Using scores to identify patients at risk of short term mortality at arrival to the acute medical unit: A validation study of six existing scores

“Early warning scores” (EWS) have been developed to quantify levels of vital sign abnormality. However, many scores have not been validated. The aim of this study was to validate six scores that all rely on vital signs: Rapid Acute Physiology Score (RAPS), Rapid Emergency Medicine Score (REMS) and the National Early Warning Score (NEWS) and the Goodacre, Groarke and Worthing physiological scores. A posthoc single-center observational cohort study of prospectively collected vital signs on acutely admitted medical patients to a Danish hospital. All adult patients arriving at an acute medical unit at a 450-bed regional teaching hospital were included. Upon arrival, we registered initial vital signs and only the first presentation in the study period was included. Patients were included from 1 June to 31 October 2012. All-cause 24-h mortality and overall in-hospital mortality were used as endpoints. A discriminatory power above 0.8 was considered acceptable. 5784 patients were included with a median age of 67 (49–78) years, 32 (0.6%) died within 24h and 161 (2.8%) while admitted. Discriminatory power for 24h mortality was above 0.8 for all scores (except the Groarke score (0.587)) and highest for the Worthing score (0.847). The discriminatory power for predicting overall in-hospital mortality was highest for the Goodacre and Worthing scores (0.810 and 0.800 respectively) but below 0.8 for the remaining scores. The Goodacre score and the Worthing physiological score have good discriminatory power at identifying patients at increased risk of 24-h and in-hospital mortality in our setting. •Out of six scores, five had acceptable discriminatory power for 24-h mortality.•Only two scores had acceptable discriminatory power for in-hospital mortality.•All scores had good calibration.•Not all clinical risk prediction scores are good at predicting mortality.