An adaptive allocation for continuous response using Wilcoxon–Mann–Whitney score

Optimal response and covariate-adaptive biased-coin designs for clinical trials with continuous multivariate or longitudinal responses

“…Examples include the play-the-winner (PW) design (Zelen, 1969), the randomized play-the-winner (RPW) design (Wei and Durham, 1978), the success driven design (Durham et al, 1998) and the drop-the-loser (DL) rule (Ivanova, 2003). Related designs for continuous responses, using non-parametric methods to discretise the problem, include Rosenberger (1993) and Bandyopadhyay and Biswas (2004).…”

Multi-treatment optimal response-adaptive designs for phase III clinical trials

“…Relatively fewer designs are available for continuous responses. These include the linear rank test statistic based design (Rosenberger, 1993), link function based design (Bandyopadhyay & Biswas, 2001), Wilcoxon score based design (Bandyopadhyay & Biswas, 2004), utility based design (Atkinson & Biswas, 2005), and the continuous dropthe-loser design (Ivanova, Biswas, & Lurie, 2006), among others. Real life applications of adaptive allocation designs are due to Bartlett et al (1985), Biswas and Dewanji (2004), Rout et al (1993) and Tamura et al (1994) among others.…”

A new response-adaptive design for continuous treatment responses for phase III clinical trials

“…For example, the play-the-winner (PW) rule (Zelen, 1969), the randomized play-the-winner (RPW) rule (Wei and Durham, 1978), the generalized Pó lya urn (GPU) design (Wei, 1979), and the drop-the-loser (DL) rule (Ivanova, 2003) are all for binary treatment responses. The designs for continuous responses are the linear rank test statistic based design (Rosenberger, 1993), link function based design (Bandyopadhyay and Biswas, 2001), Wilcoxon score based design (Bandyopadhyay and Biswas, 2004), and utility function based design (Atkinson and Biswas, 2005), among others. Real life applications of response-adaptive designs are due to Bartlett et al (1985), Rout et al (1993), Tamura et al (1994), Biswas and Dewanji (2004), among others.…”

“…Note that, apart from the designs given by Rosenberger (1993) and Bandyopadhyay and Biswas (2004), all other designs assume specific parametric models of the response distributions or at least assume the existence of the first two moments. But the existence of the first two moments cannot be always ensured for the response distributions under clinical trials.…”