Open any deviation log in a regulated facility and you will find the same corrective action repeated down the page: personnel were retrained on the relevant procedure. Training records attached. CAPA closed.

Twelve months later, the effectiveness check comes due. The same deviation has recurred - same process step, different operator, sometimes the same operator. The CAPA that looked closed now looks like evidence that the root cause was never addressed. And an investigator reading the file sees exactly that.

Why retraining fails as a corrective action

Retraining fails for a simple reason: it treats the symptom as the cause. When a deviation traces back to personnel, the finding is almost never that the person was missing information. It is that the system around them allowed - or produced - the failure. The procedure and the training said different things. The sign-off meant someone watched, not that someone could perform. The trainer taught their own version of the task. The new hire was released to independent work because the schedule needed them, not because anyone verified they were ready.

Sitting a person back in front of the same slides that did not create capability the first time does not fix any of that. It creates a record of attendance. Attendance is not competence.

What the effectiveness check is actually asking

A CAPA effectiveness check asks one question: did the action taken prevent recurrence? For a training-related corrective action, that translates into three harder questions:

  • Can the person now perform the task correctly, independently, and under normal production conditions?
  • Is there a record that proves it - not a signature saying they were present, but evidence they demonstrated the performance?
  • Would a different person, trained the same way, reach the same result?

If the corrective action was “retrained the staff,” the honest answer to all three is usually unknown. Unknown does not survive an effectiveness check, and it does not survive an investigator’s follow-up question.

What holds up instead

Corrective actions that survive share a structure. They define the standard the person must perform to, in observable terms. They rebuild the training so it teaches to that standard - including coached practice, not just information transfer. They verify performance with a demonstration measured against the standard, performed independently. And they document that verification as the record of competence.

The difference shows up in the file. “Retrained on SOP-123, records attached” invites the next question. “Operator requalified on SOP-123: demonstrated independent performance of the full inspection sequence against the defined standard on [date], verified by qualified trainer, record attached” ends the conversation.

This is also where regulatory expectation has moved. FDA’s Quality Management System Regulation, effective February 2, 2026, incorporates ISO 13485 and expects manufacturers to evaluate whether training was effective - not whether it was delivered. An attendance log is evidence that someone was present. It is not evidence of competence.

The practical test

Before closing any personnel-related CAPA, ask the question the effectiveness check will ask a year from now: if this exact situation arises again tomorrow, does the system now produce a different outcome - and can we prove it?

If the answer depends on hoping the retraining stuck, the CAPA is not closed. It is deferred.

We build qualification systems where competence is demonstrated, documented, and defensible - the kind of evidence that holds up when the effectiveness check comes due. If a finding on your desk cites training, that is exactly the conversation we should have.