计算机科学

计算机科学 ›› 2025, Vol. 52 ›› Issue (9): 88-95.doi: 10.11896/jsjkx.250300012

• 智能医学工程 • 上一篇    下一篇

基于动态超图与药物处方信息融合的时序健康事件预测

吴晗禹1,2, 刘天赐1,2, 矫拓成3, 车超1,2   

  1. 1 大连大学先进设计与智能计算省部共建教育部重点实验室 辽宁 大连 116622
    2 大连大学软件工程学院 辽宁 大连 116622
    3 沈阳建筑大学中新国际工程学院 沈阳 110168
  • 收稿日期:2025-03-03 修回日期:2025-07-01 出版日期:2025-09-15 发布日期:2025-09-11
  • 通讯作者: 车超(chechao@dlu.edu.cn)
  • 作者简介:(wuhanyu@s.dlu.edu.cn)
  • 基金资助:
    国家自然科学基金(62076045);高等学校学科创新引智基地(D23006);大连市重大基础研究项目(2023JJ11CG002);国家外国专家项目(D20240244)

DHMP:Dynamic Hypergraph-enhanced Medication-aware Model for Temporal Health EventPrediction

WU Hanyu1,2, LIU Tianci1,2, JIAO Tuocheng3, CHE Chao1,2   

  1. 1 Key Laboratory of Advanced Design and Intelligent Computing Ministry of Education,Dalian University,Dalian,Liaoning 116622,China
    2 School of Software Engineering,Dalian University,Dalian,Liaoning 116622,China
    3 Zhongxin International College of Engineering,Shenyang Jianzhu University,Shenyang 110168,China
  • Received:2025-03-03 Revised:2025-07-01 Online:2025-09-15 Published:2025-09-11
  • About author:WU Hanyu,born in 1999,postgraduate.His main research interests include medical informatics and data mining.
    CHE Chao,born in 1981,Ph.D,professor.His main research interests include medical informatics,natural language processing and mining.
  • Supported by:
    National Natural Science Foundation of China(62076045),111 Project(D23006),Dalian Major Projects of Basic Research(2023JJ11CG002) and National Foreign Expert Project of China(D20240244).

PDF (PC)

11

摘要： 时序健康事件预测是医疗人工智能领域的核心挑战之一。针对电子健康记录中药物与诊断复杂关联的建模难题,提出了DHMP模型。首先,通过动态子图学习机制,有效捕捉疾病演变的局部特征;其次,设计多超图融合架构,首次实现药物协同作用与诊断关联的联合建模;最后,开发时间感知注意力算法,精准解析诊疗记录中的长期依赖关系。在MIMIC-III和MIMIC-V两大临床数据集上的实验表明,DHMP模型将诊断预测准确率提升至26.68%,风险预测AUC达到90.65%,显著优于现有最佳方法。临床医生评估显示,模型预测结果与医学认知的一致性达89%,所提模型为智能辅助诊断提供了可靠工具。

关键词: 动态子图学习, 图神经网络, 药物相互作用, 时序健康事件预测, 临床决策支持

Abstract: Temporal health event prediction remains a fundamental challenge in medical Al.To address the critical problem of modeling complex medication-diagnosis relationships in EHR data this paper propose the DHMP model.Firstly,a dynamic subgraph learning mechanism captures local di-sease progression patterns.Secondly,a novel multi-hypergraph fusion architecture jointly models drug interactions and diagnosis associations.Finally,a temporal attention algorithm deciphering long-term depen-dencies in clinical records.Extensive experiments on MIMIC-III and MIMIC-IV datasets demonstrate that DHMP model has state-of-the-art performance,achieving 26.68% w-F1 in diagnosis prediction and 90.65% AUC in risk prediction.Clinical evaluation shows 89% consistency between model predictions and medical expertise,proving its reliability for decision support.

Key words: Dynamic subgraph learning, Graph neural network, Drug interaction, Temporal health event prediction, Clinical decision support

中图分类号: 

