Journal of the American College of Surgeons
-
The American College of Surgeons NSQIP risk calculator (RC) uses regression to make predictions for fourteen 30-day surgical outcomes. While this approach provides accurate (discrimination and calibration) risk estimates, they might be improved by machine learning (ML). To investigate this possibility, accuracy for regression-based risk estimates were compared to estimates from an extreme gradient boosting (XGB)-ML algorithm. ⋯ XGB-ML provided more accurate risk estimates than regression in terms of discrimination and calibration. Differences in calibration between regression and XGB-ML were of substantial magnitude and support transitioning the RC to XGB-ML.
-
Multimorbidity in surgery is common and associated with worse postoperative outcomes. However, conventional multimorbidity definitions (≥2 comorbidities) label the vast majority of older patients as multimorbid, limiting clinical usefulness. We sought to develop and validate better surgical specialty-specific multimorbidity definitions based on distinct comorbidity combinations. ⋯ Our new multimorbidity definitions identified far more specific, higher-risk pools of patients than conventional definitions, potentially aiding clinical decision-making.