Privacy Preserving Record Linkage

Module 1	Introduction	Unit 1 CODI Introduction
Module 2	Getting Started	Unit 1 How to Use this Toolbox
Module 3	Audiences/Roles	Unit 1 Researchers	Unit 2 Network partners	Unit 3 Legal	Unit 4 Implementers	Unit 5 Participants
Module 4	People & Processes	Unit 1 CODI Approach	Unit 2 CODI Roles & Scope	Unit 3 Clinical and Community Linkages Assessment	Unit 4 Governance Model
Module 5	Standards	Unit 1 Data Models	Unit 2 FHIR Implementation Guides
Module 6	Tools	Unit 1 Growthcleanr	Unit 2 GrowthViz	Unit 3 CODI Prevalence Queries	Unit 4 Privacy Preserving Record Linkage	Unit 5 CODI Use Cases
Module 7	Implementations	Unit 1 CODI Implementations
Module 8	Resources	Unit 1 CODI References	Unit 2 CODI Resources
Module 9	Appendices	Unit 1 Appendix A: CODI Downloads

Who should use Privacy Preserving Record Linkage?

Organizations looking to share data and link data where personal identifiable information (PII) is not shared should use Privacy Preserving Record Linkage (PPRL).

Roles interested in PPRL include:

What is PPRL?

PPRL is a process of matching records containing personal identifiable information (PII) across organizations without a shared, unique identifier. It often requires those organizations to exchange information with each other or a third party to participate in a matching process. Matching occurs by comparing shared PII to see if there are similarities in demographic attributes such as name, sex, date of birth, or address. PPRL is an alternative set of techniques to solve the issue of identifying matching without exchanging PII directly. PPRL takes place through a series of prescribed steps that makes it nearly impossible for an outside party to recover the PII, but still allows for the establishing of links across organizations.¹⁰

Why use PPRL?

PPRL allows you to link an individual’s data across different organizations and sectors without directly exchanging PII. PPRL can link an individual’s health information across clinical and community organizations to provide a more holistic view of a person’s or household’s health. Using PPRL minimizes the security risks in sharing or transmitting sensitive data.¹⁰

Description of PPRL tools

This PPRL toolkit contains the software packages that data owners use to extract and garble their data (“Data Owner Tools”) and that the linkage agent uses to match the hashed data to create shared individual and household identifiers (“Linkage Agent Tools”). The PPRL Implementation Guide explains how the PPRL process works, including specific steps of the process and how data flows between the different partners in the CODI network.

PPRL toolkit

- PPRL Overview: This section of tools includes an overview and introduction to PPRL, a review of CODI roles and the PPRL Process Flow for the Data Owner.
- PPRL Implementation Guide: This PDF implementation guide explains how to perform the CODI PPRL process and provides a comprehensive introduction to PPRL and explains in detail how the Data Owner Tools work.
- Data Owner Tools: These tools are a software package specifically for use by CODI data owners and data partners to garble their data, which is then sent to a Linkage Agent—a third-party organization that links the hashes. It includes the details on the PPRL process from installation to extraction to garbling to mapping LinkIDs to Patient IDs.
- Linkage Agent Tools: A set of tools for the CODI linkage agent to use to accept garbled input from data owners/partners, perform matching and generate individual and household linkage identifiers.

Additional documentation about PPRL

CODI PPRL Report: Initial results from CODI’s PPRL tool indicated lower than expected birthdate concordance for matched patients, and further analyses also suggested issues with linkage quality. To better define and fix PPRL issues, the CODI project team developed a quality assurance (QA) toolkit. The toolkit contains data quality checks at four stages of the PPRL process.