Statistical methods have long been fundamental to drug development, and advancements in the last few decades in computing power have opened the door to more widespread use of Bayesian methods in clinical trials. Interest in Bayesian methods is growing \u2013 in particular due to what these approaches enable.<\/p>\n
So why aren\u2019t more clinicians using Bayesian methods?<\/p>\n
<\/p>\n
To answer this question, the Medical Outreach Team of the Drug Information Association Bayesian Scientific Working Group (DIA BSWG), including Cytel\u2019s Natalia Muehlemann, surveyed over 300 researchers and clinicians to determine their ability to interpret results from commonly used conventional and Bayesian analyses. They find in their recently published article in Therapeutic Innovation & Regulatory Science that, \u201cmost of the surveyed clinicians who answered additional questions either misunderstand results from both conventional and Bayesian analyses, or experience uncertainty when interpreting them. When presented with accurate interpretations from these results, clinicians generally conclude that Bayesian results are more useful than conventional ones.\u201d<\/p>\n
One conclusion the DIA BSWG Medical Outreach Team finds is that there is a lack of knowledge and familiarity among clinical researchers on these methods, as well as a lack of educational tools that are targeted to this group (rather than those created for statisticians). This has become a key barrier to using Bayesian methods in drug development.<\/p>\n