The Med Board · a Backstaff Board for clinical AI

Prove every clinical AI is reliable.
Without a single chart ever leaving the health system.

Your health system runs a stack of AI documentation, decision-support, and triage tools. The Med Board measures every one of them — how distinct each tool actually is, whether its answers are drifting, whether its outputs stay clinically grounded — then does two things with that measurement. It shows you which tools are behaviorally identical, so you can stop paying for the same capability twice. And it emits one signed, tamper-evident record per output documenting how the tool performed against its FDA PCCP change-control gate and your §1557 nondiscrimination duty — carrying none of the chart. The model is measured inside your network; only the proof crosses the boundary.

00 ·

You can’t govern what you can’t measure.

HHS §1557 · HIPAA review mapping · FDA SaMD / PCCP · ONC HTI-1 · CMS Utilization-Management

A hosted clinical AI tool is a black box. It documents, it triages, it recommends — and every one of those outputs runs through rules that put the responsibility on the health system, not the vendor. The FDA’s PCCP framework demands you control and monitor change in a SaMD tool. HHS §1557 demands the tool not discriminate across protected classes. ONC HTI-1 demands decision-support transparency, and CMS scrutinizes AI in utilization-management coverage decisions. The measurement is designed so the chart never enters the audit surface; only the signed record crosses the boundary.

None of those questions can be answered by a vendor’s 510(k) summary or a marketing claim. They require a measurement of how the tool actually behaves — taken on your fleet, repeatable by an adversary, and produced without shipping a single chart to anyone.

01 ·

What the Med Board measures.

One reading · two payoffs: a smaller bill and a defensible record

The Med Board points the Astrolabe at your fleet of clinical AI tools and grades each one on three axes. The same measurement saves money and survives a review.

Distinctness
Are two tools answering identically?
cut the duplicate licenses
Drift
Is the tool degrading over time?
catch silent model swaps
Faithfulness
Do the outputs stay grounded?
FDA PCCP · HHS §1557

The smaller bill. Health systems rarely run one AI tool — they run several, often with overlapping mandates bought by different service lines. When the Med Board clusters them by behavior, the tools that produce the same outputs collapse into the same group. Every duplicate is a license you can drop without losing a capability. The measurement names which ones.

The defensible record. The tools that survive get graded against the doctrine — and every grade is signed. When a surveyor, a payer, or a plaintiff’s expert asks how you knew the tool was reliable, you have a record, not an assurance.

02 ·

Proof for every output.

Supervision evidence, not policy promises · FDA PCCP · HHS §1557

Every time a clinical tool produces an output, the Med Board emits one signed, tamper-evident record of the output's behavior against your reliability bar — cryptographically tied to a behavioral measurement of the model, but carrying none of the underlying chart. For a CMIO, that is how you evidence a tool's behavior against its PCCP and §1557 gates without the proof ever touching PHI: the record crosses the boundary, the chart never does. When a regulator, a payer, or your own quality committee asks how the health system supervised its AI, this is the record you hand them.

The anchors are the ones your compliance office already works from: the FDA Predetermined Change Control Plan (Dec 2024 final guidance) for SaMD change control, HHS §1557 nondiscrimination, and HIPAA — with the evidence record showing PHI stayed inside the health system while each gate was documented. For EU deployments, the same chain can support Article 12-style record-keeping continuity.
03 ·

The doctrine it maps to.

Ten canonical atoms · every link is the reference home for one concept

The Med Board is not a separate model. It is the Astrolabe reading shaped for clinical practice: the contracts it asserts at runtime, the doctrinal anchors it scores against, and the substrate primitives that make the record byte-for-byte reproducible by a third party.

The audit happens in code. audit_medicine(responses, clinical_band=…) checks all five mapped review areas — HHS §1557, HIPAA data-minimization posture, FDA SaMD/PCCP, ONC HTI-1, and CMS Utilization-Management — before it returns; then it scores every doctrinal axis and emits the record through the Port — the same boundary the Law Board, the School Board, and the War Board ship through. Want the full corpus? The complete pin-and-state reference lives at the Medicine corpus surface.

