VitaSignal
Governance Decision Package
Output of the Clinical AI Documentation Readiness Evaluation
A Governance Decision Package is the written output of a VitaSignal Readiness Evaluation Scoping Conversation. It organizes the evidence about one proposed clinical AI use case and its supporting documentation pathway so that leadership can make a human-owned decision: Proceed, Modify, Pause, or Stop.
This package does not approve, clear, validate, or deploy any AI tool. It is a decision-support record for governance, quality, and informatics review.
1. Use-case boundary
One named AI use case, intended user, workflow, decision owner, and deadline. Recorded in writing before any analysis is discussed.
2. Documentation pathway review
Required fields, event time, recorded time, provenance, missingness, duplication, and mapping variation for the named use case only.
3. Evidence and limitations
What the health system already knows, what remains unresolved, what cannot be evaluated, and what additional evidence is required. Negative and blocked findings are preserved.
4. Decision record
Proceed, Modify, Pause, or Stop, with accountable owners, stopping rules, and next evidence requirements. The decision is owned by the health system.
What the package can record
- Regulatory baseline, selected optional standard, and rationale.
- Source profile, template, terminology, and version.
- Validator or test-tool name, version, date, fixtures, failures, warnings, and unresolved findings.
- Mapping loss, ambiguity, unsupported elements, and downstream effect.
- Separate structural, semantic, analytical, security, workflow, and clinical-evidence gates.
- Whether evidence was human-authored, AI-generated, AI-reviewed, mixed, or unknown.
- Generator and reviewer identities, shared dependencies, disagreement, abstention, escalation, and reference-standard requirements.
Agreement between an AI generator and an AI reviewer is not treated as independent validation. Consequential conclusions require an appropriate reference standard or qualified human review.
Supporting reference: a scholarly letter in the published literature, not an empirical VitaSignal validation study. View the letter.
Possible outcomes
Proceed
The documentation pathway appears sufficient to support the next bounded evaluation step under the stated constraints.
Modify
The use case is promising but the scope, data pathway, or governance conditions need adjustment before testing.
Pause
Existing evidence is insufficient; additional data, documentation, or stakeholder input is needed first, or an unresolved blocker must be cleared.
Stop
The question is already answered, the pathway is not feasible, or the use case falls outside VitaSignal's boundaries.
Discuss a current governance decision
If your health system is weighing one clinical AI use case and needs a structured read-only review of the documentation pathway that supports it, start with a Readiness Evaluation Scoping Conversation.
VitaSignal is a research-stage governance decision-support prototype. Not FDA cleared or approved. Not for clinical use, diagnosis, treatment, or patient-care decisioning.
Scope and fees
Engagement structure
Every VitaSignal engagement is scoped and quoted in writing. The 30-minute Readiness Evaluation Scoping Conversation is free. Feasibility review and the Clinical AI Documentation Readiness Evaluation are separately scoped and contracted.
Stage 0
Readiness Evaluation Scoping Conversation
Free 30-minute qualification and scoping call. No data commitment and no custom written deliverable.
Stage 1
Feasibility review
Optional fixed-fee feasibility review, quoted for the specific opportunity.
Stage 2
Readiness evaluation
Site-specific, phase-gated evaluation on a retrospective extract or an approved read-only pathway, separately scoped and contracted. Scope and duration are set in the written Statement of Work and depend on approvals, data readiness, and the named use case. Silent-mode observation is not included; it is a possible later phase only under separate written authorization once all applicable technical, privacy, security, governance, contractual, and evidence gates are met.
Follow-on work
Separately scoped
Nothing is included by default. Evidence maintenance or any later evaluation is separately scoped, authorized, and contracted after the readout.