Clinical Phenotyping with an Outcomes-driven Mixture of Experts for Patient Matching and Risk Estimation

被引：1

作者：

Hurley N.C. ^{[1
]}

Dhruva S.S. ^{[2
]}

Desai N.R. ^{[3
]}

Ross J.R. ^{[3
]}

Ngufor C.G. ^{[4
]}

Masoudi F. ^{[5
]}

Krumholz H.M. ^{[3
]}

Mortazavi B.J. ^{[6
]}

机构：

[1] Nc Hurley, 3112 Tamu, Texas A&m University, College Station

[2] University of California San Francisco, 4150 Clement St, Bldg 200, San Francisco

[3] Center for Outcomes Research and Evaluation, Yale University, 195 Church St, New Haven, 06510, CT

[4] Mayo Clinic, 200 First St SW, Rochester, 55905, MN

[5] 3112 Tamu, Texas A&m University, College Station

来源：

ACM Transactions on Computing for Healthcare | 2023年 / 4卷 / 04期

关键词：

Cardiology; cardiovascular outcomes; machine learning; medical information systems; mixture of experts;

D O I：

10.1145/3616021

中图分类号：

学科分类号：

摘要：

Observational medical data present unique opportunities for analysis of medical outcomes and treatment decision making. However, because these datasets do not contain the strict pairing of randomized control trials, matching techniques are to draw comparisons among patients. A key limitation to such techniques is verification that the variables used to model treatment decision making are also relevant in identifying the risk of major adverse events. This article explores a deep mixture of experts approach to jointly learn how to match patients and model the risk of major adverse events in patients. Although trained with information regarding treatment and outcomes, after training, the proposed model is decomposable into a network that clusters patients into phenotypes from information available before treatment. This model is validated on a dataset of patients with acute myocardial infarction complicated by cardiogenic shock. The mixture of experts approach can predict the outcome of mortality with an area under the receiver operating characteristic curve of 0.85 ± 0.01 while jointly discovering five potential phenotypes of interest. The technique and interpretation allow for identifying clinically relevant phenotypes that may be used both for outcomes modeling as well as potentially evaluating individualized treatment effects. © 2023 Copyright held by the owner/author(s).

引用

共 41 条

[1]

Ahmad T., Pencina M.J., Schulte P.J., O'Brien E., Whellan D.J., Pina I.L., Kitzman D.W., Lee K.L., O'Connor C.M., Michael Felker G., Clinical implications of chronic heart failure phenotypes defined by cluster analysis, J. Amer. Coll. Cardiol., 64, 17, pp. 1765-1774, (2014)

[2]

Beaulieu-Jones B.K., Greene C.S., Et al., Semi-supervised learning of the electronic health record for phenotype stratification, J. Biomed. Info., 64, pp. 168-178, (2016)

[3]

Borah B.J., Moriarty J.P., Crown W.H., Doshi J.A., Applications of propensity score methods in observational comparative effectiveness and safety research: Where have we come and where should we go?, J. Compar. Effect. Res., 3, 1, pp. 63-78, (2014)

[4]

Chapfuwa P., Li C., Mehta N., Carin L., Henao R., Survival cluster analysis, Proceedings of the ACM Conference on Health, Inference, and Learning, pp. 60-68, (2020)

[5]

Dhruva S.S., Ross J.S., Mortazavi B.J., Hurley N.C., Krumholz H.M., Curtis J.P., Berkowitz A., Masoudi F.A., Messenger J.C., Parzynski C.S., Et al., Association of use of an intravascular microaxial left ventricular assist device vs intra-aortic balloon pump with in-hospital mortality and major bleeding among patients with acute myocardial infarction complicated by cardiogenic shock, J. Amer. Med. Assoc., 323, 8, pp. 734-745, (2020)

[6]

[7]

Dhruva S.S., Ross J.S., Mortazavi B.J., Hurley N.C., Krumholz H.M., Curtis J.P., Berkowitz A.P., Masoudi F.A., Messenger J.C., Parzynski C.S., Et al., Use of mechanical circulatory support devices among patients with acute myocardial infarction complicated by cardiogenic shock, J. Amer. Med. Assoc. Netw. Open, 4, 2, (2021)

[8]

Ghassemi M., Naumann T., Schulam P., Beam A.L., Chen I.Y., Ranganath R., A review of challenges and opportunities in machine learning for health, Proceedings of the AMIA Summits on Translational Science, (2020)

[9]

Guo R., Cheng L., Li J., Richard Hahn P., Liu H., A survey of learning causality with data: Problems and methods, ACM Comput. Surveys, 53, 4, pp. 1-37, (2020)

[10]

Holt G.B., Potential Simpson’s paradox in multicenter study of intraperitoneal chemotherapy for ovarian cancer, J. Clin. Oncol., 34, 9, (2016)

← 1 2 3 4 5 →