crap [Two-Stage / GS Designs]
Hi Mauricio,
» » If I’m in the right mood I’ll write letter to ANVISA.
»
» PLEASE!!!!!
»
» Or make your official contribution on the website:
» http://formsus.datasus.gov.br/site/formulario.php?id_aplicacao=52824
No promises…
I played around around with published (and unpublished) methods. I used the noncentral t-distribution, whereas in the papers the shifted central t-distribution was used for speed reasons. One degree less in the sample size estimation because the stage-term is used in the pooled analysis. 100,000 simulations for the average total sample size E[N] and 1 mio for the empiric Type I Error. Narrow grid for CV (10–80%, step 2%), and n1 (12–72, step 2). The power/TIE surfaces are highly nonlinear; generally the maximum inflation is observed at a combination of low CV and small n1. The TIE is given for these locations (in the papers a wider grid with step sizes of 10% and 12 was used).
In the original methods no minimum stage 2 size; for the ANVISA I forced it to ≥ 50% n1. SLF refers to a manuscript by the usual Simul-Ants (Schütz, Labes, Fuglsang) we didn’t finish (rests in peace in my “dead dogs”-folder)…
TIE which is significantly >0.05 in red (limit of the binomial test 0.05036). I don’t understand why in some scenarios the TIE is lower with a minimum n2.
Counterintuitive.
R-code:
» » If I’m in the right mood I’ll write letter to ANVISA.

»
» PLEASE!!!!!

»
» Or make your official contribution on the website:
» http://formsus.datasus.gov.br/site/formulario.php?id_aplicacao=52824
No promises…
I played around around with published (and unpublished) methods. I used the noncentral t-distribution, whereas in the papers the shifted central t-distribution was used for speed reasons. One degree less in the sample size estimation because the stage-term is used in the pooled analysis. 100,000 simulations for the average total sample size E[N] and 1 mio for the empiric Type I Error. Narrow grid for CV (10–80%, step 2%), and n1 (12–72, step 2). The power/TIE surfaces are highly nonlinear; generally the maximum inflation is observed at a combination of low CV and small n1. The TIE is given for these locations (in the papers a wider grid with step sizes of 10% and 12 was used).
In the original methods no minimum stage 2 size; for the ANVISA I forced it to ≥ 50% n1. SLF refers to a manuscript by the usual Simul-Ants (Schütz, Labes, Fuglsang) we didn’t finish (rests in peace in my “dead dogs”-folder)…
Name Method Type GMR power alpha CV n1 E[N] TIE min.n2 E[N] ANVISA comp
SLF B 1 0.90 0.8 0.0272 0.20 12 40.8 0.04997 6 40.8 0.04999 higher
SLF B 1 0.90 0.9 0.0268 0.22 16 60.3 0.04985 8 60.3 0.04977 lower
Potvin B 1 0.95 0.8 0.0294 0.24 12 29.8 0.04876 6 29.9 0.04879 higher
Potvin-SLF B 1 0.95 0.8 0.0302 0.24 12 29.5 0.04999 6 29.6 0.05020 higher
Fuglsang B 1 0.95 0.9 0.0284 0.22 12 31.7 0.04960 6 31.7 0.04958 lower
Fuglsang-SLF B 1 0.95 0.9 0.0286 0.22 12 31.6 0.04999 6 31.6 0.05032 higher
Montague D 2 0.90 0.8 0.0280 0.20 12 40.3 0.05180 6 40.3 0.05181 higher
Montague-SLF D 2 0.90 0.8 0.0268 0.18 12 32.7 0.04998 6 32.7 0.04980 lower
Fuglsang C/D 2 0.90 0.9 0.0269 0.18 12 41.8 0.05021 6 41.8 0.05011 lower
Fuglsang-SLF C/D 2 0.90 0.9 0.0266 0.18 12 42.0 0.04995 6 42.0 0.04967 lower
Potvin C 2 0.95 0.8 0.0294 0.22 12 24.9 0.05143 6 24.9 0.05136 lower
Potvin-SLF C 2 0.95 0.8 0.0282 0.10 16 16.0 0.05010 8 16.0 0.05010 equal
Fuglsang C/D 2 0.95 0.9 0.0274 0.10 16 16.0 0.05010 8 16.0 0.05010 equal
Fuglsang-SLF C/D 2 0.95 0.9 0.0275 0.20 12 25.8 0.04962 6 25.8 0.04985 higher
TIE which is significantly >0.05 in red (limit of the binomial test 0.05036). I don’t understand why in some scenarios the TIE is lower with a minimum n2.
Counterintuitive.

