OA2062. Home injuries risk of hospitalization: A Machine Learning approach on the EU-IDB data

Abstract Background Home injuries have a wide spectrum of severity and in different Countries the healthcare organization impacts differently on the need for inpatient care. Machine Learning (ML) parametric modelling consents to make repeatable predictions about this risk. Methods ML techniques were applied to the Full Data Set (FDS) of the European Injury DataBase (EU-IDB) which provides detailed information on the external causes and diagnoses of injury observed at the Emergency Departments (ED). More than 1,200,000 cases of home injuries were observed in the period 2008-22 in 19 Countries. LASSO (Least Absolute Shrinkage and Selection Operator) cross-validated linearized regression technique was used for variable selection and parameter regularization. Inpatients were considered those admitted or transferred to another hospital or deceased during hospitalization. Cross validation was performed randomly assigning the records on 5 folds. On the selected variables a cross-validated logistic model was performed on 5 folds which were randomly sampled assigning 80% of records to training and 20% to testing samples. Results The strongest predictors of hospital admission risk selected by the model were in order of importance: EUROCOST-39 diagnoses categories, Recording Country, Direct Object Causing Injury, Age Group, Mechanism of Injury, Sex of Patient, Month of Attendance. EUROCOST-39 categories represent 64.8% of the explained variability, Recording Country 17.8%, Direct Object 8.4% and Age Group 7.5%. The estimated Odds Ratios of admission risk for instance in the median test sample (fold 5) are: Recording Country = 77.9 (95%CI: 73.1-83.0); EUROCOST-39 = 65.0 (95%CI: 61.2-69.0. The model accuracy is AUC 87.5% Conclusions ML approach enhances the prediction accuracy of hospital admission main factors. A combination of more disabling injury, Country organization, older age and exposure to uneven surfaces (for falls) or chemicals increases enormously the risk of hospitalization. Topic Home injury, Machine learning, LASSO