{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,14]],"date-time":"2026-04-14T07:26:52Z","timestamp":1776151612745,"version":"3.50.1"},"reference-count":26,"publisher":"Oxford University Press (OUP)","issue":"1","license":[{"start":{"date-parts":[[2022,10,10]],"date-time":"2022-10-10T00:00:00Z","timestamp":1665360000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/academic.oup.com\/pages\/standard-publication-reuse-rights"}],"funder":[{"DOI":"10.13039\/100000133","name":"Agency for Healthcare Research and Quality","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100000133","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100006093","name":"Patient-Centered Outcomes Research Institute","doi-asserted-by":"publisher","award":["K12HS026379"],"award-info":[{"award-number":["K12HS026379"]}],"id":[{"id":"10.13039\/100006093","id-type":"DOI","asserted-by":"publisher"}]},{"name":"National Institutes of Health\u2019s"},{"DOI":"10.13039\/100006108","name":"National Center for Advancing Translational Sciences","doi-asserted-by":"publisher","award":["KL2TR002492"],"award-info":[{"award-number":["KL2TR002492"]}],"id":[{"id":"10.13039\/100006108","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100006108","name":"National Center for Advancing Translational Sciences","doi-asserted-by":"publisher","award":["UL1TR002494"],"award-info":[{"award-number":["UL1TR002494"]}],"id":[{"id":"10.13039\/100006108","id-type":"DOI","asserted-by":"publisher"}]},{"name":"The University of Minnesota Office of the Vice President of Research"},{"DOI":"10.13039\/100000070","name":"National Institute of Biomedical Imaging and Bioengineering","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100000070","id-type":"DOI","asserted-by":"publisher"}]},{"name":"MIDRC","award":["75N92020C00008"],"award-info":[{"award-number":["75N92020C00008"]}]},{"name":"MIDRC","award":["75N92020C00021"],"award-info":[{"award-number":["75N92020C00021"]}]},{"DOI":"10.13039\/100000001","name":"National Science Foundation","doi-asserted-by":"publisher","award":["#1928481"],"award-info":[{"award-number":["#1928481"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Division of Electrical, Communication & Cyber Systems"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2022,12,13]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:sec>\n                  <jats:title>Objective<\/jats:title>\n                  <jats:p>Federated learning (FL) allows multiple distributed data holders to collaboratively learn a shared model without data sharing. However, individual health system data are heterogeneous. \u201cPersonalized\u201d FL variations have been developed to counter data heterogeneity, but few have been evaluated using real-world healthcare data. The purpose of this study is to investigate the performance of a single-site versus a 3-client federated model using a previously described Coronavirus Disease 19 (COVID-19) diagnostic model. Additionally, to investigate the effect of system heterogeneity, we evaluate the performance of 4 FL variations.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Materials and methods<\/jats:title>\n                  <jats:p>We leverage a FL healthcare collaborative including data from 5 international healthcare systems (US and Europe) encompassing 42 hospitals. We implemented a COVID-19 computer vision diagnosis system using the Federated Averaging (FedAvg) algorithm implemented on Clara Train SDK 4.0. To study the effect of data heterogeneity, training data was pooled from 3 systems locally and federation was simulated. We compared a centralized\/pooled model, versus FedAvg, and 3 personalized FL variations (FedProx, FedBN, and FedAMP).<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>We observed comparable model performance with respect to internal validation (local model: AUROC 0.94 vs FedAvg: 0.95, P\u2009=\u2009.5) and improved model generalizability with the FedAvg model (P\u2009&amp;lt;\u2009.05). When investigating the effects of model heterogeneity, we observed poor performance with FedAvg on internal validation as compared to personalized FL algorithms. FedAvg did have improved generalizability compared to personalized FL algorithms. On average, FedBN had the best rank performance on internal and external validation.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Conclusion<\/jats:title>\n                  <jats:p>FedAvg can significantly improve the generalization of the model compared to other personalization FL algorithms; however, at the cost of poor internal validity. Personalized FL may offer an opportunity to develop both internal and externally validated algorithms.