Findings indicate that CLEO provides a controllable and empirically auditable framework for supporting cross-institutional research when real-world medical data cannot be directly aggregated.
Abstract
The sharing of patient-level structured data is strictly constrained by privacy regulations and governance, creating “data silos” that hinder multi-center research. We propose the CLEO (Clean-Learn-Evaluate-Optimize), a closed-loop framework that integrates a Gaussian Mixture Model generator with Q-learning-based optimization to formalize data synthesis as a Markov Decision Process. Experimental results on a multicenter intracranial aneurysm dataset show that CLEO achieved a combined score of 0.9232 ± 0.0124, outperforming the evaluated representative comparison methods, including TVAE, Gaussian Copula, and CTGAN. In downstream TSTR evaluation, models trained on CLEO-generated data achieved an average AUC of 0.7376, indicating that the synthetic data retained useful clinical decision signals. However, Macro-F1 results also suggest that minority-class prediction remains affected by the imbalanced class distribution. Empirical privacy auditing showed a Nearest Neighbor Adversarial Accuracy score of 0.4940, suggesting low observed re-identification risk under the adopted nearest-neighbor audit setting. These findings indicate that CLEO provides a controllable and empirically auditable framework for supporting cross-institutional research when real-world medical data cannot be directly aggregated.
Abstract Background The development of robust medical AI for knowledge discovery and decision support commonly necessitates large-scale datasets from multiple institutions. However, such data aggregation is severely constrained by privacy regulations and the inherent risk of sensitive information leakage, making it difficult to navigate the utility-privacy trade-off. Objective We aimed to design a secure multiparty deep learning system that enables privacy-preserving modeling from distributed medical time-series data without centralizing raw information or exposing model parameters. Our goal was to achieve predictive accuracy comparable to nonsecure models while providing strong security and efficiency. Methods We developed a framework using threshold homomorphic encryption to securely train recurrent neural networks on distributed longitudinal data. To improve the efficiency, we proposed an optimized encrypted matrix multiplication scheme, a secure ciphertext refresh protocol, and used lightweight encryption parameters and low-degree approximated activation polynomials. The system was evaluated on 4 real-world intensive care unit datasets for tasks like mortality and sepsis prediction. Results The system demonstrated practical efficiency, requiring approximately 1 minute per training iteration for processing 125 local batches over 39 variables and 48 time steps, and scaling well with data size and participant number. Securely trained models achieved predictive performance that was comparable to, and in some cases superior to, nonsecure centralized models, highlighting their ability to learn generalizable patterns in different unseen data distributions. For example, on the PhysioNet Challenge 2012 dataset, our secure model achieved an area under the curve (AUC) of 0.8480, outperforming the nonsecure baseline AUC of 0.8404. Conclusions This work provides a viable and efficient solution for cross-institutional, privacy-preserving analysis of longitudinal medical data. The framework successfully bridges the utility-privacy gap, facilitating safer collaborative research and enabling robust knowledge discovery and decision support while adhering to strict data protection standards.
Yao Lu, Yu Tian, Tianshu Zhou et al.· JMIR Formative Research· 0 citations
Several studies use privacy-enhanced technologies (PETs) like federated learning (FL) for decentralized heart disease detection. However, conventional FL algorithms frequently suffer from model instability issues driven by dataset heterogeneity across medical institutions. This study implements a robust FL framework to address this heterogeneity while simultaneously improving communication cost efficiency within the standard federated averaging (FedAvg) protocol. We employ a deep learning neural network algorithm evaluated on the UCI Heart Disease dataset, which is distributed across multiple local clients in an independent and identically distributed (IID) manner to establish a controlled baseline. The proposed algorithm evaluates predefined local epoch (E) variations, precisely calculating the total communication costs required to reach convergence relative to conventional centralized training accuracy thresholds. Furthermore, we identify the optimal trade-off point between network communication costs and client hardware computational overhead by utilizing the multi-objective weighted-sum method (WSM). Empirical test results demonstrate that systematically increasing the local epoch configuration significantly reduces convergence communication costs, albeit at the expense of increased client computational load. The optimal architectural configuration is identified at E = 20, yielding a minimum balanced WSM score of 0.246. Other configurations cause severe system imbalance, heavily dominating either network communication cost or client computational load. Notably, our optimized approach achieves a 72.5% communication cost efficiency, significantly outperforming the 30% baseline established in prior research.
This study compares two methods of generating tabular synthetic data: synthpop, based on transparent and interpretable inferential methodologies, and Conditional Tabular-Generative Adversarial Networks (CT-GANs), which leverage deep learning approaches to reproduce complex multivariate distributions.
Flavia Mayer, Maria Laura Fazio, M. Cutillo et al.· Epidemiologia & Prevenzione· 0 citations
Cardiovascular disease risk prediction models often rely on data from a single institution or centrally pooled datasets. Extending these models across institutions could be limited by privacy regulations and constraints on sharing patient-level data. Federated learning enables collaborative model development without transferring sensitive patient data, but its application in healthcare remains challenging because datasets often differ in size, population characteristics, and outcome definitions. In this study, we present a federated deep learning approach for privacy-preserving cardiovascular disease risk prediction that integrates two population-based cohorts with different characteristics: Lifelines, including 148,230 participants meeting the study inclusion criteria with self-reported outcomes, and the Rotterdam Study, including a smaller cohort of 10,155 participants with digitally linked clinical outcomes. Model performance was primarily evaluated on the Rotterdam Study because of its complete follow-up. Deep survival models trained using federated learning achieved higher predictive performance than models trained locally without federation. For the Rotterdam Study, the C-statistic increased from 0.728 (95% CI: 0.717-0.739) to 0.739 (95% CI: 0.728-0.749). For Lifelines, the C-statistic increased from 0.783 (95% CI: 0.775-0.791) to 0.787 (95% CI: 0.780-0.792). These findings suggest that federated deep learning across heterogeneous cohorts can improve cardiovascular disease risk prediction while preserving the privacy of individual-level patient data.
Hyunho Mo, Djura Smits, Mahlet A. Birhanu et al.· 0 citations
Results show that FedDP-PALD generates private synthetic representations that preserve useful decision performance while strongly resisting membership inference, and introduces Differentially Private Prototype Mixture Aggregation (DP-PMA), which clips class-level latent prototypes and adds calibrated Gaussian noise before combining them on the server to maintain differential privacy.
M. Sk., K. Hossain, M. M. H. Shawon· arXiv.org· 0 citations
With the rapid growth of digital healthcare data and increasing concerns over data security and regulatory compliance, the need for privacy-preserving collaborative learning has become more urgent than ever. Nowadays, Cardiovascular Disease (CVD) has become the leading cause of global mortality, while traditional centralized medical model training is facing problems like severe data privacy barriers and data island. This work designs a privacy-compliant federated learning architecture to realize clinical heart disease risk prediction, FedAvg algorithm will be used to enable cross-institutional collaborative training without sharing raw patient data. At the same time, the research will adopt K-Means based non-IID to simulate real-world medical data heterogeneity. Experimental results show that the proposed framework achieves competitive performance compared with centralized training, the optimal test accuracy is 0.8704, exceeding the result of conventional centralized training. Therefore, this framework can provide a feasible privacy-preserving solution for cross-hospital clinical collaboration and offer a practical approach for future distributed medical risk prediction.
Jiazhide Liu· Mathematical Modeling and Al...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
