ICU scoring system better predicts the risk of death for trauma patients
Research Activities, February 2010, No. 354
Care providers in intensive care units (ICUs) can calculate a patient's risk of dying with the Acute Physiologic and Chronic Health Evaluation (APACHE) II, which is based on 12 clinical and biochemical parameters, the patient's age, and pre-existing illnesses. Some deem this score as invalid for trauma patients in ICUs because those patients are typically younger and healthier than other ICU patients. However, a new study finds that APACHE II is actually a better predictor of death in ICU trauma patients than two commonly used trauma scores, the injury severity score (ISS) and the Trauma and Injury Severity Score (TRISS).
Researchers studied 1,019 trauma patients' scores from APACHE II, ISS, and TRISS. APACHE II was more accurate than the other two at predicting both death and how long the patient would stay in the ICU. In fact, TRISS resulted in 144 unexpected survivors and 91 unexpected deaths. APACHE II is a better predictor, because its scores are based upon deviations from normal physiology, with the greater the deviation resulting in a higher score and worse outcomes, the authors state. Conversely, ISS is based on anatomic grading of injury severity and TRISS uses physiologic variables collected during admission.
The authors caution against using APACHE II as a "one size fits all" measure for assessing care quality in ICUs or gauging physician performance. They recommend that future benchmarking initiatives instead focus on population and sub-population scores to more accurately predict risk of death among ICU patients. This study was funded in part by the Agency for Healthcare Research and Quality (HS13833).
See "Revisiting the validity of APACHE II in the trauma ICU: Improved risk stratification in critically injured adults," by Lesly A. Dossett, M.D., M.P.H., Leigh Ann Redhage, M.D., Robert G. Sawyer, M.D., and Addison K. May, M.D., F.A.C.S., F.C.C.M., in the September 2009 Injury 40(9), pp. 993-998.