Local bias assessment
What it is
The Local bias assessment is the artifact that satisfies RUAIH focus area 3 — Risk and bias reduction — at the control level.
Who signs it
The quality or equity officer, jointly with the AI committee.
A document without a signature is a draft. An assessor is checking that someone with authority put their name to it.
What an assessor asks for
Performance broken out by the subgroups you actually serve, on your own population, with a date.
What goes in it
- Subgroups assessed, and the rationale for the selection
- Data source and period
- Performance metric by subgroup with confidence intervals
- Impact ratios where a selection decision is involved
- Interpretation — what an observed difference does and does not imply
- Actions taken, or the documented reason none were
- Re-assessment date
The most common failure
Bias work done once, at selection, and never repeated. A model fair on the population you had at go-live can drift as your population, referral patterns or documentation habits change.
Where this sits
See the RUAIH crosswalk for how this control maps across CHAI’s playbooks and the NIST AI RMF.
Written and reviewed by Neel Chauhan, MD MBA, physician-executive and founder of the Healthcare AI Institute. Last reviewed 2026-07-30.
Specified from the published RUAIH focus areas and the September 2025 Joint Commission and CHAI guidance, plus the failure modes we see most often in practice.
The Institute accepts no vendor sponsorship, holds no vendor equity and takes no referral fees.