Adaptive sequential testing for multiple comparisons
Ping Gao, Lingyun Liu, Cyrus Mehta
Journal of Biopharmaceutical Statistics (Impact Factor: 0.72). 06/2014; DOI: 10.1080/10543406.2014.931409
ABSTRACT: We propose a Markov process theory based adaptive sequential testing procedure for multiple comparisons. The procedure can be used for confirmative trials involving multi-comparisons, including dose selection, or population enrichment. Dose or sub-population selection and sample size modification can be made at any interim analysis. Type I error control is exact.