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
Randomized Controlled Trial
. 2017 Feb;24(1):63-70.
doi: 10.1080/09286586.2016.1255764. Epub 2016 Dec 16.

A Bayesian Analysis of a Randomized Clinical Trial Comparing Antimetabolite Therapies for Non-Infectious Uveitis

Erica N Browne  1 , Sivakumar R Rathinam  2 , Anuradha Kanakath  3 , Radhika Thundikandy  2 , Manohar Babu  3 , Thomas M Lietman  1   4   5   6 , Nisha R Acharya  1   4   5
Affiliations
Randomized Controlled Trial

A Bayesian Analysis of a Randomized Clinical Trial Comparing Antimetabolite Therapies for Non-Infectious Uveitis

Erica N Browne et al. Ophthalmic Epidemiol. 2017 Feb.

Abstract

Purpose: To conduct a Bayesian analysis of a randomized clinical trial (RCT) for non-infectious uveitis using expert opinion as a subjective prior belief.

Methods: A RCT was conducted to determine which antimetabolite, methotrexate or mycophenolate mofetil, is more effective as an initial corticosteroid-sparing agent for the treatment of intermediate, posterior, and pan-uveitis. Before the release of trial results, expert opinion on the relative effectiveness of these two medications was collected via online survey. Members of the American Uveitis Society executive committee were invited to provide an estimate for the relative decrease in efficacy with a 95% credible interval (CrI). A prior probability distribution was created from experts' estimates. A Bayesian analysis was performed using the constructed expert prior probability distribution and the trial's primary outcome.

Results: A total of 11 of the 12 invited uveitis specialists provided estimates. Eight of 11 experts (73%) believed mycophenolate mofetil is more effective. The group prior belief was that the odds of treatment success for patients taking mycophenolate mofetil were 1.4-fold the odds of those taking methotrexate (95% CrI 0.03-45.0). The odds of treatment success with mycophenolate mofetil compared to methotrexate was 0.4 from the RCT (95% confidence interval 0.1-1.2) and 0.7 (95% CrI 0.2-1.7) from the Bayesian analysis.

Conclusions: A Bayesian analysis combining expert belief with the trial's result did not indicate preference for one drug. However, the wide credible interval leaves open the possibility of a substantial treatment effect. This suggests clinical equipoise necessary to allow a larger, more definitive RCT.

Keywords: Antimetabolite; Bayesian; clinical trial; non-infectious uveitis; statistics.

PubMed Disclaimer

Conflict of interest statement

Conflicts of interest: No conflicting relationship exists for any author. The sponsor or funding organization had no role in the design or conduct of this research.

Figures

Figure 1
Figure 1
Uveitis specialists were surveyed on which drug, methotrexate or mycophenolate mofetil, is better for treatment of noninfectious intermediate, posterior, or pan-uveitis and were asked to provide an estimate of the relative decrease in effectiveness of one drug compared to the other with a 95% credible interval. The figure displays the point estimate with 95% credible interval provided by the 10 specialists and the cumulative group belief.
Figure 2
Figure 2
Plots of the prior distribution (blue), the likelihood of the trial results (red), and the posterior distribution (black) from a Bayesian analysis of a randomized clinical trial comparing methotrexate to mycophenolate mofetil for treating noninfectious uveitis. The prior distribution was created by expert opinion collected through an online survey. The likelihood function was from the frequentist result of the trial. Curves are plotted on the log-odds ratio scale, and points along the x-axis are labeled with the corresponding odds ratio.
See this image and copyright information in PMC

References

    1. Pressman AR, Avins AL, Hubbard A, Satariano WA. A comparison of two worlds: How does Bayes hold up to the status quo for the analysis of clinical trials? Contemp Clin Trials. 2011;32(4):561–568. - PMC - PubMed
    1. Moatti M, Zohar S, Facon T, Moreau P, Mary JY, Chevret S. Modeling of experts’ divergent prior beliefs for a sequential phase III clinical trial. Clin Trials. 2013;10(4):505–514. - PubMed
    1. Austin PC, Brunner LJ, Hux JE. Bayeswatch: an overview of Bayesian statistics. J Eval Clin Pract. 2002;8(2):277–286. - PubMed
    1. Brophy JM, Joseph L. Placing trials in context using Bayesian analysis. GUSTO revisited by Reverend Bayes. JAMA. 1995;273(11):871–875. - PubMed
    1. Abrams K, Ashby D, Errington D. Simple Bayesian analysis in clinical trials: a tutorial. Control Clin Trials. 1994;15(5):349–359. - PubMed

Publication types