Regulation — healthcare-specific
HHS Section 1557 (42 USC 18116) — non-discrimination in AI-driven clinical decisions
The 2024 HHS Final Rule extended Section 1557's non-discrimination mandate explicitly to "patient care decision support tools" — the language HHS chose for clinical AI. If your ED-triage AI or inpatient early-warning system produces different outcomes across protected classes or clinical patient groups, this rule is now the one that governs the exposure.
What the regulation actually says
"… an individual shall not, on the ground prohibited under [Title VI, Title IX, the Age Discrimination Act, or Section 504] be excluded from participation in, be denied the benefits of, or be subjected to discrimination under, any health program or activity, any part of which is receiving Federal financial assistance …"
42 U.S.C. §18116(a) — the underlying statute
"A covered entity must make reasonable efforts to identify uses of patient care decision support tools in its health programs or activities that employ input variables or factors that measure race, color, national origin, sex, age, or disability … and must make reasonable efforts to mitigate the risk of discrimination resulting from the tool's use."
45 CFR 92.210 (2024 Final Rule) — the operational language for clinical AI
What this means in plain English
Two obligations for any hospital using clinical AI:
- Identify where AI-driven clinical decision support uses inputs that correlate with protected classes (race, color, national origin, sex, age, disability). "Age 65+ with 3+ comorbidities" is such an input.
- Mitigate the risk of discriminatory outcomes. Aggregate-metric monitoring does not satisfy this — the rule is about differential outcomes across patient groups.
The rule does not require the AI to be perfect. It requires the operator to be able to demonstrate reasonable efforts to detect and mitigate the class of failure the Snapshot surfaces: silent group-differential.
What triggers the exposure in the sample
H3 (age 65+, 3+ comorbidities) urgent-intervention lane rate dropped from 43% to 9%. Age is a protected class under Section 1557. A 34-percentage-point group-differential in the wrong direction that occurred silently is precisely the "adverse-impact pattern reasonably anticipated" trigger for OCR investigation.
The finding does not depend on intent. Disparate impact is the standard. The moment a hospital knows the pattern exists, "reasonable efforts to mitigate" starts running.
What the $499 Snapshot shows against this rule
- Per-patient group distributional-shape analysis — the "identify" obligation
- Documented threshold + measured differential + severity classification — the "reasonable efforts" audit trail
- 3 fix-first items scoped to that AI surface — the "mitigate" starting point
- Independent-verifier signature — stance that survives OCR question "was your assessment independent"
See the lane-shift chart that produces the finding →
$499 Snapshot. 3 business days.
Independent-verifier determination scoped to your hospital's AI surface + 3 fix-first items + signed declaration.
Buy $499
Snapshot credit applies to Baseline ($2,500) or Enterprise Attestation ($35-55K) upgrade within 30 days.