Objective: To describe the epidemiology of emergency department (ED) presentations for burns among children aged 0–5 years and to identify key predictors of hospitalization using a decision tree–based prediction framework.
Material and Methods: This retrospective, single-center observational study included children aged 0–5 years who presented to the ED with burn injuries between July 2019 and July 2024. Demographic and clinical variables, including burn etiology, depth, anatomical location, percentage of total body surface area affected, and Abbreviated Burn Severity Index (ABSI), were analyzed. A decision tree model was developed using SHapley Additive exPlanations (SHAP)–guided variable ranking and selection, and its performance was evaluated on a held-out test set.
Results: Among 329 patients, 47.7% were hospitalized (including those transferred for inpatient care). Scald burns were the predominant etiology (87.8%). Hospitalization was most strongly associated with second-degree burns, head and neck involvement, and an ABSI score ≥3. The decision tree model demonstrated good discrimination for hospitalization (area under the receiver operating characteristic curve [AUC], 0.841).
Conclusion: In early childhood burns, second-degree burn depth, head and neck involvement, and higher ABSI scores were the most influential factors associated with hospitalization decisions. This classification and regression tree (CART)–based model provides explicit decision rules derived from routinely available variables; however, its clinical utility and net benefit require external validation and prospective impact assessment.
Keywords: ABSI score, decision tree model, emergency care, hospitalization predictors, pediatric burns, scald injuries
, Halil Emre Koyuncuoğlu1 
