Inspection with attributes . Result not to be considered for average.

[former_member1244710]

Dear QM GURUS,

My client has a scenario where there are multiple values of MIC to be recorded. we have configured the same through sampling procedure .Each MIC has about 30 values .all the values are then averaged into the summary screen .

At the time of individual results recording some of the values may have attributes such as non determinable or <100 or <10.we have configured the necessary attributes .but the when an attribute is used the system should not be ask for the actual value to be entered.the actual value should be allowed as blank and it should not be considered in the average in the summary screen .

for eg ..results can be recorded for an MIC in the following way (we shuld be able to record them this way)

1) 10

2) 12

3) 11

4) blank         <10

Average should not be of 4 but it should be only of 3 so (10+12+11)/3 =11.

currently the system does not allow the result to remain blank when we use the attribute and if we put it as 10 or 0 it is considered in the average.

Kindly suggest the solution ASAP along with config and master settings. correct answer would be readily rewarded with full points .

Ram                      

[Former Member]

There is no std way or any customization for this.

but you can explore Exit :

QEEM0029 - User-Exit for Characteristic Overview Subscreen         

or

QEEM0030  -User-Exit for Subscreen: Characteristic Single Screen

[former_member1244710]

Sujit, ..thanks for your reply and sorry for the delay in reply ...

As we are working on exploring the user exits given by you , we realize that there is no export value for the average (summarized) in the user exits and hence it seems quite challenging .I would love to have your inputs on the same.

In the meanwhile , we just observed the attributes of type "INVALID" that exactly creates the same effect as needed . The only problem is the number of inspected values on the summary screen is shown as ( total recorded values - with those with attributes invalid).

Do you recommend using the attribute Invalid ..? Our users are a little hesitant with this , but we can always use the text that doesn't show..invalid ..Let me know your views on this solution .(specially as I feel there must be a reason ..you said customization is not possible )

I have a related question..how many additional attributes can we create?...I need 15 attributes but looking at the limited customer attributes (and it is strange ...that SAP provides such a limitation) it looks like it is possible to create only about 12 attributes ...

As a last option if nothing works ...I intend to create only one attribute as "QUALITY ATTRIBUTE"(copied from invalid) and then users can specify...the additional details of the attributes in inspection description ..In our COA we can try to pick up the field from the inspection description ..but the problem invalid attribute is not counted in the inspected column in the summarized screen .

Do let me know your views in all this ...if it is okay with you ..Can I have your cell no ..I shall call you tomorrow .

Thank you

Ram .

[Former Member]

Hi

Are you doing this for certificate,? If yes, then why don't you make change in output of certificate keeping result recording std one !

Just a suggestion !

[former_member1244710]

Hi ,

Nope... not only for a certificate ..Can't restrict it  only in the output as there has to be integrity between average shown on the RR screen and the certificate . I am working in a regulated environment with GAMP 5 Validation ..

Thanks for your suggestions.

[former_member42743]

Ok.  I might be a bit confused but it seems to me you are doing apples to oranges here.  In one case you want to record <= a value, like 10.  and then you want to average the values above 10 together.  What is your logic there?  Lets say I have six samples.  If I have three values below the detectable limits of 10, and three values of 12, 14 and 16, what is the average?  SAP would consider it an average of 0,0,0,12,14, and 16. or a value of 7, which could then be reported as < 10.

But it sounds like from your posting you want to report the average of 12, 14 and 16 which would be 14.  But that would be incorrect since you measured three of the six units as not detectable. 

Now if the unit is not tested at all, then you should configure an attribute like invalid with text that says full sample size not recorded and than an explanation why.  If you are in a GMP environment you can't just leave a value blank.  You have to explain why a sample result is missing.  If the result is non-detectable, then according to the sensitivity of the test the value IS zero.  And the value should be included in the average.  I wonder if the sensitivity of the test is properly defined?  For instance, if you don't have the sensitivity to report a value between 1-9, than how do you have the sensitivity to report the differeenc between 12 and 13?  Shouldn't your allowed values in this instance be rounded to ten?  I.e. 10, 20, 30.. etc?

And yes.. I'm currently working in an FDA regulated environment and have worked with several pharma's over the years.

FF