Applying the National Surgical Quality Improvement Program risk calculator to patients undergoing colorectal surgery: theory vs reality

BACKGROUND: Discussing potential morbidity and mortality is essential to informed decision making and consent. The American College of Surgery National Surgical Quality Improvement Program developed an online risk calculator (RC) using patient-specific information to determine operative risk. STUDY DESIGN: Colorectal procedures at our independent academic medical center from 2010 to 2011 were evaluated. The RC's predicted outcomes were compared with observed outcomes. Statistical analysis included Brier score, Wilcoxon sign rank test, and standardized event ratio. RESULTS: There were 324 patients included. The RC's Brier score was .24 (.015-.219) for predicting mortality and morbidity, respectively. The observed event rate for surgical site infection and any complication was higher than the RC predicted (standardized event ratio 1.9 CI [1.49 to 2.39] and 1.39 CI [1.14 to 1.68], respectively). The observed length of stay was longer than predicted (5.6 vs 6.6 days, P < .001). CONCLUSIONS: The RC underestimated the surgical site infection and overall complication rates. The RC is a valuable tool in predicting risk for adverse outcomes; however, institution-specific trends may influence actual risk. Surgeons and institutions must recognize areas where they are outliers from estimated risks and tailor risk discussions accordingly. (C) 2016 Elsevier Inc. All rights reserved.