swapnil.kuche
★    

2011-01-20 07:47
(4816 d 09:14 ago)

(edited by swapnil.kuche on 2011-01-20 11:20)
Posting: # 6452
Views: 23,402
 

 QA Audit Procedure [GxP / QC / QA]

Dear All,

Most of Quality Assurance departments used to claim that 20% raw data (Inprocess & Retrospective) and final bioequivalence report will be audited during course of conduct of the study.
My question is
  1. Do you feel it's ideal procedure to be followed?
  2. If yes. Do you have any reference regulatory guidance or any literature for the same supporting above mentioned statement?
    :ponder:
Thanks
Swapnil
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2011-01-20 13:05
(4816 d 03:56 ago)

@ swapnil.kuche
Posting: # 6453
Views: 22,218
 

 QA Audit Procedure

Dear Swapnil!

❝ Most of Quality Assurance departments used to claim that 20% raw data […] will be audited during course of conduct of the study.

❝ 1. Do you feel it's ideal procedure to be followed?


No. I don’t think that it makes sense to leave 80% unaudited. IMHO not less than 100% should be audited.

Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
ElMaestro
★★★

Denmark,
2011-01-20 14:02
(4816 d 02:59 ago)

@ swapnil.kuche
Posting: # 6454
Views: 22,123
 

 QA Audit Procedure

Dear Swapnil,

❝ Most of Quality Assurance departments used to claim that 20% raw data (Inprocess & Retrospective) and final bioequivalence report will be audited during course of conduct of the study.

❝ My question is

❝ 1. Do you feel it's ideal procedure to be followed?


Your 20% is not uncommon. Whether or not it is ideal depends on a lot of factors, incl. previous history of noncompliance. Let's say you start out with a much higher percentage than 20 in a young organisation but your qualified auditors don't ever find anything = a mature QMS and you can probably argue the percetange to go down. If on the other hand your auditors ID areas with noncompliance then these can with good argumentation be subject to more intensive audit than others. Along the same lines, if for example nobody left the bioanalytical dept. for six years but 30% of the study nurses in the clin. dept. quit last year then you have a reason for paying closer attention to that.
It sounds a little dumb but I think you should audit to the extent necessary, whether it be 1% or 99%. How to find out which percentage is right for you is a science in itself (risk analyses are popular at the moment; greedy course organisers will be happy to skin you alive).

Best regards,
EM.

Pass or fail!
ElMaestro
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2011-01-20 15:13
(4816 d 01:48 ago)

@ ElMaestro
Posting: # 6455
Views: 22,454
 

 QA Audit Procedure

Dear ElMaestro & Swapnil!

❝ It sounds a little dumb but I think you should audit to the extent necessary, whether it be 1% or 99%.


OK.

❝ How to find out which percentage is right for you is a science in itself (risk analyses are popular at the moment; greedy course organisers will be happy to skin you alive).


:-D

Well, Swapnil has asked for raw data audits. I think even more important is checking critical points, which cannot be corrected afterwards:
  • Administration
  • Transfer of plasma after centrifugation (potential mix-up)
IMHO both points should be covered either by a bar-code system or performed following the four-eye principle.

When it comes to data-transfer again it depends… When we switched from manual data-entry (CDSs ⇒ calibration ⇒ analytical results ⇒ biostatistical database) to electronic data flow within a LIMS we evaluated two studies in parallel during validation. We expected less than 1% of transfer errors (mainly transposed digits like 3657.45 instead of 3567.45) but were surprised to find ≈2%. Didn’t affect the studies’ outcome (maybe in the 6th decimal), but still interesting. Once the data-flow is validated, I think it's sufficient to check whether it was followed according to the SOPs (audit trail, hardcopies, :blahblah:).
However, if data come from external sources I don’t get the point why the data-import (analytics, lab values ⇒ statistics) should not be checked 100%. Takes about one hour for an average study. I have seen strange things (mostly caused by the nightmare Excel). Data in some (!) cells were formatted as text instead of numbers. The import seemed to work, but set text values to missing (the nice ‘.’ in SAS). Bad luck if you checked only 20%.

Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
Ohlbe
★★★

France,
2011-01-20 21:26
(4815 d 19:35 ago)

@ Helmut
Posting: # 6463
Views: 22,219
 

 QA Audit Procedure

Dear all,

Several comments:
  • First, let's not mix quality control of the data and quality assurance. Helmut, you mentioned a 100 % QA review of the data: I would say a 100 % QC is needed, but not necessarily a 100 % review by QA ! I am not aware of any specific percentage of data to be reviewed by QA.

