CoChat
Sign Up
Cited by 226Privacy-Preserving Technologies in DataArtificial Intelligence in Healthcare and EducationEthics in Clinical Research

Privacy-first health research with federated learning

Open access
Sadilek, Adam, Liu, Luyang, Nguyen, Dung, Kamruzzaman, Methun, Serghiou, Stylianos, Rader, Benjamin, Ingerman, Alex, Mellem, Stefan +10 more·2021·npj Digital Medicine
DOI 10.1038/s41746-021-00489-2PMC PMC8423792

Published version of a preprint

Open this paper in CoChat — free accountOpen PDF

Abstract

Privacy protection is paramount in conducting health research. However, studies often rely on data stored in a centralized repository, where analysis is done with full access to the sensitive underlying content. Recent advances in federated learning enable building complex machine-learned models that are trained in a distributed fashion. These techniques facilitate the calculation of research study endpoints such that private data never leaves a given device or healthcare system. We show—on a diverse set of single and multi-site health studies—that federated models can achieve similar accuracy, precision, and generalizability, and lead to the same interpretation as standard centralized statistical models while achieving considerably stronger privacy protections and without significantly raising computational costs. This work is the first to apply modern and general federated learning methods that explicitly incorporate differential privacy to clinical and epidemiological research—across a spectrum of units of federation, model architectures, complexity of learning tasks and diseases. As a result, it enables health research participants to remain in control of their data and still contribute to advancing science—aspects that used to be at odds with each other.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1038/s41746-021-00489-2

Cites

13 of 13 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

Flagged first, then by citations

Research works printed but not linked: 21

  1. [1]Zhu W Kairouz P Sun H McMahan B Li W Federated heavy hitters with differential privacy. Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics PMLR 2020 108 3837 3847
  2. [2]Hanzely, Filip, et al. Lower Bounds and Optimal Algorithms for Personalized Federated Learning. Advances in Neural Information Processing Systems 33 (2020).
  3. [5]Choudhury O Predicting adverse drug reactions on distributed health data using federated learning AMIA Annu. Symp. Proc. 2020 2019 313 322 32308824 PMC7153050
  4. [9]Sweeney L., Abu A. & Winn J. Identifying participants in the personal genome project by name. SSRN Electron J (2013). published online May. 10.2139/ssrn.2257732. · DOI
  5. [11]Chicco D Jurman G Machine learning can predict survival of patients with heart failure from serum creatinine and ejection fraction alone BMC Med. Inf. Decis. Mak. 2020 20 16 10.1186/s12911-020-1023-5 PMC6998201 32013925 · DOI
  6. [12]Johnson AEW MIMIC-III, a freely accessible critical care database Sci. Data 2016 3 160035 10.1038/sdata.2016.35 27219127 PMC4878278 · DOI · Paper page
  7. [13]Rugge M Zorzi M Guzzinati S SARS-CoV-2 infection in the Italian Veneto region: adverse outcomes in patients with cancer Nat. Cancer 2020 1 784 788 10.1038/s43018-020-0104-9 35122051 · DOI
  8. [15]Lee J Privacy-preserving patient similarity learning in a federated environment: development and analysis JMIR Med. Inform. 2018 6 e20 10.2196/medinform.7744 29653917 PMC5924379 · DOI · Paper page
  9. [17]Li T Sahu AK Talwalkar A Smith V Federated learning: challenges, methods, and future directions IEEE Signal Process Mag. 2020 37 50 60
  10. [18]Hsu, T.M.H., Qi, H. & Brown, M. Federated visual classification with real-world data distribution. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part X 16 (pp. 76-92). Springer International Publishing (2020).
  11. [19]Gong Y Fang Y Guo Y Private data analytics on biomedical sensing data via distributed computation IEEE/ACM Trans. Comput Biol. Bioinform. 2016 13 431 444 10.1109/TCBB.2016.2515610 26761861 · DOI
  12. [20]Geyer R. C., Klein T., Nabi M. Differentially private federated learning: a client level perspective. arXiv 2017; published online Dec. http://arxiv.org/abs/1712.07557 (Accessed 23 Nov 2020). · arXiv
  13. [23]Zhu L., Liu Z. & Han S. Deep leakage from gradients. arXiv. (2019). published online June. https://arxiv.org/abs/1906.08935v2 (accessed Nov 2020). · arXiv
  14. [24]Thakkar O., Ramaswamy S., Mathews R. & Beaufays F. Understanding unintended memorization in federated learning. arXiv (2020). published online June. http://arxiv.org/abs/2006.07490 (Accessed 23 Nov 2020). · arXiv
  15. [25]Carlini N., Liu C., Erlingsson Ú., Kos J. & Song D. The secret Sharer: evaluating and testing unintended memorization in neural networks. In: Proceedings of the 28th USENIX Security Symposium. 267–284 (2019).
  16. [26]Bonawitz K. et al. Practical secure aggregation for privacy preserving machine learning. In Proceedings of ACM Conference on Computer and Communications Security (ACM CCS). (2017).
  17. [27]Smith J. W., Everhart J. E., Dickson W. C., Knowler W. C. & Johannes R. S. Using the ADAP learning algorithm to forecast the onset of diabetes mellitus. In: Proceedings of the Annual Symposium on Computer Applications in Medical Care. American Medical Informatics Association. 261–265 (1988).
  18. [28]Fiebig L Avian influenza A(H5N1) in humans: new insights from a line list of World Health Organization confirmed cases, September 2006 to August 2010 Eurosurveillance 2011 16 19941 10.2807/ese.16.32.19941-en 21871222 · DOI
  19. [29]Harris PNA Risk factors for relapse or persistence of bacteraemia caused by Enterobacter spp.: a case–control study Antimicrob. Resist. Infect. Control 2017 6 14 10.1186/s13756-017-0177-0 28127422 PMC5251334 · DOI
  20. [30]Oldenburg CE Safety of azithromycin in infants under six months of age in Niger: a community randomized trial PLoS Negl. Trop. Dis. 2018 12 e0006950 10.1371/journal.pntd.0006950 30419040 PMC6258425 · DOI
  21. [31]Ohene S-A Extra-pulmonary tuberculosis: a retrospective study of patients in Accra, Ghana PLoS ONE 2019 14 e0209650 10.1371/journal.pone.0209650 30625188 PMC6326428 · DOI

