This section is written to be handed to your institution’s IRB or privacy office.

Two-step, non-reversible de-identification.

• Step 1 (at your institution). Remove direct identifiers and replace them with an opaque study identifier. Keep the table mapping that identifier to your real participants at your institution. It is never transmitted.
• Step 2 (at the Data Coordination Center). The Center derives the Repository’s own subject identifier from your opaque ID using a keyed one-way transform, validates that the dataset meets the HIPAA Safe Harbor standard, and can assist with any remaining Safe Harbor de-identification where your IRB permits us to receive the data. The Repository’s subject identifier cannot be reversed to your opaque ID, and the Repository never holds the map back to your real participants.

The single most important assurance for a privacy office: the Cannabis and Health Data Repository never receives your participants’ names or original identifiers, and never receives the key that links research identifiers to real people. Direct identifiers are removed, dates and geographic detail are reduced to permissible granularity, and each participant carries an opaque, irreversible research identifier. The resource is therefore structurally unable to re-identify individuals, and recipients additionally agree by contract not to attempt re-identification.

Free text and indirect identifiers. Narrative free-text fields and any indirect identifiers that carry disclosure risk are handled separately from the openly released de-identified data. Where governance permits, they may be routed to a restricted tier: retained for research value but never released openly, and made available to a researcher only under additional controls and a separate restricted-use agreement. If applying de-identification would unduly harm the analytic value of your data, do not strip it silently: contact us and we will discuss a restricted-use arrangement instead. (The handling of free-text and PHI-bearing narrative fields, and the restricted-tier serving mechanics, are pending governance review and not yet finalized.)

Contributors retain the IRB approvals and participant consent obtained for the data they collected. Before you contribute, confirm three things:

1. You hold IRB approval for the study that produced the data.
2. Your consent permits sharing the data for secondary research.
3. Your consent describes how participant confidentiality is maintained.

If your consent language is ambiguous on secondary sharing, contact us before submitting; this is easier to resolve up front than after data has been transferred.