<img alt="" src="https://secure.lote1otto.com/219869.png" style="display:none;">
Skip to content

Bayesian Methods

leftside leftside clark-van-der-beken-IQeBZm8F-M4-unsplash

Bayesian methods are a form of statistical analysis where a hypothesis called a prior is systematically updated as more data comes in. This makes Bayesian methods ideal for small sample trials, clinical trials using historical data and real-world solutions, or clinical trials that  require flexible learning.  These computationally intensive statistical methods were once only accessible to small number of specialized statisticians working in top-20 pharmaceuticals. For over a decade, statisticians at Cytel have been committed to ensuring that clinical solutions requiring Bayesian statistics are made accessible to all who need them.

How do I know if I need Bayesian methods?

Bayesian methods are used as solutions for a wide variety of clinical trial challenges. The adoption of historical and external data into a clinical trial to supplement regulatory submissions is a growing use of Bayesian methods. Similarly, in therapeutic areas like oncology, where standard of care might evolve during the course of a clinical trial, statisticians trained in Bayesian methods can help salvage data collected during a clinical trial.

The nature of statistics is such that many common problems can be solved using both Frequentist or Bayesian methods, but with different investment of time and resources depending on the situation. Cytel’s consulting and RWE teams are trained in both the traditional Frequentist paradigm as well as Bayesian solutions for a vast range of scenarios.

At Cytel we believe that statisticians ought to be well-versed in both Bayesian and Frequentist statistical paradigms, so that they can apply the best solutions for our customers’ specific needs. Through our hybrid approach we will ensure that our customers receive tailored solutions that build on the best of both paradigms.

stock_abstract_architecture_4

Cytel Scientists are at the forefront of Bayesian Innovation. They have taken a lead in developing:

 

  • Early phase dose-finding methods like Hi3+3;

  • Uses of meta-analytics priors in clinical trial design;

  • The use of synthetic and external control arms, complex statistical designs like basket and umbrella trials;

  • New methods for Real world evidence like quantitative bias analysis;

Bayesian Clinical Trial Design

Software

Cytel provides both software and services that can guide you through the design of the perfect Bayesian clinical trial. Our suite of East Bayes clinical trial design software provides validated software that enables all statisticians to consider Bayesian tools. We equip statisticians with rigorously tested tools for complex Bayesian innovations, many of which are innovations first published by Cytel scientists. 

Statistical Consulting

Our expert statistical consultants have experience designing Bayesian clinical trials from early phase to regulatory submission, even designing basket trials, synthetic control arms and other complex innovative trial designs. Their designs are supported by insights from premier network meta-analyses

"Adaptive Bayesian techniques are the most natural way of including historical and in-trial data to inform and adapt the course of the trial and provide a more intuitive interpretation of the results. Adaptive trials use Bayesian techniques because they enable considered in-trial changes."

Dr. Kyle Wathen, VP of Scientific Strategy and Innovation

parrish-freeman-0d-z8cJGIR4-unsplash

Bayesian Solutions

Bayesian methods are often used as solutions for difficulties and opportunities that might arise during a clinical trial. At Cytel we believe that statisticians ought to be well-versed in both Bayesian and Frequentist statistical paradigms, so that they can apply the best solutions for our customers’ specific needs. Through our hybrid approach we will ensure that our customers receive tailored solutions that build on the best of both paradigms.

Cytel statistical consultants can help you:

  • Build approaches for flexible learning into your clinical trial.
  • Assimilate historical, real-world or other forms of external data into your clinical trial.
  • Strengthen your Phase III regulatory submission with the addition of external data (i.e. utilize a variety of hybrid methods).
  • Determine how to incorporate model-based and other modern modes of dose-finding.
  • Build strategic elements like in-trial learning, resource management and other objectives into your design using Bayesian methods.  
leftside leftside william-daigneault-7c8z4KEvgds-unsplash
CYL2000_BrandGuide_R01v3-04

Intelligence

A hybrid approach that yields the best of both Bayesian and Frequentist methods depending on client needs.

CYL2000_BrandGuide_R01v3-04

Power

Technology that allows statisticians the tools to create their own Bayesian designs.

CYL2000_BrandGuide_R01v3-04

Confidence

Technology and services from an industry leader in Bayesian statistical methods for clinical trial design and implementation. 

Contact Us

Fill out the form below to get in touch