The company I work for recently passed our ISO registration audit, but we did have one weak area - Preventive Action. We do preventive action in several different areas such as FMEA's, employee ideas, first article reviews, etc. The problem is that each of these are pretty much stand alone systems with preventive action built into their function. Each of these areas may only generate one or two true Preventive Actions in a year. The current system requires that someone says to themselves, "Hey, this sounds like a preventive action, I should tell someone so we can document it and get credit for it at the next audit." In those cases, we document the PA in our PA/CA software.

I have considered creating a Preventive Action Task Force, but that would seem to be very inefficient. Somehow, we need to come up with a way to formalize the capture and documentation of preventive actions.

My plea is for some suggestions on ways you have seen this done successfully.

Thanks,

This may be helpful. We had the same problem in 1999. We incoporated PA as a part of CA. What I mean is that when you perform a Corrective Action and there are other areas that can be identified with the possibility of the same thing happening that is like or similar to the CA problem, you identify the action you take as PA for the like or similar issues.

Confused? I was until we tried it. And it worked. Our auditor told us that this was GOOD Preventive Action.

I hope it works for you.
This is my first post. Good Luck!

Danno

I agree that including in the CA program is a good option. We do the same thing here. But there is a word of caution with this. It would be a good idea to check with your auditor before implementing this to make sure they don't have a problem. Some do and some don't.

There have been a couple of pretty good debates about corrective and preventive action and how to classify them. There are some people that would say the system described here would still be corrective action since it is in response to something. So just make sure that your auditor won't give you problems first. Otherwise, this is a good way to go.

Isnt performing FMEA a preventive action, even if the review itself does not generate further actions to reduce risks, the very identification of low or no risk through FMEA is a preventive action - well it is to me.

I would agree Martin, but we can get into all sorts of tangles if we look at terminology. For example, if you do react to the RPN to lower them, is that preventive, or corrective? We've been down that road before.

chermann, I think it is important that you have a system that works for you, not your auditor. If your current system is based on:

"Hey, this sounds like a preventive action, I should tell someone so we can document it and get credit for it at the next audit."

Then at best your actions are an exercise in futility.

db

By virtue of the fact that FMEA should be done at the product and process design stage I would say that any action to reduce RPN is preventive as you have not made anything yet. FMEA's conducted further down the line, where incidences of occurence are based on true data is, as you say, corrective by nature.

Just to clarify, we are a job shop and only a handful of our customers require FMEAs (we are not a direct supplier to automotive).

Being a job shop, we develop processes specific to each part number our customers send to us. We have 50-100 new part numbers show up every day. We have 2 technicians who review the part and the process for each of these orders and create a process program (anodize, electroless nickel, etc.) This review process is not typically a formal FMEA, but these guys do incorporate preventive actions in this process. For example, if a part has alot of blind holes, extra rinsing and a note to rinse thoroughly might be added to the process program to "prevent" possible problems with solution bleed out. It would be nearly impossible to track each part number specifically to see if your preventive action was effective, especially considering a large percentage of these orders (part numbers) may never show up again.

chermann

To my mind what you describe are perfect examples of preventive actions, and they are often called contract review, or design control/review, or process planning. In fact every part of our system is a preventive action isnt it. ISO9001:1994 used to say in its opening paragraphs somewhere that the whole standard was based on prevention, and many of the clauses contained comments such as 'in order to prevent.......' etc, etc, etc.

Donkeys posts ago I suggested that a PA procedure only had to bullet point all the things that you do in your QMS, as it is all preventive.

chermann said:
. It would be nearly impossible to track each part number specifically to see if your preventive action was effective, especially considering a large percentage of these orders (part numbers) may never show up again.

Do you keep records of those process plans? Do you keep records of customer returns? No return, no problem. Maybe a result of the PA's built into your process and maybe not. Proving a PA is effective is not easy. But proving that problems aren't reported is easy. Make the assumption that the PA is as effective as can be demonstrated by lack of returns.