  • There are several aspects to QA. One is study-related: that can be in-process audit of critical steps in a given study, and of course checking the report against raw data, or consistency of the raw data. Another aspect is system-based and study-independent (though you will probably pick a study as an example): that will be periodic audits of critical points (once again investigational product dispensing or dosing, sample processing...). These audits will be part of an annual quality audit plan aiming to check that the system works, that SOPs are followed, and to propose possible improvements to the system.
The combination of system-based audits and trial-based audits helps to check the reliability of the data and to (hopefully) increase the confidence in the system and its reliability.

Regards
Ohlbe

Regards
Ohlbe
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2011-01-21 05:23
(4815 d 11:38 ago)

@ Ohlbe
Posting: # 6465
Views: 22,006
 

 QC # QA

Dear Ohlbe!

100% agree. I must confess that I probably never in depth comprehended the subtile difference between QC and QA.
Thanks for the clarification!

Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
swapnil.kuche
★    

2011-01-21 06:32
(4815 d 10:29 ago)

@ Helmut
Posting: # 6466
Views: 22,040
 

 QC # QA

Dear Helmut, Ohlbe & ElMaestro,

Thanks for most comprehensive disscussion on my post. Through this disccusion i think you all elaborate ideal QA practices to be implemented. (As there is no such guidance is available for extend of QA audit). At the end I think it is important for any QA system to assure error free data with proper implementation of QMS system.
Similarly this disscussion will clear difference between QA & QC responsibilities.


Regards,
Swapnil
ElMaestro
★★★

Denmark,
2011-01-21 14:39
(4815 d 02:22 ago)

@ Helmut
Posting: # 6472
Views: 22,308
 

 QC # QA

Dear all,

❝ 100% agree. I must confess that I probably never in depth comprehended the subtile difference between QC and QA.


...and why oh friggin why are the GLP docs so completely unclear regarding QC...

EM.

Pass or fail!
ElMaestro
Ohlbe
★★★

France,
2011-01-23 16:18
(4813 d 00:43 ago)

@ ElMaestro
Posting: # 6482
Views: 22,037
 

 QC # QA

Ahoy Captain !

❝ ...and why oh friggin why are the GLP docs so completely unclear regarding QC...


True ! GCP make a clear difference between QC (monitoring) and QA (auditing), but there is nothing about QC in GLP.

Regards
Ohlbe

Regards
Ohlbe
ElMaestro
★★★

Denmark,
2011-01-25 14:15
(4811 d 02:46 ago)

@ Ohlbe
Posting: # 6488
Views: 21,939
 

 ISO17025 is King

Ahoy Ohlbe,

❝ True ! GCP make a clear difference between QC (monitoring) and QA (auditing), but there is nothing about QC in GLP.


<rant>

And while we are at it:
It is incomprehensible why traceability plays such a minor role in the GLP document.
Furthermore, to be honest ISO17025 is a much much better document as criteria for bioanalytics than GLP. I think the regulators should eliminate the GLP sentence from the EU BE guideline and mention ISO17025 in stead.

</rant>

Pass or fail!
ElMaestro
joy_fm
☆    

Indonesia,
2014-01-15 08:55
(3725 d 08:06 ago)

@ Ohlbe
Posting: # 12171
Views: 19,597
 

 QC # QA

Dear all,
so what is the conclusion ? What's the difference between QA and QC in daily activities ? Because both of department (QC and QA) check the same raw data to ensure that report reflect the raw data.

best regards
Joy_fm
riteshsrmc
☆    

India,
2014-01-16 13:55
(3724 d 03:06 ago)

@ joy_fm
Posting: # 12178
Views: 19,523
 

 QC # QA

Hi Joy.

QC: Dependent (May be biased)
QA: Independent: Unbiased (should not be biased)
Mahesh M
★    

India,
2014-01-16 14:30
(3724 d 02:31 ago)

@ riteshsrmc
Posting: # 12179
Views: 19,509
 

 QC # QA

Dear Joy.

QA is for ensuring quality in the processes.
QA aims to prevent mistake/ error with a focus on the process/ Regulatory requirements/SOP. It is a proactive quality process.

QC is for ensuring quality in raw data (study specific). Activities focus on identifying mistake/ error in the raw data.

Regards
mahesh M
ElMaestro
★★★

Denmark,
2014-01-16 14:35
(3724 d 02:26 ago)

@ joy_fm
Posting: # 12180
Views: 19,496
 

 QC # QA

Tricky, tricky with these definitions.

