老年失能预警模型的专家共识

通过采用多元异构大数据分析技术，结合专家知识系统及人工神经网络模型，在完成老年人群失能分类评估的基础上，充分考虑结合社会学因素、疾病因素、以及生物样本多组学分子生物学指标等因素，基于机器学习挖掘不同指标与失能结果之间的潜在关联，完成重要指标筛选，量化失能风险，建立老年失能精准风险预警模型。本专家共识从老年失能背景、数据采集及选择、使用图神经网络直接在因果结构上建模因素间的时序关系等总结了中国老年失能预警模型构建的过程，供相关医师参考

By leveraging multi-heterogeneous big data analysis technology, integrating expert knowledge systems and artificial neural network models, and completing the disability classification assessment of elderly populations, this study fully considers sociological factors, disease-related factors, and multi-omics molecular biology indicators derived from biological samples. It extracts potential correlations between various indicators and disability outcomes through machine learning, screens key indicators, quantifies disability risk, and establishes a precise risk early warning model for elderly disability. This expert consensus summarizes the workflow of constructing elderly disability early warning models in China, including the background of elderly disability, data collection and selection, and the application of graph neural networks (GNNs) to directly model the temporal relationships among factors based on causal structures, for reference by relevant clinicians.