Also, is it possible to make something formal? Such as a table showing some possible problems that could occur in one column and then some preventive actions listed that would address that sort of problem listed in another, and then the type of part that this kind of problem is most likely to occur in? I don't mean create one for each part, but something that is used over and over for all parts that are made. Something like that could show that there is some PA planning in effect and is a good reference guide for those who put together those process programs. It may also brainstorm additional PA's that hadn't been thought of previously.
Lastly I would then suggest that you have a program that when a new part with new risks is being made, that there is a review of the chart to ensure that there is nothing additional to add, and that you have a method for updating and modifying the chart.

Employee suggestions for anything and everything that impacts the business should also go through your PA/CA process. If it saves money, improves efficiency, prevents something bad from happening, etc. it is all good. This can be done as simply as informing everyone by email and/or bulletin board that Susie Q in Accounting is accepting employee suggestions by email or note. And then having Susie Q log these into a simple Access database or Excel spreadsheet. Set aside a time for review of the suggestions and log in the actions taken.

Sorry so longwinded. Does any of this sound feasible/reasonable?

Yes, we do keep records of the process, it is stored in the computer and used as the default when that part shows up again. We also formally document customer complaints and returns. I can take a look at the feasibility of looking at no returns as the measure of effectiveness. As I mentioned, the real challenge for us is that we do not "make" the parts. Our customers send us their raw parts for us to add a coating and return it to them. If a customer changes from part # 1234A to 1234B, our system assumes it is a completely different part. With 50 new parts a day, 250 new parts a week, 13,000 new parts a year, it boggles the mind to try to track the effectiveness of a single preventive action implemented on part #1234A.

We do have an employee suggestion process, but that really seems to wane with time. Every once in a while, someone breathes a little life into it, but it is usually short lived. When we do get an idea, it is reviewed by the Continuous Improvement Team and assigned to someone. We do get a few PA's from that process.

I do like your suggestion on creating a table for the process rather than the part. I will use that as a starting point and see where it leads, thanks and longwindednesss is is fine by me. Everyone on this sight seems to offer very helpful and useful advice. I appreciate your feedback.

Chermann,
Excellent example of preventive action. You can apply this formally, to an FMEA or informally to your process flow sheets. Whatever is needed to meet your customer requirements.

"It would be nearly impossible to track each part number specifically to see if your preventive action was effective, especially considering a large percentage of these orders (part numbers) may never show up again."

You do not need to track each part number. You could random audit part numbers or group similar preventive actions and cnduct a random audit.

:bigwave:

Hello Folks!

This is my first posting on the forum and I hope you will be of help.

My company achieved ISO 9001:2000 accreditation in late November of last year and it was suggested to us by our accreditation body auditor that we should look at a new format for our Corrective/Preventive action Procedures.

I have drafted proposals for Senior Management explaining what the QA Dept would like to put in place to achieve continual improvement and effective Corrective and Preventive Actions.

The proposals contain details of how to apply a risk assessment style to the processes involved and using risk assessment analysis to pinpoint opportunities for both corrective and preventive actions.

Does anyone else here use a similar approach and if so what kind of results were found from this style of approach?

Absolutely.

We use Design and Process FMEA's, which is basically just a risk analysis process.

Are you aware of FMEA ?

FMEA - Failure Mode Effect Analysis.

Yes, I am aware of it and have based my proposals around it but don't tell the management, as they are afraid of things that are abbreviated like that;)

How well does this "Risk Assessment" approach work in practice?

Forgive me if my questions seem so basic but I'm just recently qualified (IQA Diploma) and although I have a good grasp of theory (alledgedly!) I am now learning the day to day practice of managing a QMS.

Hi calqig and welcome to the Cove. :bigwave:

You show good sense by not using the abbrevations, because your managements reaction is a very common one.

As for using FMEA "in the wild" it's a very good tool if used right. It's not hard to do. I'm a great fan of it, but to be honest there are some drawbacks (as with any tool): It is not a "quick" method and it's pretty labour-intensive.

/Claes

:bigwave:

Thanks for the welcome and your reply.

The proposals I've drafted are based on a continual improvement loop and FMEA and are basically as follows:

Trend analysis feeds info into the corrective action system and the information gathered from the corrective action system feeds into the preventive action system forming a database of known problems and their causes (which we in the QA dept refer to as the knowledge). Within both corrective and preventive action systems a simple method of damage/risk assessment is used to prioritise which problem or potential problem should be tackled first. The results of these actions are used to effect improvement of the effected process or processes. The records of all these actions are added to the knowledge database and so the loop continues.

