Construction of a prediction model for perioperative blood transfusion in elderly hip fracture patients: a retrospective study.

OBJECTIVES: Perioperative blood transfusion in elderly patients is associated with various complications. This study aimed to develop a clinical prediction model for intraoperative blood transfusion in elderly patients undergoing hip fracture surgery to minimize transfusion requirements and related risks. METHOD: Patients who underwent hip surgery between January 2021 and December 2023 at Xiangyang Central Hospital (Hubei, China) were retrospectively included in this study. Relevant factors for postoperative blood transfusion were analyzed using univariate and multivariate logistic regression analysis. A nomogram was established with the identified relevant factors, and its performance was evaluated using the area under the subject operating characteristic curve, calibration curve, and decision curve analysis. RESULT: A total of 1092 patients were included in this study for evaluation. Through logistic regression analysis, we identified surgery type, preoperative hemoglobin level, platelet count, osteoporosis, and cerebrovascular disease as relevant factors for postoperative blood transfusion. The nomogram was established, and the areas under the curve of the subjects' work characteristics were found to be 0.730 (95% confidence interval [CI]: 0.686-0.773) and 0.726 (95% CI: 0.671-0.781) for the training and test sets, respectively. CONCLUSION: In this study, we identified surgery type, preoperative hemoglobin level, platelet count, osteoporosis, and cerebrovascular disease as the risk factors for postoperative blood transfusion in elderly hip fracture patients. The nomogram, established based on these risk factors, showed an acceptable discriminative ability to predict intraoperative blood transfusion in elderly patients during hip fracture surgery. TRIAL REGISTRATION: This clinical trial was registered at http://www.chictr.org.cn/ (ChiCTR2400085204, June 03, 2024).