MP54-20 APPLICATION OF ARTIFICIAL INTELLIGENCE TO IMPROVE PATIENT SELECTION FOR THE KAISER PERMANENTE HEALTHY STONE POPULATION MANAGEMENT PROGRAM

Abstract

You have accessJournal of UrologyStone Disease: Epidemiology & Evaluation II (MP54)1 Sep 2021MP54-20 APPLICATION OF ARTIFICIAL INTELLIGENCE TO IMPROVE PATIENT SELECTION FOR THE KAISER PERMANENTE HEALTHY STONE POPULATION MANAGEMENT PROGRAM Reza Goharderakhshan, Nikhil Crain, Drew Clausen, Dennis Walsh, Gary Chien, and Ronald Loo Reza GoharderakhshanReza Goharderakhshan More articles by this author , Nikhil CrainNikhil Crain More articles by this author , Drew ClausenDrew Clausen More articles by this author , Dennis WalshDennis Walsh More articles by this author , Gary ChienGary Chien More articles by this author , and Ronald LooRonald Loo More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000002084.20AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: The Healthy Stone (HS) program follows the AUA kidney stone medical management guidelines with the goals of reducing both stone recurrence and additional treatments in high risk patients. We previously presented on the predictive accuracy of a machine learning algorithm to assess the risk of stone recurrence. This algorithm was used to risk stratify eligible patients diagnosed with a kidney stone who participated in the HS program vs. those who did not. We examined the impact of our Healthy Stone program for each risk category over a three year follow up period. METHODS: The predictive algorithm was retrospectively applied to two groups of patients who remained within the Kaiser Permanente health plan with three year follow up data. The treatment group (N=537) included patients enrolled in the HS program. The control group (N=1984) included patients diagnosed with kidney stones in the same period as the treatment group who did not enroll in the HS program. The kidney stone recurrence risk was computed for all patients. They were then stratified into four risk groups (lowest, low, moderate, high). The volume of kidney stone related surgical procedures and Emergency Department (ED) encounters was examined for both groups. RESULTS: The AI algorithm identified 242 (45%) high risk patients in the treatment group, and 397(20%) high risk patients in the control group. In the high-risk group, there was a reduction in ED visits (36%) (Figure 1) and surgical procedures (44%) (Figure 2) for patients managed by the HS program over those not enrolled in the program. In the lower risk groups, the impact of the HS program on ED visits and surgeries was less significant. CONCLUSIONS: Our study demonstrates a significant reduction in recurrent stone treatment encounters for high-risk patients who were enrolled in the Healthy Stone program vs. those who were not. Using an AI algorithm may augment our clinical decision making for patients who are most likely to benefit from a prevention program and therefore improve resource utilization. Source of Funding: None © 2021 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 206Issue Supplement 3September 2021Page: e957-e957 Advertisement Copyright & Permissions© 2021 by American Urological Association Education and Research, Inc.MetricsAuthor Information Reza Goharderakhshan More articles by this author Nikhil Crain More articles by this author Drew Clausen More articles by this author Dennis Walsh More articles by this author Gary Chien More articles by this author Ronald Loo More articles by this author Expand All Advertisement Loading ...