OBJECTIVES: The Charlson Comorbidity Index (CCI), although intended to predict mortality risk, is widely used as a means of controlling for comorbidities. Charlson et al. (2008; CCI2008) added items to better predict resource use and alternatively Quan et al. (2011; CCI2011) removed or updated the weighting of comorbidities to reflect modern estimates of contributions to mortality risk.  The current study assessed the predictive utility of these three CCI variants across categories of health outcomes. METHODS: Data were analyzed from the 2013 U.S. National Health and Wellness Survey (n=75,000), an annual, cross-sectional, general health survey of adults (aged ≥18).  CCI scores were calculated from a weighted sum of self-reported diagnoses.  Health outcomes included: health-related quality of life measured via the SF-36v2 (mental and physical health status, and health utilities); Work Productivity and Activity Impairment (WPAI-GH) questionnaire-based measures; and resource use (visits to the emergency room, hospital, and traditional providers in the last 6 months).  CCI scores were correlated with health outcomes (Pearson correlations for HRQoL, Spearman’s rho for WPAI-GH and resource use).  RESULTS: All correlations were significant, p<.001, with higher CCI scores associated with worse outcomes.  CCI, CCI2008, and CCI2011 correlations ranged from lows of -.06, -.20, and -.07, respectively (with mental health status), to -.34, -.37, and -.30 (with physical health status), .27, .36, and .24 (with activity impairment), and .33, .39, and .25 (with provider visits). CONCLUSIONS: Across all outcomes, CCI2008 exhibited the strongest correlations, with the original CCI exhibiting slightly higher correlations than CCI2011 in all but two cases.  The utility of the CCI depends on variance explained in the outcomes plus the anticipated independence of included comorbidities vis-à-vis predictors of interest in multivariable models.