目的 构建并验证危重患儿桡动脉置管失败的风险预测模型。方法 选取2023年2月—2024年4月广东省某三级甲等专科医院收治的405例患儿为建模组，2024年5月—8月收治的111例患儿作为验证组。收集患儿的一般资料、诊疗情况、实验室检查及穿刺信息，通过单因素及Logistic回归分析构建危重患儿桡动脉置管失败风险预测模型，绘制列线图并评价效果。结果 全组516例患儿中，置管失败287例，总失败率为55.62%（287/516）。Logistic回归分析表明，身高（长）（OR=0.978）、急性生理学与慢性健康状况评分系统Ⅱ（OR=1.095）、操作者动脉置管年限（OR=0.737）、镇静（OR=0.415）、动脉搏动情况[动脉搏动一般（OR=3.292）、动脉搏动消失（OR=7.984）]及血红蛋白（OR=0.492）是危重儿童桡动脉置管失败的独立预测因子（P<0.05）。该模型建模组和验证组中的受试者操作特征曲线下面积分别为0.823和0.774。Hosmer-Lemeshow检验显示模型校准度良好（建模组χ²=7.123，P=0.624；验证组χ²=9.040，P=0.434）。临床决策曲线显示，该模型具有较好的临床实用性。结论 该研究构建的风险预测模型具有良好的预测性能，有助于护士早期识别桡动脉置管困难患儿，为制订个性化置管方案提供参考依据。

Objective Construct and validate a risk prediction model for failure of radial artery catheterization in critically ill children. Methods A total of 405 children admitted to a Grade A tertiary specialized hospital in Guangdong Province from February 2023 to April 2024 were enrolled as the modeling cohort，and 111 children admitted from May to August 2024 were selected as the validation cohort. General data，clinical diagnosis and treatment information，laboratory test results，and catheterization records of the children were prospectively collected. Univariable analysis and logistic regression analysis were performed to construct a risk prediction model for radial artery catheterization failure in critically ill children. A nomogram was plotted and the model efficacy was evaluated. Results Among the 516 children in the group，287 cases experienced catheterization failure，with a total failure rate of 55.62%（287/516）.Logistic regression analysis identified body height（length）（OR=0.978），APACHE Ⅱ score（OR=1.095），operator’s catheterization experience（OR=0.737），sedation（OR=0.415），radial artery pulsation status（moderate pulse：OR=3.292，absent pulse：OR=7.984） and hemoglobin level（OR=0.492） as independent predictors of catheterization failure. The area under the receiver operating characteristic curve of the model in the modeling cohort and the validation cohort were 0.823 and 0.774，respectively. The Hosmer-Lemeshow test indicated good model calibration（modeling cohort：χ²=7.123，P=0.624；validation cohort：χ²=9.040，P=0.434）. The clinical decision curves demonstrated that the model had good clinical utility. Conclusion The risk prediction model established in this study exhibits good predictive performance，which facilitates early identification of children with difficult radial artery catheterization by nurses and provides a reference for formulating individualized catheterization protocols.