Statistical analysis and estimation of clinical effect based on Bayes factor for otorhinolaryngology

The article discusses the use of Bayesian methods for statistical inference as an alternative to the traditional method of testing hypotheses based on the significance level. Illustrative examples of solving traditional statistical problems of hypothesis testing in otorhinolaryngology based on the calculation and interpretation of the Bayes factor are presented. As two tasks for illustration, we used the tasks of assessing the impact of sauna visits on indicators of nasal flow in patients with allergic rhinitis and the task of assessing the association of the season of the year and the frequency of birth of children with hearing impairment. Although in the form of an article it is impossible to fully describe and explain all the mathematical terms and their origin for understanding the logic of Bayesian methods, we tried to explain what they mean without reference to the mathematical manuals. As Bayesian methods are increasingly used in statistical applications, a basic understanding of how to calculate them should be part of the toolkit of every medical researcher, and how to interpret it, of every practitioner who is interested in modern results of clinical trials. All calculations used in the article are accompanied by an R-code, so they can easily be reproduced, the text of the article can be used as step-by-step instructions for their implementation.