Jump to content


Photo

SAS codes for BE assessment in PHX Winnonlin

SAS Winnonlin BE Connect

  • Please log in to reply
5 replies to this topic

#1 Mastan

Mastan

    Newbie

  • Members
  • Pip
  • 3 posts

Posted 07 December 2018 - 05:50 PM

Hello Friends,

I would like to use SAS software in Phoenix Winnonlin using Phoenix Connect to assess the bioequivalence for a crossover (PROC GLM) and replicate BE studies (PROC MIXED).

In this regard, if anyone had some experience, kindly share any project example or SAS codes to have a better understanding?

Best regards,
Mastan

#2 Ana Henry

Ana Henry

    Advanced Member

  • Members
  • PipPipPip
  • 220 posts

Posted 07 December 2018 - 06:23 PM

Hello,

Here is an old project from 2009 (done in version 6.0) that did BE in Phoenix and in SAS and compared results showing that they were the same.   In order to have SAS work correctly you need to set up the Edit>Preferences>SAS to indicate the location of the SAS executable.

 

Hope it helps,

Ana

Attached Files


Edited by Ana Henry, 07 December 2018 - 09:36 PM.


#3 Mastan

Mastan

    Newbie

  • Members
  • Pip
  • 3 posts

Posted 08 December 2018 - 02:38 AM

Hello Ana,

Thank you for your kind response.
I think this SAS code is reflecting PROC mixed and there is a slight variation in the outcome compared to WNL BE.

Best regards,
Mastan

#4 Simon Davis

Simon Davis

    Advanced Member

  • Administrators
  • 1,116 posts

Posted 08 December 2018 - 08:18 AM

HI Mastan, can you perhaps be a little more specific about your concern/ what type/size of difference you found.  This topic has been discussed many times both here and elsewhere e.g.

https://forum.bebac....sc=DESC#top1826

Ultimately SAS has a lot of different options possible to code, and it's the responsibility of the coder to ensure that they are using the appropriate code for their problem.

Phoenix's defaults and GUI try to guide you however ultimately the user should be confident in the statistical hypotheses being assessed.

  Simon.



#5 Mastan

Mastan

    Newbie

  • Members
  • Pip
  • 3 posts

Posted 09 December 2018 - 08:25 AM

Hi Simon,

Thank you for sharing this link which is useful to recollect the possible differences and codes impact on BE outcome.

There was a difference observed with the present SAS code used in the project scenario example (recommend by Ana) tried with an another data file (titles are same, only numerical data is different) of 2 to 4 digits difference (WNL Vs SAS = 98.02 - 116.48 Vs 94.32 Vs 113.54). Perhaps, the specific reasons to be explored.

Best regards,
Mastan

#6 Simon Davis

Simon Davis

    Advanced Member

  • Administrators
  • 1,116 posts

Posted 11 December 2018 - 08:59 AM

Mastan, that is a notable difference can you provide the Phoenix project, ideally with the corresponding SAS code included so we can see what might be the cause?  Remove the compound name/details if you need to.

 

 Simon.







Also tagged with one or more of these keywords: SAS Winnonlin, BE, Connect

0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users