A prospective evaluation of the cumulative illness rating scale.

Assessment of overall physical health is an important yet little studied problem in geriatric research. The Cumulative Illness Rating Scale (CIRS) was among the first instruments that attempted to summarize the overall severity of illness based on clinical information. This study evaluated the CIRS in a prospective longitudinal study of 181 elderly (mean age +/- SD = 79 +/- 7.4) subjects undergoing comprehensive geriatric assessment in an outpatient unit. The CIRS was found to correlate negatively with activities of daily living (r = -0.49, p = 0.0001), instrumental activities of daily living (r = -0.34, p = 0.0001), patient morale (r = -0.30, p = 0.0001), and positively with days in hospital (r = 0.21, p = 0.0075) and number of medications (r = 0.31, p = 0.0001). Mean CIRS scores for subjects who died during follow-up were significantly higher than the scores for survivors (p less than 0.01). In logistic regression, CIRS was a significant predictor of death, yet it did not improve that prediction over information contained in measures of activities of daily living. In separate logistic analyses, CIRS and age predicted acute care hospital days during follow-up, while ADL or IADL predicted the use of nursing home services. Although the CIRS appears to be a reliable method of summarizing medical information and to have some external validity, in its present form it does not provide additional prognostic information.