Domain generalization for enhanced predictions of hospital readmission on unseen domains among patients with diabetes.

Autor: Hai AA; Computer and Information Sciences, Temple University, Philadelphia, PA, United States of America., Weiner MG; Weill Cornell Medicine, New York, NY, United States of America., Livshits A; Lewis Katz School of Medicine, Temple University, Philadelphia, PA, United States of America., Brown JR; Departments of Epidemiology and Biomedical Data Science, Geisel School of Medicine at Dartmouth, Hanover, NH, United States of America., Paranjape A; Lewis Katz School of Medicine, Temple University, Philadelphia, PA, United States of America., Hwang W; Department of Public Health Sciences, Penn State College of Medicine, Hershey, PA, United States of America., Kirchner LH; Department of Population Health Sciences, Geisinger, Danville, PA, United States of America., Mathioudakis N; Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, MD, United States of America., French EK; Division of Endocrinology and Metabolism, University of Pittsburgh, Pittsburgh, PA, United States of America., Obradovic Z; Computer and Information Sciences, Temple University, Philadelphia, PA, United States of America., Rubin DJ; Lewis Katz School of Medicine, Temple University, Philadelphia, PA, United States of America. Electronic address: daniel.rubin@tuhs.temple.edu.
Jazyk: angličtina
Zdroj: Artificial intelligence in medicine [Artif Intell Med] 2024 Dec; Vol. 158, pp. 103010. Date of Electronic Publication: 2024 Nov 10.
DOI: 10.1016/j.artmed.2024.103010
Abstrakt: A prediction model to assess the risk of hospital readmission can be valuable to identify patients who may benefit from extra care. Developing hospital-specific readmission risk prediction models using local data is not feasible for many institutions. Models developed on data from one hospital may not generalize well to another hospital. There is a lack of an end-to-end adaptable readmission model that can generalize to unseen test domains. We propose an early readmission risk domain generalization network, ERR-DGN, for cross-domain knowledge transfer. ERR-DGN internalizes the shared patterns and characteristics that are consistent across source domains, enabling it to adapt to a new domain. It transforms source datasets to a common embedding space while capturing relevant temporal long-term dependencies of sequential data. Domain generalization is then applied on domain-specific fully connected linear layers. The model is optimized by a loss function that integrates distribution discrepancy loss to match the mean embeddings of multiple source distributions with the task-specific loss. A model was developed using electronic health record (EHR) data of 201,688 patients with diabetes across urban, suburban, rural, and mixed hospital systems to enhance 30-day readmission predictions among patients with diabetes on 67,066 unseen patients at a rural hospital. We also explored how model performance varied by the number of sites and over time. The proposed method outperformed the baseline models, yielding a 6 % increase in F1-score (0.79 ± 0.006 vs. 0.73 ± 0.007). Model performance peaked with the inclusion of three sites. Performance of the model was relatively stable for 3 years then declined at 4 years. ERR-DGN may be a proficient tool for learning data from multiple sites and subsequently applying a hospitalization readmission prediction model to a new site. Including a relatively small number of varied sites may be sufficient to achieve peak performance. Periodic retraining at least every 3 years may mitigate model degradation over time.
Competing Interests: Declaration of competing interest We hereby confirm that this work has not been published previously and is not under consideration for publication elsewhere. The study protocol was approved by the Johns Hopkins Medicine (JHM) and PennState Health Institutional Review Boards (IRB). The other three participating institutions ceded oversight to the JHM single IRB. The publication has been approved by all authors and the responsible authorities. If accepted, it will not be published elsewhere in the same form, or any other form in any other language. This paper is a substantial extension of our previously published paper at Artificial Intelligence in Medicine, AIME 2023 [1].
(Copyright © 2024 Elsevier B.V. All rights reserved.)
Databáze: MEDLINE
Externí odkaz: