External Validation of the Comorbid Operative Risk Evaluation Score Among Medicare Beneficiaries Undergoing Major Gastrointestinal Surgery.

BACKGROUND: Accurate risk adjustment is essential in gastrointestinal surgery, particularly among older adults. The surgery-specific Comorbid Operative Risk Evaluation (CORE) score demonstrated greater in-hospital mortality discrimination than conventional comorbidity indices in its derivation study, but its performance has not been independently evaluated among Medicare beneficiaries undergoing major gastrointestinal surgery. METHODS: This retrospective cohort study used Medicare claims to identify beneficiaries aged 66 years or older undergoing major gastrointestinal surgery from 2016 to 2022, including esophageal, gastric, hepatopancreatobiliary, biliary, colonic, and rectal procedures. CORE was calculated without re-estimation using the algorithm and Stata package developed by Chervu et al. Patients were stratified into cohort-derived quartiles. Multivariable regression assessed mortality, morbidity, readmission, non-home discharge, and length of stay, adjusting for patient, operative, and hospital characteristics. Discrimination was compared among CORE, the van Walraven-weighted Elixhauser score, and CCI with incremental performance assessed after adding each measure to a base covariate model. RESULTS: Among 711,438 patients, median CORE score was 57.5. Higher CORE quartiles were associated with older age, greater comorbidity burden, fewer malignant indications, and more urgent/emergent surgery. Outcomes worsened stepwise across quartiles: from Q1 to Q4, in-hospital mortality increased from 0.41% to 9.33%, non-home discharge from 29.47% to 71.27%, prolonged length of stay from 7.82% to 47.19%, and 30-day readmission from 11.62% to 18.83%. After adjustment, Q4 CORE was associated with in-hospital mortality (aOR 24.66), 30-day mortality (aOR 12.36), 90-day mortality (aOR 8.19), and 30-day post-discharge complications (aOR 1.62). CORE outperformed CCI for in-hospital mortality discrimination (AUROC 0.791 vs 0.625) but underperformed relative to Elixhauser (0.808; P<0.001). CONCLUSIONS: Higher CORE scores were associated with progressively worse postoperative outcomes after major gastrointestinal surgery. CORE outperformed CCI but demonstrated lower mortality discrimination than Elixhauser, with variable performance across non-mortality outcomes. These findings support CORE as a useful surgery-specific comorbidity measure while highlighting the need for outcome-specific validation against established risk indices.