04 ·

Next step.

One scoping call · one Phase 0 audit · one signed evidence record

The first engagement is a scoping conversation. Tool inventory, hosting topology, the clinical band (decision-support, ambient-documentation, utilization-management, samd, or predictive-dsi), the doctrine anchor (HHS §1557 / FDA SaMD-PCCP / ONC HTI-1 / CMS UM / system-specific), and the target review date. We return a Phase 0 scope and an evidence-record delivery date. No PHI changes hands at any stage.

——— evidence domain · medicine · HHS §1557 · HIPAA review mapping · FDA SaMD / PCCP · ONC HTI-1 · CMS Utilization ———
Medicine AI workflow · the corpus colossum

Medicine AI workflow.

69 compliance pins. 50 state constituencies. Every regulatory anchor and every state buyer has a canonical Planisphere page framed for clinical AI. Pick the pin the evaluator names; the bundle lands the cite-anchor next to every SHA. This is the corpus colossum for health-system AI governance.

69 pins· 50 states· HHS §1557 · HIPAA · FDA PCCP
CC.01

Which compliance pins apply to clinical AI?

69 compliance regimes · each one a Planisphere page framed for this fleet.
NIST AI RMF
NIST AI RMF for clinical AI
reg 09
OMB M-25-21/22
OMB M-25-21/22 for clinical AI
reg 09
DoW RAI S&IP
DoW RAI S&IP for clinical AI
reg 09
EO 14179
EO 14179 for clinical AI
reg 09
ISO/IEC 42001
ISO/IEC 42001 for clinical AI
reg 09
FIPS 140-3
FIPS 140-3 for clinical AI
reg 09
SHA-256 / FIPS 180-4
SHA-256 / FIPS 180-4 for clinical AI
reg 09
SLSA v1.0
SLSA v1.0 for clinical AI
reg 09
SBOM
SBOM for clinical AI
reg 09
SCITT
SCITT for clinical AI
reg 09
C-SCRM
C-SCRM for clinical AI
reg 09
NIST SP 800-53 Rev. 5
NIST SP 800-53 Rev. 5 for clinical AI
reg 09
CMMC
CMMC for clinical AI
reg 09
SSDF SP 800-218
SSDF SP 800-218 for clinical AI
reg 09
ATO
ATO for clinical AI
reg 09
DoW IL2 / IL4
DoW IL2 / IL4 for clinical AI
reg 09
DFARS 252.204-7012
DFARS 252.204-7012 for clinical AI
reg 09
HIPAA
HIPAA for clinical AI
reg 09
HHS §1557
HHS §1557 for clinical AI
reg 09
FDA PCCP / SaMD
FDA PCCP / SaMD for clinical AI
reg 09
FRE 702
FRE 702 for clinical AI
reg 09
ABA Model Rules
ABA Model Rules for clinical AI
reg 09
FERPA
FERPA for clinical AI
reg 09
Title VI
Title VI for clinical AI
reg 09
NIST AI RMF · GOVERN
NIST AI RMF · GOVERN for clinical AI
reg 09
NIST AI RMF · MAP
NIST AI RMF · MAP for clinical AI
reg 09
NIST AI RMF · MEASURE
NIST AI RMF · MEASURE for clinical AI
reg 09
NIST AI RMF · MANAGE
NIST AI RMF · MANAGE for clinical AI
reg 09
NIST AI 600-1 GenAI Profile
NIST AI 600-1 GenAI Profile for clinical AI
reg 09
OMB M-25-21
OMB M-25-21 for clinical AI
reg 09
OMB M-25-22
OMB M-25-22 for clinical AI
reg 09
High-Impact AI
High-Impact AI for clinical AI
reg 09
SSDF · Protect Software (PS)
SSDF · Protect Software (PS) for clinical AI
reg 09
ISO/IEC 42001 · Annex A
ISO/IEC 42001 · Annex A for clinical AI
reg 09
NIST 800-53 · AU
NIST 800-53 · AU for clinical AI
reg 09
NIST 800-53 · CA
NIST 800-53 · CA for clinical AI
reg 09
FedRAMP
FedRAMP for clinical AI
reg 09
DoW IL5 / IL6
DoW IL5 / IL6 for clinical AI
reg 09
RMF (SP 800-37)
RMF (SP 800-37) for clinical AI
reg 09
ITAR
ITAR for clinical AI
reg 09
EAR
EAR for clinical AI
reg 09
SCIF / Air-Gap Posture
SCIF / Air-Gap Posture for clinical AI
reg 09
CMS
CMS for clinical AI
reg 09
ONC HTI-1
ONC HTI-1 for clinical AI
reg 09
DSI Source Attributes
DSI Source Attributes for clinical AI
reg 09
FDA PCCP
FDA PCCP for clinical AI
reg 09
FDA SaMD 510(k)/De Novo
FDA SaMD 510(k)/De Novo for clinical AI
reg 09
GMLP
GMLP for clinical AI
reg 09
HIPAA Security Rule
HIPAA Security Rule for clinical AI
reg 09
FSMB AI / state boards
FSMB AI / state boards for clinical AI
reg 09
FRE 707 (proposed)
FRE 707 (proposed) for clinical AI
reg 09
ABA Formal Op. 512
ABA Formal Op. 512 for clinical AI
reg 09
AI Citation Sanctions
AI Citation Sanctions for clinical AI
reg 09
State Bar AI Opinions
State Bar AI Opinions for clinical AI
reg 09
Section 504
Section 504 for clinical AI
reg 09
IDEA
IDEA for clinical AI
reg 09
COPPA
COPPA for clinical AI
reg 09
OCR AI Guidance
OCR AI Guidance for clinical AI
reg 09
SOPIPA
SOPIPA for clinical AI
reg 09
Proctoring Fairness
Proctoring Fairness for clinical AI
reg 09
NYC Local Law 144
NYC Local Law 144 for clinical AI
reg 09
Colorado AI Act
Colorado AI Act for clinical AI
reg 09
Texas TRAIGA
Texas TRAIGA for clinical AI
reg 09
Utah AI Policy Act
Utah AI Policy Act for clinical AI
reg 09
Illinois HB 3773
Illinois HB 3773 for clinical AI
reg 09
California AI Transparency
California AI Transparency for clinical AI
reg 09
EU AI Act
EU AI Act for clinical AI
reg 09
EU GPAI Code of Practice
EU GPAI Code of Practice for clinical AI
reg 09
UK Pro-Innovation AI
UK Pro-Innovation AI for clinical AI
reg 09
CC.02

