Mp02-19 Automated Abstraction Algorithm for Radical Prostatectomy Surgical Outcomes

You have accessJournal of UrologyHealth Services Research: Quality Improvement & Patient Safety I (MP02)1 May 2024MP02-19 AUTOMATED ABSTRACTION ALGORITHM FOR RADICAL PROSTATECTOMY SURGICAL OUTCOMES Maximilian J. Rabil, Michael Jalfon, Dylan Heckscher, Victoria Kong, Aleksandra Golos, Rhys Richmond, Adam Chess, Michael S. Leapman, and Jaime A. Cavallo Maximilian J. RabilMaximilian J. Rabil , Michael JalfonMichael Jalfon , Dylan HeckscherDylan Heckscher , Victoria KongVictoria Kong , Aleksandra GolosAleksandra Golos , Rhys RichmondRhys Richmond , Adam ChessAdam Chess , Michael S. LeapmanMichael S. Leapman , and Jaime A. CavalloJaime A. Cavallo View All Author Informationhttps://doi.org/10.1097/01.JU.0001008600.97797.3b.19AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: The monetary and work-hour cost of quality metric reporting is significant, with the annual cost of data reporting to external organizations estimated to exceed $5 million and requiring over 108,000 work hours for a single academic hospital.1 We hypothesized that a novel institutional electronic medical record (EMR)-based automated algorithm for surgical outcomes following robot-assisted laparoscopic radical prostatectomy (RALP) would demonstrate>90% sensitivity and specificity and significant inter-rater reliability (IRR) with National Surgical Quality Improvement Program (NSQIP) data abstraction. METHODS: We developed an algorithm to automatically abstract RALP surgical outcomes from the EPICTM EMR and report them via a filterable dashboard. All RALPs performed between January 2013-May 2023 by urologists at our hospital system were included. Pathology results were abstracted through exact-text recognition from pathology reports, while NSQIP-defined surgical outcomes were abstracted using ICD-10 codes, CPT codes, and other EMR intermediate variables. Sensitivity, specificity, and IRR as measured by Cohen's weighted kappa statistic between the algorithm and the NSQIP-trained data abstraction team were assessed. RESULTS: 927 cases were mutually tracked. Sensitivity of the algorithm was>90% for all outcomes except rectal injury (0/3, 0%); specificity was>97% for all outcomes. IRR was highest for mortality (k=1.00, CI: 1.00-1.00) and lowest for both ureteral obstruction and dialysis (k=0.00, CI: 0.00-0.00). IRR was fair for renal insufficiency (k=0.32, CI: 0.19-0.44), sepsis (k=0.28, CI: 0.06-0.50), and prolonged NGT/NPO (k=0.39, CI: 0.11-0.66); moderate for UTI (k=0.50, CI: 0.34-0.65) and stage (k=0.53, CI: 0.30-0.76); and substantial for: urine leak (k=0.60, CI: 0.43-0.76), surgical margins (k=0.94, CI: 0.86-1.02), pneumonia (k=0.80, CI: 0.42-1.19), and C. difficile infection (k=0.67, CI: 0.05-1.28). CONCLUSIONS: We developed a novel EMR-based algorithm that matched or exceeded the accuracy of NSQIP abstraction in the identification of post-prostatectomy outcomes. High sensitivity and specificity, and substantial agreement between the algorithm and NSQIP indicates that automated abstraction is a tenable replacement for trained manual abstraction. Broader application of this algorithm may facilitate standardization and cost reduction for real-time national outcome and quality metric benchmarking. 1 Saraswathula, A, et al. JAMA 2023; 329(21): 1840-1847. Source of Funding: None © 2024 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 211Issue 5SMay 2024Page: e20 Advertisement Copyright & Permissions© 2024 by American Urological Association Education and Research, Inc.Metrics Author Information Maximilian J. Rabil More articles by this author Michael Jalfon More articles by this author Dylan Heckscher More articles by this author Victoria Kong More articles by this author Aleksandra Golos More articles by this author Rhys Richmond More articles by this author Adam Chess More articles by this author Michael S. Leapman More articles by this author Jaime A. Cavallo More articles by this author Expand All Advertisement PDF downloadLoading ...