Using serial severity scores to predict death in ICU patients: a validation study and review of the literature

Abstract

Purpose of review: The present study describes the use of serial severity scores to predict death in ICU patients and compares the results with previously published literature on this topic.

Recent findings: Predicting mortality in critically ill patients has tremendous significance and methods to do so accurately have been studied for decades. The ability to accurately predict death impacts medical therapies, triaging, end-of-life care, and many other aspects of ICU care. There are many methods in existence to help physicians predict mortality, but most are not very accurate on an individual basis. The main tools available are severity scores, published outcomes data, and personal experience and all of them have significant limitations. One strategy that has been shown to be effective in accurately predicting death is to use serial severity scoring during the patient's ICU admission. Recently, a retrospective study done on a large cohort of ICU patients at a single institution showed very high specificity in predicting death by using serial acute physiology, age, and chronic health evaluation (APACHE III) scores on days 1 and 3. The authors of this article sought to validate this study in a different institution using a slightly different model that was easier to use and might increase sensitivity. The results of this small study are presented with a review of the literature on the use of serial scores to predict death in ICU patients.

Summary: Over the years multiple studies have shown that systems using serial severity scores can predict death in ICU patients with very high but not perfect accuracy. The clinical use of these systems remains low however and ultimately their main utility may be in research.