1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.
Dismiss Notice
You must be a registered member in order to post messages and view/download attached files in this forum.
Click here to register.

Thinking through a process validation

Discussion in 'Qualification & Validation (Also 21 CFR Part 11)' started by Candi1024, Mar 31, 2016.

  1. Candi1024

    Candi1024 Well-Known Member

    Joined:
    Jul 30, 2015
    Messages:
    129
    Likes Received:
    83
    Trophy Points:
    27
    Location:
    Pennsylvania
    As the title says, I've been assigned to look at a process and decide what, if any validation needs to be done.

    It is a final acceptance test. A part is manually moved to verify it bends to the proper angles. The unit is placed on a board with markings to show the correct angles. This board is not calibrated, maybe it should be? With visual checks every year or 2 to verify the marks are still readable?

    Torque on a knob is also checked with a calibrated torque wrench. We will be recording the ID number of the wrench.

    Since these functions are not checked again before shipping, my boss wants to know what kind of PQ or what not should be done. I don't see a need to do any PQ, maybe maybe a R&R with the angle board or the torque wrench? I'm not even sure that is really needed.

    Thoughts?
     
  2. hogheavenfarm

    hogheavenfarm Well-Known Member

    Joined:
    Jul 30, 2015
    Messages:
    220
    Likes Received:
    160
    Trophy Points:
    42
    You need to look at your packaging and shipping methods also, there could be conditions that may affect the product (temp, position/nesting orientation, etc.)
    The idea being to safeguard the item at least to the destination point. As far as function, you may be able to select the 'maximum' angle and simply check all units to that , but without any more data it is hard to visualize what we are talking about. With a maximum function check, all internal angles would be ok, and you would need to validate the jig or gauge.
     
  3. yodon

    yodon Well-Known Member

    Joined:
    Aug 3, 2015
    Messages:
    198
    Likes Received:
    115
    Trophy Points:
    42
    You would typically consider process validation for those processes whose output is not 100% verified. In your case, it sounds like the process is the test - so what may be more appropriate is a test method validation, if anything. How do you know that the angles on the board are the right ones? How do you know the part is place properly so as not to influence the bend observation (probably getting into your R&R here)? How do you know the torque target is right (and what the tolerable range is)?

    You also should consider the risk when approaching this. Does a life depend on the part being able to bend to the proper angles or is it just a nice to have as long as it's pretty close?
     
  4. Ronen E

    Ronen E Well-Known Member

    Joined:
    Jul 31, 2015
    Messages:
    133
    Likes Received:
    70
    Trophy Points:
    27
    Hi,

    I agree with Don that test method validation is the name of the game. Terminology like PQ is less applicable because you're not dealing with (production) Process Validation.

    Please note that both unaided visuals and manual torque testing are prone to problematic GR&R results. With torque testing, the goal is applying the force slowly and uniformly (the best is mechanising it, though it may be an overkill in your case). Pay lots of attention to personnel training & qualification for the specific tests.

    Cheers,
    Ronen.
     
  5. Candi1024

    Candi1024 Well-Known Member

    Joined:
    Jul 30, 2015
    Messages:
    129
    Likes Received:
    83
    Trophy Points:
    27
    Location:
    Pennsylvania
    I'm going to do the GR&R on the torque measurement tool.

    The tool uses two different fixtures to measure at two locations. Do I need to do the R&R for each location? I don't need to do it for each fixture do I? (as in there are other identical fixtures used at different stations). Or maybe I do unless I am somehow verifying the fixtures are the same?
     
  6. Ronen E

    Ronen E Well-Known Member

    Joined:
    Jul 31, 2015
    Messages:
    133
    Likes Received:
    70
    Trophy Points:
    27
    If you can confidently rationalise why the fixtures don't affect the measurement, maybe that will be enough. Otherwise, it comes to similarity between fixtures. If they are made based on a drawing, try to determine what dimensions / characteristics are significant for the torque test and how the "population" of fixtures distributes in these. If you validate the extremes and maybe also a "nominal" fixture you should be fine I think.

    If the test is carried out at two different locations you have to assess and qualify the different operators at both locations. You might be surprised - sometimes there's a big difference in applying the "same" method in, say, manufacturing and a QC lab.

    Cheers,
    Ronen.
     
  7. Candi1024

    Candi1024 Well-Known Member

    Joined:
    Jul 30, 2015
    Messages:
    129
    Likes Received:
    83
    Trophy Points:
    27
    Location:
    Pennsylvania
    I meant two different locations on the part, and each point uses a different sized fixture. There are two work tables right next to each other on the manufacturing floor.
     
  8. Ronen E

    Ronen E Well-Known Member

    Joined:
    Jul 31, 2015
    Messages:
    133
    Likes Received:
    70
    Trophy Points:
    27
    Sorry. It goes to demonstrate the shortcomings of written-only communications...

    Generally, each part location or type of fixture would constitute a different test and thus the need for validation should be determined for each, unless they are very similar in nature and you can provide a sound written justification why separate validations would be redundant.
    Other than that my previous answer holds.

    If you want to share photos or drawings confidentially just send me a PM and we can take it from there. The more specific a discussion, the more applicable the advice that can be given.

    Cheers,
    Ronen.