A sequential procedure for comparing two experimental treatments with a control

Vincent, E., Todd, S. ORCID: https://orcid.org/0000-0002-9981-923X and Whitehead, J. (2002) A sequential procedure for comparing two experimental treatments with a control. Journal of Biopharmaceutical Statistics, 12 (2). pp. 249-265. ISSN 1054-3406 doi: 10.1081/BIP-120015747

Abstract/Summary

A procedure is described in which patients are randomized between two experimental treatments and a control. At a series of interim analyses, each experimental treatment is compared with control. One of the experimental treatments might then be found sufficiently superior to the control for it to be declared the best treatment, and the trial stopped. Alternatively, experimental treatments might be eliminated from further consideration at any stage. It is shown how the procedure can be conducted while controlling overall error probabilities. Data concerning evaluation of different doses of riluzole in the treatment of motor neurone disease are used for illustration.