My proposal:
QA: Any activity related to assurance of quality: Training, SOP updates, QC, QMS maintenance and so on.
QC: The specific act of ensuring four eyes and/or check on source documentation.
Audit: The act of being an outright paranoid and unfair pain in the neck to someone who is in all likelihood completely innocent and doing a good job. :-D

EM.
joy_fm
☆    

Indonesia,
2014-01-16 15:55
(3724 d 01:06 ago)

@ ElMaestro
Posting: # 12182
Views: 19,575
 

 QC # QA

Dear riteshsrmc, Mahesh M, ElMaestro
thank you for the explanation. Very helpful :-)

But, I'm still confused by this statement in a previous post from Ohlbe

❝ and of course checking the report against raw data, or consistency of the raw data…


If QC has checked 100% raw data, why QA should check again?
and how to ensure consistency of the raw data, if QA do random check only as an external auditor do?
Because in OECD, only describes QA program and not about the QC Program


Best regards
Joy_fm


Edit: Standard quotes restored, post linked. [Helmut]
Mahesh M
★    

India,
2014-01-17 13:01
(3723 d 04:00 ago)

@ joy_fm
Posting: # 12185
Views: 19,517
 

 QC # QA

Hi Joy,

"If QC has checked 100% raw data, why QA should check again?"

Yes you are right because QA work as independent body means QA not repored to laboratory manager or analysist but QC reported to laboratory manager/ Scientest.

In practical "QC department is healping hand of analyst or lab manager to detact error or mistake"

Regards
Mahesh M
joy_fm
☆    

Indonesia,
2014-01-17 14:42
(3723 d 02:19 ago)

@ Mahesh M
Posting: # 12187
Views: 19,512
 

 QC # QA

Dear Mahesh M,
Thank u, I absolutely agree with your opinion.

Dear Helmutz,
Can u explain about this statement

❝ However, if data come from external sources I don’t get the point why the data-import (analytics, lab values ⇒ statistics) should not be checked 100%. Takes about one hour for an average study.


In my opinion, it will take at least two days to check 100% raw data. How to check 100% raw data in one hour ? What's your checking point ?


Best regards
Joy_fm


Edit: Standard quotes restored, post linked. [Helmut]
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-01-17 15:16
(3723 d 01:45 ago)

@ joy_fm
Posting: # 12188
Views: 19,532
 

 Waste of time

Hi Joy,

❝ Dear Helmutz,


Who?

❝ ❝ However, if data come from external sources I don’t get the point why the data-import (analytics, lab values ⇒ statistics) should not be checked 100%. Takes about one hour for an average study.


❝ In my opinion, it will take at least two days to check 100% raw data. How to check 100% raw data in one hour ? What's your checking point ?


Let’s say you have 900 data values in two tables (from the analytical and bio­statistical reports). Assuming an eight-hours working day (including a 30 minutes lunch-break) it would take you one minute to compare each data pair‽ I would say my four seconds are already rather slow.
And if it really would take you two days, so what?

PS: Please see this post.

Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
joy_fm
☆    

Indonesia,
2014-01-20 08:38
(3720 d 08:23 ago)

@ Helmut
Posting: # 12198
Views: 19,443
 

 QC # QA

Dear Helmut,
Thank u for the explanation. I'm sorry, because I wrote your name with "Z".

Best regards
Joy_fm
Ohlbe
★★★

France,
2014-01-17 16:27
(3723 d 00:34 ago)

@ Mahesh M
Posting: # 12189
Views: 19,657
 

 QC # QA

Dear Manesh,

❝ Yes you are right because QA work as independent body means QA not repored to laboratory manager or analysist but QC reported to laboratory manager/ Scientest.


Sorry but I totally disagree with this statement. Actually I would put it the other way round:
  • QC results can be discussed with whoever made a mistake, so that the mistake gets corrected, but do not necessarily need to be reported to management, unless there are systematic errors requiring corrective and preventative action;
  • QA audit results have to be reported to management. See OECD GLP §2.2.e:
The responsibilities of the quality assurance personnel include, but are not limited to, the following functions. They should: […] promptly report any inspection results in writing to management and to the study director […].

("Inspection" in this case should be understood as QA audit, not regulatory inspection), as shown by § 2.2.c).

