Performance of the Functional Comorbidity Index (FCI) in Prognostic Models for Risk Adjustment in Patients With Back Pain

Background: The Functional Comorbidity Index (FCI) is a comorbidity measure associated with physical function and may contribute to risk adjustment models in rehabilitation settings, but an International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) of the FCI has not been tested in outpatient settings. Objective: This study examines the ability of an ICD-9-CM– based FCI to predict function, health-related quality of life, and overall health care use. Design: Cohort study. Setting and Participants: This was a secondary analysis of 5155 adults ≥65 years of age with a new back pain visit from the Back pain Outcomes using Longitudinal Data cohort study. Independent Variables: We measured 18 comorbidities with an ICD-9-CM version of the FCI using diagnosis codes 12 months prior to an index visit. Main Outcome Measurements: Outcomes included the Roland Morris Disability Questionnaire (RMDQ, 0-24), health-related-quality-of-life (EQ5D, 0-1), and total health care use (sum of all relative value units [RVUs]) measured at baseline and 12 months after the index visit. Linear regression and generalized linear models estimated the association between the FCI and each outcome and to examine goodness of fit. We used a 10-fold cross-validation to develop and compare predictive models with and without the FCI. Results: There were 1398 participants (27%) with two or more comorbidities. Adjusted estimates show that for every one unit increase in FCI, RMDQ increased by 1.0 (95% confidence interval [CI] 0.8 to 1.1) and R² = 0.093; EQ5D decreased by 0.023 (95% CI −0.028 to −0.019) and R² = 0.076; and mean total RVUs increased by 13% (95% CI 1.09 to 1.17). Cross-validation showed that FCI contributed to small improvements in the performance of predictive models. Conclusion: An ICD-9-CM version of the FCI is associated with long-term function, health-related quality of life, and total health care use among older adults with back pain; however, it explains only a small proportion of the variance.