Abstract
A distributed data network architecture allows sensitive individual-level and institution-level information to be stored locally under the direct control of participating data partners. It enables multidatabase comparative safety and effectiveness studies of rare exposures, rare outcomes, or specific patient populations, while providing strong protection for patient privacy and data security. In this chapter, we describe the design, development, implementation, strengths, and challenges of distributed data networks. We discuss the methodologic and data issues unique to distributed data networks, and progress that has been accomplished to address these issues. We also examine the design and analytic considerations associated with using a common data model, a common protocol, or both, in distributed data network studies. We conclude with a discussion about some of the future directions for distributed data networks.
Original language | English |
---|---|
Title of host publication | Pharmacoepidemiology |
Publisher | Wiley-Blackwell |
Pages | 617-638 |
Number of pages | 22 |
ISBN (Electronic) | 9781119413431 |
ISBN (Print) | 9781119413417 |
DOIs | |
Publication status | Published - 21 Oct 2019 |
Bibliographical note
Publisher Copyright:© 2020 John Wiley & Sons Ltd.
Keywords
- Common data model
- Common protocol
- Disease risk score
- Distributed data network
- Distributed regression
- Distributed research network
- Meta-analysis
- Propensity score