Performance of the APOP-screener for predicting in-hospital mortality in older COVID-19 patients: a retrospective study

Validation of the acutely presenting older patient screener for short term mortality prediction in older patients hospitalized for COVID-19

PURPOSE

The aim of this study was to validate the acutely presenting older patient (APOP) screener, routinely used on the Emergency Department to predict risk of adverse outcomes in older people, for prediction of in-hospital mortality and 30-days-mortality in older patients hospitalized for COVID-19.

METHODS

Patients ≥ 70 years from a multicenter cohort hospitalized for COVID-19 with measured APOP risk were included. External validation analysis of the APOP screener for in-hospital mortality and 30-days-mortality was performed including discrimination and calibration.

RESULTS

389 patients (median age 80 (IQR 75-85) years, 41.4% female, 138 APOP high risk) were included. APOP high risk patients more often lived institutionalized, (26% vs. 4%; p < 0.001), had more comorbidities (Charlson Comorbidity Index 2 (1-3) vs. 2 (0-3); p = 0.002) and were less often fit (Clinical Frailty Scale 1-3 17% vs. 62%; p < 0.001). 84 patients died in hospital and 114 within 30 days. APOP high risk patients had a higher risk of in-hospital-death [OR 1.6 (95% CI 1.0-2.6)] and death within 30 days [OR 2.7 (95% CI 1.7-4.2)]. The APOP screener discriminated poorly for in-hospital mortality [AUC 0.56 (95% CI 0.48-0.63)] and for 30-days-mortality [AUC 0.62 (95% CI 0.55-0.68)]. Calibration plots revealed overestimation of the screener for both mortality risks.

CONCLUSION

The APOP screener had a poor predictive performance for in-hospital mortality and 30-days-mortality in older people hospitalized for COVID-19. Screening tools routinely used on the ED may not be useful to predict mortality in different than usual clinical circumstances such as during a pandemic of a novel disease.

Prevalence of Frailty in European Emergency Departments (FEED): an international flash mob study

INTRODUCTION

Current emergency care systems are not optimized to respond to multiple and complex problems associated with frailty. Services may require reconfiguration to effectively deliver comprehensive frailty care, yet its prevalence and variation are poorly understood. This study primarily determined the prevalence of frailty among older people attending emergency care.

METHODS

This cross-sectional study used a flash mob approach to collect observational European emergency care data over a 24-h period (04 July 2023). Sites were identified through the European Task Force for Geriatric Emergency Medicine collaboration and social media. Data were collected for all individuals aged 65 + who attended emergency care, and for all adults aged 18 + at a subset of sites. Variables included demographics, Clinical Frailty Scale (CFS), vital signs, and disposition. European and national frailty prevalence was determined with proportions with each CFS level and with dichotomized CFS 5 + (mild or more severe frailty).

RESULTS

Sixty-two sites in fourteen European countries recruited five thousand seven hundred eighty-five individuals. 40% of 3479 older people had at least mild frailty, with countries ranging from 26 to 51%. They had median age 77 (IQR, 13) years and 53% were female. Across 22 sites observing all adult attenders, older people living with frailty comprised 14%.

CONCLUSION

40% of older people using European emergency care had CFS 5 + . Frailty prevalence varied widely among European care systems. These differences likely reflected entrance selection and provide windows of opportunity for system configuration and workforce planning.

Comprehensive analysis of clinical data for COVID-19 outcome estimation with machine learning models

COVID-19 is a global threat for the healthcare systems due to the rapid spread of the pathogen that causes it. In such situation, the clinicians must take important decisions, in an environment where medical resources can be insufficient. In this task, the computer-aided diagnosis systems can be very useful not only in the task of supporting the clinical decisions but also to perform relevant analyses, allowing them to understand better the disease and the factors that can identify the high risk patients. For those purposes, in this work, we use several machine learning algorithms to estimate the outcome of COVID-19 patients given their clinical information. Particularly, we perform 2 different studies: the first one estimates whether the patient is at low or at high risk of death whereas the second estimates if the patient needs hospitalization or not. The results of the analyses of this work show the most relevant features for each studied scenario, as well as the classification performance of the considered machine learning models. In particular, the XGBoost algorithm is able to estimate the need for hospitalization of a patient with an AUC-ROC of 0 . 8415 ± 0 . 0217 while it can also estimate the risk of death with an AUC-ROC of 0 . 7992 ± 0 . 0104 . Results have demonstrated the great potential of the proposal to determine those patients that need a greater amount of medical resources for being at a higher risk. This provides the healthcare services with a tool to better manage their resources.