Regards
Ohlbe
mittyri
★★  

Russia,
2014-01-20 15:29
(3720 d 01:32 ago)

@ Helmut
Posting: # 12205
Views: 19,492
 

 QC According to ICH E3

Dear Helmut & All,

According to the ICH E3 Guideline there could be an Appendix 16.1.10. "Documentation of inter-laboratory standardisation methods and quality assurance procedures if used" in the Clinical Study Report.
In our BEQ reports we include in this chapter only the information about method preparation and validation & QA (if any). The information about QC series is usually included to the chapter PK evaluation of the main report.

Could you explain your practice: have you used to include in this chapter QC data?

Kind regards,
Mittyri
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-01-20 16:20
(3720 d 00:41 ago)

@ mittyri
Posting: # 12206
Views: 19,490
 

 QC According to ICH E3

Hi Miityri,

❝ Could you explain your practice: have you used to include in this chapter QC data?


In my reports I use to give in this section certificates of the clinical laboratory (e.g., atten­dance in national interlaboratory comparison programmes, ISO 900x, accredited methods) covering parameters measured in pre-/post study exams.
I never referred to the bioanalytical method used for PK there.

Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
mittyri
★★  

Russia,
2014-01-22 19:44
(3717 d 21:17 ago)

@ Helmut
Posting: # 12229
Views: 19,499
 

 BE Report: Recommendations to the structure

Dear Helmut,

❝ I never referred to the bioanalytical method used for PK there.


Could you specify, where do you include method preparation and validation?

In accordance with your consent I would broaden the question:
BE studies are very specific, and typical Report structure (from E3) should be deeply modified. Sometimes keen discussions are conducted in our department about it. Many men, many minds... I cannot find any recommendations to the structure of BE Report. Please advise

Kind regards,
Mittyri
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-01-22 21:53
(3717 d 19:08 ago)

@ mittyri
Posting: # 12231
Views: 19,515
 

 BE creative!

Hi Miityri,

❝ ❝ I never referred to the bioanalytical method used for PK there.

❝ Could you specify, where do you include method preparation and vali­dation?


I give both the analytical and statistical reports as parts of the appendix. These reports contain their respective protocols as appendices within. Haven’t seen any information about method development in the E3 report so far.

❝ In accordance with your consent I would broaden the question:

❝ BE studies are very specific, and typical Report structure (from E3) should be deeply modified.


Absolutely. I hate to read reports, where you have headings on top of empty pages – or worse –

“this page contains no text”

printed across. Also brilliant: All of E3’s headings and some followed by “(not applic­able)”… E3 is almost twenty years old and it was not written having BE in mind.

❝ Sometimes keen discussions are conducted in our department about it. Many men, many minds...


Don’t you have NLYW in the department?

❝ I cannot find any recommendations to the structure of BE Report.


There are a few basic recommendations in national guidelines. They are all completely different but may inspire you. However, as a starter I recommend ICH’s Q&A on E3. Some quotes:

For example, pharmacokinetic […] results could be placed in appropriately iden­­­ti­­fied subsections of the efficacy and safety results sections, or they could be placed in new, appropriately identified results sections.
A rationale is not necessary if sections presented in E3 are re-ordered, re­named, or deleted (if warranted by the study design) or if new sections are added.
It should be noted that E3 was developed for submission of adequate and well-controlled clinical effectiveness studies. Nevertheless, the basic principles de­scribed can be applied to other kinds of trials, such as clinical pharma­co­logy studies […], recognizing that not all sections or data presentations may be appropriate or needed for these other types of trials. Sponsors are en­cour­aged to adapt the recommendations in the Guideline as needed (e.g., by de­let­ing sections that are not relevant or adding needed sections that are not mentioned in the Guideline).
It is appropriate to create new headings in the CSR and new Appendices for these topics. The Guideline provides for and focuses on Efficacy and Safety variables known at the time. Other topics should be well referenced in the CSR body and clearly identified in the Table of Contents.
Current submission options include:

  1. Stand alone reports
    These can be placed in “parallel” with the main clinical study report in the eCTD. For example, a clinical pharmacology study might have the clinical study report, a PK report, and an assay validation report.

(my emphases)

Be creative rather than formalistic. BTW, I don’t write anything about statistics under 16.1.9. ;-)

Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
UA Flag
Activity
 Admin contact
22,957 posts in 4,819 threads, 1,638 registered users;
77 visitors (0 registered, 77 guests [including 7 identified bots]).
Forum time: 17:01 CET (Europe/Vienna)

Nothing shows a lack of mathematical education more
than an overly precise calculation.    Carl Friedrich Gauß

The Bioequivalence and Bioavailability Forum is hosted by
BEBAC Ing. Helmut Schütz
HTML5