Double Entry of Values in Results Recording

[cristin_charbonneau]

Hi Experts.

Does anyone know how to prompt the user doing results recording to re-enter the values (as verification) as they go?  I've been asked to look into this concept but I feel that using authorizations to control who enters results versus who closes results versus who completes the UD should be sufficient to satisfy QA concerns.

Thoughts?

Cristin

[former_member42743 · Accepted solution]

I don't know of a way to enter double results for verification.  The only thought I have on this is to use a sample size of 2 instead of 1.  Record the results.  (the average is what will be used by the system).

Then program a user exit. Maybe one of these:

QEEM0006	User exit: add. functions after closing inspection characs
QEEM0007	User exit: add. functions after completing partial samples

In the exit, have the program check the individual values recorded against the mean value calulated by the system.  All three should be the same.  Have it throw an error to the user if they don't match.

The problem I see is building logic into the program to decide which characteristics to check.  You might only want to check select or critical characteristics.  While I'm usually against hard coding values into exits, I think I might create a special sampling procedure for this. I.e. like a FIXED-2 but maybe CKFIX-2.  So the program does the check only on characteristics with a sampling procedure of CHFIX-2.  All other characteristics are ignored.  Keep in mind that by recording two values now for a characteristic means they have to use a different recording screen.  They won't be able to make all their entries on the first, main screen.  This is why I'd want to do this only for critical MIC's.

I don't think the programming would be all that difficult to do.  I'm not sure if the actual mean will be available at the time of the user exit so you may have to have the program just check the two values.  It should be easy, value 1 + value 2 = x   Then x/value 1 = y.  If y <> 2.0000 = error.

BTW: If you do this, you might want to create an attribute to be recorded for the second value.  Like an V = verification entry.  This would indicate that the second result is simply a reentry of the first value and not an actual, real second test.  Any SOP's or documentation may need to be updated as well to reflect this type of new procedure.

Personally, I agree with you that the other reviews you have should be enough. But maybe this might be warranted for a few very select critical MIC's or ones that are highly subject to keyboarding error. For instance, values that are very small.  If values are typically recorded in ranges of .0001 - .0009 it could be very easy to leave off a zero and enter ..001 or .008 instead of .0001 or .0008.

FF

[cristin_charbonneau · Accepted solution]

Well that's a creative solution.   I'll give that a try and see how it works.  Then I will present it to them and if they really feel they must have it maybe they can choose the more critical ones as you say.

I'm not concerned about the results recording screen since I am using single results with a serial number effectively forcing them to use the "unit to be inspected" screen anyway.

Thanks as always,

Cristin

Message was edited by: Cristin Charbonneau Actually, this might not work.  I can't have a sample size of 2 when I am only receiving 1.  I need to only receive and inspect 1 unit at a time. I could change my MIC to allow for larger scope while keeping the sample procedure as fixed to 1? But then in the exit I guess I need to count how many results they recorded for the critical MICs to make sure they did the double entry and go from there?

[michelle_cannon · Accepted solution]

Hi Cristin

I don't 100% understand your business scenario but here are two other suggestions.

1. Implement the EHP3 digital signature functionality. This will allow the original results recorder to enter results & execute digital signature. Then the reviewer/verifier must execute their digital signature in QEDS which forces them to verify they opened & read the results. Finally the UD person executes their digital signature in QA11.

The down size of this solution is that digital signature must be done on ALL characteristics within the inspection lot & not just the ones that you want second verification on. But then again digital signature is not a bad thing if you are a FDA/Regulated Authority company.

2. You can use the user exits Mr Firefighter suggested and to zone into specific characteristics you could use the Info Field 1 in the characteristic (QPMK_DUMMY10) in QS21 to denote that characteristics with a symbol present in this field must obtain a second verification result entry before it goes to status "5" and allows closure of the characteristic.

With Development there is the possibility of developing a new screen to appear during results recording. When a characteristic has the Info Field 1 populated, if the user goes to do results recording it will popup a new screen that will force them to put in two entries (because you have put the number 2 in the Info Field 1 of the characteristic) although the sample procedure is only for n=1. They will not be able to close the characteristics unless the result is recorded twice in the pop-up box. This is an alternative to Firefighters suggestion but in case you must have the sampling procedure stay as n=1. The reason I mention this is in case you have sample management active & by having a sample size of 2 it will force your samplers to take two samples when in reality you only need one for testing so you are potentially wasting money taking extra samples.

Of course this is all development but very possible.

Good luck & let us know what was your final solution.

M

[former_member42743 · Accepted solution]

Yeah.. I've been accused of being to creative at times!

I will have to look into this a bit. But I know you can record multiple results for the same characteristic even though your quanitity is 1.

This is required for tests like total solids where test variability requires two runs of the same sample.  So two pans are weighed, sample placed in pan, cooked in oven, pans re-weighed and the two results are entered and the average is used. 

I think maybe you need to select "independent multiple sample" in the sampling procedure and then enter in the number of samples, (new field after you select radio button and hit enter).

But you'd have to test it.  I haven't done one of those for quite a while.

Michelle has some good ideas as well!

FF

[Martin_H · Accepted solution]

The only limitation to this is that even "independent multiple sample" allows the creation of a new sample only if the recorded NC of the first sample is equal or above d.

Regards

MH

[cristin_charbonneau]

Thanks to all of you for your suggestions. 

We are still in development and have prototyped the following solution. Using a flag digit ("2") in Info Field 1 of the characteristic (per Michelle's suggestion) ABAP implemented an implicit enhancement to generate a popup screen to prompt the user to re-enter the value. Incorrect duplicate entry results in an error and the user cannot close the characteristic.  (Per Fire Fighter) this will only be implemented for characteristics considered critical to the business (and numeric only - we didn't attempt qualitative).

This is our mock-up:

We still intend to use authorizations to ensure that the person entering the results is not the person who closes or completes the UD.

Thanks!

Cristin