糖尿病足感染风险预测模型的系统评价

A systematic review of risk prediction models for diabetic foot infections

目的系统评价糖尿病足患者发生感染的风险预测模型,并探讨其对临床护理工作的启示。方法系统检索中国知网、维普数据库、万方数据库、中国生物医学文献数据库、CINAHL、PubMed、Web of Science、Cochrane Library、Embase、ProQuest及Ovid等中英文数据库中的相关研究,检索时限为建库至2023年12月。由2名研究者独立筛选文献,并提取模型曲线下面积及其95%置信区间(confidence interval,CI)、校准方法和预测因子,使用预测模型研究偏倚风险评估工具对模型进行质量评价。使用RevMan 5.4软件对模型预测因子的预测价值进行Meta分析。结果共纳入7篇文献,包含8个风险预测模型,所有模型均报告了曲线下面积(0.748~0.922),6个模型进行了模型校准。纳入文献总体适应性良好,1篇文献模型低适应性,7篇文献均存在高偏倚风险,主要集中在预测因子和分析方面。Meta分析结果显示,合并后的模型曲线下面积为0.831(95%CI:0.780~0.883)。糖尿病病程(OR=2.460,95%CI:1.850~3.270)、血管疾病(OR=3.110,95%CI:1.070~9.000)、神经病变(OR=3.550,95%CI:1.920~6.560)是糖尿病足感染的预测因子。结论当前糖尿病足感染风险预测模型尚处在发展阶段,建议未来对现有模型进一步优化和进行外部验证。

Objective To systematically evaluate the risk prediction models of diabetic foot infections,and to explore its implications for clinical nursing care.Methods Relevant studies in the databases of CNKI,VIP,Wanfang,CBM,CINAHL,PubMed,Web of Science,Cochrane Library,Embase,ProQuest,and Ovid were searched for the timeframe of database construction to December 2023.Literature was screened independently by 2 investigators and the area under the model curve(AUC)and its 95%confidence interval(CI),calibration method and predictors were extracted,and the quality of the model was evaluated using the Predictive Modeling Study Risk of Bias Assessment Tool.Meta-analysis of the predictive value of model predictors was performed using RevMan 5.4 software.Results 7 studies were included,containing 8 risk prediction models,all of which reported the AUC(0.748-0.922),and 6 models underwent model calibration.The overall fit of the included studies was good,among which 1 model was low fit,and all 7 studies were at high risk of bias,mainly focusing on predictors and analysis.Meta-analysis showed that the area under the curve of the combined models was 0.831(95%CI:0.780~0.883);the disease duration(OR=2.460,95%CI:1.850~3.270),vascular disease(OR=3.110,95%CI:1.070~9.000),and neuropathy(OR=3.550,95%CI:1.920~6.560)were effective predictors of diabetic foot infection.Conclusion The current diabetic foot infection risk prediction model is still in the developmental stage,and further optimization and external validation of the existing model is recommended for the future.