<\/jats:p>\n               <\/jats:sec>","DOI":"10.1093\/jamia\/ocac188","type":"journal-article","created":{"date-parts":[[2022,10,10]],"date-time":"2022-10-10T20:10:50Z","timestamp":1665432650000},"page":"54-63","source":"Crossref","is-referenced-by-count":28,"title":["Evaluation of federated learning variations for COVID-19 diagnosis using chest radiographs from 42 US and European hospitals"],"prefix":"10.1093","volume":"30","author":[{"given":"Le","family":"Peng","sequence":"first","affiliation":[{"name":"Department of Computer Science and Engineering, University of Minnesota , Minneapolis, Minnesota, USA"}]},{"given":"Gaoxiang","family":"Luo","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Engineering, University of Minnesota , Minneapolis, Minnesota, USA"}]},{"given":"Andrew","family":"Walker","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Engineering, University of Minnesota , Minneapolis, Minnesota, USA"}]},{"given":"Zachary","family":"Zaiman","sequence":"additional","affiliation":[{"name":"Department of Computer Science, Emory University , Atlanta, Georgia, USA"}]},{"given":"Emma K","family":"Jones","sequence":"additional","affiliation":[{"name":"Department of Surgery, University of Minnesota , Minneapolis, Minnesota, USA"}]},{"given":"Hemant","family":"Gupta","sequence":"additional","affiliation":[{"name":"Fairview Health Services , Minneapolis, Minnesota, USA"}]},{"given":"Kristopher","family":"Kersten","sequence":"additional","affiliation":[{"name":"Nvidia Corporation , Santa Clara, California, USA"}]},{"given":"John L","family":"Burns","sequence":"additional","affiliation":[{"name":"The School of Medicine, Indiana University , Indianapolis, Indiana, USA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4803-3632","authenticated-orcid":false,"given":"Christopher A","family":"Harle","sequence":"additional","affiliation":[{"name":"Department of Health Outcomes and Biomedical Informatics, University of Florida , Gainesville, Florida, USA"}]},{"given":"Tanja","family":"Magoc","sequence":"additional","affiliation":[{"name":"University of Florida College of Medicine , Gainesville, Florida, USA"}]},{"given":"Benjamin","family":"Shickel","sequence":"additional","affiliation":[{"name":"Department of Medicine, University of Florida , Gainesville, Florida, USA"},{"name":"Intelligent Critical Care Center, University of Florida , Gainesville, Florida, USA"}]},{"given":"Scott D","family":"Steenburg","sequence":"additional","affiliation":[{"name":"Department of Radiology and Imaging Sciences, Indiana University School of Medicine , Indianapolis, Indiana, USA"}]},{"given":"Tyler","family":"Loftus","sequence":"additional","affiliation":[{"name":"Intelligent Critical Care Center, University of Florida , Gainesville, Florida, USA"},{"name":"Department of Surgery, University of Florida , Gainesville, Florida, USA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5193-1663","authenticated-orcid":false,"given":"Genevieve B","family":"Melton","sequence":"additional","affiliation":[{"name":"Department of Surgery, University of Minnesota , Minneapolis, Minnesota, USA"},{"name":"Fairview Health Services , Minneapolis, Minnesota, USA"},{"name":"Center for Learning Health System Sciences, University of Minnesota , Minneapolis, Minnesota, USA"},{"name":"Institute for Health Informatics, University of Minnesota , Minneapolis, Minnesota, USA"}]},{"given":"Judy Wawira","family":"Gichoya","sequence":"additional","affiliation":[{"name":"Department of Radiology, Emory University , Atlanta, Georgia, USA"}]},{"given":"Ju","family":"Sun","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Engineering, University of Minnesota , Minneapolis, Minnesota, USA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8079-5565","authenticated-orcid":false,"given":"Christopher J","family":"Tignanelli","sequence":"additional","affiliation":[{"name":"Department of Surgery, University of Minnesota , Minneapolis, Minnesota, USA"},{"name":"Center for Learning Health System Sciences, University of Minnesota , Minneapolis, Minnesota, USA"},{"name":"Institute for Health Informatics, University of Minnesota , Minneapolis, Minnesota, USA"}]}],"member":"286","published-online":{"date-parts":[[2022,10,10]]},"reference":[{"key":"2023101001493059000_ocac188-B1","author":"Health Insurance Portability and Accountability Act of 1996","year":"2022"},{"key":"2023101001493059000_ocac188-B2","author":"General Data Protection