Two regions: different populations [General Sta­tis­tics]

posted by Helmut Homepage – Vienna, Austria, 2020-08-31 11:26 (92 d 15:27 ago) – Posting: # 21900
Views: 2,617

Hi Brus,

» But, if α is not adjusted, since there are two reference products, even though they are from different regions, would there be no multiplicity issues?

Consider ElMaestro’s post.
The current regulatory thinking is that an α adjustment is not required since the populations of patients are different (as are the local reference products). If you are concerned about tourists, go ahead and adjust.* If you employ a lower α, any agency will accept that. However, depending on the variability expect roughly 25% larger sample sizes.
BTW, what if tourists switch from one regions’ reference to the other one’s? There is no guarantee that they are BE (see this post). Given, generally we don’t have to worry (see the references in this post).



library(PowerTOST)
# defaults: theta0 = 0.95, targetpower = 0.8
CV  <- seq(0.15, 0.4, 0.05)
res <- data.frame(CV = CV,
                  alpha1 = 0.050, n1 = NA, # unadjusted
                  alpha2 = 0.025, n2 = NA) # Bonferroni
for (j in 1:nrow(res)) {
  res$n1[j] <- sampleN.TOST(alpha = 0.050, CV = CV[j], details = FALSE,
                            print = FALSE)[["Sample size"]]
  res$n2[j] <- sampleN.TOST(alpha = 0.025, CV = CV[j], details = FALSE,
                            print = FALSE)[["Sample size"]]
}
res$penalty     <- sprintf("%+.2f%%", 100 * (res$n2 - res$n1) / res$n1)
names(res)[2:5] <- rep(c("alpha", "n"), 2)
print(res, row.names = FALSE)

   CV alpha  n alpha  n penalty
 0.15  0.05 12 0.025 16 +33.33%
 0.20  0.05 20 0.025 24 +20.00%
 0.25  0.05 28 0.025 36 +28.57%
 0.30  0.05 40 0.025 50 +25.00%
 0.35  0.05 52 0.025 66 +26.92%
 0.40  0.05 66 0.025 82 +24.24%


Dif-tor heh smusma 🖖
Helmut Schütz
[image]

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

Complete thread:

Activity
 Admin contact
21,226 posts in 4,427 threads, 1,482 registered users;
online 10 (1 registered, 9 guests [including 3 identified bots]).
Forum time: Wednesday 01:54 CET (Europe/Vienna)

A big data-analyst is an expert
in producing misleading conclusions from huge datasets.
It is much more efficient to use a statistician,
who can do the same with small ones.    Stephen Senn

The Bioequivalence and Bioavailability Forum is hosted by
BEBAC Ing. Helmut Schütz
HTML5