Counterfactual and Factual Reasoning over Hypergraphs for Interpretable Clinical Predictions on EHR
Counterfactual and Factual Reasoning over Hypergraphs for Interpretable Clinical Predictions on EHR
复制标题
DOI：
--
复制
发表时间：
2022-11
期刊：
影响因子：
--
通讯作者：
Ran Xu;Yue Yu;Chao Zhang;Carl Yang;C. Yang;Xu Zhang;Ali Ho Yang
中科院分区：
文献类型：
--
作者：
Ran Xu;Yue Yu;Chao Zhang;Carl Yang;C. Yang;Xu Zhang;Ali Ho Yang
电子健康记录建模对于数字医学至关重要。然而，现有的模型忽略了医学代码之间的高阶相互作用及其对下游临床预测的因果关系。为了解决这些限制，我们提出了一个新的框架CACHE，提供有效的和有见地的临床预测的基础上超图表示学习和反事实和事实推理技术。在两个真实的EHR数据集上的实验表明了CACHE的上级性能。领域专家的案例研究说明了CACHE在产生对正确预测的临床有意义的解释方面的首选能力。
Electronic Health Record modeling is crucial for digital medicine. However, existing models ignore higher-order interactions among medical codes and their causal relations towards downstream clinical predictions. To address such limitations, we propose a novel framework CACHE, to provide effective and insightful clinical predictions based on hypergraph representation learning and counterfactual and factual reasoning techniques. Experiments on two real EHR datasets show the superior performance of CACHE. Case studies with a domain expert illustrate a preferred capability of CACHE in generating clinically meaningful interpretations towards the correct predictions.