Development of a Surgery-specific Comorbidity Score for Use in Administrative Data

Area of Science:

• Surgical outcomes research
• Health informatics
• Predictive modeling in medicine

Background:

• Existing comorbidity indices (Elixhauser and Charlson) are outdated, developed decades ago.
• These indices were primarily based on non-surgical patient cohorts.
• There is a need for a modern comorbidity score tailored for surgical research.

Purpose of the Study:

• To develop and validate a novel comorbidity score for surgical database research.
• To improve the prediction of in-hospital mortality in surgical patients.
• To create a tool that is more relevant to current surgical practices and patient populations.

Main Methods:

• Utilized 2019 National Inpatient Sample (NIS) data for 699,155 adult patients across 62 operations.
• International Classification of Diseases, 10th Revision (ICD-10) codes were mapped to Clinical Classifications Software Refined (CCSR) groups.
• Machine learning algorithms identified important CCSRs for predicting in-hospital mortality, which informed logistic regression and the CORE score calculation.
• Model performance was assessed using Area Under the Receiver Operating Characteristic (AUROC) curves and validated across multiple datasets.

Main Results:

• The CORE score demonstrated superior prediction of in-hospital mortality compared to the Elixhauser Comorbidity Index (ECI) across NIS, State Inpatient Databases (SID), and institutional data.
• CORE also significantly outperformed the Charlson Comorbidity Index (CCI) in predicting mortality across all tested datasets.
• AUROC values for CORE consistently exceeded those for ECI and CCI (e.g., NIS: CORE 0.90 vs. ECI 0.84 vs. CCI 0.76).

Conclusions:

• The novel Comorbid Operative Risk Evaluation (CORE) score shows enhanced predictive accuracy for in-hospital mortality in surgical patients.
• CORE is a more effective tool for outcome-based research compared to traditional comorbidity indices like ECI and CCI.
• This score has the potential to refine risk assessment and improve research in surgical populations.