Federated and Differentially Private Learning for Electronic Health Records

TL;DR

This paper evaluates the privacy and effectiveness of federated learning combined with differential privacy for clinical risk prediction using electronic health records across multiple hospitals.

Contribution

It provides a comparative analysis of centralized and federated learning with differential privacy in healthcare, highlighting challenges in maintaining privacy in federated settings.

Findings

Differential privacy is easier to implement in centralized training.

Federated learning with differential privacy faces significant challenges.

Federated approach enables decentralized data use without sharing sensitive information.

Abstract

The use of collaborative and decentralized machine learning techniques such as federated learning have the potential to enable the development and deployment of clinical risk predictions models in low-resource settings without requiring sensitive data be shared or stored in a central repository. This process necessitates communication of model weights or updates between collaborating entities, but it is unclear to what extent patient privacy is compromised as a result. To gain insight into this question, we study the efficacy of centralized versus federated learning in both private and non-private settings. The clinical prediction tasks we consider are the prediction of prolonged length of stay and in-hospital mortality across thirty one hospitals in the eICU Collaborative Research Database. We find that while it is straightforward to apply differentially private stochastic gradient descent to achieve strong privacy bounds when training in a centralized setting, it is considerably more difficult to do so in the federated setting.