Skip to content

Guides

Guides to health data trust

Three long-form guides, each written to be read on its own and each dated. The newest update is from 16 November 2026.

  1. Clinicians in the loop: which records need one, and what the record must carry

    A clinician in the loop judges the answer; the record under it has to be proven first. How the DTI sandbox routes records to clinical review, and what they carry.

    Published 16 November 2026

    Read the guide
  2. How to verify a record before AI training

    Ten checks a record must pass before it enters an AI training set, each one a dimension the Data Trust Index scores at intake: integrity, provenance, consent, recency, quality, concordance, validation, breadth, stability and the floor.

    Published 14 September 2026 · updated 16 November 2026

    Read the guide
  3. Who certifies data for AI?

    As of 14 September 2026 no single body certifies data for AI. Data licensing, NCQA Data Aggregator Validation, ISO 8000, the Data & Trust Alliance provenance standards and per-record scoring with the Data Trust Index, side by side by what each certifies, who does it, what you receive, and what it says about one record.

    Published 14 September 2026

    Read the guide