What do you folks think, am I following the correct course of action?

Just a thought here. What type of Prioritization process are you using? And who review it? I have seen managers that will use a frequency chart and take on the big number. I just went through this last Friday, the biggest hitter was misspelling in the title block of the prints (20 times in a month) which resulted in 30 minutes of rework and total cost of around $600 applied to Cost of NonQuality. However, he had 6 occurances of wrong vectors on the print which resulted in 45 hours of rework and a cost of $18,000 applied to our CONQ metric. It's the old 80/20 priciple being proved again. So remember, if it is just occurances that are looked at, you may miss the target.
Isn't it great how theory and practice can be so different! :biglaugh:

You're not located by Largs are you?
:bigwave:

Hi Randy!:bigwave:

We are based in Bellshill, North Lanarkshire. The otherside of Glasgow.

We don't propose to use a specific number to trigger actions, we propose to use a rating system to prioritise which action should be tackled first, the higher the rating the higher the priority.

Sounds good to me calgig. The biggest problem I've found has been with sustaining the effort. The Plan, Do, Check, Act cycle is great if you can keep the influx of ideas and people. All to often I see the same people doing all the problem solving / improvment initiatives.

Oh, I have some great memories of Glasgow, and some not so great memories of headaches after leaving Glasgow. Must have been something about the trains, it couldn't have been the lager! I worked at the sub base (UK) and lived in Largs for a while.

:bigwave:

Randy,
Sorry I've not been able to reply to you for a while.

Glasgow does have a great reputation for being a good place for a night out! :biglaugh:

Were you in the US Navy or the USMC then?

Back to business now.
The proposals I tabled to management have now been accepted and we are going to embark on a pilot project which shall apply them for the first time within the Sales and Contracts Dept.

I now have the task of explaining the process to the guys in our Quality Improvement Group who are being teamed up with an Internal Auditor to carry out the assessments. I will also need to brief the Internal Auditors on these processes.
So I'm going to be really busy for the next couple of weeks!

Gaboring, the method you use for PA is actually CA Impact. The problem has happened but just not in that area. Many auditors will not accept this as PA. Especially the QS auditors. This is the method that we used at one time. Just some FYI. If they accept it great!!

the method you use for PA is actually CA Impact
I partially agree. If you apply the action "across the board" it can be preventive. You are looking for a systemic or process based problem not just stopping the leak.
If the solution only addresses the current issue then IMO, you are right gp. However, if how they apply the findings goes into poke yoke, etc. then it can be viewed as preventive and continuous improvement.

It is really in the application of the knowledge gained.

Were you in the US Navy or the USMC then?
I was in the Navy. I was a member of the security detachment that worked on the British sub base.

Hi, this is my first post but have been hanging around "the Cove" for quite some time. I have found this place to be very informative and thought provoking so THANKS!

Anyway, I wanted to post b/c I am working on restructuring our Corrective/Preventive Action program. We have been told by our ISO auditors to split the two procedures because they are in fact different approaches. Now that I have dug into it, I have to agree.

The way I see it is this: Corrective Action is reactive to a problem that has already occurred. agreed?

Preventive Action is meant to prevent non-conformances - so how do you formalize this approach? From what I learned from our ISO and FDA auditors (we are a medical device manufacturer) I designed our Preventive Action procedure to be a review of quality data on a regular basis, including customer returns, customer feedback, monitoring data, management review, audit results etc. The goal of the review is to identify potential unsatisfactory trends or non-conformances etc. Up till now, our Preventive Action system was basically what chermann described with the words "Preventive Action" thrown into our Corrective Action procedure. Everyone is really excited when a Preventive Action is identified. It just doesn't happen very often because it just isn't a regular process....

Now that I have these procedures split; they have shaken out quite differently in the process and it has helped our employees understand it better as well.

My main problem has been with verification of effectiveness. Most of our auditors want us to define a length of time to monitor for Corrective Actions for verification of effectiveness which we are going along with.. however, this just does not seem to fit with Preventive Actions (if the problem hasn't occurred and you are trying to prevent it how do you know your action is effective just because it doesn't occur after???).
What I am proposing is that some kind of analysis be done for verification of effectiveness depending upon the action taken.

