Risk Factors for Postoperative Death in Elderly Patients with Hip Fracture and Establishment of Prediction Model

WANG Heping; XING Zhichao; ZHONG Kaihua

doi:10.3969/j.issn.1004-437X.2026.13.019

Abstract：Objective To analyze risk factors associated with postoperative mortality after hip fracture in elderly patients and develop a predictive model,so as to provide evidence for clinical assessment of postoperative mortality risk.Methods Four hundred elderly patients with hip fractures were selected from the Department of Orthopedics of Central Hospital of Zhoukou from May 2020 to October 2024,and they were randomly divided into a modeling and a verification group at a 3∶1 ratio.Based on patient survival within 1 year after surgery,the modeling group was further subdivided into a survival group(249 cases)and a death group(51 cases).Independent risk factors for 1-year postoperative mortality in elderly hip fracture patients were identified to develop a prognostic prediction model.The discriminative ability of the model was assessed by receiver operating characteristic(ROC)curve in the modeling and validation groups.Results Logistic regression showed that age ≥ 75 years,preoperative comorbidities＞2,preoperative prognostic nutritional index ≥ 45,American Society of Anesthesiologists(ASA)grade Ⅲ-Ⅳ,surgical delay,and postoperative complications were independent risk factors for mortality within 1 year after elderly hip joint surgery.The Hosmer-Lemeshow goodness-of-fit test was used to assess the regression model for 1-year postoperative mortality risk in elderly patients undergoing hip surgery,with a test P=0.411,indicating acceptable model calibration.The area under the curve(AUC)for the modeling group was 0.835(95％CI:0.776-0.894),with a sensitivity of 78.43％,specificity of 75.90％,and a cut-off point of 0.236;the validation group AUC was 0.793(95％CI:0.667-0.919),with a sensitivity of 81.25％and specificity of 77.38％.Conclusion The risk model for postoperative mortality in elderly patients with hip fracture has favorable discriminative validity and can be applied to identify high-risk individuals for 1-year postoperative death among such patients.