The ICD-10 Charlson Comorbidity Index predicted mortality but not resource utilization following hip fracture

Objectives: To evaluate the performance of the Charlson Comorbidity Index (CCI) in the prediction of mortality, 30-day readmission, and length of stay (LOS) in a hip fracture population using algorithms designed for use in International Classification of Diseases, 10th Revision (ICD-10)--coded administrative data sets.

Study design and setting: Hospitalization and death data for 47,698 New South Wales residents aged 65 years and over, admitted for hip fracture, were linked. Comorbidities were ascertained using ICD-10 coding algorithms developed by Sundararajan (2004) and Quan (2005). Regression models were fitted, and area under the receiver operating curve (AUC) and Akaike information criterion were assessed.

Results: Both algorithms had acceptable discrimination in predicting in-hospital (AUC, 0.72-0.76), 30-day (0.72-0.75), and 1-year mortality (0.69-0.75) but poor ability to predict 30-day readmission (0.54-0.57) or LOS (adjusted R(2), 0.007-0.045). The Quan algorithm provided better model fit than the Sundararajan algorithm. Models incorporating comorbidities as individual variables performed better than the Charlson weighted or updated Quan weighted score. Including a 1-year lookback period increased predictive ability for 1-year mortality only.

Conclusion: The CCI is a valid tool for predicting mortality but not resource utilization after hip fracture. We recommend the use of the Quan algorithm rather than Sundararajan algorithm and to model individual conditions rather than categorized weighted scores.