Risk Factors for Emergency Department Short Time Readmission in Stratified Population

Abstract

Background. Emergency department (ED) readmissions are considered an indicator of healthcare quality that is particularly relevant in older adults. The primary objective of this study was to identify key factors for predicting patients returning to the ED within 30 days of being discharged. Methods. We analysed patients who attended our ED in June 2014, stratified into four groups based on the Kaiser pyramid. We collected data on more than 100 variables per case including demographic and clinical characteristics and drug treatments. We identified the variables with the highest discriminating power to predict ED readmission and constructed classifiers using machine learning methods to provide predictions. Results. Classifier performance distinguishing between patients who were and were not readmitted (within 30 days), in terms of average accuracy (AC). The variables with the greatest discriminating power were age, comorbidity, reasons for consultation, social factors, and drug treatments. Conclusions. It is possible to predict readmissions in stratified groups with high accuracy and to identify the most important factors influencing the event. Therefore, it will be possible to develop interventions to improve the quality of care provided to ED patients.

Machine Learning Readmission Risk Modeling: A Pediatric Case Study

“…Recent reviews and comparative studies [1, 4, 6] have found that studies on adult readmissions reported low values of area under ROC Curve (AUC aka c-statistic) ranging between 0.56 and 0.72. One way to improve prediction results is to carry out stratified studies, i.e., building specific predictive models for specific patient categories [41].…”

Modelling hospital readmissions under frailty conditions for healthy aging

“…Recent works favor the application of predictive machine‐learning approaches, formulating readmission prediction as a binary classification problem (Artetxe et al, ; Ottenbacher et al, ). Instances of classifier models used for readmission prediction are support vector machines (SVM) (Besga et al, ; Futoma et al, ; Zheng et al, ; Cui et al, ) deep learning (Reddy & Delen, ; Xiao et al, ), artificial neural network (Ottenbacher et al, ), and naïve Bayes (Vukicevic et al, ; Shameer et al, ).…”

“…Most of the studies about readmissions are focused in high‐risk populations of middle age adults (Artetxe et al, ; Kansagara et al, ), with specific chronic conditions so that readmission is potentially preventable or predictable, the latter implying that it is possible to manage resource planning in order to minimize the impact of unexpected readmissions. There are few studies that consider explicitly the frail aged individuals in their cohorts (Besga et al, ). However, frailty is an increasingly prevalent condition as the population is aging in all developed countries (Rodriguez‐Mañas & Fried, ), but is also an issue of growing concern for so‐called underdeveloped countries (Aboderin & Beard, ).…”

“…Readmission prediction in the case of adult patients has been tackled with diverse statistical approaches (Artetxe et al, 2018;Kansagara et al, 2011) such as logistic regression (Garmendia et al, 2017;Ottenbacher et al, 2001) and survival analysis (Garmendia et al, 2019). machines (SVM) (Besga et al, 2015;Futoma et al, 2015;Zheng et al, 2015;Cui et al, 2018) deep learning (Reddy & Delen, 2018;Xiao et al, 2018), artificial neural network (Ottenbacher et al, 2001), and naïve Bayes (Vukicevic et al, 2015;Shameer et al, 2017).…”

Programa educativo sobre registro da pressão arterial em serviço hospitalar de emergência: um estudo de intervenção

“…Até o presente momento, não foram encontrados estudos brasileiros que evidenciassem, com exatidão, as caraterísticas da população atendida em unidades de emergência; em contrapartida, diversos são os estudos internacionais que discutem a estreita relação da readmissão do paciente com o cumprimento do exercício profissional e com a qualidade da assistência prestada (BESGA et al, 2015;ADAMS, 2016;ATKINS;KANSAGARA, 2016).…”