ISO 4.14.2 d) application of controls to ensure that corrective action is taken and that it is effective.

Does this indicate that there should be a:
1) plan for corrective action
2) schedule for corrective action
3) any other time limit for completion of corrective actions?

How do you control that corrective action is taken if you have no proposed close out date?

Thanks,

AJP

AJ - Don't know for sure why - but have always "read- TIMLEY" - into this one - sure setting dates is fine and within x days normally works out ..but some reasonably take longer than others .

Regards
Jim

We've got a time limit for coming up with a corrective action plan, but now we're discussing what should be in the plan. Specifically, a completion date for the corrective action.

The assignee is best able to assess how long the action items will take. However, should s/he include that information in the plan?

My argument is that in order to implement controls some due date for action items must be set up.

Can you have controls on the corrective actions without some schedule involved, minimum - projected/required completion date?

I'd say no, but I may be missing something.

Thanks,

AJP

[This message has been edited by AJPaton (edited 26 July 2000).]

AJ you have a good basis for argument - I see a number of variables & possible good/bad outcomes to each side of the situation.

Originally posted by AJPaton:

The assignee is best able to assess how long the action items will take. However, should s/he include that information in the plan?

I for one would tend to agree - the assignee should take origional responsibility and include completion due timing as long as it is "REASONABLE" and "TIMLEY" to the task at hand. A free pass to set a date possibly years down the road is however not a good thing.

The responsibility for accepting the date supplied should/could be agreed and communicated to the assignees superior.

Sad part is that when you start "Setting Dates/timelines" that may/could be missed - what action steps would be included in the plan/procedures to cover a "missing the date issue" ?

It COULD turn into a "if I wait long enough the situation will go away item" and if nothing has been covered on the "what will happen if I just ignore it anyway" syndrome. - That would tend to give ammunition to anyone that really believes this ISO document stuff is "just bull" anyway.

IMHO PROJECTED could work - as long as there is a mechinisim in planning/procedure for moving the projected date due to unseen circumstances that come to light after the origional planning stage.... That way the assignees supierior would retain some semblance of "control" over what is or is not happening/when with the issue.

The final outcome of what works...where depends heavily on the individual workplace environment, commitment and attitudes.

Regards
Jim

[This message has been edited by Marc Smith (edited 28 July 2000).]

We see setting a reply due date for CAR's issued works ok, but setting a date for actual closure of CAR is more difficult. We had an understanding that a CAR could not be closed until the CA has been validated. Now recently, during an audit, we were told by our auditor that we did not have to wait to varify, in fact we could close the CAR if we accepted the reply, trusted the CA to take place from the CAR. If the CA did not work, we could just issue a new CAR.
I guess my guestion would be, is this a true statement made by the autitor that CAR'S can be closed before we validate the CA ??
Thanks for any input.

Bryan

Bryan,

Our current take on this is that the Corrective Action can be closed when the action items are completed.

However, we then have an "Effectiveness Review" date that is set to check on the efficacy of the corrective action. That's how we set a control on the effectiveness, which I mentioned in my opening for this topic.

Hope this helps.

AJP

I think the 'typical' approach is to consider the CA open until effectiveness is confirmed. However, in some instances this may take months or more. Typically a few weeks or so is sufficient time before confirming effectiveness in a high volume production operation.

For long term complex CA I close them when action is complete and schedule a follow up audit when there is too long a period between...... For simple stuff I followup first, then close.....there is an option box on the CAR for choosing....

Your auditor should not be giving you advice that detailed/specific...it is only his opinion...You can be flexible, as these wounds for CA/PA are self inflicted typically

[This message has been edited by barb butrym (edited 21 August 2000).]

Hi,

Ok, this is how we do it:

We have a rather big database application for audits and corrective actions.

First of all we allow two weeks to come up with a suggested action and deadline.

Every issue gets a deadline based on an estimate of the time needed to complete the agreed action. ( Deadlines are moved if necessary, but passing the deadline without reporting the status means that you'll find your issue on the delayed actions list on our intranet the very next day... People tend to dislike that ;-) ).

When an issue is closed, we're still not quite finished with it: The application adds a follow up date ( Default 90 days later ). We then proceed to make certain that the action taken has been effective. If not, we reopen the issue for further action.

I firmly believe that time limits are neccessary in this field.

Claes Gefvenberg

Is there a difference between corrective action and problem analysis? If so, what criteria would you use?

The reason I ask is because our current system mandates that a Problem Analysis (PA) be opened and processed for every single instance of non-conformance. As Barb mentioned in an earlier message, these "wounds" are typically self-inflicted. The problem is, my company is "wounding" itself to the point of total ineffectiveness.

Dos anybody have any suggestions or comments? Any help would be greatly appreciated!

Regards

Qualman

P.S. To the gentleman from the great state of Texas (you know who you are) who referred me to this site, THANK YOU!

[This message has been edited by qualman (edited 29 August 2000).]

[This message has been edited by qualman (edited 29 August 2000).]

IMHO>>>>>The connection is root cause analysis and problem analysis...not necessarily the CA system itself. A good CA system will include both...

The trick is to stay away from overkill (bloody paperwork...LOL), and be effective (ergo ... add value) in the documenting of it all so as not to get caught in the ..."Oh Sh*t, now I have to document this thing, so lets skip it" game.

It takes a lot of insight and working the system to get it right and the shame is that once you have it you are still not done...it will still progress as continuous improvement and change some more....AND further more....you cannot count on it working anywhere else you go...its unique to the culture of the company and its players.

A word of advise, for what its worth....Don't ever say always....or for that matter never...it will come back to haunt you !!!

how do you define nonconformance in your procedure?

