A comparison of two comorbidity indices for predicting inpatient rehabilitation outcomes.

Abstract

BACKGROUND: Comorbid conditions are important in health care. The best comorbidity index for predicting the impact of comorbidities on rehabilitation outcomes has not been determined. AIM: Compare the associations of comorbidity measured using the Charlson Comorbidity Index (CCI) and the Cumulative Index Rating Scale (CIRS) with key rehabilitation outcomes. Aim was to determine whether either of these comorbidity indices helped explain the variation in key rehabilitation outcomes. DESIGN: Prospective open-cohort study. SETTING: Inpatient rehabilitation ward, Melbourne, Australia. POPULATION: Adults admitted for inpatient rehabilitation (N.=280). METHODS: The main outcomes were demographic (e.g. age, gender, discharge destination) and clinical outcomes (reason for rehabilitation, length of stay, LOS, Functional Independence Measure, CCI and CIRS). A series of regression analyses were performed to determine the influence of comorbidity on three dependent variables: 1) LOS in rehabilitation; 2) the change in Functional Independence Measure-motor score between rehabilitation discharge and admission; 3) the discharge destination (home vs. other). RESULTS: The mean age was 57.7 years, there were slightly more females (51%), most (95%) patients previously lived at home with family or other relatives (63%). The most common reason for rehabilitation was orthopedic or other conditions (52%) and most (80%) people were discharged home. The median LOS was 27 days. There were 100 (35.7%) patients who had no comorbidity recorded using the CCI, 112 (40.0%) had one comorbidity and 26 (9.3%) who had three or more. All patients had at least one comorbidity recorded with the CIRS, and 264 (94.3%) had 3 or more comorbidities. There was little or no difference between the CCI or CIRS in terms of their ability to explain the variance in LOS (adjusted R2=0.38 with and without comorbidities), change in disability during rehabilitation (adjusted R2=0.31-0.33 with and without comorbidities) or the discharge destination (AUC=0.72 without comorbidities; 0.73-0.74 with comorbidities) beyond that accounted for by demographic and clinical information. CONCLUSIONS: Neither the CIRS nor the CCI in our patient sample provide additional information that explains the impact of comorbidities on key rehabilitation outcomes. CLINICAL REHABILITATION IMPACT: Further research is needed to determine the most appropriate measure of comorbidity of relevance to inpatient rehabilitation outcomes.