Preventive Action (PA) and Corrective Action (CA) - One or Two Procedures?

[Craig H.]

C Emmons:

I would have to say that looking for a problem (we don't expect one, and have no reason to think there is one) is definately PA. So, yes, I would buy that argument.

FWIW

Craig

 

[Sean Kelley]

2 procedures here simply from a standard numbering standpoint. They are handled the same however they are opened for different reasons. A CA is opened when you are below the goal and the need is there to address the situation. PA is opened when a trend over time is negative but not necessarily below the goal.

 

[Kevin Mader]

In reading Tom’s post (with Energy’s quote), it reminded me that from a QS9000 perspective, the view on what an extended CA is and PA is weren’t always different. If someone has access to the original requirements, before section 2 and 3 were merged, I believe that you will find that the definition of a PA included the extended application of a CA. This changed after 2 and 3 were combined, whereby PA dealt more specifically with prevention of potential problems. I believe the sanctioned interpretations were also along these lines.

The question on whether or not FMEAs are preventive action or not has been an ongoing debate amongst some groups for sometime now. Me: I always thought they were PA by definition of why they were performed and in general, when they were performed which is as part of the Design/Design Transfer process. It certainly can be run as part of an investigation process for a CA, thus changing the light a bit. Nonetheless, from the perspective of Edward Link, who cares so long as a CI effort is being made? For the most part, it is just semantics. But not completely.

What kind of organization do you work for? One that reacts to problems when they pop up (detect/correct)? Or, do you work for an organization that tries to avoid problems (predict/prevent)? This is perhaps where the distinction between CI activities has the most merit. An organization reacting to trends within control limits makes me feel better than an organization who only reacts when a point is outside of the limits. It also shows me that the organization has made an effort to draw a useful distinction between what a CA and PA might be (perhaps this is another ongoing debate). I sense a forward thinking organization, here. Yes, I think I might do business with them.

Good discussion folks!!

Kevin

 

[bluloo]

While I'm a bit late to the party, we use a single procedure for CA and PA. It too is entirely manufacturing centric.

What I see as problematic is that our system is very much a reactive process. Our higher lever CAPA SOPs state that we respond with a CA or PA in the event of a non-conformance.

FMEA and Hazard Analysis fall under Risk Management in our organization. While Hazard Analysis could be considered part of a wider PA progeam, it is used during Design Review by Development only. FMEA is our only current, formalized tool to evaluate failure modes as a PA.

We are ISO certified, have many CE marked products and are number 1 in our Medical Device category, worldwide.
Despite our "successes" in the marketplace I have never liked the reactive PA view here. I was repeatedly told that PA is in response to a defect only, separate from continuous improvement (which we are only now beginning to touch upon from a practical standpoint).
Additionally we do not formally consider the role of Quality beyond manufacturing, another view that I find "incorrect" and short-sighted.

It's nice to hear another set of perspectives.

cheers

 

[mshell]

I am in the process of developing our procedure now. I am using one procedure for both CA & PA. I am not sure if this will meet all of the requirements as it is short and to the point. I have also referenced CA/PA in all other procedures for any nonconformances that occur. In your opinion, is this enough to meet the requirements?

 

[David Hartman]

I am in the process of developing our procedure now. I am using one procedure for both CA & PA. I am not sure if this will meet all of the requirements as it is short and to the point. I have also referenced CA/PA in all other procedures for any nonconformances that occur. In your opinion, is this enough to meet the requirements?

CA/PA may be the "result" of the Management Review process as well. Furthermore I do have a few questions related specifically to the following section of your procedure.