Cited by

Sorted by citations

Authors
Sadilek, A., Liu, L., Nguyen, D., Kamruzzaman, M., Serghiou, S., Rader, B., Ingerman, A., Mellem, S., Kairouz, P., Nsoesie, E. O., MacFarlane, J., Vullikanti, A., Marathe, M., Eastham, P., Brownstein, J. S., Arcas, B. A. Y., Howell, M. D., & Hernandez, J.
Affiliations
  • Google (United States) (US)
  • University of Virginia (US)
  • Boston University (US)
  • Boston Children's Hospital (US)
Corresponding author
  • Sadilek, Adam
  • Hernandez, John
ORCID
Venue
npj Digital Medicine
Publisher
Springer Science and Business Media LLC
Published
2021-09-07
Volume
4
Issue
1
Pages
132-132
ISSN
2398-6352 (ISSN-L)
DOI
10.1038/s41746-021-00489-2
PMC
PMC8423792
OpenAlex
W3197552266
Cited by
226
References
13
Citations by year
2026: 61 · 2025: 61 · 2024: 41 · 2023: 39 · 2022: 23 · 2021: 1
FWCI
9.19
Type
journal article
Language
en
Open access
Licence
Locations
  • npj Digital Medicine (publishedVersion, cc-by)
  • medRxiv (acceptedVersion, cc-by-nd)
  • Europe PMC (PubMed Central) (publishedVersion, other-oa)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • National Science Foundation (1955797, 1633028, 1443054, 2028004, 2027541, 1917819, 1916805, 1918656)
  • National Institutes of Health (CNS-2028004, OAC-1916805, OAC-2027541, CCF1917819, CCF-1918656, U01CK000589, IIS-1633028)
  • Centers for Disease Control and Prevention (U01CK000589)
  • National Institute of General Medical Sciences (R01GM109718)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
  • Health Sciences › Medicine › Public Health, Environmental and Occupational Health › Ethics in Clinical Research
Keywords
federated learning, differential privacy, health research, privacy-preserving machine learning, clinical research, epidemiological studies, multi-site study, data decentralization, model generalization

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review