Privacy-Preserving Federated Learning Framework for Multi-Source Electronic Health Records Prognosis Prediction
Abstract
:1. Introduction
- Feature Heterogeneity: Different healthcare institutions collect varying sets of clinical measurements and patient information, resulting in non-overlapping feature spaces. Traditional FL approaches assume shared feature spaces, leading to information loss or incompatibility when features differ across institutions.
- Task Diversity: Medical institutions often have different prediction targets based on their clinical focus (e.g., readmission prediction, mortality risk, length of stay), requiring models that can extract relevant features for diverse downstream tasks.
- Limited Knowledge Transfer: Existing approaches either sacrifice performance for privacy or fail to effectively transfer knowledge across heterogeneous data sources, especially when feature spaces only partially overlap.
- We introduce a multi-channel embedding structure that accommodates both shared and institution-specific private features, allowing each institution to utilize its complete feature set without compromise.
- We propose a secure feature recalibration module that allows each institution to emphasize features most relevant to their specific prediction tasks, enabling personalized predictions while still leveraging shared knowledge.
- We implement shared low-level feature extraction layers that enable knowledge transfer across institutions while maintaining privacy, allowing institutions to benefit from collective insights without exposing sensitive information.
- We propose a secure federated clinical prediction framework, MultiProg, which builds privacy-preserving health representations by safely leveraging data from multiple institutions with various prediction tasks while maintaining strict data privacy. MultiProg helps to improve the performance of health evaluation (e.g., the prognosis of COVID-19 patients) effectively for all collaborators under data insufficiency scenarios, especially at the early stage of an emerging pandemic, without compromising sensitive patient information.
- We design a privacy-aware adaptive feature recalibration mechanism that securely adjusts the importance of different clinical features. This approach not only suppresses non-existent or less relevant features, but also enhances critical features for patients in diverse health conditions. Through shared feature extraction channels, the framework enables embedding of patients with different recorded features into a unified clinical feature space, while protecting each institution’s data characteristics and confidentiality.
- We validate our framework’s effectiveness through secure collaboration between multiple medical institutions (i.e., Tongji Hospital in China, HM Hospitals in Spain, and a private nephrology center) on various prediction tasks (i.e., length of stay prediction for COVID-19 patients, and mortality risk prediction for chronic kidney disease patients). Experimental results demonstrate improved prediction performance across all participating institutions while maintaining data privacy. The source code is available at GitHub https://github.com/anonymous20250128/MultiProg (accessed on 28 January 2025).
2. Related Work
2.1. Clinical Background
2.2. Solutions for the Data Scarcity Problem
2.3. Federated Learning for Healthcare
3. Problem Formulation
4. Methodology
4.1. Sequential Medical Records Representation
4.2. Multi-Institutional Federated Learning with Various Feature Sets
4.3. Multi-Channel Feature Recalibration
4.4. Prediction Layer
Algorithm 1 Multitask collaborative training method |
|
5. Experiments
5.1. Medical Institution Collaborators
- TJH [1]: comprises anonymized EHR data from 485 COVID-19 patients admitted to Tongji Hospital, China, between 10 January and 24 February 2020. The dataset includes 74 lab tests and vital signs, all of which are numerical features, as well as two demographic features (age and gender).
- CDSL [33]: This dataset is derived from the HM Hospitales EHR system in Spain and consists of anonymized records of 4479 patients admitted with a confirmed or suspected diagnosis of COVID-19. CDSL offers a rich variety of medical features, including comprehensive details on diagnoses, treatments, admissions, ICU stays, diagnostic imaging tests, laboratory results, and patient discharge or death status.
- ESRD [34]: The end-stage renal disease (ESRD) dataset comprises data from 656 peritoneal dialysis patients, including 13,091 visit records collected over a 12-year period, from 1 January 2006 to 1 January 2018. This longitudinal dataset features patients’ baseline information, visit records, and clinical outcomes, offering a unique perspective on long-term peritoneal dialysis treatment and patient progression.
5.2. Experimental Setup
5.2.1. Tasks and Evaluation Metrics
5.2.2. Baseline Approaches
- RNN [37] is the most popular framework to learn the abstract embedding of variable-length time series.
- GRU [38] is the basic gated recurrent unit network.
- LSTM [39] is a variant of the recurrent neural network, capable of learning long-term dependencies.
- RETAIN [16] is the deep-based reverse time attention model for analyzing EHR data. It utilizes a two-level neural attention module to attend important clinical visits and features.
