We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ME

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Methodology

Title: Optimal multiple testing and design in clinical trials

Abstract: A central goal in designing clinical trials is to find the test that maximizes power (or equivalently minimizes required sample size) for finding a true research hypothesis subject to the constraint of type I error. When there is more than one test, such as in clinical trials with multiple endpoints, the issues of optimal design and optimal policies become more complex. In this paper we address the question of how such optimal tests should be defined and how they can be found. We review different notions of power and how they relate to study goals, and also consider the requirements of type I error control and the nature of the policies. This leads us to formulate the optimal policy problem as an explicit optimization problem with objective and constraints which describe its specific desiderata. We describe a complete solution for deriving optimal policies for two hypotheses, which have desired monotonicity properties, and are computationally simple. For some of the optimization formulations this yields optimal policies that are identical to existing policies, such as Hommel's procedure or the procedure of Bittman et al. (2009), while for others it yields completely novel and more powerful policies than existing ones. We demonstrate the nature of our novel policies and their improved power extensively in simulation and on the APEX study (Cohen et al., 2016).
Subjects: Methodology (stat.ME); Statistics Theory (math.ST); Applications (stat.AP)
MSC classes: 62K05
Cite as: arXiv:2104.01346 [stat.ME]
  (or arXiv:2104.01346v2 [stat.ME] for this version)

Submission history

From: Ruth Heller [view email]
[v1] Sat, 3 Apr 2021 08:54:13 GMT (9766kb,D)
[v2] Mon, 31 Jan 2022 09:54:03 GMT (7952kb,D)

Link back to: arXiv, form interface, contact.