LINKAGE HONEST BROKER SERVICES FOR THE N3C DATA ENCLAVE

National COVID-19 Cohort Collaborative (N3C): The N3C Data Enclave is a secure platform storing harmonized clinical data provided by more than 50 contributing members. The Enclave hosts over 651 million clinical observations on over 6.5 million persons, including over 2.1 million COVID cases, amounting to more than 7.3 billion rows of data. Harmonization, anonymity, and security is accomplished through the N3C privacy-preserving record linkage (PPRL). The PPRL uses a de-identified, software-generated token applied by a data contributor to each patient record. A linkage honest broker holds the de-identified tokens and provides a service matching token generated across disparate data sets without knowledge of the identity of the patients. The de-identified tokens are held separately from data residing within the data enclave. As an illustration of the PPRL process, Regenstrief generates a series of de-identified tokens using software provided by Datavant, Inc. The tokens are provided to a data provider, such as Hospital Network Alpha (HNA). HNA strips identifying PII from its clinical records and replaces the PII with the tokens. HNA then uploads the de-identified clinical records to N3C. If researchers require additional information about those records, they provide the token to Regenstrief, who passes it on to HNA without any knowledge of the underlying clinical data. HNA then provides the appropriate response. The linkage honest broker function is critical to the PPRL, which in turn is critical to continued use of the N3C Data Enclave by researchers. The salient characteristic of this requirement is the vendor?s ability to provide uninterrupted linkage honest broker support (harmonization, anonymity, and security) for the N3C Data Enclave.