- M3Care [40] is an end-to-end model compensating the missing information of the patients with missing modalities to perform clinical analysis.
- AICare [34] consists of a multi-channel feature extraction module and an adaptive feature importance recalibration module to build the health status embedding for each patient individually.
5.2.3. Implementation Details
5.3. Quantitative Analysis
6. Discussion
- Limitations and future directions in privacy protection: Our approach builds upon a traditional federated learning framework, inheriting its fundamental privacy guarantees such as secure aggregation and differential privacy. However, when dealing with heterogeneous data, specific privacy challenges remain. In particular, when data distributions across different institutions exhibit significant variations, models may become more vulnerable to membership inference attacks even with differential privacy applied [46]. In the future, we plan to explore hybrid schemes combining local differential privacy with hierarchical encryption [47], tailoring privacy protection levels for scenarios with extreme data heterogeneity.
- Missing data handling: We handle missing data using standard mean imputation for continuous features and mode imputation for categorical features. More sophisticated approaches could further improve model performance when dealing with sparse or incomplete patient records. For example, incorporating uncertainty estimation for imputed values could be integrated into our feature recalibration module, as demonstrated by Nazabal et al. in their probabilistic approach to missing data [48].
- Extended ethical and regulatory considerations: While our method complies with existing ethical standards and regulations (such as HIPAA and GDPR), regulatory gray areas may emerge in multi-institutional collaborative settings. Particularly when international institutions collaborate, differences in data protection regulations across countries may lead to compliance challenges. Additionally, our current framework lacks ethical considerations for specific disease categories (such as rare diseases), where patient identities might be exposed even with anonymized data.
7. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Conflicts of Interest
Appendix A. Online Resources
- The Tongji Hospital COVID-19 Collaborator is introduced at https://www.nature.com/articles/s42256-020-0180-7 (accessed on 24 October 2023).
- The HM Hospitals COVID-19 Collaborator is introduced at https://www.hmhospitales.com/prensa/notas-de-prensa/comunicado-covid-data-save-lives (accessed on 15 October 2023).
References
- Yan, L.; Zhang, H.T.; Goncalves, J.; Xiao, Y.; Wang, M.; Guo, Y.; Sun, C.; Tang, X.; Jing, L.; Zhang, M.; et al. An interpretable mortality prediction model for COVID-19 patients. Nat. Mach. Intell. 2020, 2, 283–288. [Google Scholar]
- Rajkomar, A.; Oren, E.; Chen, K.; Dai, A.M.; Hajaj, N.; Hardt, M.; Liu, P.J.; Liu, X.; Marcus, J.; Sun, M.; et al. Scalable and accurate deep learning with electronic health records. NPJ Digit. Med. 2018, 1, 18. [Google Scholar] [PubMed]
- Miotto, R.; Li, L.; Kidd, B.A.; Dudley, J.T. Deep patient: An unsupervised representation to predict the future of patients from the electronic health records. Sci. Rep. 2016, 6, 26094. [Google Scholar]
- Johnson, A.E.; Ghassemi, M.M.; Nemati, S.; Niehaus, K.E.; Clifton, D.A.; Clifford, G.D. Machine Learning and Decision Support in Critical Care. Proc. IEEE 2016, 104, 444–466. [Google Scholar]
- Kelly, C.J.; Karthikesalingam, A.; Suleyman, M.; Corrado, G.; King, D. Key challenges for delivering clinical impact with artificial intelligence. BMC Med. 2019, 17, 195. [Google Scholar]
- U.S. Department of Health and Human Services Pandemic Influenza Plan: 2017 Update. Available online: https://www.cdc.gov/pandemic-flu/?CDC_AAref_Val=https://www.cdc.gov/flu/pandemic-resources/pdf/pan-flu-report-2017v2.pdf (accessed on 23 November 2024).