Regulation","year":"2022"},{"key":"2023101001493059000_ocac188-B3","author":"Li","year":"2019"},{"key":"2023101001493059000_ocac188-B4","author":"Zhao","year":"2018"},{"key":"2023101001493059000_ocac188-B5","author":"Li","year":"2018"},{"key":"2023101001493059000_ocac188-B6","author":"Li","year":"2021"},{"key":"2023101001493059000_ocac188-B7","author":"Wang","year":"2020"},{"key":"2023101001493059000_ocac188-B8","author":"Huang","year":"2020"},{"key":"2023101001493059000_ocac188-B9","author":"Fallah","year":"2020"},{"key":"2023101001493059000_ocac188-B10","author":"NVIDIA Clara: An Application Framework Optimized for Healthcare and Life Sciences Developers","year":"2022"},{"issue":"4","key":"2023101001493059000_ocac188-B11","doi-asserted-by":"crossref","DOI":"10.1148\/ryai.210217","article-title":"Performance of a chest radiograph ai diagnostic tool for COVID-19: a prospective observational study","volume":"4","author":"Sun","year":"2022","journal-title":"Radiol Artif Intell"},{"issue":"7","key":"2023101001493059000_ocac188-B12","doi-asserted-by":"crossref","first-page":"610","DOI":"10.1038\/s42256-021-00338-7","article-title":"AI for radiographic COVID-19 detection selects shortcuts over signal","volume":"3","author":"DeGrave","year":"2021","journal-title":"Nat Mach Intell"},{"key":"2023101001493059000_ocac188-B13","author":"MONAI","year":"2021"},{"key":"2023101001493059000_ocac188-B14","author":"ImageNet","year":"2021"},{"key":"2023101001493059000_ocac188-B15","author":"BIMCV-COVID19","year":"2021"},{"key":"2023101001493059000_ocac188-B16","author":"Kingma","year":"2014"},{"key":"2023101001493059000_ocac188-B17","author":"PyTorch: An Imperative Style, High-Performance Deep Learning Library","year":"2022"},{"issue":"10","key":"2023101001493059000_ocac188-B18","first-page":"YE01","article-title":"Requirements for minimum sample size for sensitivity and specificity analysis","volume":"10","author":"Bujang","year":"2016","journal-title":"J Clin Diagn Res"},{"issue":"1","key":"2023101001493059000_ocac188-B19","doi-asserted-by":"crossref","first-page":"32","DOI":"10.1002\/1097-0142(1950)3:1<32::AID-CNCR2820030106>3.0.CO;2-3","article-title":"Index for rating diagnostic tests","volume":"3","author":"Youden","year":"1950","journal-title":"Cancer"},{"issue":"3","key":"2023101001493059000_ocac188-B20","doi-asserted-by":"crossref","first-page":"837","DOI":"10.2307\/2531595","article-title":"Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach","volume":"44","author":"DeLong","year":"1988","journal-title":"Biometrics"},{"issue":"2","key":"2023101001493059000_ocac188-B21","doi-asserted-by":"crossref","first-page":"153","DOI":"10.1007\/BF02295996","article-title":"Note on the sampling error of the difference between correlated proportions or percentages","volume":"12","author":"Mc","year":"1947","journal-title":"Psychometrika"},{"issue":"10","key":"2023101001493059000_ocac188-B22","doi-asserted-by":"crossref","first-page":"1735","DOI":"10.1038\/s41591-021-01506-3","article-title":"Federated learning for predicting clinical outcomes in patients with COVID-19","volume":"27","author":"Dayan","year":"2021","journal-title":"Nat Med"},{"key":"2023101001493059000_ocac188-B23","author":"Tackling the Objective Inconsistency Problem in Heterogeneous Federated Optimization","year":"2022"},{"key":"2023101001493059000_ocac188-B24","author":"Bonawitz","year":"2019"},{"key":"2023101001493059000_ocac188-B25","author":"Federated Multi-Task Learning","year":"2022"},{"issue":"3","key":"2023101001493059000_ocac188-B26","doi-asserted-by":"crossref","first-page":"50","DOI":"10.1109\/MSP.2020.2975749","article-title":"Federated learning: challenges, methods, and future directions","volume":"37","author":"Li","year":"2020","journal-title":"IEEE Signal Process Mag"}],"container-title":["Journal of the American Medical Informatics Association"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/jamia\/advance-article-pdf\/doi\/10.1093\/jamia\/ocac188\/46416986\/ocac188.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/jamia\/article-pdf\/30\/1\/54\/47829262\/ocac188.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/jamia\/article-pdf\/30\/1\/54\/47829262\/ocac188.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,10,10]],"date-time":"2023-10-10T01:50:36Z","timestamp":1696902636000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/jamia\/article\/30\/1\/54\/6754819"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,10,10]]},"references-count":26,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2022,10,10]]},"published-print":{"date-parts":[[2022,12,13]]}},"URL":"https:\/\/doi.org\/10.1093\/jamia\/ocac188","relation":{},"ISSN":["1067-5027","1527-974X"],"issn-type":[{"value":"1067-5027","type":"print"},{"value":"1527-974X","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2023,1,1]]},"published":{"date-parts":[[2022,10,10]]}}}