ACS NSQIP Risk Calculator Shows Strong Performance in High-Risk Patients

The ACS NSQIP Risk Calculator demonstrated excellent calibration and good discrimination in predicting mortality and morbidity across 21 patient subsets, regardless of risk factors. Observed mortality rates varied significantly, from 0.04% to 60.01%, while morbidity ranged from 3.33% to 59.17%. Both the current machine learning algorithm (xgb) and a potential future algorithm (catb) showed consistent accuracy (ape < 10% and auc > 0.7), proving the tool’s value in surgical decision-making for high-risk populations.