- Ji, Y.; Ma, Z.; Peppelenbosch, M.P.; Pan, Q. Potential association between COVID-19 mortality and health-care resource availability. Lancet Glob. Health 2020, 8, e480. [Google Scholar]
- Lee, C.; Zame, W.; Yoon, J.; Van Der Schaar, M. Deephit: A deep learning approach to survival analysis with competing risks. In Proceedings of the AAAI Conference on Artificial Intelligence, New Orleans, LA, USA, 2–7 February 2018; Volume 32. [Google Scholar]
- Yuan, Y.; Xun, G.; Suo, Q.; Jia, K.; Zhang, A. Wave2vec: Learning deep representations for biosignals. In Proceedings of the 2017 IEEE International Conference on Data Mining (ICDM), New Orleans, LA, USA, 18–21 November 2017; pp. 1159–1164. [Google Scholar]
- Zheng, K.; Gao, J.; Ngiam, K.Y.; Ooi, B.C.; Yip, W.L.J. Resolving the bias in electronic medical records. In Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Halifax, NS, Canada, 13–17 August 2017; pp. 2171–2180. [Google Scholar]
- Esteban, C.; Staeck, O.; Baier, S.; Yang, Y.; Tresp, V. Predicting clinical events by combining static and dynamic information using recurrent neural networks. In Proceedings of the 2016 IEEE International Conference on Healthcare Informatics (ICHI), Chicago, IL, USA, 4–7 October 2016; pp. 93–101. [Google Scholar]
- Suresh, H.; Gong, J.J.; Guttag, J. Learning Tasks for Multitask Learning: Heterogenous Patient Populations in the ICU. arXiv 2018, arXiv:1806.02878. [Google Scholar]
- Heo, J.; Lee, H.B.; Kim, S.; Lee, J.; Kim, K.J.; Yang, E.; Hwang, S.J. Uncertainty-aware attention for reliable interpretation and prediction. In Proceedings of the Advances in Neural Information Processing Systems, Montreal, Canada, 3–8 December 2018; pp. 909–918. [Google Scholar]
- Baytas, I.M.; Xiao, C.; Zhang, X.; Wang, F.; Jain, A.K.; Zhou, J. Patient subtyping via time-aware LSTM networks. In Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Halifax NS Canada, 13–17 August 2017; pp. 65–74. [Google Scholar]
- Lee, W.; Park, S.; Joo, W.; Moon, I.C. Diagnosis Prediction via Medical Context Attention Networks Using Deep Generative Modeling. In Proceedings of the 2018 IEEE International Conference on Data Mining (ICDM), Singapore, 17–20 November 2018; pp. 1104–1109. [Google Scholar]
- Choi, E.; Bahadori, M.T.; Sun, J.; Kulas, J.; Schuetz, A.; Stewart, W. Retain: An interpretable predictive model for healthcare using reverse time attention mechanism. In Proceedings of the Advances in Neural Information Processing Systems, Barcelona, Spain, 5–10 December 2016; pp. 3504–3512. [Google Scholar]
- Ma, T.; Xiao, C.; Wang, F. Health-ATM: A Deep Architecture for Multifaceted Patient Health Record Representation and Risk Prediction. In Proceedings of the 2018 SIAM International Conference on Data Mining, San Diego, CA, USA, 3–5 May 2018; pp. 261–269. [Google Scholar]
- Qiao, Z.; Zhao, S.; Xiao, C.; Li, X.; Qin, Y.; Wang, F. Pairwise-Ranking based Collaborative Recurrent Neural Networks for Clinical Event Prediction. In Proceedings of the IJCAI, Stockholm, Sweden, 13–19 July 2018; pp. 3520–3526. [Google Scholar]
- Pham, T.; Tran, T.; Phung, D.; Venkatesh, S. Deepcare: A deep dynamic memory model for predictive medicine. In Proceedings of the Pacific-Asia Conference on Knowledge Discovery and Data Mining, Auckland, New Zealand, 19–22 April 2016; pp. 30–41. [Google Scholar]
- Choi, E.; Xiao, C.; Stewart, W.; Sun, J. Mime: Multilevel medical embedding of electronic health records for predictive healthcare. In Proceedings of the Advances in Neural Information Processing Systems, Montréal, Canada, 3–8 December 2018; pp. 4547–4557. [Google Scholar]
- Ma, F.; You, Q.; Xiao, H.; Chitta, R.; Zhou, J.; Gao, J. Kame: Knowledge-based attention model for diagnosis prediction in healthcare. In Proceedings of the 27th ACM International Conference on Information and Knowledge Management, Turin, Italy, 22–26 October 2018; pp. 743–752. [Google Scholar]
- Gupta, P.; Malhotra, P.; Vig, L.; Shroff, G. Transfer Learning for Clinical Time Series Analysis using Recurrent Neural Networks. arXiv 2018, arXiv:1807.01705. [Google Scholar]
- Gupta, P.; Malhotra, P.; Vig, L.; Shroff, G. Using Features from Pre-trained TimeNet for Clinical Predictions. In Proceedings of the 3rd International Workshop on Knowledge Discovery in Healthcare Data at IJCAI, Stockholm, Sweden, 12–15 July 2018. [Google Scholar]
