ABSTRACT
In one of their publications, the Pharmaceutical Research Manufacturer Association (PhRMA) Working Group on Adaptive Design defines an adaptive design as a clinical trial design that uses accumulating data to decide on how to modify aspects of the study as it continues, without undermining the validity and integrity of the trial (Gallo et al. 2006). PhRMA’s definition emphasizes the feature by design adaptations rather than ad hoc adaptations. Based on adaptations applied, Chow, Chang, and Pong (2005) classified adaptive trials into different types of adaptive design including adaptive group sequential design (see e.g., Kelly, Stallard, and Todd 2005; Kelly et al. 2005), adaptive dose finding design (see, e.g., Bornkamp et al. 2007; Krams et al. 2007), and (two-stage) adaptive seamless design (see, e.g., Maca et al. 2006; Chow 2008). An adaptation is referred to as a modification or a change made to trial procedures and/or statistical methods before, during, and after the conduct of a clinical trial. By definition, adaptations that are commonly employed in clinical trials can be classified into the categories of prospective adaptation, concurrent adaptation, and retrospective adaptation. Prospective adaptations include, but are not limited to, stopping a trial early due to safety, futility, or efficacy at interim analysis, dropping the losers (or inferior treatment groups), sample size reestimation, and so on. Thus, prospective adaptations are usually referred to as by design adaptations described in the PhRMA white paper (Gallo et al. 2006). Concurrent (or ad hoc) adaptations include, but are not limited to, any modifications made to trial procedure such as inclusion/exclusion criteria, evaluability criteria, dose/regimen, and treatment duration, and so on. Retrospective adaptation is usually referred to as changes in the statistical analysis plan, which often occur prior to database lock or unblinding. In practice, prospective and ad hoc adaptations are implemented by study protocol and protocol amendments, respectively. Note that for significant (or major) adaptations (changes) in the statistical analysis plan, a protocol amendment is necessarily implemented. It also should be noted that by design adaptations are not flexible as an adaptive
4.1 Introduction ...................................................................................... 4-1 4.2 Moving Target Patient Population ................................................. 4-2 4.3 Statistical Inference with Covariate Adjustment ......................... 4-3
4.5 Concluding Remarks ...................................................................... 4-18
design means to be. On the other hand, ad hoc adaptations via protocol amendments reflects real clinical practice, which give clinical investigators a lot of flexibility for identifying best clinical benefits of the test treatment under investigation. Retrospective adaptations derive the most appropriate statistical methods for data analysis without undermining the validity and integrity of the trial. In this chapter, our emphasis will be placed on the impact of protocol amendments in clinical trials utilizing adaptive design methods.
