机器学习方法构建健康素养预测模型的范围综述  

Health literacy prediction models based on machine learning methods:a scoping review

作　　者：潘祥 童莺歌[1] 李怡萱 倪珂 程雯倩 辛蒙雨 胡钰滢 PAN Xiang;TONG Yingge;LI Yixuan;NI Ke;CHENG Wenqian;XIN Mengyu;HU Yuying(School of Nursing,Hangzhou Normal University,Hangzhou,Zhejiang 311121,China)

机构地区：[1]杭州师范大学护理学院,浙江杭州311121

出　　处：《预防医学》2025年第2期148-153,共6页China Preventive Medicine Journal

基　　金：教育部规划基金项目(23YJAZH136)。

摘　　要：目的对应用机器学习方法构建的健康素养预测模型的种类、构建方法和预测效果进行范围综述,为该类模型的改进和应用提供参考。方法检索中国知网、万方数据知识服务平台、维普中文科技期刊数据库、PubMed和Web of Science,收集建库至2024年5月1日发表的应用机器学习方法构建健康素养预测模型研究文献。采用预测模型偏倚风险评估工具进行文献质量评价,对纳入文献的基本特征、模型构建方法、数据来源、缺失值处理、预测因子和预测效果等进行综述。结果检索获得文献524篇,最终纳入22篇,发表时间为2007—2024年。涉及48个健康素养预测模型,其中25个偏倚风险为高风险,占52.08%,主要问题集中在缺失值处理、预测因子选择和模型评价方法。模型构建方法包括回归模型、基于树的机器学习方法、支持向量机和神经网络模型。预测因子主要包括个人、人际关系、组织和社会/政策4个层面的因素,年龄、文化程度、经济水平、健康状况和互联网使用的出现频率较高。14篇文献进行了模型内部验证,4篇进行了外部验证。42个模型报告了受试者操作特征曲线下面积,范围为0.52~0.983,区分度良好。结论应用机器学习方法构建的健康素养预测模型展现了较好的预测能力,但研究在偏倚风险、数据处理和验证规范性等方面存在不足。Objective To conduct a scoping review on the types,construction methods and predictive performance of health literacy prediction models based on machine learning methods,so as to provide the reference for the improve⁃ment and application of such models.Methods Publications on health literacy prediction models conducted using ma⁃chine learning methods were retrieved from CNKI,Wanfang Data,VIP,PubMed and Web of Science from inception to May 1,2024.The quality of literature was assessed using the Prediction Model Risk of Bias ASsessment Tool.Basic characteristics,modeling methods,data sources,missing value handling,predictors and predictive performance were re⁃viewed.Results A total of 524 publications were retrieved,and 22 publications between 2007 and 2024 were finally enrolled.Totally 48 health literacy prediction models were involved,and 25 had a high risk of bias(52.08%),with ma⁃jor issues focusing on missing value handling,predictor selection and model evaluation methods.Modeling methods in⁃cluded regression models,tree-based machine learning methods,support vector machines and neural network models.Predictors primarily encompassed factors at four aspects:individual,interpersonal,organizational and society/policy as⁃pects,with age,educational level,economic status,health status and internet use appearing frequently.Internal valida⁃tion was conducted in 14 publications,and external validation was conducted in 4 publications.Forty-two models report⁃ed the areas under the receiver operating characteristic curve,which ranged from 0.52 to 0.983,indicating good discrim⁃ination.Conclusion Health literacy prediction models based on machine learning methods perform well,but have deficiencies in risk of bias,data processing and validation.

关 键 词：健康素养 预测模型 机器学习 范围综述 

分 类 号：R193[医药卫生—卫生事业管理]