R-code:
library(Power2Stage)
even.n2 <- function(n1, pct) {
ceiling(n1 * (1 + pct/100) / 2) * 2 - n1
}
alpha0 <- 0.05 # for type 2 designs
# locations of TIE (narrow grid)
CV <- c(0.24, 0.24, 0.22, 0.10, 0.22, 0.22, 0.10, 0.20, 0.20,
0.22, 0.20, 0.18, 0.18, 0.18)
n1 <- c(12, 12, 12, 16, 12, 12, 16, 12, 12, 16, 12, 12, 12, 12)
min.n2 <- even.n2(n1, 50)
cond <- data.frame(Name = c(rep(c("Potvin", "Potvin-SLF"), 2),
rep(c("Fuglsang", "Fuglsang-SLF"), 2),
rep("SLF", 2), "Montague", "Montague-SLF",
"Fuglsang", "Fuglsang-SLF"),
Method = c(rep("B", 2), rep("C", 2), rep("B", 2),
rep("C/D", 2), rep("B", 2), rep("D", 2),
rep("C/D", 2)),
Type = c(rep(1, 2), rep(2, 2), rep(1, 2), rep(2, 2),
rep(1, 2), rep(2, 2), rep(2, 2)),
GMR = c(rep(0.95, 8), 0.90, 0.90, rep(0.90, 2),
rep(0.90, 2)),
power = c(rep(0.80, 4), rep(0.90, 4), 0.80, 0.90,
rep(0.80, 2), rep(0.90, 2)),
alpha = c(0.0294, 0.0302, 0.0294, 0.0282, 0.0284, 0.0286,
0.0274, 0.0275, 0.0272, 0.0268, 0.0280, 0.0268,
0.0269, 0.0266),
CV = CV, n1 = n1, stringsAsFactors = FALSE)
res <- cbind(cond, ASN = NA, TIE = NA, min.n2 = min.n2, ASN.1 = NA,
ANVISA = NA, comp = "equal", stringsAsFactors = FALSE)
for (j in 1:nrow(cond)) {
ifelse (cond$Type[j] == 1, method <- "B", method <- "C")
x1 <- power.tsd(method = method, alpha0 = alpha0,
alpha = rep(cond$alpha[j], 2), n1 = cond$n1[j],
GMR = cond$GMR[j], CV = cond$CV[j],
targetpower = cond$power[j])
x2 <- power.tsd(method = method, alpha0 = alpha0,
alpha = rep(cond$alpha[j], 2), n1 = cond$n1[j],
GMR = cond$GMR[j], CV = cond$CV[j],
targetpower = cond$power[j], theta0 = 1.25)
res$ASN[j] <- round(x1$nmean, 1)
res$TIE[j] <- signif(x2$pBE, 4)
y1 <- power.tsd(method = method, alpha0 = alpha0,
alpha = rep(cond$alpha[j], 2), n1 = cond$n1[j],
GMR = cond$GMR[j], CV = cond$CV[j],
targetpower = cond$power[j], min.n2 = res$min.n2[j])
y2 <- power.tsd(method = method, alpha0 = alpha0,
alpha = rep(cond$alpha[j], 2), n1 = cond$n1[j],
GMR = cond$GMR[j], CV = cond$CV[j],
targetpower = cond$power[j], min.n2 = res$min.n2[j],
theta0 = 1.25)
res$ASN.1[j] <- round(y1$nmean, 1)
res$ANVISA[j] <- signif(y2$pBE, 4)
}
names(res)[c(9, 12)] <- rep("E[N]", 2)
res$comp[which(res$ANVISA > res$TIE)] <- "higher"
res$comp[which(res$ANVISA < res$TIE)] <- "lower"
print(res[order(res$Type, res$GMR, res$power, res$Name, res$Method,
decreasing = c(FALSE, FALSE, TRUE, FALSE, TRUE)), ],
row.names = FALSE)
—
Dif-tor heh smusma 🖖
Helmut Schütz
![[image]](https://static.bebac.at/img/CC by.png)
The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
Dif-tor heh smusma 🖖
Helmut Schütz
![[image]](https://static.bebac.at/img/CC by.png)
The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
Complete thread:
- ANVISA guidelines for two-stage design Mauricio Sampaio 2020-02-11 20:52 [Two-Stage / GS Designs]
- Brainless copy-and-paste-devil Helmut 2020-02-12 02:26
- Brainless copy-and-paste-devil Mauricio Sampaio 2020-02-12 17:33
- ANVISA guidelines for two-stage design ElMaestro 2020-02-12 08:27
- ANVISA guidelines for two-stage design nobody 2020-02-12 08:49
- ANVISA guidelines for two-stage design Helmut 2020-02-12 13:12
- ANVISA guidelines for two-stage design Mauricio Sampaio 2020-02-12 17:04
- ANVISA guidelines for two-stage design Mauricio Sampaio 2020-02-12 17:11
- ANVISA guidelines for two-stage design ElMaestro 2020-02-12 21:54
- crap Helmut 2020-02-15 17:53
- crap Mauricio Sampaio 2020-02-16 05:01
- crapHelmut 2020-02-16 15:43
- crap d_labes 2020-02-16 19:34
- Explained crap remains crap Helmut 2020-02-17 12:49
- Proposed changes Mauricio Sampaio 2020-02-17 13:50
- Proposed changes Helmut 2020-02-17 14:16
- Proposed changes Mauricio Sampaio 2020-02-17 13:50
- Explained crap remains crap Helmut 2020-02-17 12:49
- crap d_labes 2020-02-16 19:34
- crapHelmut 2020-02-16 15:43
- crap Mauricio Sampaio 2020-02-16 05:01
- crap Helmut 2020-02-15 17:53
- Brainless copy-and-paste-devil Helmut 2020-02-12 02:26