ashwinim13
☆    

India,
2008-12-16 14:15
(5581 d 02:48 ago)

Posting: # 2930
Views: 19,655
 

 SAS Code for Outlier Detection in 2 way Crossover Study [Outliers]

Dear all

I would like to know the how do we find out outlier in Bioequivalence studies with the Lund's Test. What will be the SAS Code.
Thanks to everybody in advance...

With kind regards
atish_azad
☆    

2008-12-17 07:22
(5580 d 09:41 ago)

@ ashwinim13
Posting: # 2931
Views: 18,427
 

 SAS Code for Outlier Detection in 2 way Crossover Study

Dear Ashwinim 13,

Kindly refer book on Pharmaceutical Statistic practical and clinical application by Sanford Bolton for outlier detection by Lunds method. Hope this will help.

Regards,
Atish
ashwinim13
☆    

India,
2008-12-17 10:31
(5580 d 06:32 ago)

@ atish_azad
Posting: # 2932
Views: 18,221
 

 SAS Code for Outlier Detection in 2 way Crossover Study

Dear Atish,
Thanks for your prompt response.

I have checked Pharmaceutical Statistic practical and clinical application by Sanford Bolton for outlier detection at page no 308, code is mentioned but my doubt is at page no 309 it is written
"This model has 12 parameters & 36 observations.
Could you please explain this.

Thanks in advance
Regards
Ashwinim 13

--
Edit: Full quote removed. Please see this post! [Ohlbe]
d_labes
★★★

Berlin, Germany,
2008-12-17 10:33
(5580 d 06:30 ago)

@ ashwinim13
Posting: # 2933
Views: 18,693
 

 Lund's test obsolete

Dear Ashwinim13,

see Helmut's post here for the applicability of Lund's test.
You will recognize that Lund's test is outdated in BE studies.

BTW: Search, search and once again search. Hint: Search box is in the upper right corner :-D.

Regards,

Detlew
statistician
☆    

India,
2010-05-25 13:07
(5056 d 04:56 ago)

@ ashwinim13
Posting: # 5372
Views: 17,823
 

 SAS Code for Outlier Detection in 2 way Crossover Study

dear friends i am new to this forum i would like to know about the use of grubbs test in ba/be studies when to use it and i need sas grubbs outlier detection program for 2x2 cross over design. i will be thankfull if anyone clear my doubts.
thank you,
RK
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2010-05-28 14:10
(5053 d 03:53 ago)

@ statistician
Posting: # 5393
Views: 17,675
 

 Grubbs' test

Dear RK,

❝ i would like to know about the use of grubbs test in ba/be studies when to

❝ use it


:not really: I would not recommend it... There are better alternatives.
Have a look at the Outliers category and Search

❝ and i need sas grubbs outlier detection program for 2x2 cross over design.


:google:

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
KR
★    

India,
2010-05-31 17:18
(5050 d 00:45 ago)

@ Helmut
Posting: # 5404
Views: 17,804
 

 Grubbs' test

Dear HS,

I would like to know why u r not recommending the Grubbs' outlier test in bioequivalence? If this test is not strong enough then which is the better outlier test for bioequivalence study? Can we use Lund's outlier test instead of Grubbs' test as a better option?

Thanks in advance,

KR


Edit: Full quote removed. Please see this post! [Helmut]
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2010-05-31 19:59
(5049 d 22:04 ago)

@ KR
Posting: # 5406
Views: 17,325
 

 Quoting and editing

Dear KR,

please don’t quote the entire text of the previous post all the time (read this post).
We (the team of moderators) had to edit all (!!) of your previous posts...
Using the [image] would have displayed the mess. See also here. You may [image]Edit your post within 24 hours after posting – no need to open a new one (which I deleted this time).
Maybe it’s high time to read the forum’s policy.

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
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2010-06-08 17:15
(5042 d 00:47 ago)

@ KR
Posting: # 5474
Views: 17,805
 

 Outlier tests (EMA)

Dear KR and all,

at last week’s EGA workshop one of the issues was following statement (Section 4.1.10) of the BE-GL

For the acceptance interval to be widened the bioequivalence study must be of a replicate design where it has been demonstrated that the within-subject variability for Cmax of the reference compound in the study is >30%. The applicant should justify that the calculated intra-subject variability is a reliable estimate and that it is not the result of outliers.