Any help on this one?
Thanks in advance.

Hi Cowgirl,

Glad you made your first post. Keep it up. Your challenge today could be my headache tomorrow.

Check out this thread. I posted a poll about this very subject. Could be of some use to you.



Regards,
CarolX

Thanks Carol,

I have since decided to define our verification of effectiveness for preventive actions similar to the way we are verifying corrective actions:

CA: "Verification of Effectiveness may be achieved by monitoring the process/problem to have been corrected for a specified time frame not to be less than two months. All data collected supporting the monitoring period will be included in the CA/PA report.
a) The time period length is dependent upon frequency and duration of process in which the problem occurred and agreed upon by the MRB.
b) If Verification of Effectiveness can be achieved through sufficient data, testing and/or problem resolution analysis ensuring correction of the problem, the MRB may waive the minimum two month time period. "

PA: will be the same by verifying that the unsatisfactory trend declines or stops within a specified time frame or if the effectiveness can be verified through data, testing or resolution analysis the time frame requirement can be waived.

We've also added some other enhancements to our CA system that define criticality levels to problems which in return define the time period in which certain actions must take place. This was all done as a corrective action addressing our timelieness in completing Corrective Actions!!

yes, it is a headache


:vfunny:

Hello again folks!:bigwave:

As Randy said, we are applying the proposals, "across the board", i.e. the data gathered from corrective action provides info which has a feed into preventive action and which in turn feeds into continuous improvement and the information collected is used as a resource for corrective action and so the loop goes on and on...

Hello to Cowgirl!

Cowgirl We use aseperate FM for PA.

Potential Problem:
Potential Consequence:
Potential Cost of Not Reacting:
Estimated Cost of Reacting:
Do this affect our reputation:
Steps to deal wit potential problem:
who responsible:
due date:
Application of Control to Ensure Effectiviness (measurements)
Result of Effectiviness( based on above data)
Signed:
Date:
Approved:

Hi
if the auditor asking for the evidence for prevention action, is it showing the FMEA is good enough ? for improvemnet activity..can it consider as prevention action ? pls advised

Thanks Carol,

I have since decided to define our verification of effectiveness for preventive actions similar to the way we are verifying corrective actions:

CA: "Verification of Effectiveness may be achieved by monitoring the process/problem to have been corrected for a specified time frame not to be less than two months. All data collected supporting the monitoring period will be included in the CA/PA report.
a) The time period length is dependent upon frequency and duration of process in which the problem occurred and agreed upon by the MRB.
b) If Verification of Effectiveness can be achieved through sufficient data, testing and/or problem resolution analysis ensuring correction of the problem, the MRB may waive the minimum two month time period. "

PA: will be the same by verifying that the unsatisfactory trend declines or stops within a specified time frame or if the effectiveness can be verified through data, testing or resolution analysis the time frame requirement can be waived.

We've also added some other enhancements to our CA system that define criticality levels to problems which in return define the time period in which certain actions must take place. This was all done as a corrective action addressing our timelieness in completing Corrective Actions!!

yes, it is a headache


:vfunny:


It doesn't have to be a headache. A preventive action is like a corrective action, just done before the failure actually manifests. (I know there is much discussion on this forum about this, but just read the standard.)

You can verify a corrective action, if you can determine the cause has been eliminated. That is the definition. Well, a preventive action can be verified the same way - was the root cause effectively eliminated?

It can be done very simply and naturally.

Hi
if the auditor asking for the evidence for prevention action, is it showing the FMEA is good enough ? for improvemnet activity..can it consider as prevention action ? pls advised

Yes. Action taken on an FMEA with regards to RPN is normally preventive in nature. As long as a process is stable, any further action taken can be classed as an improvement.

Yes. Action taken on an FMEA with regards to RPN is normally preventive in nature. As long as a process is stable, any further action taken can be classed as an improvement.


While I agree that FMEAs are preventive in nature, I don't believe that was the intent of clause 8.5.3. This has already been debated much on this forum (just do a search for Preventive Action), so I don't want to begin a new debate.

But, I would suggest reading cl 7.5.1.1 and cl 8.5.3/8.5.2. I think the pattern described in cl 8.5.3 is much more similar to the pattern in 8.5.2.

Best of luck.