Bioequivalence and Bioavailability Forum

Dear Detlew!

❝ First: Thanx for this valuable information! Indeed I have a 2-stage study under way.

You are not alone.

❝ Second: From Potvin et.al.:

❝ "If the individual ln-transformed data are to be used in the analysis instead of the differences, then the error term derived from the GLM ANOVA model including

❝ sequence2, stage1, period(stage)5, treatment6, subject(sequence × stage)4

❝ will give the appropriate s2 term."

❝ Thus the emphasis should be at sequence × stage .

Yep. Doesn’t the order of factors influence the analysis in SAS as well (aka type III limbo)?

❝ ❝ Note 1: The CI in the pooled analysis is not affected if compared to a model without subject(sequence × stage)...

❝

❝ Typo? If you fit without subject(sequence × stage) you would omit totally the subjects effects! This will heavily affect the CI I guess.

❝ If you meant sequence × stage you are totally right.

Sure – bloody typo; corrected.

❝ Since stage and sequence are between-subjects effects the additional inclusion of the sequence × stage interaction is only a further breakdown of the subjects effect. The question is why the mighty oracle want us to include such a term .

Maybe Señor García-Arieta’s footprints? See the quoted (d) in this post.

❝ BTW: Where did your p-value came from?

Grml; duno…

Model Specification and User Settings
       Dependent variable : Response
                Transform : LN
              Fixed terms : int+Stage+Sequence+Sequence*Stage+Subject(Sequence*Stage)+
                            Period(Stage)+Treatment

Class variables and their levels
                  Subject : 1  2  3  4  5  6  7  8  9  10
                            11 12 13 14 15 16 17 18 19 20
                    Stage : 1  2
                 Sequence : RT TR
                   Period : 1  2
                Treatment : R  T

Diagnostics
       Total Observations :           40
        Observations Used :           40
 Obs. Missing Model Terms :            0
              Residual SS :     0.780224
              Residual df :           17
        Residual Variance :    0.0458956

Partial Tests of Model Effects
            Hypothesis  Numer_DF  Denom_DF      F_stat     P_value
----------------------------------------------------------------------
                   int         1        17   1088.59       0
                 Stage         1        17     17.3504     0.000648307
              Sequence         1        17      0.369037   0.551559
        Sequence*Stage         1        17      2.22777    0.153875
Sequence*Stage*Subject        16        17     24.9524     1.04115E-08
          Stage*Period         2        17      0.967076   0.400153
             Treatment         1        17      0.0454252  0.833759

Now what? Will the mighty oracle not accept pooling since stages are highly significant different? Would not be the first time ignoring Potvin et al.

• Does not require poolability criteria (or at least should know whether results from both stages are poolable before sample analysis, i.e. base poolability on study conduct such as subject demographics, temporal considerations, use of same protocol, use of same site, etc., rather than a statistical test of poolability).

(my emphases)

P.S.: In the mixed-effects model I used in my previous studies [fixed: Sequence, Stage, Period(Stage), Treatment and random: Subject(Sequence × Stage)] p for stage is 0.403286…