Livingston raises 2 concerns. He questions the use of the Charlson index comorbidity score as an adequate surrogate in the adjustment for patient risk factors when assessing operative mortality. He proposes the inclusion of individual variables in adjustment that may specifically influence perioperative mortality to more accurately reflect patient risk factors that may confound analysis of outcomes of colon resection between THs and non-THs. We concur that the Charlson index was originally developed for longitudinal mortality studies and that its predictive accuracy in accounting for inpatient mortality has been subject to scrutiny.1,2 Previous accounts, however, that specifically examined volume-outcome relationships have found that use of the Charlson index with its published weights yields identical results to using the Charlson index with empirically derived weights for each procedure or by the inclusion of additional comorbid conditions individually inserted into each respective model.3,4 We assert that the comorbid conditions used to derive the Charlson index including history of myocardial infarction, congestive heart failure, chronic pulmonary disease, diabetes mellitus, or liver disease are factors that may influence both perioperative and long-term outcomes and, furthermore, are likely supported by tests of goodness of fit.
Hayanga A, Kaiser H. Teaching Hospital Status and Operative Mortality in the United States—Reply. Arch Surg. 2010;145(10):1025-1026. doi:10.1001/archsurg.2010.210