CAPA effectiveness review: what’s the point, and how do you make it useful?

“We’ve had that deviation on line 4 again. — Didn’t we have it already? — Yes, three months ago. We changed the part, then closed the CAPA, as usual.” CAPA effectiveness review is written in black and white in GMP. Yet on the shop floor, it ranges from complete absence to systems so heavy that they lose their meaning.

Not every CAPA deserves an effectiveness review

Should the effectiveness of every CAPA be assessed? No. A low-severity, rare, easy-to-detect problem doesn’t justify a formal, time-consuming assessment. You don’t bring out the heavy artillery to check that nobody uses a black pen in the batch record any more.

On the other hand, when severity is high, when recurrence threatens the robustness of the system or when the problem is hard to detect, the effectiveness review becomes essential. Good sorting relies on three criteria: severity, occurrence and detectability. They make it possible to target the few dozen critical cases that justify an in-depth assessment, and to follow the others through trends.

Define effectiveness when designing the CAPA

Under time pressure, teams often decide on an action, update a procedure, train staff, without anticipating how they will prove it worked. The assessment then rests on vague criteria. How do you show that a simple update to a work instruction stopped the problem from coming back?

The question “how will we know this action worked?” must be asked at the same time as the action itself. If you can’t answer it, it’s often because the action doesn’t really address the cause.

Implementation and effectiveness: two different steps

Checking that a procedure has been issued, that a machine has been modified or that training has taken place is checking implementation. Many organisations stop there. Effectiveness is something else: no recurrence, a measurable improvement, a real reduction in risk. Confusing the two explains why so many CAPAs are closed in due form, yet ineffective.

Recurrence analysis: a good start, not an end

On a site that was doing nothing, tracking deviation recurrence is already considerable progress: results are made objective, and you spot the CAPAs that changed nothing (“look, the reminder to operators three months ago wasn’t enough”). A periodic deviation review brings out trends and common causes that are invisible day to day.

But simply noting that “nothing has come back” remains a reactive logic: you keep your fingers crossed. It doesn’t measure the gains achieved, and it says nothing about rare problems.

And when the CAPA didn’t work?

The most common practices are to close it anyway, noting that effectiveness hasn’t been proven, or to open a new deviation. Both muddy the system. The goal of a CAPA isn’t to carry out an action, but to eliminate a problem. If the action fails, a new analysis should extend the same CAPA, until its effectiveness is demonstrated.

Move forward step by step

On a site that is already stretched, imposing effectiveness reviews across the board would mostly produce incomplete files. It’s better to prioritise: invest energy in a few critical cases, achieve tangible results, then widen the scope. Effectiveness review shouldn’t generate paperwork: it should turn problems into learning.