Automated algorithm for surgical outcomes abstraction for cystectomy

Cystectomy is a potentially morbid procedure that presents a prime opportunity for quality improvement. We hypothesized that a novel electronic medical record (EMR)-based automated algorithm would demonstrate >90% sensitivity and specificity with significant agreement with institutional National Surgical Quality Improvement Program (NSQIP) data abstraction.

We developed an EMR-based algorithm that automatically abstracts surgical outcomes and quality metrics for cystectomy. All cystectomies performed between January 2013 and March 2025 by urologists in our health system were included. Outcomes were abstracted using CPT/ICD-10 codes and EMR-based variables. Sensitivity and specificity were compared with NSQIP abstraction. Agreement between the algorithm and NSQIP abstractors was analyzed with Cohen's kappa.

Two hundred forty-six cases were mutually tracked. Sensitivity of the algorithm was ≥90%, and specificity was ≥96% for all outcomes with at least one observed event. Kappa could not be assessed for anastomotic bowel leak and rectal injury, as no positive events were identified in the matched cohort. Among assessable outcomes, kappa ranged from 0.14 (renal insufficiency) to 1.00 (mortality and stroke), with substantial agreement for readmission (k = 0.89), prolonged ventilation (k = 0.91), cardiac complications (k = 0.85), and C. difficile infection (k = 0.78).

This EMR-based algorithm for cystectomy surgical outcomes and quality metrics matches or exceeds the sensitivity and specificity of NSQIP abstraction. Automated abstraction enables immediate access to quality improvement data, potentially reducing resource utilization. Quality improvement programs may benefit from integration of EMR-based automated algorithms for cystectomy outcomes reporting; however, external validation across institutions and EMR systems is required before broader implementation.