  • TP391
[1]LU C,REDDY C K,NING Y.Self-supervised graph learningwith hyperbolic embedding for temporal health event prediction[J].IEEE Transactions on Cybernetics,2021,53(4):2124-2136.
[2]YANG X M,ZHAN X C,CHENG H L,et al.Diagnosis Prediction Based on Graph Convolutional Network and Attention Mechanism[J].Computer Science,2023,50(S2):599-604.
[3]WANG M X,XIONG Y.Disease Diagnosis Prediction Algorithm Based on Contrastive Learning[J].Computer Science,2023,50(7):46-52.
[4]ZHANG X,PENG X,GUAN H,et al.Fusion of Dynamic Hypergraph and Clinical Event for Sequential Diagnosis Prediction[C]//2023 IEEE 29th International Conference on Parallel and Distributed Systems(ICPADS).2023:1620-1627.
[5]LU C,HAN T,NING Y.Context-aware health event prediction via transition functions on dynamic disease graphs[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2022:4567-4574.
[6]CHOI E,BAHADORI M T,SUN J,et al.Retain:An interpretable predictive model for healthcare using reverse time attention mechanism[C]//Advances in Neural Information Processing Systems.2016.
[7]MA F,CHITTA R,ZHOU J,et al.Dipole:Diagnosis prediction in healthcare via attention-based bidirectional recurrent neural networks[C]//Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.2017:1903-1911.
[8]LU C,REDDY C K,CHAKRABORTY P,et al.Collaborativegraph learning with auxiliary text for temporal event prediction in healthcare[J].arXiv:2105.07542,2021.
[9]NGUYEN P,TRAN T,WICKRAMASINGHE N,et al.Deepr:a convolutional net for medical records[J].IEEE Journal of Biomedical and Health Informatics,2016,21(1):22-30.
[10]LI H,LI W H,CHEN W,et al.Diagnostic Prediction Based on Node2vec and Knowledge Attention Mechanism[J].Computer Science,2021,48(S2):630-637.
[11]LU H,UDDIN S.Disease prediction using graph machine learning based on electronic health data:a review of approaches and trends[C]//Healthcare.2023.
[12]ZHANG X,PENG X,CHEN W,et al.Multi-Channel Hypergraph Network for Sequential Diagnosis Prediction in Healthcare[C]//2024 27th International Conference on Computer Supported Cooperative Work in Design(CSCWD).2024:2937-2942.
[13]WU J,DONG Y,GAO Z,et al.Dual attention and patient similarity network for drug recommendation[J].Bioinformatics,2023,39(1):btad003.
[14]LI Y,FENG L,WANG X.Clinical Assisted Diagnosis Based on Heterogeneous Graph Medical Record Attention Network[J].Computer Engineering and Science,2023,45(9):1701-1710.
[15]SHANG J,MA T,XIAO C,et al.Pre-training of graph augmented transformers for medication recommendation[J].arXiv:1906.00346,2019.
[16]JIANG J,WEI Y,FENG Y,et al.Dynamic hypergraph neural networks[C]//IJCAI.2019:2635-2641.
[17]XU R,ALI M K,HO J C,et al.Hypergraph transformers for ehr-based clinical predictions[J].AMIA Summits on Translational Science Proceedings,2023,2023:582-591.
[18]KATZ D,BAPTISTA J,AZEN S,et al.Obtaining confidence intervals for the risk ratio in cohort studies[J].Biometrics,1978,34(3):469-474.
[19]HIDALGO C A,BLUMM N,BARABÁSI A L,et al.A dynamic network approach for the study of human phenotypes[J].PLoS Computational Biology,2009,5(4):e1000353.
[20]LU H,UDDIN S.A disease network-based recommender system framework for predictive risk modelling of chronic diseases and their comorbidities[J].Applied Intelligence,2022,52(9):10330-10340.
[21]XU Z,ZHANG Q,YIP P S F.Predicting post-discharge self-harm incidents using disease comorbidity networks:A retrospective machine learning study[J].Journal of Affective Disorders,2020,277:402-409.
[22]CHURCH K,HANKS P.Word association norms,mutual information,and lexicography[J].Computational linguistics,1990,16(1):22-29.
[23]FENG Y,YOU H,ZHANG Z,et al.Hypergraph neural net-works[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2019:3558-3565.
[24]KIM E S,KANG W Y,ON K W,et al.Hypergraph attention networks for multimodal learning[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2020:14581-14590.
[25]JOHNSON A E,POLLARD T J,SHEN L,et al.MIMIC-III,a freely accessible critical care database[J].Scientific Data,2016,3(1):1-9.
[26]JOHNSON A E,BULGARELLI L,SHEN L,et al.MIMIC-IV,a freely accessible electronic health record dataset[J].Scientific Data,2023,10(1):1.
[27]BAI T,ZHANG S,EGLESTON B L,et al.Interpretable representation learning for healthcare via capturing disease progression through time[C]//Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mi-ning.2018:43-51.
[28]YANG Z,LIN Y,XU Y,et al.Interpretable Disease Prediction via Path Reasoning over medical knowledge graphs and admission history[J].Knowledge-Based Systems,2023,281:111082.
[29]LUO J,YE M,XIAO C,et al.Hitanet:Hierarchical time-aware attention networks for risk prediction on electronic health records[C]//Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining.2020:647-656.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市渝北区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发