Predicting outcome in intensive therapy units--a comparison of Apache II with subjective assessments

In a prospective study 568 patients admitted to a mixed medical and surgical intensive therapy unit (ITU) were assessed using the Apache II severity of illness score to predict outcome. Their outcome was also predicted subjectively by a doctor and nurse on admission. There were 260 deaths in the group. The subjective predictions were compared with the Apache II predictions using logistic regression analysis and receiver-operating-characteristic curve measurement. The subjective assessments were found to be a more powerful predictor of outcome in this group of patients than the Apache II scores and predicted risk of death. Although the predictions could be successfully applied to the population as a whole, none of the tests were suitable for predicting outcome on an individual patient.