Know before
you commit.

Protocol readiness and synthetic trial data, before the trial is real. Every claim cited to its source, every decision owned by a person.

Book a protocol preflight (opens in a new tab)
Sentences light up, beams carry them to fact stones; an unsupported fact stays dark until a person decides.

The window everyone
else skips

  1. Concept
  2. Protocol draft
  3. Antevra No real data yet; decisions still cheap
  4. Protocol final
  5. Study start
  6. Data: where most study software begins

Reviews: a register,
not a summary

A summary tells you what a document says. A record tells you who decided what, on what evidence, and what is still unanswered.

  1. A source bundle in: protocol, SAP, registry, label, publications
  2. Facts and assumptions out, each cited to a source revision
  3. Evidence status set by inspectable rules, which can only downgrade
  4. Conflicts and questions routed to named owners
  5. Append-only decisions, then a gate with no override
  6. The Readiness Pack
How Reviews works

Register · STUDY-A · protocol v0.3

F-07 PK samples at 0.5 h post-doseSupported

A-03 Washout of 7 days is sufficientAssumption · owner: clin pharm

C-01 SAP §6.1 and protocol §9.2 disagree on a visit windowConflict · open

Readiness Pack exportGated: 1 open conflict

A seed star grows DM, VS, PC and EX stacks; the same seed always grows the same data.

Simulations: data
before first patient

An SDTM-shaped synthetic dataset generated from the protocol’s own extracted design, reproducible from a seed, with a validation report, an assumptions report and a CRF traceability spec. Build and test SDTM and ADaM programs months earlier. No patient data is used.

For building and testing before real data exists. Not a replacement for real data.

How Simulations works

One sentence, traced four ways

  1. Protocol sentence “PK samples at 0.5 h post-dose” · §9.2
  2. Design value timepoint = PT0.5H
  3. CRF field PC.collection_time
  4. SDTM variable PC.PCTPT, PCTPTNUM

Also built,
and demonstrable

Protocol Design Risk. A finished protocol in; a ranked, cited risk report out, with ICH M11 section detection.

Oncology Endpoint Studio. RECIST derived from tumor measurements, two readers, visit windows, two named PFS censoring conventions.

Assurance. A Context of Use and a seven-step credibility dossier, with the model and prompt pinned for each stage.

What to expect

No patient data. Protocols, documents and synthetic data only, by design.

20 to 40 minutes for a full Reviews analysis. Partial re-analysis keeps your decisions.

No accuracy figure. We publish 30 golden eval cases and what is not measured instead. Verification

Bring one protocol.See its register of facts and assumptions and an analyzable synthetic dataset. Preflight (opens in a new tab)