Demonstration prepared from ComplianceWorxs material. Not commissioned customer work. Illustrative supplier: “[Supplier]”, an eQMS vendor. No product claims are made.
Every action on the CAPA is complete. Then someone asks: why did the evidence you had support calling this effective?
The procedure was revised, the analysts were retrained, and the effectiveness check has a date and a signature. The CAPA is closed.
A complete task list answers a different question. It shows the work was finished. It does not show why Quality was entitled to stop watching.
Why this Question Lands on Investigations
Investigation failures are among the most frequently cited drug manufacturing observations. In FDA’s FY2025 inspection observation data for drugs, 21 CFR 211.192 (investigations of discrepancies and failures) was cited 164 times. Only 211.22(d), the quality control unit’s written procedures, was cited more often.
FDA inspection observations for drugs, October 2024 through September 2025. The counts are citations, not companies.
No standard citation in that data reads “CAPA effectiveness.” In our reading, a weak effectiveness check is more likely to be cited as a failure to investigate or to follow written procedures than under a heading of its own.
What the Regulation Requires, and What It Does Not
Be exact about the source of each expectation. Inspectors are.
What it requires
Under 21 CFR 211.192, unexplained discrepancies and batch failures must be thoroughly investigated. The investigation must extend to other batches of the same product and other products that may have been associated with the failure. A written record must include the conclusions and follow-up.
What it does not say
In our reading, 211.192 does not set acceptance criteria, an observation period or a number of batches for deciding that a corrective action worked. Those are judgments the company makes. This is an editorial reading, not a regulatory determination.
When the standard is a judgment, the record of the judgment is the evidence.
What the Closure Decision Has to Show
A closure record that can be read by someone who was not in the room should show five things.
- 1
The original condition
What happened, what was affected, the risk, and the cause or control gap the CAPA was meant to address.
- 2
The connection from action to cause
How each action addresses the cause. Training fits a knowledge gap the investigation supports. If the cause was a weak procedure or a missing review step, the record should show which control changed.
- 3
The effectiveness method, set before the check
What would count as working, and what records would be examined. The population and the observation period should give the failure a realistic chance to recur.
- 4
The evidence and the judgment
What was reviewed, over what period, what was found, and why that supported the decision. Adverse or ambiguous signals are weighed in the record, not left out of it.
- 5
The authority
Who evaluated the evidence, who authorized closure, what residual risk or limit remains, who owns the monitoring, and what would reopen the CAPA. A signature shows approval. The rationale shows the judgment.
Where Closure Records Go Quiet
- Completion treated as effectiveness. A finished action plan may still contain no evidence that the changed control works.
- Training treated as proof. A training record shows training happened. It does not show routine performance.
- No meaningful observation period. “No recurrence” means little without a defined population, a window, and a realistic chance for the event to happen again.
- Evidence disconnected from cause. A thick package can still be weak if it never tests the gap the investigation identified.
- Rationale absent. A date and an approval record an administrative event. The rationale records the Quality decision.
A Worked Example
A site revises a reconciliation step, retrains analysts, and adds a second-person review after a reconciliation discrepancy was missed in batch review. The example below is illustrative. It is not drawn from any company’s record and is not an FDA-required format.
| Element | Example record |
|---|---|
| Original condition | Reconciliation discrepancy missed during batch review |
| Cause and control gap | Workflow did not require independent confirmation before final sign-off |
| Actions | Procedure revised; reviewer checkpoint added; analysts and reviewers trained |
| Effectiveness method | Every reconciliation event for eight weeks after implementation; second-person checks verified; deviation log reviewed for repeat or related failures |
| Evidence and findings | 26 post-change records reviewed. Checkpoint recorded in all 26. No repeat discrepancy of the defined type found in the period |
| Limit and residual risk | Eight weeks and 26 events show the checkpoint operated. Whether that is enough depends on risk and on how often the failure could recur. An adjacent manual handoff stays under monthly monitoring |
| Authority | QA manager documents the assessment; site Quality head authorizes closure, recording the limit and the monitoring owner |
Questions to Put to Your Own Closure Records
Take one recently closed CAPA and ask:
- Is the original condition clear enough to reconstruct?
- Does each action address the identified cause?
- Was the effectiveness method defined before closure?
- Did the population and period give the failure a real chance to recur?
- Does the evidence test the risk the CAPA was meant to control?
- Were adverse or ambiguous findings reviewed and recorded?
- Are residual uncertainty and continued monitoring owned?
- Who made and authorized the decision, and why?
A “no” on any of them marks a point at which a closure that looks complete cannot be explained.
Sources: 21 CFR 211.192; FDA Inspection Observations FY2025 (Drugs). The questions, five-element structure and worked example are editorial applications of those sources, not regulatory requirements. Source references should be verified against the current primary text by the client’s technical, legal and regulatory reviewers before this material is published or relied on.