Bioequivalence and Bioavailability Forum

Proc Mixed for >2 treatments [Software]

posted by H_Rotter – Germany, 2008-01-08 14:20 (6742 d 14:03 ago) – Posting: # 1480
Views: 14,828

Hi Matz!

❝ Hai All

  ^^ -- that's German for 'shark'

❝ Anybody please tell me what is the advantage of using Proc MIXED insted of Proc GLM in the SAS program, for the analysis of bioequivalence studies with missing values (subjects not completed some periods).

See this post by Jaime and Helmut's as well.
So if you have more than two treatments in your study you may use Proc MIXED.
In a 2×2 study you will get the same results - but Proc MIXED is more time consuming (iterative instead of explicit).

❝ Is it required to specify in the protocol that which procedure (GLM or MIXED) is use for the statistical analysis.

IMHO that's a good idea. Simply specify it as a decision tree:
No drop-outs (complete data from all periods) → Proc GLM
Drop-outs (incomplete data) → Proc MIXED

Regards,
Hermann

Complete thread:

• Proc Mixed mathews 2008-01-08 07:47 [Software] 
  • Proc Mixed for >2 treatmentsH_Rotter 2008-01-08 13:20
    • Proc Mixed for >2 treatments mathews 2008-01-17 08:02
      • Proc Mixed for >2 treatments H_Rotter 2008-01-17 23:29
        • Proc Mixed for >2 treatments mathews 2008-01-18 12:56
          • Proc Mixed for >2 Periods H_Rotter 2008-01-18 13:25