WinnieH
☆

Sweden,
2023-01-04 14:33
(79 d 14:23 ago)

Posting: # 23417
Views: 784

## Questions on under­standing the 4-way RSABE method SAS code [RSABE / ABEL]

Dear all,

I am a new learner and I am trying on understanding the 4-way RSABE method SAS code provided by FDA:
/*Intermediate analysis – ilat*/
proc mixed data=scavbe;
class seq;
model ilat =seq/ddfm=satterth;
estimate 'average' intercept 1 seq 0.5 0.5/e cl alpha=0.1;
ods output CovParms=iout1;
ods output Estimates=iout2;
ods output NObs=iout3;
run;
data ilat;
set iout2;
pointest=exp(estimate);
x=estimate**2-stderr**2;
boundx=(max((abs(lower)),(abs(upper))))**2;
/*Intermediate analysis – dlat*/
proc mixed data=scavbe;
class seq;
model dlat=seq/ddfm=satterth; estimate 'average' intercept 1 seq 0.5 0.5/e cl alpha=0.1;
ods output CovParms=dout1;
ods output Estimates=dout2;
ods output NObs=dout3;
run;
data dlat;
merge dout1(keep=estimate) dout2(keep=df);
s2wr=estimate/2;
dfd=df;
run;
/*Criterion Bound */
data results;
merge ilat(keep=pointest x boundx) dlat(keep=s2wr dfd);
theta=((log(1.25))/0.25)**2;
y=-theta*s2wr;
boundy=y*dfd/cinv(0.95,dfd);
sWR=sqrt(s2wr);
critbound=(x+y)+sqrt(((boundx-x)**2)+((boundy-y)**2));
run;

I found some discussion about the S2wr and it is really helpful to me. About the ilat, my understanding is that the model is a simple linear model and the estimate is the least square mean. My R code example is as follows:
mod<-lm(ilat~SEQ, dataset)
result <- confint(emmeans(mod, specs=~1), level = 0.9)
pointest <- exp(result$emmean) x <- (result$emmean)^2- (result$SE)^2 boundx <-(max(c(abs(result$lower.CL), abs(result\$upper.CL))))^2

My question is 1) how to interpret x and critbound=(x+y)+sqrt(((boundx-x)**2)+((boundy-y)**2))? Is there any related explanation about them? 2) Has anyone tried to reproduce this FDA code in R before? I would like to get some hints from you

Thank you very much and wish you a happy new year!

Bests,
Winnie
Helmut
★★★

Vienna, Austria,
2023-01-04 15:02
(79 d 13:54 ago)

@ WinnieH
Posting: # 23418
Views: 699

## Questions on understanding the 4-way RSABE method SAS code

Hi Winnie,

❝ I am a new learner and I am trying on understanding the 4-way RSABE method SAS code provided by FDA: […]

❝ I found some discussion about the S2wr and it is really helpful to me. About the ilat, my understanding is that the model is a simple linear model and the estimate is the least square mean.

I don’t speak SAS but isn’t it a restricted maximum likelihood estimate?

❝ My R code example is as follows: […]

I’m short in time right now; compare it to Detlew’s script in base R.

❝ My question is 1) how to interpret x and critbound=(x+y)+sqrt(((boundx-x)**2)+((boundy-y)**2))? Is there any related explanation about them?

Haha, ‘the unknown x’.  I’m aware of only one reference,* which is – regrettably – not in the public domain.
• The statistical approach we use is very similar to that proposed by Tothfalusi,
Endrenyi, et al. 2001, with a minor difference (use of an unbiased estimator for
(μTμR)2).

• Schuirmann D. U.S. FDA Perspective: Statistical Aspects of OGD’s Approach to Bioequivalence (BE) As­sess­ment for Highly Variable Drugs. Presentation at the 2nd conference of The Global Harmonisation Ini­ti­a­tive (GBHI). Rockville. September 15–16, 2016.

Dif-tor heh smusma 🖖🏼 Довге життя Україна!
Helmut Schütz

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
WinnieH
☆

Sweden,
2023-01-04 16:12
(79 d 12:44 ago)

@ Helmut
Posting: # 23419
Views: 665

Hi Helmut,

Best regards,
Zhe