Developing a Frailty Index from routinely collected data in the Emergency Department among of hospitalized patients

Aim: To develop a Frailty Index from data routinely collected during the standard clinical assessment in the Emergency Department and to test its predictive capacity for adverse outcomes. Findings: The Frailty Index presented in this study is significantly correlated with age and associated with the length of hospital stay and the risk of institutionalization. The same association was not found between age and the two outcomes. Message: The use of the Frailty Index may support decisions from the very first contact of the individual with the hospital system, without burdening the clinical practice. Purpose: The number of frail patients admitted to Emergency Departments is increasing. The so-called Frailty Index based on the age-related accumulation of deficits models is often perceived as excessively burdening or not feasible in busy clinical settings due to its quantitative nature. We wanted to prove the possibility of generating a Frailty Index in the Emergency Department from data that are routinely collected during the standard clinical practice in this setting and to test its predictive capacity for adverse events. Methods: A retrospective analysis of the medical records of 110 hospitalized patients (mean age = 67.4 ± 18.9 years; women 41.8%) admitted to our Emergency Department during 6 days of 2017. A 41-item Frailty Index was computed from vital signs, physical examination, anamnestic diseases, and blood tests routinely collected by Emergency Department physicians. The length of the subsequent hospital stay and the institutionalization of the patient at the hospital discharge were the dependent variables of interest. Results: Median length of stay was 11.0 (interquartile range, IQR = 6.0–16.0) days. Institutionalization rate at discharge was 18.2%. The median Frailty Index was 0.22 (IQR = 0.17–0.30). The Frailty Index was significantly correlated with age (Spearman’s r = 0.44, p < 0.001) and resulted significantly associated with length of stay and institutionalization. The receiver operating characteristics areas under the curve were 0.731 (Confidence Interval, 95%CI 0.601–0.860, p = 0.001) and 0.726 (95%CI 0.610–0.841, p < 0.001) in the prediction of institutionalization and prolonged hospital stay, respectively. No statistically significant association of age with a length of stay (p = 0.75) nor institutionalization (p = 0.09) was reported. Conclusions: The standard multidimensional assessment conducted at the Emergency Department admission has all the necessary features to generate a meaningful clinical Frailty Index, potentially supporting decisions since the first contact of the individual with the hospital system.