The MDS-CHESS scale: a new measure to predict mortality in institutionalized older people

OBJECTIVES: To develop a scale predicting mortality and other adverse outcomes associated with frailty. DESIGN: Observational study based on Minimum Data Set (MDS) 2.0 and mortality data. SETTING: Ontario chronic hospitals. PARTICIPANTS: All chronic hospital patients (N = 28,495) assessed with the MDS 2.0 after mandatory implementation in July 1996 followed until May 1999. MEASUREMENTS: MDS 2.0 assessments done as part of normal practice mainly by registered nurses or multidisciplinary teams in a chronic hospital. Mortality data are available from the accompanying discharge tracking form. RESULTS: The MDS-Changes in Health, End-stage disease and Symptoms and Signs (CHESS) score is a composite measure addressing changes in health, end-stage disease, and symptoms and signs of medical problems. It is a strong predictor of mortality (P <.0001) independent of the effects of age, sex, activities of daily living impairment, cognition, and do-not-resuscitate orders. It is also strongly associated with physician activity, complex medical procedures, and pain (P <.001 for each dependent variable). CONCLUSIONS: The CHESS score provides a useful new MDS-based test to predict mortality and to measure instability in health as a clinical outcome.