- McMahan, B.; Moore, E.; Ramage, D.; Hampson, S.; y Arcas, B.A. Communication-efficient learning of deep networks from decentralized data. In Proceedings of the Artificial Intelligence and Statistics, Fort Lauderdale, FL, USA, 20–22 April 2017; pp. 1273–1282. [Google Scholar]
- Liu, B.; Yan, B.; Zhou, Y.; Yang, Y.; Zhang, Y. Experiments of federated learning for covid-19 chest x-ray images. arXiv 2020, arXiv:2007.05592. [Google Scholar]
- Sheller, M.J.; Reina, G.A.; Edwards, B.; Martin, J.; Bakas, S. Multi-Institutional Deep Learning Modeling Without Sharing Patient Data: A Feasibility Study on Brain Tumor Segmentation; Springer: Cham, Switzerland, 2018. [Google Scholar]
- Sheller, M.J.; Edwards, B.; Reina, G.A.; Martin, J.; Bakas, S. Federated learning in medicine: Facilitating multi-institutional collaborations without sharing patient data. Sci. Rep. 2020, 10, 12598. [Google Scholar]
- Murphy, S.N.; Weber, G.; Mendis, M.; Gainer, V.; Chueh, H.C.; Churchill, S.; Kohane, I. Serving the enterprise and beyond with informatics for integrating biology and the bedside (i2b2). J. Am. Med. Inform. Assoc. 2010, 17, 124–130. [Google Scholar] [PubMed]
- Huang, L.; Shea, A.L.; Qian, H.; Masurkar, A.; Deng, H.; Liu, D. Patient clustering improves efficiency of federated machine learning to predict mortality and hospital stay time using distributed electronic medical records. J. Biomed. Inform. 2019, 99, 103291. [Google Scholar]
- Purushotham, S.; Meng, C.; Che, Z.; Liu, Y. Benchmark of Deep Learning Models on Large Healthcare MIMIC Datasets. arXiv 2017, arXiv:1710.08531. [Google Scholar]
- Ma, L.; Zhang, C.; Wang, Y.; Ruan, W.; Wang, J.; Tang, W.; Ma, X.; Gao, X.; Gao, J. Concare: Personalized clinical feature embedding via capturing the healthcare context. In Proceedings of the AAAI Conference on Artificial Intelligence, New York, NY, USA, 7–12 February 2020; Volume 34, pp. 833–840. [Google Scholar]
- Kingma, D.P.; Ba, J. Adam: A method for stochastic optimization. arXiv 2014, arXiv:1412.6980. [Google Scholar]
- Hospitales, H. Covid Data Save Lives. 2020. Available online: https://www.hmhospitales.com/prensa/notas-de-prensa/comunicado-covid-data-save-lives (accessed on 5 June 2024).
- Ma, L.; Zhang, C.; Gao, J.; Jiao, X.; Yu, Z.; Zhu, Y.; Wang, T.; Ma, X.; Wang, Y.; Tang, W.; et al. Mortality prediction with adaptive feature importance recalibration for peritoneal dialysis patients. Patterns 2023, 4, 100892. [Google Scholar]
- Kim, M.; Hwang, K.B. An empirical evaluation of sampling methods for the classification of imbalanced data. PLoS ONE 2022, 17, e0271260. [Google Scholar]
- Davis, J.; Goadrich, M. The relationship between Precision-Recall and ROC curves. In Proceedings of the 23rd International Conference on Machine Learning, Pittsburgh, PA, USA, 25–29 June 2006; pp. 233–240. [Google Scholar]
- Gao, J.; Yang, C.; Heintz, J.; Barrows, S.; Albers, E.; Stapel, M.; Warfield, S.; Cross, A.; Sun, J. MedML: Fusing medical knowledge and machine learning models for early pediatric COVID-19 hospitalization and severity prediction. Iscience 2022, 25, 104970. [Google Scholar]
- Chung, J.; Gulcehre, C.; Cho, K.; Bengio, Y. Empirical evaluation of gated recurrent neural networks on sequence modeling. arXiv 2014, arXiv:1412.3555. [Google Scholar]
- Data, M.C.; Pirracchio, R. Mortality prediction in the icu based on mimic-ii results from the super icu learner algorithm (sicula) project. In Secondary Analysis of Electronic Health Records; Springer: Berlin/Heidelberg, Germany, 2016; pp. 295–313. [Google Scholar]
- Zhang, C.; Chu, X.; Ma, L.; Zhu, Y.; Wang, Y.; Wang, J.; Zhao, J. M3care: Learning with missing modalities in multimodal healthcare data. In Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, Washington, DC, USA, 14–18 August 2022; pp. 2418–2428. [Google Scholar]
- Paszke, A.; Gross, S.; Massa, F.; Lerer, A.; Bradbury, J.; Chanan, G.; Killeen, T.; Lin, Z.; Gimelshein, N.; Antiga, L.; et al. Pytorch: An imperative style, high-performance deep learning library. Adv. Neural Inf. Process. Syst. 2019, 32, 8024–8035. [Google Scholar]
- Falcon, W.A. Pytorch Lightning. GitHub 2019. Available online: https://github.com/PyTorchLightning/pytorch-lightning, (accessed on 15 December 2024).
