Objectives: This study was designed to determine if comorbidity added more information than knowing only the patient's age in predicting survival and length of hospital stay.
Methods: The authors compared the relative predictive validity of three comorbidity indices: the Cumulative Illness Rating Scale, the Charlson Index, and a count of International Classification of Diseases, 9th Revision, Clinical Modification medical diagnoses in relation to survival and length of hospital stays in patients with spinal cord injury. The sample consisted of 330 longitudinally followed spinal-cord injured patients admitted between January 1989 and December 1990 who were followed for an additional 18 months.
Results: During the follow-up, 25 (7.5%) patients died and 249 (75.5%) were readmitted to hospital with a median of one admission (range, 1-8). The corresponding lengths of hospital stay ranged from 0 to 548 days, with a median of 7 days.
Conclusions: Patients who died were not significantly older but had higher comorbidity scores. Using patients alive at the end of the follow-up period, linear regression models were fit to the data to determine if comorbidity added more information regarding length of hospital stay than knowing only the patient's age. In the model that included only age as an independent variable, there was a significant relation between age and length of stay (F(1,303) = 5.2; P = 0.012). The R2 value for this model was 0.017. In further models that included age and each of the three comorbidity scores (separately) as the independent variables, the model that included age and the Cumulative Illness Rating Scale yielded the highest R2 value (R2 = 0.062). This study is among the first to compare three different measures of comorbidity and documents that comorbidity provides more information than knowing only the patient's age in relation to survival and length of hospital stay.