<b>Systematic Review of the Intraoperative Hypothermia Risk Prediction Models in Total Joint Arthroplasty Patients</b>

Machine learning has been widely applied to the prediction of intraoperative hypothermia (IH) risk in surgical patients. Compared to traditional scoring systems, machine learning methods can more comprehensively identify risk factors for hypothermia, providing effective clinical guidance for nursing. The incidence of hypothermia in total joint arthroplasty (TJA) patients remains high, warranting attention. Although prediction models developed by researchers have shown promising results, the quality of these models still requires scrutiny. This review aims to evaluate existing intraoperative hypothermia risk prediction models in TJA patients, focusing on the development quality and predictive performance of these models.机器学习已被广泛应用于手术患者术中低体温（intraoperative hypothermia, IH）风险预测领域。相较于传统评分系统，机器学习方法能够更全面地识别低体温相关危险因素，为临床护理提供有效的指导依据。全关节置换术（total joint arthroplasty, TJA）患者的低体温发病率仍处于较高水平，值得临床关注。尽管研究者开发的相关预测模型已展现出良好的应用效果，但此类模型的质量仍有待严格评估。本综述旨在对现有针对全关节置换术患者的术中低体温风险预测模型进行系统性评价，重点关注这些模型的开发质量与预测性能。