A secure distributed logistic regression protocol for the detection of rare adverse drug events

Abstract

BackgroundThere is limited capacity to assess the comparative risks of medications after they enter the market. For rare adverse events, the pooling of data from multiple sources is necessary to have the power and sufficient population heterogeneity to detect differences in safety and effectiveness in genetic, ethnic and clinically defined subpopulations. However, combining datasets from different data custodians or jurisdictions to perform an analysis on the pooled data creates significant privacy concerns that would need to be addressed. Existing protocols for addressing these concerns can result in reduced analysis accuracy and can allow sensitive information to leak.ObjectiveTo develop a secure distributed multi-party computation protocol for logistic regression that provides strong privacy guarantees.MethodsWe developed a secure distributed logistic regression protocol using a single analysis center with multiple sites providing data. A theoretical security analysis demonstrates that the protocol is robust to plausible collusion attacks and does not allow the parties to gain new information from the data that are exchanged among them. The computational performance and accuracy of the protocol were evaluated on simulated datasets.ResultsThe computational performance scales linearly as the dataset sizes increase. The addition of sites results in an exponential growth in computation time. However, for up to five sites, the time is still short and would not affect practical applications. The model parameters are the same as the results on pooled raw data analyzed in SAS, demonstrating high model accuracy.ConclusionThe proposed protocol and prototype system would allow the development of logistic regression models in a secure manner without requiring the sharing of personal health information. This can alleviate one of the key barriers to the establishment of large-scale post-marketing surveillance programs. We extended the secure protocol to account for correlations among patients within sites through generalized estimating equations, and to accommodate other link functions by extending it to generalized linear models.

A Query Workflow Design to Perform Automatable Distributed Regression Analysis in Large Distributed Data Networks

“…Lastly, we have developed a workflow with three different levels of automation and the ability to review files prior to download, upload, and transmission to facilitate data partner acceptability of DRA. Future work would be to enhance our workflow to include other secure multiparty computation protocols with these functionalities [1634]. …”

“…This makes DRA a highly desirable analytic method within DDNs. Although researchers have successfully performed DRA in relatively small or simulated multi-database settings [12131415161718192021], routine implementation of the analytic method in practice is challenging. This is because convergence of some regression models common to biomedical research (e.g., logistic and Cox regression) is an iterative process that requires frequent exchanges of intermediate statistics among data partners and an analysis center (Figure 1).…”

EXpectation Propagation LOgistic REgRession (EXPLORER): Distributed privacy-preserving online model learning

“…In this paper, we focus on the horizontally partitioned data for distributed logistic regression learning in Bayesian paradigm. During the past decade, numerous privacy-preserving/secure distributed frequentist regression models for horizontally partitioned data [20, 21, 22, 23, 24, 25, 26] have been studied. For example, the DataSHIELD framework [20] provides a secure multi-site regression solution without sacrificing the model learning accuracy.…”

“…studied numerous SMC based distributed regression model [21, 22, 23, 24, 25]. Unfortunately, as mentioned by El Emam et.al in [26], aforementioned approaches can still potentially leak sensitive personal information. Therefore, the authors [26] proposed a secure distributed logistic regression protocol to offer stronger privacy/security protection.…”

“…Unfortunately, as mentioned by El Emam et.al in [26], aforementioned approaches can still potentially leak sensitive personal information. Therefore, the authors [26] proposed a secure distributed logistic regression protocol to offer stronger privacy/security protection. The computational complexity of the above protocol grows exponentially with the increase of site number.…”