Outcome of intensive care. An application of a predictive model

Abstract

The development of a scoring system to estimate equivalency of illness between individuals has been undertaken. A model has been formulated to calculate probability of survival from each of 225 potential "conditions" apt to occur in patients admitted to intensive care areas. The presence or absence of each factor was noted on admission to a university hospital ICU. The relation between conditions noted in observations on a sample of patients, and survival allows assignment of a weight to each complication on the basis of which a "Condition Index Score" (CIS) or prognosis index can be objectively calculated. Potential uses of CIS are to: (1) establish objective criteria for admission to and discharge from intensive care, and for transfer to tertiary care centers; (2) compare quality of care (outcome vs. CIS) between different intensive care facilities; (3) serve as basis for multi-institutional studies concerning critically ill patients; (4) compare outcomes in groups of patients with equal CIS who are and are not treated in ICUs; and (5) establish appropriate numbers of critical care beds for any hospital or area by CIS criteria.