非酒精性脂肪性肝病预测模型的特征分析  

Characteristics Analysis of Predictive Models of Nonalcoholic Fatty Liver Disease

作　　者：高小莲[1] 肖明中[2,3] 陶军秀 艾亚婷[1] GAO Xiaolian;XIAO Mingzhong;TAO Junxiu;AI Yating(Hubei University of Chinese Medicine,Wuhan 430065,China;Hubei Provincial Hospital of TCM,Wuhan 430061,China;Hubei Provincial Institute of TCM,Wuhan 430074,China)

机构地区：[1]湖北中医药大学,湖北省武汉市430065 [2]湖北省中医院,湖北省武汉市430061 [3]湖北省中医药研究所,湖北省武汉市430074

出　　处：《中国全科医学》2021年第30期3888-3892,3899,共6页Chinese General Practice

基　　金：2020年湖北省自然科学基金创新群体项目(2020CFA023);2015年度国家中医临床研究基地业务建设第二批科研专项(JDZX2015168)。

摘　　要：作为疾病风险评估的量化工具,疾病预测模型有助于识别高危人群,为健康管理提供参考依据,而非酒精性脂肪性肝病(NAFLD)预测模型研究尚未引起中国肝病健康管理领域研究者的充分关注。本文针对性地在PubMed、Web of Science、中国知网等数据库进行了文献查阅,得到8种NAFLD预测模型,分别是脂肪肝指数(FLI)、肝脂肪变性指数(HSI)、肝脂肪百分比、Framingham脂肪变性指数(FSI)、ZJU指数、NAFLD筛查评分(NSS)、Young Jin Park模型、Mika Aizawa模型。通过归纳上述8种NAFLD预测模型的建模方法、模型的质量表现、模型的表达及应用现状发现,Logistic回归分析是主要的建模方法,模型区分度较好,模型多采取内部验证,NAFLD预测模型的研究还存在较大的开发空间。As a quantitative tool of risk estimate,the disease prediction model is helpful to identify high-risk groups and provide evidence for health management.However,the research on nonalcoholic fatty liver disease(NAFLD)prediction models has not yet attracted sufficient attention in the field of liver disease management in China.Literature review of the NAFLD prediction model was conducted using PubMed,Web of Science,and CNKI databases.Eight NAFLD prediction models were obtained,which are fatty liver index,hepatic steatosis index,percentage of liver fat,Framingham steatosis index,ZJU index,NAFLD Screening Score,Young Jin Park model,and Mika Aizawa model.Modeling methods,performance assessment,expression and application status of the models are summarized in this paper,where Logistic regression is found to be the main modeling method,as the model discrimination is effective and internal validation is mostly adopted.Overall,the NAFLD prediction model is worthy of further exploration.

关 键 词：非酒精性脂肪性肝病 预测 模型 LOGISTIC模型 

分 类 号：R575.5[医药卫生—消化系统]