids in personalized kidney stone management strategies.
• ML offers a promising avenue for proactive intervention in recurrent stone formers.