Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Mar;79(1):86-97.
doi: 10.1111/biom.13581. Epub 2021 Nov 8.

A novel statistical test for treatment differences in clinical trials using a response-adaptive forward-looking Gittins Index Rule

Helen Yvette Barnett  1 , Sofía S Villar  1 , Helena Geys  2 , Thomas Jaki  1   3
Affiliations

A novel statistical test for treatment differences in clinical trials using a response-adaptive forward-looking Gittins Index Rule

Helen Yvette Barnett et al. Biometrics. 2023 Mar.

Abstract

The most common objective for response-adaptive clinical trials is to seek to ensure that patients within a trial have a high chance of receiving the best treatment available by altering the chance of allocation on the basis of accumulating data. Approaches that yield good patient benefit properties suffer from low power from a frequentist perspective when testing for a treatment difference at the end of the study due to the high imbalance in treatment allocations. In this work we develop an alternative pairwise test for treatment difference on the basis of allocation probabilities of the covariate-adjusted response-adaptive randomization with forward-looking Gittins Index (CARA-FLGI) Rule for binary responses. The performance of the novel test is evaluated in simulations for two-armed studies and then its applications to multiarmed studies are illustrated. The proposed test has markedly improved power over the traditional Fisher exact test when this class of nonmyopic response adaptation is used. We also find that the test's power is close to the power of a Fisher exact test under equal randomization.

Keywords: allocation probability; inference; nonmyopic; power; testing for superiority.

PubMed Disclaimer

Figures

Figure 1
Figure 1
An illustration of the null distribution and two alternative distributions of Q, the total number of blocks for which the allocation probability to the experimental arm is greater than 0.5, for K = 10, B = 2 and nz = 2
Figure 2
Figure 2. Comparison of power for N = 40 & B = 2; rejection criteria adjusted for type I error rate
Figure 3
Figure 3. Comparison of power for N = 80 & B = 2; rejection criteria adjusted for type I error rate
Figure 4
Figure 4. Comparison of power for N = 160 & B = 2; rejection criteria adjusted for type I error rate
See this image and copyright information in PMC

References

    1. Attarian S, Vallat JM, Magy L, Funalot B, Gonnaud PM, Lacour A, et al. An exploratory randomised double-blind and placebo-controlled phase 2 study of a combination of baclofen, naltrexone and sorbitol (PXT3003) in patients with Charcot-Marie-Tooth disease type 1A. Orphanet Journal of Rare Diseases. 2014;9:1–15. - PMC - PubMed
    1. Gittins JC. Bandit processes and dynamic allocation indices. Journal of the Royal Statistical Society Series B. 1979;41:148–177.
    1. Gittins JC, Glazebrook K, Weber R. Multi-Armed Bandit Allocation Indices. 2nd edition. Chichester, UK: John Wiley & Sons, Inc; 2011.
    1. Hardwick JP, Stout QF. Bandit strategies for ethical sequential allocation. Computing Science and Statistics. 1991;23:421–424.
    1. He X, Madigan D, Yu B, Wellner J. Statistics at a crossroads: Who is for the challenge? Technical report, The National Science Foundation. 2019

Publication types