When a temporary control becomes permanent.

Some of the most interesting inefficiencies I have found in regulated environments did not start as bad decisions. Quite often, they started as very sensible ones.

Years ago, there was an environmental monitoring result that moved out of trend. It is important to make that distinction: it was out of trend, not out of specification. There was no specification failure, but the trend was enough of a signal to investigate what was happening and to consider what might happen if it continued moving in the wrong direction.

Temporary controls can be the right immediate response

While the investigation was ongoing, additional monitoring was introduced as an immediate corrective action. More locations were sampled and the sampling frequency was increased while the underlying cause was investigated and the CAPA developed. At that point, the logic made perfect sense.

The increased monitoring did not remove the cause of the problem, but that was not its purpose. It increased the likelihood that any further deterioration would be detected before the situation potentially developed towards an out-of-specification condition. At the same time, the investigation and CAPA were intended to reduce the probability that the problem would occur again. In risk terms, we were working on both sides: reducing occurrence through the permanent action while temporarily increasing detectability until we knew the situation was under control.

I still think that was a good decision.

The permanent action worked, but the temporary control remained

The deviation was investigated, the cause was addressed and the CAPA was implemented. Eventually, an effectiveness check was performed and, because the issue had not recurred, the CAPA was considered effective. That was also perfectly reasonable, except that nobody went back to the additional monitoring.

It simply continued.

Years later, during an internal audit, an auditor asked a surprisingly simple question: why was this area being sampled more frequently than comparable areas?

That question sent people back through the history. There was a reason, and once the records were reconstructed the story made sense. The additional monitoring had been introduced as an immediate corrective action while the real problem was being investigated. It had provided additional assurance at a moment when that assurance was useful.

The condition that had justified it, however, no longer existed.

The CAPA effectiveness check had demonstrated that the underlying issue had been addressed, but it had not included a second question: if the permanent action is now effective, do we still need the temporary measure that protected us while we were getting there?

Nobody had deliberately decided to keep it, and nobody had deliberately decided to stop it. As a result, it stayed.

I find this example interesting because there is no obvious villain in the story. The initial response was appropriate, the investigation was performed, a CAPA was implemented and its effectiveness was assessed. Every individual step could look perfectly defensible when viewed on its own, while the overall process had quietly accumulated work it no longer needed.

More control is not automatically more value

Extra controls are very easy to describe as conservative. More monitoring sounds safer, another sample sounds safer and another check sounds safer. Sometimes those additional controls genuinely are safer, and removing a necessary control simply in the name of efficiency would be the opposite of what I am arguing for.

Controls are not free, though.

In this case, the additional environmental monitoring required people to perform the sampling. It required materials, laboratory capacity, documentation, review and trending. It introduced additional handling and additional activity in the monitored environment, while every sample created data that somebody had to assess and retain. Every additional activity also creates its own opportunities for something to go wrong.

That does not mean the activity was dangerous or inappropriate. It means that once the original reason for doing it had disappeared, continuing it was no longer automatically the conservative choice.

Effectiveness checks need an exit question

So perhaps an effectiveness check should not end when we have demonstrated that the CAPA worked. Perhaps there is one more question to ask.

If the permanent solution is now effective, what did we put in place temporarily that should no longer be there?

A quality system should know when to add a control, but it should also know when that control has finished its job.

What am I missing?