"6.2. CAR/PAR Analysis and Closure
The responsible department head will determine the root cause, identify potential solutions, records all possible solutions on the CAR/PAR, implement the most effective solution, complete the supplier sections of the CAR/PAR and return it to the originator for further processing prior to the response due date. If it is not possible to implement the solution within the thirty (30) day period, the supplier will indicate expected implementation date on the CAR/PAR. The originator will review the proposed action and notify the ISO Coordinator as to the need for a review date. The ISO Coordinator will assign a review date to the CAR/PAR taking into consideration the expected implementation date, assemble a cross-functional team and assess the effectiveness of the action taken on the review date. If the action is deemed effective by the team, the CAR/PAR will be closed. Otherwise, the CAR/PAR will be returned to the responsible department head for further action. This process will continue until the action is deemed effective"

Q1: What value do you gain from having "all possible solutions" listed on the CAR/PAR? Along with that: Who determines if in-fact these are ALL the possible solutions?

Q2: If you are allowing 30 days to "respond" to the CAR/PAR (per 6.1), are you allowing an additional 30 days for implementation? (It appears that you are in-fact assuming that your going to receive the response prior to the end of the 30 days allowed, and are then allowing the remaining time for implementation - but what happens if it takes the full 30 days for response?)

Q3: If the "response" is "unacceptable" what steps occur? And how long is the supplier given to provide an "acceptable" response?

Q4: If the action is NOT effective, how long is allowed for implementing "effective" action?

Scenario to be concerned about: Response on day 30, response unacceptable, second response received on day 60, implementation on day 90, action ineffective, revised response on day 120, implementation on day 150. Conclusion and closure 5 months after action requested. For complex issues 5 months may be acceptable, but are you willing to allow that for issues that should be easy to fix?

 

[mshell]

ddhartma,

Thanks for the critique. It has opened my eyes to some gaps in the procedure and I will begin working on those.

As for listing all possible solutions, I had an auditor tell me that we should list the potential solutions, select the best one and try the others if that one did not work however, he did not say all solutions. I will be removing the "all".

As for the 30 days, I will be rewording that section of the procedure to include extension requests and update intervals.

Thanks again,

Mshell

 

[Peter Fraser]

While I'm a bit late to the party, we use a single procedure for CA and PA. It too is entirely manufacturing centric.

What I see as problematic is that our system is very much a reactive process. Our higher lever CAPA SOPs state that we respond with a CA or PA in the event of a non-conformance.

FMEA and Hazard Analysis fall under Risk Management in our organization. While Hazard Analysis could be considered part of a wider PA progeam, it is used during Design Review by Development only. FMEA is our only current, formalized tool to evaluate failure modes as a PA.

Despite our "successes" in the marketplace I have never liked the reactive PA view here. I was repeatedly told that PA is in response to a defect only, separate from continuous improvement (which we are only now beginning to touch upon from a practical standpoint).

I'm late too! And I share your concern about the interpretations you mention. My view is that CA is triggered by something happening (or not happening) - ie an event or an incident - and PA is triggered by you realising that something might happen (or not happen) - ie a thought. It sounds to me that your PAs are CAs?

The fact that the standard says that you must have a “documented procedure” to explain how you address PA means that you need to describe how you identify what might go wrong and how you avoid it. But it does not even need to be in a single document. It could (should?) be addressed in your processes for planning, allocation of responsibilities, product design, process design, management review ... It is neither a single process nor a single action. The identification and reduction of risk should be part of an organisation's operating strategy. Not just in manufacturing, not reactive, not for just for "quality", but for the sake of staff, safety, profits, and ongoing improvement ... In fact, it is an intrinsic part of good "process management".

 

[Groo3]

Under the 1994 standard, we had one document covering both. About a year - year and a half ago, we decided to split these out into two documents... We think this helps our employees better understand the differences between the two. We felt this change would also help us get away from the old habit of reacting to problems only after they occur. After splitting into two documents, we then realized that there were already so many places that preventive action was already occurring... we were just doing a poor job of acknowledging / recognizing it... it helped open a lot of eyes all around.

 

[JRKH]

Little diferent

We have a 2 procedures. A CA procedure and a PA procedure. But the form for the PA is called a "Process Modification Request" and feeds into the improvement facet of the system.

I got a headache trying to figure out what was corrective and what was preventive and what was improvement.

James