Comorbidity as a correlate of length of stay for hospitalized patients with acute chest pain.

OBJECTIVE

To determine whether comorbid medical conditions as measured with the Charlson Comorbidity Index are independent correlates of length of stay after adjusting for other clinical and socioeconomic data.

DESIGN

Prospective cohort study.

SETTING

Urban teaching hospital.

PATIENTS

All 1,261 patient aged 30 years or more who were admitted to this hospital after coming to the emergency department with acute chest pain between October 1990 and May 1992.

MEASUREMENTS AND OUTCOMES

Clinical data including comorbid medical conditions used in the Charlson index were prospectively recorded by the evaluating physician at the time of admission or by a research nurse who was blinded to the subsequent events. History of myocardial infarction was excluded from the calculation of the Charlson index score. Charlson index scores were 0 to 1 for 921 patients (73%), 2 to 3 for 263 (21%), and greater than 3 for 77 (6%). Unadjusted means (+/- SD) lengths of stay in these groups were 4.4 +/- 5.2, 5.2 +/- 5.9, and 7.5 +/- 9.3 days, respectively. In multiple linear regression analysis, compared with Charlson index scores of 0 to 1, scores of 2 to 3 and greater than 3 were significant (p < .01) independent correlates of the log transformation of length of stay after adjusting for clinical data from the initial presentation and subsequent course (model R2 = .510). In an analysis restricted to the 795 patients without clinical complications, a Charlson index score greater than 3 was an independent correlate of length of stay compared with scores of 0 to 1 (p < .01). Individual comorbid conditions were not significant correlates of length of stay after controlling for Charlson index score.

CONCLUSIONS

In this population of patients with acute chest pain, comorbidity as measured with the Charlson index was independently associated with length of stay after adjustment for other clinical data. After adjusting for the Charlson index, no separate comorbid condition was significantly correlated with length of stay. These findings suggest that the Charlson index can be used to adjust for comorbidities in analyses of length of stay for patients with this condition.