- Gao, J.; Zhu, Y.; Wang, W.; Wang, Z.; Dong, G.; Tang, W.; Wang, H.; Wang, Y.; Harrison, E.M.; Ma, L. A comprehensive benchmark for COVID-19 predictive modeling using electronic health records in intensive care. Patterns 2024, 5, 100951. [Google Scholar]
- Zhu, Y.; Wang, W.; Gao, J.; Ma, L. PyEHR: A Predictive Modeling Toolkit for Electronic Health Records. 2023. Available online: https://github.com/yhzhu99/pyehr (accessed on 22 November 2024).
- Loshchilov, I.; Hutter, F. Decoupled weight decay regularization. arXiv 2017, arXiv:1711.05101. [Google Scholar]
- Nasr, M.; Shokri, R.; Houmansadr, A. Comprehensive privacy analysis of deep learning: Passive and active white-box inference attacks against centralized and federated learning. In Proceedings of the 2019 IEEE Symposium on Security and Privacy (SP), San Francisco, CA, USA, 20–22 May 2019; pp. 739–753. [Google Scholar]
- Xu, R.; Baracaldo, N.; Zhou, Y.; Anwar, A.; Ludwig, H. Hybridalpha: An efficient approach for privacy-preserving federated learning. In Proceedings of the 12th ACM Workshop on Artificial Intelligence and Security, London, UK, 15 November 2019; pp. 13–23. [Google Scholar]
- Nazabal, A.; Olmos, P.M.; Ghahramani, Z.; Valera, I. Handling incomplete heterogeneous data using vaes. Pattern Recognit. 2020, 107, 107501. [Google Scholar]
TJH | CDSL | ESRD |
---|---|---|
All | ||
albumin | ALBUMINA | Albumin |
Urea | UREA | Urea |
hemoglobin | Hemoglobina | Hemoglobin |
... | ... | ... |
TJH-CDSL | ||
HCO3- | HCO3 | - |
Eosinophil | Eosinfilos | - |
... | ... | |
TJH-ESRD | ||
WBC | - | WBC |
Potassium | - | Potassium |
... | ... | |
CDSL-ESRD | ||
- | CLORO | Cl |
... | ... |
Methods | TJH | CDSL | ESRD | |||
---|---|---|---|---|---|---|
AUPRC (↑) | AUROC (↑) | AUPRC (↑) | AUROC (↑) | AUPRC (↑) | AUROC (↑) | |
RNN | 98.08 ± 1.46 | 98.52 ± 1.16 | 50.83 ± 4.07 | 79.56 ± 2.08 | 40.17 ± 4.63 | 49.09 ± 4.48 |
GRU | 98.12 ± 1.70 | 98.34 ± 1.41 | 80.21 ± 4.01 | 95.87 ± 0.95 | 42.70 ± 5.20 | 50.63 ± 4.47 |
LSTM | 98.62 ± 1.54 | 99.01 ± 1.20 | 64.60 ± 4.86 | 89.94 ± 1.55 | 65.07 ± 7.00 | 77.72 ± 3.88 |
RETAIN | 98.78 ± 1.18 | 99.13 ± 0.81 | 75.88 ± 4.12 | 93.66 ± 1.17 | 65.24 ± 6.08 | 75.32 ± 3.84 |
AICare | 99.14 ± 0.82 | 99.11 ± 0.82 | 83.42 ± 3.71 | 95.78 ± 1.00 | 69.11 ± 6.08 | 76.30 ± 4.08 |
M3Care | 97.20 ± 2.63 | 98.36 ± 1.57 | 71.63 ± 4.69 | 92.22 ± 1.49 | 70.42 ± 6.01 | 75.75 ± 4.27 |
