In the context of registries and clinical studies, de-identify refers to removing information from a dataset that can directly or indirectly be used to determine the identity of someone. De-identification may also be accomplished by storing the personally identifiable information (PII) of a participant in a separate secure database from other data from the participant. In this case, each participant is assigned an identifier such as a code, algorithm, or pseudonym unrelated to any PII. This assigned identifier is attached to the other data from the participant to allow data to be re-identified if needed. Similarly, de-identified data may be accessed or exported from a dataset using an assigned identifier.
NIST National Technology Laboratory Computer Security Resource Center: Guide to Protecting the Confidentiality of Personally Identifiable Information (PII)
NIH: De-identifying Protected Health Information Under the Privacy Rule
Rare Diseases Registry Program (RaDaR): Protect Human Research Participants