[This message has been edited by barb butrym (edited 29 August 2000).]

the definition is ok...but is it a reality to require CA on each one? I like the document (look for trends...then act) and do selective root cause/problem analysis accordingly.

No ... it does not indicate lack of commitment.. to me and most of my peers

in fact the original would indicate a system for show, and less focused on CI...(my humble opinion only...) how can you focus on real improvements when so much time is wasted?

The trick is to stay away from overkill (bloody paperwork...LOL), and be effective (ergo ... add value) in the documenting of it all so as not to get caught in the ..."Oh Sh*t, now I have to document this thing, so lets skip it" game.


We (my company) define nonconformance as: "a product or material that does not conform to the customer and/or any internal requirements or specifications." Too broad?

Here's the question: is it possible to have a CA system in which every situation of N/C is documented, but not every N/C requires a Problem Analysis? The problems still get fixed (the customer is happy) but the root cause for every single situation isn't investigated.

If so, would you then look at that company and come to the conclusion that they didn't have a commitment to Continuous Improvement?


Qualman

Barb:

Okay, now we're talkin'...things are definitely starting to look a little better!

If I sound worried, its because I am - our registration audit is is less than 2 months away - just picked our registrar.

My concern is that our consultant has let too much of his personal opinion mandate how to fulfill the CA req's, instead of showing us the req's and then letting us run with implementation. This is my company's first foray into TQM, and as you might expect, it hasn't been a painless process.

I'm very hesitant to "fix" our current CA system right before our audit, but on the other hand, I risk N/C for not having all previous documented N/C's addressed. Any thoughts?

Qualman

Originally posted by qualman:

The reason I ask is because our current system mandates that a Problem Analysis (PA) be opened and processed for every single instance of non-conformance.

What are you calling "every single instance of a non-conformance?" Every reject tag written? If that's the case than I would say you do not need C/A for every n/c, but some Pareto order. You would however need to make sure the requirement of disposition is covered.

Same thing as Barb said I guess, but I would define some logical reason into which ones get a CAR or PA I guess you call it. Pareto order seems to be the tool of choice here, but you can tailor it "by quantity" or "by cost" - whichever makes sense for your biz.
If you're really going to improve, you can't work on too many at a time.

When a company designs a nonconformance -- corrective action -- preventive action system it has do do so with some thought. I recently visited a company which requird a corrective action for every nonconformance. Of course this was a QS company to boot. They were trying to work themselves out of a pretty deep hole and were redesigning the system to allow what I call a gate to determine what requires a CA and what does not.

The 'gate' is someone with the qualifications to assess whether a CA is initiated or not in response to a nonconformance, customer complaint, etc. on an individual basis. Nothing gets a corrective action response unless the 'gate' says so. Noncritical, 'new' (no history of previous related event) issues seldom 'rate' a CA response. Sometimes an abbreviated CA response is 'requested' by the 'gate'. But the point is each event is evaluated for merit, etc.

Just as important as a 'gate' is that a company define (as well as it can) what nonconforming product is and what a customer complaint is. If a customer calls and complains that the price s/he paid is too high, is this sufficient cause for a corrective action? It's a customer complaint, isn't it? If a customer calls and complains that his/her widget broke after only 2 years (1 year warranty), is this a valid complaint? If I buy a screw driver and the tip breaks when I pry off a lid from a can of old paint, is this a valid basis for a customer complaint?

There is more common sense to all this than often meets the eye.

Great concept!

What if you were to document the criteria for the gate, instead of having an individual act as the gate?

That way, the gate would be distributed across all phases of operations and departments, and not rest with one individual(s). Would the concept still work?

When I said:
> Just as important as a 'gate' is that a company define (as well as it
> can) what nonconforming product is and what a customer complaint is.

That's the "Black-and-White". But - everything is not black and white - which is where the human 'gate' comes into play. There are many times it takes a human to assess the specific situation including risk factors.

In bigger companies you may have multiple gates. You just have to look at your specific company and its needs.

In my opinion you cannot have only one - you must have both. The base criteria and the human gate.

Originally posted by Marc Smith:
The 'gate' is someone with the qualifications to assess whether a CA is initiated or not in response to a nonconformance, customer complaint, etc. on an individual basis. Nothing gets a corrective action response unless the 'gate' says so.

We use the same system. Cuts way down on CAs and therefore makes the ones processed more effective.

In my case, the "gate" is a Senior QA Tech with more than 15 years experience. And if he has any doubts (which isn't often), he asks me (QA Manager) to make the final call.

Works well.

I agree with marc's gate theory.......

short term...with an audit coming....
In the mean time use your system to correct itself to cover your past sins....issue a CA/NC for the ineffictivy of the system stating the reasons you choose from here and whats not happening. Respond with a time line for revamping the system that is realistic....use the time before the audit to get started and to show commitment to the time line and assign action items.....REissue the important ones not addressed under the new system if its not to much like shutting the barn door after the horse gets out......

I'm looking at literally hundreds of N/C's that go back at least 6 months (or more). ALL the horses have left the barn a long time ago - the barn has just about collapsed.

Do I spend the time repairing the barn and wrangling all the mustangs back into the barn (issuing CA for the 'important' N/C's while revamping our CA system) or do I do I build a new barn - start fresh, get more horses, and dump the old N/C's?

Of the two choices, which is more value-added?

1. Review them - it's not to late for a 'gate review'. Close out those you know are 'lost' or 'not worthy' of a CA and document your reasoning on each one. Document this review and the reason for it. Be ready to discuss / explain what you did and why.

2. That should yeild a few 'important' ones for you to continue with.

3. Write the 'gate' (review step) step into your system.