A class of confidence intervals for sequential phase II clinical trials with binary outcome

Shande Chen

Research output: Contribution to journalArticlepeer-review

Abstract

The phase II clinical trials often use the binary outcome. Thus, accessing the success rate of the treatment is a primary objective for the phase II clinical trials. Reporting confidence intervals is a common practice for clinical trials. Due to the group sequence design and relatively small sample size, many existing confidence intervals for phase II trials are much conservative. In this paper, we propose a class of confidence intervals for binary outcomes. We also provide a general theory to assess the coverage of confidence intervals for discrete distributions, and hence make recommendations for choosing the parameter in calculating the confidence interval. The proposed method is applied to Simon's [14] optimal two-stage design with numerical studies. The proposed method can be viewed as an alternative approach for the confidence interval for discrete distributions in general.

Original languageEnglish
Pages (from-to)479-489
Number of pages11
JournalJournal of Applied Statistics
Volume38
Issue number3
DOIs
StatePublished - 1 Mar 2011

Keywords

  • Binomial parameter
  • Coverage probability
  • Discrete distribution
  • Expected length
  • Simon's optimal design
  • Two-stage design

Fingerprint

Dive into the research topics of 'A class of confidence intervals for sequential phase II clinical trials with binary outcome'. Together they form a unique fingerprint.

Cite this