Systematic Review of Privacy-Preserving Distributed Machine Learning From Federated Databases in Health Care

Abstract

Big data for health care is one of the potential solutions to deal with the numerous challenges of health care, such as rising cost, aging population, precision medicine, universal health coverage, and the increase of noncommunicable diseases. However, data centralization for big data raises privacy and regulatory concerns.Covered topics include (1) an introduction to privacy of patient data and distributed learning as a potential solution to preserving these data, a description of the legal context for patient data research, and a definition of machine/deep learning concepts; (2) a presentation of the adopted review protocol; (3) a presentation of the search results; and (4) a discussion of the findings, limitations of the review, and future perspectives.Distributed learning from federated databases makes data centralization unnecessary. Distributed algorithms iteratively analyze separate databases, essentially sharing research questions and answers between databases instead of sharing the data. In other words, one can learn from separate and isolated datasets without patient data ever leaving the individual clinical institutes.Distributed learning promises great potential to facilitate big data for medical application, in particular for international consortiums. Our purpose is to review the major implementations of distributed learning in health care.

FIG 1.

Relationship between artificial intelligence, machine learning, and deep learning.

FIG 2.

Schematic representation of the processes in a transparent distributed learning network. (A) Data preparation steps. (B) Distributed learning network, which is composed of three hospitals, each of which is equipped with a learning machine that can communicate with a master machine responsible for sending model parameters and checking convergence criteria. (C) Flowchart of the distributed learning network described in B. (D) Example of an action that can be tracked by blockchain (designed and implemented according to needs agreed among network members) and keep all network participants aware of any new activity taken in the network. DB, database; FAIR, findable, accessible, interoperable, reusable.

FIG 3.

Description of findable, accessible, interoperable, reusable (FAIR) principles.

FIG 4.

Visual representation of blockchain. Adapted from Rennock et al.

FIG A1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2009 flow diagram.