Where is the canonical reference for each pin?

The schema'd reference page for each pin — for evaluators who need the canonical DefinedTerm.
NIST AI RMF
reference
OMB M-25-21/22
reference
DoW RAI S&IP
reference
EO 14179
reference
ISO/IEC 42001
reference
FIPS 140-3
reference
SHA-256 / FIPS 180-4
reference
SLSA v1.0
reference
SBOM
reference
SCITT
reference
C-SCRM
reference
NIST SP 800-53 Rev. 5
reference
CMMC
reference
SSDF SP 800-218
reference
ATO
reference
DoW IL2 / IL4
reference
DFARS 252.204-7012
reference
HIPAA
reference
HHS §1557
reference
FDA PCCP / SaMD
reference
FRE 702
reference
ABA Model Rules
reference
FERPA
reference
Title VI
reference
NIST AI RMF · GOVERN
reference
NIST AI RMF · MAP
reference
NIST AI RMF · MEASURE
reference
NIST AI RMF · MANAGE
reference
NIST AI 600-1 GenAI Profile
reference
OMB M-25-21
reference
OMB M-25-22
reference
High-Impact AI
reference
SSDF · Protect Software (PS)
reference
ISO/IEC 42001 · Annex A
reference
NIST 800-53 · AU
reference
NIST 800-53 · CA
reference
FedRAMP
reference
DoW IL5 / IL6
reference
RMF (SP 800-37)
reference
ITAR
reference
EAR
reference
SCIF / Air-Gap Posture
reference
CMS
reference
ONC HTI-1
reference
DSI Source Attributes
reference
FDA PCCP
reference
FDA SaMD 510(k)/De Novo
reference
GMLP
reference
HIPAA Security Rule
reference
FSMB AI / state boards
reference
FRE 707 (proposed)
reference
ABA Formal Op. 512
reference
AI Citation Sanctions
reference
State Bar AI Opinions
reference
Section 504
reference
IDEA
reference
COPPA
reference
OCR AI Guidance
reference
SOPIPA
reference
Proctoring Fairness
reference
NYC Local Law 144
reference
Colorado AI Act
reference
Texas TRAIGA
reference
Utah AI Policy Act
reference
Illinois HB 3773
reference
California AI Transparency
reference
EU AI Act
reference
EU GPAI Code of Practice
reference
UK Pro-Innovation AI
reference
CC.03

