儿童静脉血栓栓塞症风险预测模型的系统评价

A systematic review of risk prediction models for venous thromboembolism in children

目的:系统地评价儿童静脉血栓栓塞症(VTE)风险预测模型,为选择合适的预测模型提供客观依据,为模型的更新及新模型的构建提供参考。方法:计算机检索PubMed、Web of Science、Embase、Cochrane Library、Wiley Online Library、Scopus、ProQuest、万方、中国知网、维普网,搜集关于儿童VTE风险预测模型的构建研究,检索时限为建库至2022年1月13日。2名研究者独立筛选文献、提取资料,采用预测模型研究的偏倚风险评价工具(PROBAST)工具对纳入研究进行评价。结果:最终纳入共纳入11篇文献,共11个儿童VTE风险预测模型,其中6篇为回顾性病例对照研究,3篇为回顾性队列研究,2篇为前瞻性队列研究。10篇文献报告了受试者工作特征曲线下面积(AUC)进行内部验证,1篇研究进行了外部验证,预测性能指标AUC为0.670~0.945,10篇均>0.7,预测效能较好。预测模型中包含较多的预测因子为中心静脉导管、年龄、感染、肿瘤和手术。10篇存在较高的偏倚风险,但所有模型适用性较好。结论:儿童VTE风险预测模型的预测性能较好,但存在一定的偏倚。未来应规范报道模型构建验证的全过程,以减少方法学偏倚,应侧重于模型的外部验证和模型更新,不断改进模型预测性能,为临床提供实际可用的模型。

Objective To systematically evaluate the risk prediction model of venous thromboembolism(VTE)in children,to provide an objective basis for selecting an appropriate prediction model,and to provide a reference for model updating and new model construction.Methods Studies on VTE risk prediction model construction for children published up to January 13,2022 were retrieved from PubMed,Web of Science,Embase,Cochrane Library,Wiley Online Library,Scopus,ProQuest,WanFang,CNKI and VIP databases.Results A total of 11 articles were included,including 11 children's VTE risk prediction models,of which 6 were retrospective case-control studies,3 were retrospective cohort studies,and 2 were prospective cohort studies.Ten articles reported the area under the receiver operating characteristic curve(AUC)for internal verification,and one study carried out external verification.The AUC of the prediction performance index was 0.670-0.945,10 articles all>0.7,indicating that the prediction performance was good.The most common predictors included in the prediction model were central venous catheter,age,infection,tumor and surgery.Ten articles had a high risk of bias,but all models had good applicability.Conclusions The prediction performance of the children's VTE risk prediction model is good,but there is a certain bias.In the future,the whole process of reporting model construction and validation should be standardized to reduce methodological bias,and emphasis should be placed on external validation and model updating to continuously improve the prediction performance and provide clinically available models.