A wrong note got past review
Example · wrong-answer check · AI scribe · 20 scenarios × 8 runs
160 runs · model version recorded on eachReviewed by a named person
Compliance evidence
DCB0129 and IEC 62304 evidence, produced by testing.
The same testing that checks your AI’s safeguards produces the evidence behind your clinical safety case and your verification records. Your clinical safety officer owns the safety case. We produce the tested proof it relies on.
Evidence for DCB0129 and IEC 62304 · produced on your real product · every result traceable
A model update broke a safety check
Example · model-update re-check · same 20 scenarios re-run
160 runs re-checkedEvery difference reviewed by a named person
Illustration. Example values from a fictional project, not a client engagement.
What the standard asks
Two lines in DCB0129 this testing answers.
Manufacturers must show their safety controls actually work, including when things go wrong.
-
Verify every control
DCB0129 says the manufacturer “MUST verify the effectiveness of each clinical risk control measure”. We test each one and record the result.
-
Normal and fault conditions
Hazards are to be considered “in both normal and fault conditions”. Supplying wrong AI output on purpose creates exactly those fault conditions.
-
Every change
When your product or its AI model changes, the same tests run again, so the evidence stays current at every update.
Where the evidence goes
Behind your safety case and your technical file.
Two standards, one set of tests on your real product.
DCB0129
Test evidence behind your hazard log and clinical safety case report: each control, the fault supplied, what your product did, the model version and the reviewer.
IEC 62304
Traceable verification records for your technical file: requirement, test, result, version and date.
Compliance questions
What clinical safety officers ask first.
Do you provide a Clinical Safety Officer?
No. Your Clinical Safety Officer, or one you appoint, makes the clinical judgements and signs the safety case. We produce the tested evidence it relies on, in a form they can reference directly.
Do you write our safety case?
No. We produce the test evidence that sits behind it, in a form your clinical safety officer can reference directly.
Is this a certificate?
No. It is evidence. Decisions about your safety case and regulatory status stay with you and your regulatory advisers.
Our product has no AI. Is this still relevant?
Yes. The same approach produces evidence that each safety control works, whether or not AI is involved.
See what evidence your safety case is missing.
A free 45-minute session. We map your highest-risk controls and whether each one has been tested.