CONTEXT(CLONE)
Sign in

Review site consent forms against the protocol and master template

For: Study start-up manager or regulatory coordinator handling site and country consent form variants

Pattern: Specialist panelNeeds live modelsDesigned for 8 to 160 agents
Get early accessSign in to use

The pain today

Each site and country adapts the master consent form. After a protocol amendment, some variants still describe old procedures or risks, and required local wording is missing in others.

The ask

“I attached the current protocol, the master consent form and every site and country version. For each version, tell me what differs from the master, whether procedures, risks and data use still match the protocol, and what required elements are missing. Quote the passages.”

Plain words, as you would say it to a colleague. Edit it to fit your case before you send it.

What you attach or connect

  • Current protocol and amendment summary
  • Master informed consent form
  • Site and country consent form versions
  • Checklist of required consent elements

The unit of work

One worker task per one consent form version under one reviewer brief.

Why a swarm fits

Each version is read against the same master by reviewers with different briefs. Versions are independent and each brief needs only the matching protocol sections.

Not for

Deciding whether consent is ethically adequate, or assessing translation quality. One study with one form does not need a panel.

The decision tree

6 typed decisions, each with an action for every answer

At fixed moments in a run, the engine puts one narrow question to a decision model. The decision model never writes text: it answers yes or no with a probability, picks from listed options, or gives a score, about a small slice of the material. The engine then does exactly what this tree says, which is what makes the run auditable. The thresholds are the template's design values, not measured results.

  1. Planner, while planning

    Scope checkA choice among options

    Before work starts on a unit

    Does the version date on this consent form fall after the latest protocol amendment approval for that site?

    Sees only: The form's version footer and the site's amendment approval date

    Why: Flags forms that predate the amendment for a closer read from the start.

    • Dated after the amendmentthenAccept
    • Dated before the amendmentthenEscalate to a strong model
    • No version date foundthenMark unresolved

  2. Before workers, before a task runs

    Small worker or strong modelA choice among options

    Before a task runs

    Which reviewer brief does this form section belong to?

    Sees only: One section heading with its first paragraph

    Why: Gives each specialist only its own sections, so briefs stay narrow and cheap.

    • Procedures and risksthenAccept
    • Data use and privacythenAccept
    • Required elements and contactsthenAccept
    • Mixed or unclearthenEscalate to a strong model

  3. After workers, the judge checks

    Evidence checkYes or no, with a probability

    After a worker answers

    Does the procedure, visit or risk described in the quoted form passage appear in the quoted current protocol passage?

    Sees only: One form passage and one current protocol passage

    Why: Catches forms that still describe procedures or risks the amendment removed or changed.

    • Yes: 0.85 or higherthenAccept
    • Unsure: 0.50 up to 0.85thenEscalate to a strong model
    • No: below 0.50thenReject and retry

  4. Evidence checkA choice among options

    After a worker answers

    Does the form contain the required element named in this checklist line?

    Sees only: One checklist line and the form section where it would appear

    Why: Reports missing elements against the supplied checklist only, never against general expectations.

    • PresentthenAccept
    • Present but differs from the master wordingthenMark unresolved
    • AbsentthenMark unresolved

  5. Reconciler, while merging

    Conflict checkA choice among options

    While reconciling

    Do these two findings from different briefs concern the same passage of the same form version?

    Sees only: Two findings with their quoted passages

    Why: Keeps reviewer disagreement visible instead of merging it away.

    • Same passage, compatible findingsthenAccept
    • Same passage, contradictory findingsthenMark unresolved
    • Different passagesthenContinue

  6. Accountable person, before anything is settled

    Person decidesYes or no, with a probability

    Before anything is reported as settled

    Does the finding change what a participant is told about risks, procedures, data use or the right to withdraw?

    Sees only: One finding with form and protocol quotes

    Why: Every participant-facing change goes to the investigator and ethics committee route; the swarm approves nothing.

    Accountable: The principal investigator and the ethics committee own approval of every consent form; the sponsor's regulatory lead owns corrections.

    • Yes: 0.40 or higherthenAsk a person
    • Unsure: 0.15 up to 0.40thenAsk a person
    • No: below 0.15thenAccept

The fleet: who does what

Model tiers by role, not brands: you choose the models. Strong reasoning models plan and reconcile, small fast models do the wide work, and the judge is a decision model from a different family, so it does not share the workers' blind spots.

  1. Planner

    A strong reasoning model aligns each version to the master section by section and hands each brief its sections.

    Decisions here:1. Scope check

  2. Workers

    Small fast workers hold one brief each: procedures and risks, required elements, data protection wording, plain language.

    Designed for 8 to 160 agents, one worker task per one consent form version under one reviewer brief. Each worker receives only its own unit.

    Decisions here:2. Small worker or strong model

  3. Judge, from a different model family

    A decision model from a different family checks each reported gap against the quoted form and protocol text.

    Decisions here:3. Evidence check4. Evidence check

  4. Reconciler

    A strong reasoning model merges the briefs per version and keeps findings where two briefs disagree.

    Decisions here:5. Conflict check

  5. Accountable person

    The principal investigator and the ethics committee own approval of every consent form; the sponsor's regulatory lead owns corrections.

    Decisions here:6. Person decides

Checked before anything is accepted

  • Each finding quotes the site version and the master or protocol passage
  • Required elements are checked against the supplied checklist only
  • Translated versions are flagged for a qualified translator, not assessed for nuance

What comes back

  • Findings per consent version by brief
  • Versions still describing superseded procedures or risks
  • Missing required elements per version
  • Conflicts between reviewer briefs

What to measure

  • Findings confirmed by the regulatory coordinator
  • Ethics committee queries on reviewed forms
  • Review time per version
  • Cost per version

Names of measures only. No result is claimed for this template.

Templates open in the workspace chat with the ask filled in. Nothing runs until you send it.

Get early accessSign in to use

More in Healthcare and life sciences

Grade protocol deviations across trial sites, with evidence

For: Clinical operations lead or trial quality manager at a sponsor or contract research organisation

Deviation logs and monitoring reports pile up across sites.

Pattern: Map, verify, reduceNeeds scale6 decisionsDesigned for 40 to 400 agents

Check adverse-event narratives against their case data

For: Pharmacovigilance lead or safety writer preparing case narratives for a study report or aggregate report

Hundreds of case narratives are written from line listings.

Pattern: Cross-examinationNeeds scale5 decisionsDesigned for 30 to 300 agents