Details
| Original language | English |
|---|---|
| Pages (from-to) | 292-310 |
| Number of pages | 19 |
| Journal | Journal of Biopharmaceutical Statistics |
| Volume | 19 |
| Issue number | 2 |
| Early online date | 12 Feb 2009 |
| Publication status | Published - 2009 |
Abstract
The simultaneous comparison of proportions of success between many treatments and one control group is a common problem in randomized clinical trials or toxicity studies. In this article, three recently recommended asymptotic confidence interval approaches for the difference of proportions are adjusted for multiplicity, taking the correlation into account. The coverage probability of the resulting interval methods is compared in a simulation study using parameter settings relevant for clinical trials. For moderate to small sample sizes, a method adding two successes and two failures can be recommended. The usage of the proposed methods is illustrated by two examples; an R package is available.
Keywords
- Binomial, Comparison to control, Multiple comparisons, One-sided confidence intervals, Risk difference
ASJC Scopus subject areas
- Mathematics(all)
- Statistics and Probability
- Pharmacology, Toxicology and Pharmaceutics(all)
- Pharmacology
- Medicine(all)
- Pharmacology (medical)
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
In: Journal of Biopharmaceutical Statistics, Vol. 19, No. 2, 2009, p. 292-310.
Research output: Contribution to journal › Article › Research › peer review
}
TY - JOUR
T1 - Asymptotic simultaneous confidence intervals for many-to-one comparisons of binary proportions in randomized clinical trials
AU - Schaarschmidt, Frank
AU - Biesheuvel, Egbert
AU - Hothorn, Ludwig A.
PY - 2009
Y1 - 2009
N2 - The simultaneous comparison of proportions of success between many treatments and one control group is a common problem in randomized clinical trials or toxicity studies. In this article, three recently recommended asymptotic confidence interval approaches for the difference of proportions are adjusted for multiplicity, taking the correlation into account. The coverage probability of the resulting interval methods is compared in a simulation study using parameter settings relevant for clinical trials. For moderate to small sample sizes, a method adding two successes and two failures can be recommended. The usage of the proposed methods is illustrated by two examples; an R package is available.
AB - The simultaneous comparison of proportions of success between many treatments and one control group is a common problem in randomized clinical trials or toxicity studies. In this article, three recently recommended asymptotic confidence interval approaches for the difference of proportions are adjusted for multiplicity, taking the correlation into account. The coverage probability of the resulting interval methods is compared in a simulation study using parameter settings relevant for clinical trials. For moderate to small sample sizes, a method adding two successes and two failures can be recommended. The usage of the proposed methods is illustrated by two examples; an R package is available.
KW - Binomial
KW - Comparison to control
KW - Multiple comparisons
KW - One-sided confidence intervals
KW - Risk difference
UR - http://www.scopus.com/inward/record.url?scp=60749117479&partnerID=8YFLogxK
U2 - 10.1080/10543400802622501
DO - 10.1080/10543400802622501
M3 - Article
C2 - 19212881
AN - SCOPUS:60749117479
VL - 19
SP - 292
EP - 310
JO - Journal of Biopharmaceutical Statistics
JF - Journal of Biopharmaceutical Statistics
SN - 1054-3406
IS - 2
ER -