Comparing Charlson Comorbidity Index Scores between Anesthesiologists, Patients, and Administrative Data: A...

Area of Science:

• Medical research
• Health informatics
• Clinical epidemiology

Background:

• Comorbidities significantly influence perioperative risk assessment.
• The comparative accuracy of Charlson Comorbidity Indices (CCIs) from various data sources remains unclear.
• Understanding these differences is crucial for accurate patient risk stratification.

Purpose of the Study:

• To compare the accuracy of CCIs derived from patient self-reports, physician assessments, and hospital administrative data.
• To evaluate the predictive performance of different CCI sources for in-hospital mortality.
• To inform best practices in perioperative risk assessment.

Main Methods:

• Prospective observational study involving 1007 patients.
• Comparison of CCIs calculated from patient self-reports, physician assessments, and hospital administrative data.
• Statistical analysis including kappa statistics for agreement and AUROC for mortality prediction.

Main Results:

• Fair agreement was observed between patient self-report/physician assessment and administrative data (kappa 0.24-0.28).
• Physician assessment and patient self-report showed the best agreement (kappa 0.33).
• All three CCI sources demonstrated comparable predictive performance for in-hospital mortality (AUROC ~0.80-0.86).

Conclusions:

• CCIs derived from patient self-report, physician assessment, and administrative data exhibit similar efficacy in predicting postoperative in-hospital mortality.
• These findings suggest flexibility in data source selection for perioperative risk assessment.
• Further research may explore the integration of multiple data sources for enhanced predictive accuracy.