The Use of SCORE and GRACE Risk Tools to Assess the Length of Stay in a Cardiac Intensive Care Unit

Abstract

AbstractThe possibility of using simple and effective models to estimate the patient’s length of stay in intensive care units is decisive to support the clinical professional decisions. These models can help professionals in the stratification process and, particularly, in the identification of the necessary intervention plan to improve the patient’s health condition. In clinical practice specific prognostic scores are available and validated in the cardiovascular context. These risk tools address the primary prevention domain, as well as the secondary prevention domain, usually involving long-term (years) and short-term (months) prediction periods, respectively.The aim of this study is to investigate the capacity of available prognosis risk tools, in particular SCORE (primary tool) and GRACE (secondary tool), to estimate the length of stay in a cardiac intensive care unit. For validation purposes a dataset collected by the Centro Hospitalar Universitário de Coimbra was used, consisting of approximately 1400 patients that have been admitted into the cardiology intensive care unit. The obtained results suggested that SCORE and GRACE models are not sufficiently accurate to estimate the actual length of stay. Moreover, GRACE presents better results than SCORE, which can be justified by the employed risk factors, more specific for short-term prediction periods.KeywordsLength of stayPredictionCardiovascular risk scoresIntensive care units