FDA group model in R [Two-Stage / GS Designs]

posted by VStus – Poland, 2016-10-17 17:32 (2719 d 00:17 ago) – Posting: # 16730
Views: 34,639

Dear mittyri,
Dear zizou,

❝ The ANOVA type I is the same as in PHX, but PE and CIs are not. I'm in stuck....


Thank you very much for your help!

I have checked proposed model formula in R's lm() versus SAS output of some old study report with unbalanced dataset.
I found perfect much here for group*treatment!

1. Analysis started as FDA Group Model 1:
> anovadata <- lm(log(data2$Cmax)~group+seq+seq*group+subj%in%(seq)+prd%in%group+drug+drug*group, data=data2, na.action=na.exclude)
> anova(anovadata)
Analysis of Variance Table

Response: log(data2$Cmax)
           Df Sum Sq Mean Sq F value Pr(>F)   
group       1   0.12   0.120    1.23  0.272   
seq         1   0.05   0.046    0.47  0.497   
drug        1   0.18   0.182    1.87  0.177   
group:seq   1   0.46   0.460    4.71  0.034 * 
seq:subj   59  25.25   0.428    4.38  3e-08 ***
group:prd   2   0.33   0.164    1.68  0.195   
group:drug  1   0.01   0.007    0.07  0.797   
Residuals  59   5.76   0.098

> drop1(anovadata, test="F")
Single term deletions

Model:
log(data2$Cmax) ~ group + seq + seq * group + subj %in% (seq) +
    prd %in% group + drug + drug * group
           Df Sum of Sq   RSS  AIC F value Pr(>F)   
<none>                   5.76 -255                   
group:seq   0      0.00  5.76 -255                   
seq:subj   59     25.25 31.01 -161    4.38  3e-08 ***
group:prd   2      0.32  6.09 -252    1.64    0.2   
group:drug  1      0.01  5.77 -256    0.07    0.8             


F and P values for group:drug are exactly the same as obtained by SAS:

Param   Source        DF  F Value   P-value      Model to be used:
lCmax   Group*Treat   1   0.07   0.7966   N.S.   Remove interaction


2. I've got same F and P values (at least for some of the effects) while proceeding to FDA Group Model 2 (or EMA's only model with all effects fixed), not mentioning exactly the same outputs for MSE/ Intra-subject CV and PE / 90%CIs :

> anovadata2 <- lm(log(data2$Cmax)~group+seq+seq*group+subj%in%(seq)+prd%in%group+drug, data=data2, na.action=na.exclude)
> drop1(anovadata2, test="F")
Single term deletions

Model:
log(data2$Cmax) ~ group + seq + seq * group + subj %in% (seq) +
    prd %in% group + drug
          Df Sum of Sq   RSS  AIC F value  Pr(>F)   
<none>                  5.77 -256                   
drug       1      0.21  5.98 -254    2.15    0.15   
group:seq  0      0.00  5.77 -256                   
seq:subj  59     25.25 31.02 -163    4.45 1.8e-08 ***
group:prd  2      0.33  6.10 -254    1.71    0.19


SAS output:
Param   Effect          DF   F Value   P-value
lCmax   Group           1    0.21      0.6475
lCmax   Seq             1    0.31      0.5822
lCmax   Group*Seq       1    1.08      0.3040
lCmax   Treat           1    2.15      0.1476
lCmax   Per(Group)      2    1.71      0.1896
lCmax   Subj(Group*Seq) 59   4.45      <.0001


Maybe lm() sometimes doing exactly the same as PROC GLM?
Regards, VStus

Complete thread:

UA Flag
Activity
 Admin contact
22,957 posts in 4,819 threads, 1,638 registered users;
75 visitors (0 registered, 75 guests [including 7 identified bots]).
Forum time: 16:49 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