Enhanced risk prediction of femoral head osteonecrosis in the elderly: a comparative study of random forest and logistic regression models.

Background and aim: Osteonecrosis of the femoral head (ONFH) is a degenerative joint disorder that frequently leads to structural collapse and impaired mobility, particularly in older adults. Early detection of associated risk factors is essential for timely intervention. This study aimed to compare the predictive performance of a Random Forest (RF) algorithm and a Logistic Regression (LR) model in identifying key contributors to ONFH in elderly patients. Methods: This retrospective study included 339 patients aged ≥ 75 years who received treatment at Shanxi Bethune Hospital from January 2017 to December 2023, with complete clinical and imaging records. Variables included demographics, bone mineral density, medication and lifestyle history, comorbidities, and radiographic findings. Patients were randomly allocated into training (70%) and validation (30%) cohorts. Predictive models were developed using RF and LR, with performance assessed by accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC). Results: Both models consistently identified corticosteroid exposure, reduced bone mineral density, prior femoral fractures, and advanced age as major risk factors. The RF model demonstrated superior performance (AUC = 0.896; accuracy = 83.5%; sensitivity = 82.4%; specificity = 84.3%) compared to the LR model (AUC = 0.797; accuracy = 75.0%; sensitivity = 72.0%; specificity = 76.0%). ROC analysis confirmed the RF model's enhanced discriminative ability. Conclusion: The RF algorithm outperformed traditional logistic regression in predicting ONFH among older adults, highlighting the potential of machine learning techniques to support early risk identification and improve clinical decision-making in orthopedic care.