Skip to main content
Eligibility & QA module

TissueQA. Cited eligibility, ready for sign-off.

TissueQA is the eligibility and QA module of the DonorIQ system of record. It ingests the donor record, serology, physical assessment, your SOPs, and — optionally — your bank's own decision history, then returns a cited eligibility recommendation ready for medical director review and sign-off. Organ eligibility and tissues for research allocation are on the roadmap.

/ Structured extraction

Pull the values your SOPs care about. Skip the rest.

Configure what gets extracted per tissue type. Every field is a named, typed value, not a paragraph summary. Start from your own schema.

sample · fictional case · not real PHI

...patient presented with hemoglobin of 11.2 g/dL following admission. Prior medical history notable for hypertension (controlled). Serology performed 2026-04-18. HIV 1/2 non-reactive, HBsAg non-reactive, HBcAb reactive. Tox screen not performed at admission...

Extracted fields · per SOP schema

  • PARSED
    hgb_g_dl
    11.2
  • PARSED
    serology_date
    2026-04-18
  • WATCH
    hbcab
    reactive
  • MISSING
    tox_performed
    false

Named, typed values · skip the prose.

RULE TRIGGEREDAATB §D3.1 · NAT confirmation required

/ SOP matching

Your SOPs. One match table.

Point TissueQA at your org SOP library. Every extracted value is matched: pass, fail, or escalate, with the rule citation and source page.

Per-tissue matching

CorneaHeart valvesMSK
Org SOP §A · screening
PASS
PASS
WATCH§C1.4
Org SOP §B · serology
PASS
PASS
FAIL§D3.1
Org SOP §C · tissue rules
PASS
PASS
FAILB5 tox

One donor record · evaluated per graft against your SOPs.

Evaluation order

  1. Org SOP
  2. Standards check
  3. MD review

Org SOP overrides allowed · always logged.

/ Historical eligibility · opt-in

Your bank's history, as decision support. Not as the decision.

An optional, tenant-local pattern model built only on your own historical eligible and non-eligible cases where configured. It stays inside your tenant, is never pooled with other customers, and never feeds a shared model. When a new donor record arrives, the model surfaces similar past cases as advisory context for the reviewer.

Current SOPs always override historical pattern. When the two disagree — say a serology or risk factor that used to be acceptable is now a deferral — TissueQA shows both, flags the conflict, and requires the medical director to acknowledge before signing.

Pattern match · this donor

  • PARSED
    similar_past_cases
    23
  • PARSED
    historically_eligible
    19
  • PARSED
    historically_deferred
    4
  • PARSED
    pattern_confidence
    moderate

Advisory only · current standards override pattern.

CONFLICTCurrent AATB §C2.7 overrides historical pattern · MD ack required

Decision order

  1. Historical pattern
  2. SOP check
  3. Conflict surfaced
  4. MD sign-off

Pattern is advisory · current SOPs and reviewer judgment take precedence.

/ Citations and recorded human review

Every claim is a link. Every review is a signature.

Each AI-surfaced finding cites the exact source: page number in the donor record, the SOP clause, the regulatory section. The reviewer opens the citation, reads the source, and confirms, overrides, or escalates.

The platform records the reviewer identity, the action, and the electronic signature against the AI output. Both the model recommendation and the human decision are stored. That gives the medical director and QA a complete, reviewable record of every case.

/ Safeguards

What keeps TissueQA accountable.

Advisory outputs require accountability. These are the design constraints that keep TissueQA from operating as a black box.

Citation required
Every AI-surfaced finding must link to a specific source page, SOP clause, or regulatory section. Unsupported findings are not presented as conclusions.
Partial flags surfaced
Low-confidence extractions are labelled PARTIAL and shown to the reviewer. Nothing is silently inferred into a signed field. Uncertainty is visible before sign-off.
Reviewer override recorded
Overrides, escalations, and sign-offs are captured with reviewer identity and timestamp. The system records both the AI output and the qualified human decision.
SOP precedence
Current SOPs and qualified reviewer judgment always take precedence over AI output and historical pattern. No pattern match can override a SOP rule without reviewer acknowledgment.
Tenant-local model scope
The optional historical pattern model is scoped to that customer's own cases only. It never pools across tenants and never feeds a shared model. It is active only where separately configured.
Model version retention
The model version used for each evaluation is logged alongside the case. If a model changes, prior evaluations remain traceable to the version that produced them.

/ Standards mapped

What your SOPs typically reference.

  • AATB standards

    If your SOP references AATB, we match the citations and link them back.

  • FDA 21 CFR 1271

    HCT/P clauses your SOP cites are matched and linked back to source.

  • Your org SOPs

    Uploaded once, versioned, enforced per tissue type.

  • AABB (blood)

    If your blood-center SOP references AABB, the same matching applies.

  • Your historical decisions

    Optional tenant-local pattern model, scoped to your cases where configured. Advisory only.

Roadmap: organ eligibility, tissues for research allocation.

Ready to see TissueQA on your workflow

Bring a redacted case. We will run it.

30 minutes. We map your SOPs to TissueQA and walk a real eligibility decision end to end.