No clarification is given in the presentations, but EMA’s panellists agreed that
  • no statistical test should be performed, and
  • Box plots may be accepted.
Well, box plots are a wide field (:vomit: nonparametric by nature!). Keywords: lower/upper adjacent values, inner and outer fence,… Most people would treat values outside the lower/upper quartile ±3× the interquartile range (IQR) as severe outliers (and ±1.5× IQR as mild outliers), but even the method of calculating the quartiles and IQR is not unambigous. ;-)

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
d_labes
★★★

Berlin, Germany,
2010-06-15 13:48
(5035 d 04:15 ago)

@ Helmut
Posting: # 5512
Views: 17,259
 

 EMA Outlier

Dear Helmut,

seems to me that this meeting was overshadowed by high statistical skills. :cool:
SCRN.

Regards,

Detlew
Dr_Dan
★★  

Germany,
2010-06-15 14:36
(5035 d 03:27 ago)

@ Helmut
Posting: # 5513
Views: 17,352
 

 Outlier tests (EMA)

Dear Helmut and all
At the EGA workshop you mentioned Lorelei Lutter pointed out the contradiction in the guideline: On the one hand no outlier test is accepted in order to evaluate the results of a BE study and on the other hand it should be demonstrated that the high within-subject variability in a replicate study is not the result of outliers. The whole problem with outliers is not solved by the regulatory authorities and they refuse to discuss this. Why should I detect outliers if no outlier test is accepted and the whole dataset should be evaluated? I suggest not to perform any outlier test.
I am looking forward to further comments regarding this topic.
Kind regards
Dan

Kind regards and have a nice day
Dr_Dan
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2010-06-15 14:56
(5035 d 03:06 ago)

@ Dr_Dan
Posting: # 5515
Views: 17,566
 

 Outlier tests (EMA & Canada)

Dear Dan & all!

❝ […] Lorelei Lutter pointed out the contradiction in the guideline:


Lorelei comes from Canada, where statistical outlier tests are recommended in the current drafted guideline – THX to Eric Ormsby! ;-)

If you want to follow EMA’s GL literally you have to act like a schizophrenic:
  • no outlier test in the general case
  • demonstrate that CV>30% of HVD(P)s is not caused by outlier(s) – but not by a statistical test

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
d_labes
★★★

Berlin, Germany,
2010-06-15 17:40
(5035 d 00:23 ago)

@ Helmut
Posting: # 5517
Views: 17,247
 

 EMA Outliers present

Dear Dan & dear Helmut!

Was there any POV what to do if there are outliers, how ever identified?

Regards,

Detlew
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2010-06-15 17:59
(5035 d 00:03 ago)

@ d_labes
Posting: # 5518
Views: 17,488
 

 EMA Outliers present

Dear D. Labes!

❝ Was there any POV what to do if there are outliers, how ever identified?


Easy. In a conventional BE study you define something in the protocol – subjects may be removed before any statistical procedure takes place (e.g., subject vomits, diarrhea). Once you decided the data-set is ready to be handled to the statistician, no statistical tests whatsoever. You may eye-ball strange results, but no way to remove a subject from the analysis.
In a replicate design where you want to go with ABEL, you have to demonstrate that the CV is not caused by outliers – again eye-balling (aka box plots).

Regulators are a strange bunch. Gerald Beuerle (ratiopharm) presented an example, where due to a sample mix-up in the area of Cmax the test would pass BE, which otherwise (either after exchange of values or after removal of the subject) would fail. Even then panellist stated that this is not acceptable. Sancta simplicitas! Kersti Oselin (MHRA) went one step further and stated, that one can expect such events and should power the study to be protected against that. :crying:

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
d_labes
★★★

Berlin, Germany,
2010-06-15 18:14
(5034 d 23:48 ago)

@ Helmut
Posting: # 5521
Views: 17,150
 

 EMA Outliers present

Dear Helmut!

❝ In a replicate design where you want to go with ABEL, you have to demonstrate that the CV is not caused by outliers – again eye-balling (aka box plots).


Boxplots of intra-individual contrasts R-R? Or what else (Mixed model further not acceptable)?

Ok, now it happens that I have an/some outlier(s). Or some assessor does not follow my arguing that I have none. What then? Scaling not allowed?

Regards,

Detlew
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2010-06-15 18:29
(5034 d 23:33 ago)

@ d_labes
Posting: # 5523
Views: 17,221
 

 EMA Outliers present

Dear D. Labes!

❝ Boxplots of intra-individual contrasts R-R?


No idea; but I would say that’s reasonable.

❝ Or what else (Mixed model further not acceptable)?


Nice question. Next question.

❝ Ok, now it happens that I have an/some outlier(s). Or some assessor does not follow my arguing that I have none. What then? Scaling not allowed?


Right guess. It’s a nightmare. :not really:

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;
80 visitors (0 registered, 80 guests [including 7 identified bots]).
Forum time: 17:03 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