Every page is signed. The batch is released. An inspector will ask something the signatures do not answer: who authorized release, on what evidence, how anything open or borderline was weighed, and why release was justified.
Answers use this article’s own words and show the section they come from. This is not advice on your own record.
Your question is saved without your name or contact details so we can see what readers want to know. Please leave out names and record details.
A signed batch record proves the review happened. It does not always show why the batch was released.
The gap sits after the last signature.
The batch record review is finished. Production steps are signed. Test results are attached. Deviations are closed or referenced. The release signature is on the page.
By every internal check, the record is complete.
An inspector does not stop at complete. The inspector picks a batch with something in its history and asks why it went out the door. The record usually shows that someone released it. It often does not show what that person weighed to get there.
If the answer lives in memory, in a meeting nobody minuted, or in an email thread, the record has a hole exactly where the decision was made.
The Record Behind the Decision.
Every batch disposition leaves two kinds of paper. One records the event: what was made, tested, and signed. The other records the decision: what was weighed, what was accepted, and who took responsibility.
The first kind is usually thorough. The second kind is usually a signature, a date, and a phrase such as “reviewed and approved for release.”
The regulations point at both. 21 CFR 211.188 sets out what a batch production and control record must include, among it in-process and laboratory control results and any investigation made under 211.192. 21 CFR 211.192 requires production and control records to be reviewed and approved by the quality control unit before a batch is released or distributed, and requires unexplained discrepancies and specification failures to be thoroughly investigated, extending to other batches that may be associated, with a written record of conclusions and followup. 21 CFR 211.22(a) gives the quality control unit the responsibility and authority to approve or reject drug products.
Approve or reject is a decision. The record has to carry it.
Four questions an inspector can ask of a release.
Ask them in this order. Authority comes first, because a decision without a clear owner cannot be examined further.
2. On what evidence?
What did that person have in front of them? Review findings, the status of every deviation and OOS result, the test results including any retest. Can you separate what existed on the release date from what was assembled afterward?
3. How was remaining risk weighed?
If anything was open, borderline, or explained by assumption, what could follow from releasing, and who accepted it? Silence here reads as risk that was never considered.
4. Why was release justified?
What alternatives existed (hold, reject, extend the investigation), why were they set aside, and how does the evidence lead to release? This is the reasoning itself, and it is the part most often missing.
An illustrative release, and where it breaks.
Consider a fictional batch of a solid oral dose product. During review, an assay result for one in-process sample came back near the lower limit. A retest was run and passed. A related deviation, a short balance calibration lapse earlier in the shift, was closed with “no impact to product quality.” Every page is signed. The batch is released.
This is an illustration, not a real case. Now apply the questions.
| Question | What the record shows | What the inspector asks |
|---|---|---|
| Authorization | Released by QA on the signature date. | Was that role the release authority? Which signature was the decision? |
| Evidence | Original and retest results attached; deviation referenced as closed. | Which results did the releaser rely on? Was the original result considered, or only the passing one? |
| Risk | “No impact to product quality.” | Based on what? What was the borderline result taken to mean for the rest of the batch? |
| Justification | “Reviewed and approved for release.” | Why release and not hold? What was weighed and set aside? |
Every statement in the middle column may be true. The record still cannot answer the right-hand column without someone reconstructing it.
The mechanism: capture, pressure-test, preserve.
The reasoning that matters is formed while the release is being decided. Weeks later it can only be remembered, and memory edits.
- Capture it as the decision happens. At release, the person with authority writes down the evidence relied on, what was open or borderline, how that was weighed, and why release was acceptable. Three to five sentences, in their own words, at the time.
- Pressure-test it before release. Put the four questions to the disposition while the people involved are still available. Where an answer is thin, the release decision is not finished.
- Preserve it with the decision. Keep the reasoning attached to the release and the authorizing role, so the record on inspection day is the record that existed on release day.
This adds a short step at release. It removes the long one later, when the batch is being questioned in front of an inspector.
Your exercise this week.
This takes one person about an hour. Do not prepare for it.
- Pick one released batch that had a deviation, an OOS result, or a retest in its history. Avoid the batch you already know is well documented.
- Collect only what exists today. Pull the record as it stands in the system. Do not call the releaser first to get the story.
- Ask the four questions in order. For each, point to the place in the record where the answer is written. If you have to say “well, what happened was,” the record has not answered.
- Mark each answer present, partial, or absent. Partial counts when the answer can only be inferred from a signature, a title, or a template field.
- Hand it to a second person, blind. What they cannot find, an inspector may not find either.
Do not backfill the answer.
The record will fail somewhere. That is the point of the test.
The wrong response is to fix it now: a late-dated memo, an added paragraph in the closed record, or asking the releaser to recall their reasoning and filing it as though it were contemporaneous. A reconstruction states what someone now believes they must have thought. It is not the reasoning that existed on the release date.
Record the finding through your own quality system, label any later account clearly as a later account with its date, author, and sources, and take the question to your Quality unit. Then change what gets captured at the next release.
The executive conclusion
The batch record review ends when every page is signed. The release decision ends when someone can show who authorized it, on what evidence, with what judgment about remaining risk, and why release was justified. Ask for that record before it is asked of you.
For a related disposition record, see A Passing Retest Does Not Erase an OOS Result. To apply the same test to another completed GMP decision, see Inspection Readiness Is Not the Document Set.
Source context: Regulatory requirements: 21 CFR 211.188, 21 CFR 211.192, and 21 CFR 211.22, read on eCFR (up to date as of October 7, 2026). The fictional released batch, the four questions, and their ordering are CW editorial examples, not regulatory mandates. ComplianceWorxs does not provide regulatory or legal advice and does not judge whether any record is adequate.
