In-hospital hip fracture mortality score: predicting mortality after surgery for proximal femoral fracture in older patients

Abstract

Abstract Background/Objective/Introduction Hip fractures in older patients are associated with substantial morbidity and in-hospital mortality. Despite the clinical importance of this condition, few tools specifically estimate in-hospital mortality or account for clinical deterioration during hospitalization, particularly in middle-income settings. This study aimed to develop a dynamic prognostic score to estimate the risk of in-hospital mortality among older patients undergoing surgical treatment for hip fractures. Materials and Methods A retrospective cohort study was conducted with 1,509 patients aged 60 years and older treated at a Brazilian trauma centre between 2013 and 2023. Demographic, clinical, orthopaedic, laboratory and in-hospital evolution variables were analysed. Independent predictors of mortality were identified using binary logistic regression. A prognostic score was constructed using weights derived from the magnitude of association (odds ratios) observed in the logistic regression. Model accuracy was assessed by the area under the ROC curve (AUC). Patients were stratified into four in-hospital death risk categories: low, moderate, high, and very high. Results The in-hospital mortality rate was 3.7%. Independent predictors included dementia, acute myocardial infarction, heart failure, and abnormalities in creatinine, potassium, and INR. In-hospital events including pulmonary thromboembolism, need for blood transfusion, and postoperative ICU admission were also strongly associated with death. The final score showed good discriminative performance with an AUC of 0.819. Predicted in-hospital mortality increased progressively across the risk groups, from 3%, among low-risk patients to > 20% among very-high-risk patients. Conclusions The In-Hospital Hip Fracture Mortality Score demonstrated good predictive ability for in-hospital mortality in older surgically treated patients with hip fractures. However, it should be considered a preliminary hospital-based tool, and external validation in independent cohorts is needed before routine clinical implementation