儿童四肢骨折术中及术后发生低体温风险预测模型的构建

Construction of a Risk Prediction Model for Hypothermia during and after Surgery in Pediatric Limb Fractures

摘要: 目的：构建四肢骨折患儿术中及术后低体温风险预测模型，以优化手术室护理流程，降低四肢骨折患儿术中及术后低体温的发生率。方法：回顾性分析2022年6月至2024年10月期间在我院接受手术治疗的221例四肢骨折患儿的临床资料，通过LASSO回归筛选关键变量，并使用logistic回归识别出与低体温密切相关的预测因子，构建预测模型并将其可视化，通过ROC曲线、校准曲线和决策曲线分析评估列线图的预测性能。结果：最终纳入221例患者，其中47例发生低体温，174例未发生低体温。Logistic回归分析显示，患儿术中失血量、入室体温及手术室温度是四肢骨折患儿术中低体温发生的独立影响因素(P < 0.05)。预测模型AUC为0.895 (95% CI：0.856~0.934)，模型的一致性和预测能力均较好，临床决策曲线显示模型在不同阈值下均具有临床获益。结论：术中失血量、入室体温计手术室温度是影响四肢骨折患儿发生术中或术后低体温的独立危险因素，以此构建的骨折患儿术中低体温风险预测模型具有较高的准确性和适用性，可提供可靠的预测工具，有助于术中低体温的预防和管理。

Abstract: Objective: To develop a risk prediction model for hypothermia during and after surgery in pediatric patients with limb fractures, in order to optimize operating room nursing processes and reduce the incidence of hypothermia in these patients. Methods: A retrospective analysis was conducted on the clinical data of 221 pediatric patients who underwent surgical treatment for limb fractures at our hospital from June 2022 to October 2024. Key variables were selected using LASSO regression, and logistic regression was employed to identify predictive factors closely associated with hypothermia. A predictive model was developed and visualized. The performance of the nomogram was evaluated through ROC curve, calibration curve, and decision curve analysis. Results: A total of 221 patients were included, of which 47 developed hypothermia and 174 did not. Logistic regression analysis revealed that intraoperative blood loss, room temperature at admission, and operating room temperature were independent factors influencing intraoperative hypothermia in pediatric limb fracture patients (P < 0.05). The AUC of the predictive model was 0.895 (95% CI: 0.856~0.934), with good consistency and predictive ability. Clinical decision curve analysis showed that the model provided clinical benefit at various threshold values. Conclusion: Intraoperative blood loss, room temperature at admission, and operating room temperature are independent risk factors for intraoperative or postoperative hypothermia in pediatric limb fracture patients. The risk prediction model developed based on these factors demonstrates high accuracy and applicability, providing a reliable tool for predicting and managing intraoperative hypothermia.