Beyond chronological age: a decision tree model integrating nutritional and renal function indicators for risk prediction in pre-frail older adults with hepatolithiasis

The study included 451 pre-frail older adults and developed a decision tree model to predict hepatolithiasis risk. Using Propensity Score Matching, two balanced groups with 147 matched pairs were created. In the full-sample model, age over 66.85 years was identified as the primary factor, followed by positive urine protein, marital status, and education level, achieving 72.4% classification accuracy and AUC of 0.795. In subgroup models, additional key factors were identified including BMI, white blood cell count, serum creatinine, glucose, and bilirubin. Group 1 achieved AUC of 0.914 and Group 2 achieved AUC of 0.877. Age emerged as the dominant risk factor across all models. The study concluded that decision tree modeling combined with PSM provides an effective approach for identifying high-risk individuals, though external validation on larger samples is necessary.