To SPC or Not – That Is the Question . . . A Solid Maybe
- June 26, 2018 at 12:44 pm #56035
A little bit of background might help. I instituted a SPC program complete with flowcharts and questionnaires in response to out of control/trend signals. I thought this was going to be the best thing since sliced bread. Shortly thereafter, I saw the problem which hit me like a brick. The OOT signals were easy enough to document (GMP environment all must be documented or it did not happen) but almost impossible to explain. I found that the staff, already squeezed for time, were required to provide root cause for all of these “ticky tack” violations (minor trend rules). I liked to call this phenomenon “chasing ghosts”. We all knew that finding this variation would be downright impossible but we still had to document something because we had an OOT.
I found that one of the problems was that there were way too many control rules in place. After reading “When Should We Use Extra Detection Rules?” on this site, I now conclude that I will only use Nelson’s rule 1 and 2 and forget the rest. My company does not employ real-time SPC (not possible due to outsourced manufacturing) so we are analyzing in house testing data sometimes 1+ months after manufacturing. I’m still a little concerned about responding to signals when it is not warranted which is why I am asking for input from this crowd. I can see SPC working on the shop floor where one can monitor the process real-time and then stop it immediately but that isn’t the case here.
Does the community have any concrete suggestions for standard operating procedures in a GMP environment (grey areas are not allowed) on when to respond or not respond to signals? I want to employ value added activity but not tax an already stressed workforce (e.g. if the OOT point is touching or adjacent to the UCL/LCL then monitor future lots instead of searching for a root cause). I imagine that the solution would takes on a risk assessment sort of approach.
Thanks in advance.June 26, 2018 at 7:43 pm #202736
If I understand your description correctly, your process is not well controlled. In that case you’re going to get lots of signals. Rather than addressing each of them. look for common causes. Ask what’s causing so many signals to happen.June 27, 2018 at 4:26 am #202738
After hunting through Google it looks like you are saying that you have signals – specifically you are using the control chart rules of 9 on one side of the mean/target and 1 outside of 3 standard deviations and that you are worried about out of tolerance (OOT) results which, because they are OOT do not conform to Good Manufacturing Practice (GMP).
Item 1 – The usual practice is that one point outside of 3 standard deviations indicates your process is out of control and out of tolerance. The other rules – 2 out of 3 outside of 2 SD, 4 out of 5 outside of 1 SD, 8/9 on one side of the target/mean are usually treated as indicators that the variation observed in your process is no longer random. They are telling you to start looking for special cause variation. The measures corresponding to these data points are still representing product that is within tolerance.
Item 2 – The one thing you didn’t tell us is the frequency of the sampling. A key assumption of control charts is that the individual data points are independent of one another. If you haven’t checked this independence assumption then, based on what you have described, there is a good chance your data is auto-correlated. What this means is you will have control limits that are far too narrow and, as a result, you will get numerous false signals concerning process shift or actual loss of control.June 29, 2018 at 3:39 pm #202757
I don’t understand how you’d have Out of Control Actions to take if you’re doing SPC on data from a previous month’s info?June 29, 2018 at 5:40 pm #202758
To elaborate on my previous post, I’ve found that control charts aren’t very helpful if you have a lot of out-of-control signals. All they’re telling you is that your process is out of control. You need to work on identifying the root and common causes for variation. There’s a reason why control charts are a tool in the C phase of DMAIC and usually not in DMAI.July 1, 2018 at 10:14 am #202767
@timma Just so you are clear up front I am not a big advocate of control charts particularly attribute control charts. Those tend to be Statistical Product Control rather than Statistical Process Control. Variables Control charts because you should be able to use them on the input to a process rather than the output and they are actually Statistical Process Control. Now a lot of what was done in terms of process control can be done with controllers. Now there is some guy on here that gets all emotional over controllers and the number of adjustments. Lets understand the number of adjustments doesn’t matter. If I need an output of X and I can get it consistently out of a controller I am good with that because that consistency on the input side of a process controls the consistency on the output side. If it doesn’t then it isn’t a leverage variable.
From your comments it sounds like you have a stability issue which a control chart may help you with. It is difficult to really be able to tell without more knowledge of the process. It could also be you just calculated the control limits incorrectly or your sampling is wrong.
Just my opinion.July 2, 2018 at 10:40 am #202786
Martin K. HutchisonParticipant
For supplied parts, perhaps a lot acceptance sampling plan for KQC’s and have a defined response for when lots fail?
Otherwise fly to china and put SPC charts on their walls and train them to use them, get data sent to you electronically every night. Good luck with that.
You must be logged in to reply to this topic.