An acceptance criterion defines the comparison used to judge a result. If it is chosen only after the result is known, a passing label may reflect the choice of limit rather than evidence that a pre-existing requirement was met. Clear criteria make the decision reproducible.
A specification needs more than a number
ICH Q6A describes a specification as tests, analytical procedure references and acceptance criteria. Its drug-development context illustrates why a limit cannot be interpreted separately from what is measured and how.ICH Q6A — Specifications: Test Procedures and Acceptance Criteria for New Drug Substances and New Drug Products (opens in a new tab)
| Part | Example description |
|---|---|
| Property | Target content per original vial |
| Procedure | A named method and version |
| Limit | An explicitly inclusive numerical interval |
| Decision rule | How rounding and uncertainty are handled |
| Applicability | The material and stage to which it applies |
These are descriptive examples, not proposed Novum limits. A percentage without a measurement basis, or a maximum with no stated unit, can leave the outcome ambiguous even when it was written in advance.
The method also needs to be capable of assessing the chosen requirement. A limit below reliable quantification does not become measurable because it appears in a table.
See what changes when the limit follows the result
In an original example, a test has a predefined minimum of 95 units and returns 93. Moving the minimum to 92 makes the number pass the new comparison, but it does not show that 93 met the original requirement.
The change may look small, yet the logical question has changed. The original decision asked whether the material satisfied a requirement chosen independently of that outcome. The new decision may instead ask which requirement includes the observed number.
The same issue applies to choosing a rounding convention, averaging rule or exclusion after seeing which choice produces a pass. The numerical limit is only one part of the decision process.
Learning from data is still legitimate
Development data are necessary for understanding analytical and material behaviour. ICH Q6A discusses justifying proposed criteria using relevant development, testing and stability information, while considering analytical and manufacturing variability.ICH Q6A — Specifications: Test Procedures and Acceptance Criteria for New Drug Substances and New Drug Products (opens in a new tab)
That is different from changing a routine criterion solely to dispose of one inconvenient result. In development, the purpose is to establish a defensible requirement and method. In routine assessment, the result is judged against the applicable established framework.
New evidence can justify revising a method or criterion. The record should explain what was learned, why the change is appropriate, which materials it covers and when it applies. Any necessary independent or regulatory review depends on the context.
A scientifically justified revision may also prompt reassessment of earlier findings. That reassessment should preserve the original result and criteria, and clearly identify the new interpretation rather than silently rewriting history.
Check the evidence behind a pass label
Look for the specification or criterion version used, its relationship to the method version and any report note explaining a change. A pass label with no recoverable comparison is difficult to evaluate.
If two reports classify the same number differently, compare their criteria and decision rules before concluding that the analytical measurements conflict. Different rules can produce different labels without changing the measured value.
A useful clarification asks which criterion governed the assessment and whether it was already applicable when the work began. The answer should identify the relevant record, not just reassure the reader that the result was acceptable.
This makes the decision reviewable while leaving room for legitimate scientific improvement. Consistency comes from documented reasoning and version history, not from pretending that specifications can never change.
Sources and further detail
- ICH Q6A — Specifications: Test Procedures and Acceptance Criteria for New Drug Substances and New Drug Products (opens in a new tab)
Sections 1.2 and 3.1 read for specification definition and justification. Used for general decision reasoning, not a proposed product specification or advice on a regulatory filing.
Sources checked 19 September 2026. Worked examples are illustrative unless a supplied report is explicitly identified. This article has not undergone independent scientific peer review.