gms | German Medical Science

Bayesian inference allows to assess whether a claim about an effect (e.g., effect > 0, effect > Δ, |effect| < Δ) is justified given the (likelihood of the) data and a prior distribution that quantifies an individual’s belief in the effect before seeing the data. Accordingly, recipients of such an analysis often vary in their prior distributions. Thus, it remains unclear whether they should agree on the claim. Reverse Bayes analysis and the concept of the sufficiently skeptical prior address this problem by asking how strongly one may believe in the absence of an effect in order to be convinced otherwise by the data. To this end, a method called Region of Evidence (RoE) is presented that can be utilized for any normally distributed prior and effect estimate. RoE visualizes the impact of all the prior distributions that, if they were used, would support the claim, including those that a priori favor an effect or its absence. Since RoE only requires an effect estimate and its standard error, it can be easily applied to previously published results. The method incl. its open-source implementation in R and Stata is introduced and its utility is highlighted regarding evidence synthesis in superiority and non-inferiority trials.

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.