EVALUATION OF 2 COMPETING METHODS FOR CALCULATING CHARLSON COMORBIDITY INDEX WHEN ANALYZING SHORT-TERM MORTALITY USING ADMINISTRATIVE DATA

The performance and predictive power of the Deyo-Charlson and the Roma no-Charlson comorbidity indices were compared when shore-term mortalit y after hospitalization was the outcome of interest. These indices are commonly used to adjust for the effect of comorbidities when using ad ministrative data in comparative studies. In-hospital Medicare claim d ata for patients admitted to one of six medical categories (back pain, stroke, pneumonia, hip replacement, transurethral radical prostatecto my, or lysis of peritoneal adhesion), were selected for analyses. Logi stic regression models were employed to evaluate the relative importan ce and the explanatory power of these indices for predicting mortality 30, 90, and 180 days after admission. The contribution of each index to mortality was assessed via the likelihood ratio chi-square statisti c (G(2)), and the area under the receiver operator characteristic (ROC ) curve was used to assess predictive power. Indices were evaluated us ing weights suggested by Charlson et al. and using empirically derived weights. Both indices improved the base model equally, although the p redictive power of both indices was poor with Values of the C statisti c ranging from 0.60 to 0.78. Our results suggest limited applicability of these approaches when examining short-term mortality. A slight inc rease in predictive power was observed when indices were calculated us ing empirical weights derived from our data. (C) 1997 Elsevier Science Inc.