Which state-specific holds apply on top of federal regulators?

50 state public-sector buyers · Medicine framing per state · federal floor + state-specific holds in one bundle.
Alabama
Alabama for clinical AI
reg 08
Alaska
Alaska for clinical AI
reg 08
Arizona
Arizona for clinical AI
reg 08
Arkansas
Arkansas for clinical AI
reg 08
California
California for clinical AI
reg 08
Colorado
Colorado for clinical AI
reg 08
Connecticut
Connecticut for clinical AI
reg 08
Delaware
Delaware for clinical AI
reg 08
Florida
Florida for clinical AI
reg 08
Georgia
Georgia for clinical AI
reg 08
Hawaii
Hawaii for clinical AI
reg 08
Idaho
Idaho for clinical AI
reg 08
Illinois
Illinois for clinical AI
reg 08
Indiana
Indiana for clinical AI
reg 08
Iowa
Iowa for clinical AI
reg 08
Kansas
Kansas for clinical AI
reg 08
Kentucky
Kentucky for clinical AI
reg 08
Louisiana
Louisiana for clinical AI
reg 08
Maine
Maine for clinical AI
reg 08
Maryland
Maryland for clinical AI
reg 08
Massachusetts
Massachusetts for clinical AI
reg 08
Michigan
Michigan for clinical AI
reg 08
Minnesota
Minnesota for clinical AI
reg 08
Mississippi
Mississippi for clinical AI
reg 08
Missouri
Missouri for clinical AI
reg 08
Montana
Montana for clinical AI
reg 08
Nebraska
Nebraska for clinical AI
reg 08
Nevada
Nevada for clinical AI
reg 08
New Hampshire
New Hampshire for clinical AI
reg 08
New Jersey
New Jersey for clinical AI
reg 08
New Mexico
New Mexico for clinical AI
reg 08
New York
New York for clinical AI
reg 08
North Carolina
North Carolina for clinical AI
reg 08
North Dakota
North Dakota for clinical AI
reg 08
Ohio
Ohio for clinical AI
reg 08
Oklahoma
Oklahoma for clinical AI
reg 08
Oregon
Oregon for clinical AI
reg 08
Pennsylvania
Pennsylvania for clinical AI
reg 08
Rhode Island
Rhode Island for clinical AI
reg 08
South Carolina
South Carolina for clinical AI
reg 08
South Dakota
South Dakota for clinical AI
reg 08
Tennessee
Tennessee for clinical AI
reg 08
Texas
Texas for clinical AI
reg 08
Utah
Utah for clinical AI
reg 08
Vermont
Vermont for clinical AI
reg 08
Virginia
Virginia for clinical AI
reg 08
Washington
Washington for clinical AI
reg 08
West Virginia
West Virginia for clinical AI
reg 08
Wisconsin
Wisconsin for clinical AI
reg 08
Wyoming
Wyoming for clinical AI
reg 08
REF

Capability brief (reference).

The original Medicine AI workflow capability declaration — kept below the colossum as the long-form reference for evaluators who want it.