OPENMHRA Data IntegrityICH Q10

Development Evidence Inventory

Mark what you hold across ten classes of development evidence and get the gaps ranked by reconstructibility — whether the record can still be made later. A literature justification can be written at any point; a contemporaneous observation of what a batch actually did cannot, and that loss is silent.

OUTPUT

Gaps ranked by what stops being obtainable

TIME

~10 min

Limitations — read before you rely on this

  • This is an inventory of what you say you hold. It is not an assessment of a submission, a pathway or a regulatory status, and it makes no determination of any kind.
  • SPEQ is not a validated system. Reproduce this inventory inside your own quality system, where it can be maintained against the actual records rather than against a form.
  • The ten classes are general to regulated product development. A specific framework, product or jurisdiction can require evidence this list does not name.
  • Stage and status are self-declared, so both the ranking and the coverage figure reflect your own assessment and carry no weight with any authority.

WHAT THIS CALCULATES

Which classes of development evidence are outstanding at the stage a programme has reached, ranked by whether the record can still be created later — so what stops being obtainable is listed before what merely remains undone.

THE METHOD

due = classes where firstNeededAtStage <= stage ; rank by reconstructibility ; coverage = held / due
due
The classes of evidence normally first needed at or before the stage entered.
firstNeededAtStage
The journey stage, one to ten, by which a class of evidence is normally first relied on.
stage
Where the programme is, self-declared against the ten-stage startup journey.
reconstructibility
Whether a missing record can still be created later: no, at growing cost, or freely. This is the ranking property and the reason the tool exists.
held
Due classes marked held or partial.
coverage
Held over due, across the classes due at the stated stage. A coverage figure, never a readiness verdict.

Classes marked not applicable are excluded from coverage and still listed, so an exclusion stays a decision rather than becoming an omission.

THE INPUTS, AND WHAT THEY MEAN

Where is the programme?
The stage on the ten-step journey from bench result to postmarket. It decides which classes count as due; it does not change what any class is or how recoverable it is.
Status of each evidence class
Held, partial, not started, or not applicable, for each of the ten classes. Not applicable is a deliberate answer and is reported back rather than silently dropped, because an unexamined exclusion is the failure this inventory exists to surface.
[ E65 · EVIDENCE INVENTORY ]

Which records stop being possible if you wait?

Set where the programme is and mark what you hold. SPEQ ranks the gaps by whether the record can still be made later — because a literature justification can be written at any point, and a contemporaneous observation of what a batch actually did cannot. That loss is silent: nothing fails, the record simply never existed.

PERMANENTLY AT RISK
4 records
Due at this stage, not held, and not recreatable later. Everything else on this page can wait longer than these can.
COVERAGE OF WHAT IS DUE
0%
Held or partial, over the classes due at stage 3. A coverage figure, never a readiness verdict.
CANNOT BE RECREATED LATER (4)

Start these now. The window closes without any signal that it has.

  • Intended-use statement and its revision history

    A claim rewritten later cannot show what it said when the data was collected, and evidence is only valid for the claim it was gathered against.

  • Contemporaneous experimental records

    A record written after the fact is a recollection. The contemporaneity is the property being relied on, and it cannot be added back.

  • Material and reagent provenance

    Once a lot is consumed and its paperwork discarded, the link between the result and the material it came from is gone.

  • Instrument identity, calibration and configuration at time of use

    A calibration certificate issued today says nothing about the instrument’s state on the day the data was taken.

RECOVERABLE, AT A COST THAT GROWS (1)

Still possible, and steadily more expensive as people and context move on.

  • Rationale for development decisions taken

    The reasoning can be reconstructed while the people who held it are still reachable, and not after they leave.

ASSEMBLABLE WHEN NEEDED (1)

Real work, and deliberately last: nothing is lost by doing it later.

  • Literature and standards basis

    Published sources remain available, so this can be assembled when needed — which is exactly why it should not consume attention early.

NOT YET DUE (4)

Plan for these; they are not outstanding at the stage you set.

  • Risk analysis and the changes it drove

    A risk file can be authored later, but a file with no trace of having changed anything reads as written for the reader rather than used by the team.

  • Change history for the product and the process

    Partial history can be rebuilt from source control, orders and notebooks, at considerable effort and with gaps that stay visible.

  • Supplier selection basis and agreements

    A supplier can be re-qualified later, but work already performed under no agreement stays unqualified.

  • Who did the work, and what qualified them to

    CVs and role records can be gathered afterwards for people still with the organisation, and not for those who have left.

PROFESSIONAL EXPORT

HOW TO READ THE OUTPUT

  • The first list is the one that matters. A contemporaneous record not made today is not a task deferred — it is a record that will never exist, and nothing signals the moment the window closed.
  • The second list is real work that is still possible and steadily more expensive: decision rationale and supplier basis can be rebuilt while the people who hold them are reachable, and not after they leave.
  • The third list is deliberately last. A literature and standards basis can be assembled whenever it is needed, so it should not consume attention early — ranking it alongside the others is how founders spend their first year writing the recoverable material.
  • Coverage answers "how much of what is due do I hold", not "am I ready". Nothing here assesses a submission.

WORKED EXAMPLE

A two-person company at stage three has notebooks, no supplier agreements, and no record of which reagent lots produced its lead results.

Stage
3 — Building the evidence
Contemporaneous experimental records
Partial
Material and reagent provenance
Not started
Literature and standards basis
Not started

RESULT

Material provenance is listed as permanently at risk; the literature basis is listed last, as assemblable when needed.

The instinct is to write the literature review, because it is the piece that looks like a document. The inventory inverts that: the lots behind the lead results are being consumed now and their paperwork discarded, and once that link is gone no later effort restores it — while the literature will still be there next year.

REGULATORY BASIS

MHRA GXP Data Integrity Guidance 2018
Sets out the attributable, legible, contemporaneous, original and accurate expectations that make contemporaneity a property of the record rather than an aspiration, which is what makes it unrecoverable.
ICH Q10
Establishes that knowledge accumulated across development is a resource the quality system manages, which is the basis for treating decision rationale as evidence rather than as context.
ICH Q8(R2)
Frames development as a documented rationale linking product understanding to the eventual control strategy — the reason the reasoning behind a decision is itself a record.
PROFESSIONAL · WORKED SCENARIOS · SPEQ SYNTHESIS

See this tool applied to real cases

CHECKING ACCESS

Checking your Professional access…