MultiProg-2 | 99.68 ± 1.68 | 99.83 ± 1.34 | 84.88 ± 2.82 | 96.91 ± 0.95 | - | - |
MultiProg | 99.70 ± 1.59 | 99.78 ± 1.38 | 87.45 ± 4.50 | 97.89 ± 1.03 | 61.84 ± 6.77 | 78.34 ± 4.13 |
Methods | TJH | CDSL | ||||
---|---|---|---|---|---|---|
MSE (↓) | RMSE (↓) | MAE (↓) | MSE (↓) | RMSE (↓) | MAE (↓) | |
RNN | 38.40 ± 16.30 | 6.05 ± 1.34 | 3.74 ± 0.77 | 5.64 ± 2.07 | 2.33 ± 0.44 | 0.63 ± 0.07 |
GRU | 33.51 ± 17.14 | 5.58 ± 1.53 | 3.15 ± 0.78 | 5.64 ± 2.08 | 2.33 ± 0.44 | 0.58 ± 0.07 |
LSTM | 38.20 ± 18.70 | 5.97 ± 1.59 | 3.05 ± 0.84 | 5.70 ± 2.10 | 2.35 ± 0.44 | 0.49 ± 0.07 |
RETAIN | 44.21 ± 20.73 | 6.46 ± 1.58 | 3.79 ± 0.88 | 5.83 ± 2.01 | 2.38 ± 0.42 | 0.82 ± 0.07 |
AICare | 38.87 ± 20.40 | 5.99 ± 1.73 | 2.95 ± 0.86 | 5.49 ± 2.06 | 2.30 ± 0.44 | 0.55 ± 0.07 |
M3Care | 34.29 ± 16.92 | 5.66 ± 1.49 | 3.17 ± 0.79 | 5.69 ± 2.07 | 2.34 ± 0.44 | 0.59 ± 0.07 |
MultiProg | 33.02 ± 9.68 | 5.69 ± 0.82 | 4.38 ± 0.60 | 3.91 ± 0.97 | 1.96 ± 0.26 | 0.75 ± 0.05 |
Dataset | MultiProg Performance | Best Competitor | Statistical Comparison | |||
---|---|---|---|---|---|---|
AUPRC(↑) | AUROC(↑) | AUPRC | AUROC | p-Value | Significance | |
TJH | 99.70 ± 1.59 | 99.78 ± 1.38 | AICare (99.14 ± 0.82) | RETAIN (99.13 ± 0.81) | ***/*** | |
CDSL | 87.45 ± 4.50 | 97.89 ± 1.03 | AICare (83.42 ± 3.71) | GRU (95.87 ± 0.95) | 0.00/0.00 | ***/*** |
ESRD | 61.84 ± 6.77 | 78.34 ± 4.13 | M3Care (70.42 ± 6.01) | LSTM (77.72 ± 3.88) | ***/* |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Zhao, H.; Sui, D.; Wang, Y.; Ma, L.; Wang, L. Privacy-Preserving Federated Learning Framework for Multi-Source Electronic Health Records Prognosis Prediction. Sensors 2025, 25, 2374. https://doi.org/10.3390/s25082374
Zhao H, Sui D, Wang Y, Ma L, Wang L. Privacy-Preserving Federated Learning Framework for Multi-Source Electronic Health Records Prognosis Prediction. Sensors. 2025; 25(8):2374. https://doi.org/10.3390/s25082374
Chicago/Turabian StyleZhao, Huiya, Dehao Sui, Yasha Wang, Liantao Ma, and Ling Wang. 2025. "Privacy-Preserving Federated Learning Framework for Multi-Source Electronic Health Records Prognosis Prediction" Sensors 25, no. 8: 2374. https://doi.org/10.3390/s25082374
APA StyleZhao, H., Sui, D., Wang, Y., Ma, L., & Wang, L. (2025). Privacy-Preserving Federated Learning Framework for Multi-Source Electronic Health Records Prognosis Prediction. Sensors, 25(8), 2374. https://doi.org/10.3390/s25082374