Independent observational research · GLP-3 is investigational and not FDA-approved
Evidence Standards

What our data can and cannot say

Observational registry data is useful and limited at the same time. We publish the tier of every claim so that no reader has to guess how strong it is.

E0 — Anecdotal

Single unverified self-report with no standardized collection or follow-up.

Lowest tier. Illustrative only.

E1 — Structured self-report

Standardized definitions, units, and cadence collected prospectively in the registry.

Usable for descriptive trends with stated limitations.

E2 — Registry analysis

Pre-specified cohort analysis of longitudinal registry data with attrition and missing-data reporting.

Primary tier for R3TA research reports.

E3 — Source-verified registry analysis

Registry analysis restricted to third-party laboratory or connected-device measurements.

Highest observational tier.

E4 — Governed clinical study

Formally governed prospective study with protocol, principal investigator, and ethics oversight.

Highest R3TA evidence tier.

Governance

Versioned consent

Consent text carries a version identifier (currently v1.0). Every observation stores the consent version in force when it was collected. Changing the consent text requires re-consent before further collection.

Separation of identity and research data

Account identity is stored separately from research records. Research records are keyed by a generated research ID, and published outputs contain no free-text identifiers.

Small-cell suppression

Any aggregate group with fewer than 5 contributors is suppressed rather than rounded, including inside filtered explorer views.

Role-based access

Contributors can read and write only their own records. Reviewers can see de-identified research data for verification and safety follow-up. Database policies enforce this, not the interface.

Provenance on every value

Each observation records source (self-report, laboratory, device), collection method, verification status, and a data-quality flag. Verified-only views can be requested in the explorer.

Safety follow-up

Reported events enter a follow-up queue with an explicit status. Reports are recorded verbatim and never edited to soften them.

Withdrawal

Withdrawal halts collection immediately and marks the record withdrawn. Aggregate analyses already published are not retroactively recomputed, and this limitation is disclosed at consent.

Pre-specified endpoints

Cohort endpoints are published before analysis. Post-hoc findings are labelled as exploratory in the report.

Boundaries

R3TA does not prescribe, recommend, assign, or direct the use of any compound, and does not provide dosing guidance or medical advice. Registry findings are descriptive. Interventional research requires a separate protocol, a named principal investigator, and independent ethics review.

For researchers and collaborators →