Escalation to CAPA - Assessing if an NC warrants a CAPA

Hello
I am sorry if this topic has been dicussed before....

We currently follow the following criteria when we are assessing if an NC warrants a CAPA, if the NCs meets any of the criteria below then a CAPA is required:
The nonconforming product escaped and has a severity rating >S2
The NC is systemic in nature?
The NC can recurre with Permantant correction in place?



What we are finding is that the number of CAPAs opening is less. The ration is currently 1:16??

Do you think our criteria is robust enough?
How many times is a recurrence more than once, twice etc.

I have been asked to put in place a "simple" decision tree - any suggestions?

What are your thoughts.

Thank you

Just thinking out loud here...

Whether a CAPA is opened isn't really the point.
The point seems more like is your operation repeating mistakes of consequence?

How many you open is just paperwork.
How many issues you fix in a way that prevents recurrence has value.

Nonconforming product escaping (at least in my area) does not have a severity rating. Non-conforming product escaping is a problem needing fixing.

An issue worth mentioning to someone else is an issue worth fixing.

Recurrence once means a systemic change is needed.

Systemic change should be documented...CAPA is a fine way to document it.

Again...just thinking out loud and considering how your questions would be handled here...HTH

are these audit NCs? or non conforming material?

in theory as you continually improve your processes you will continue to perform corrective actions and move to more preventative actions. the severity of these actual and potential events will (in theory) get smaller and smaller.

in practice we want to ensure that the organization is working on the highest severity items (pareto principle). The dilemma here is how to prioritize. really high severity events don't need a high occurrence rate to require action. lower severity events would require a higher occurrence rate to require action.

management always wants some simple algorithm, but in real life this decision isn't simple.

I've used a 4 grid system where the x axis is difficulty and the y axis is severity (impact). I've found that trying to divide the grids more finely only adds complexity without adding value to the decision.

high severity low effort is priority 1
high severity high effort is priority 2
low severity low effort is priority 3
low severity high effort is rarely worked on...

Keep in mind that if you attempt to treat every NC as a CAPA (I really don't like the term CAPA, as it tends to confuse the difference between corrective actions and preventive actions), you will wear yourself out with the minutia and lose sight of the big picture. The end results is that your CAPA responses won't be of much value either.

If you try to set hard and fast criteria, again you will find yourself chasing bunnies.

Make your criteria pretty simple, such as if it is a repeat or if there is a strong likelihood of shipping bad product, and then add one more. Add to whatever you have the need for someone with the knowledge to determine its severity to review all NCs and determine if they should be escalated. Perhaps have the quality manager review them. Ideally they would sign off all NCs as to if they need escalation or not.

To determine if an NC requires a CAPA some points to consider:
Impact of the NC to be assessed and the Probability of occurrence of the NC.
If impact is high - CAPA required
impact is medium then you will need to assess the probability and then make an informed rationale decision on if CAPA is needed or not. If CAPA is not needed then providing a justification will help.
Impact is high then CAPA required for sure.
Also another good process is Trending of your low and medium NC's on a periodic basis so anything that did not enter the CAPA system if it re-occurs then you might need to re-consider.