The use of Bayes factors in biomedicine
In the last few decades, the field of biomedicine has seen significant advancements, leading to improved treatments for various diseases. Central to this progress is the proper analysis and accurate interpretation of the available data. In biomedicine, the predominant approach for this remains rooted in the traditional frequentist approach to statistical inference, even though there is increasing recognition of its drawbacks. Bayesian statistics is an alternative approach to statistical inference that remedies many of the shortcomings of frequentist inference. In particular, the Bayes factor can be used to directly contrast the evidence for two competing hypotheses, allows for optional stopping of data collection based on interim results, and can be interpreted in a straightforward way. This dissertation introduces the baymedr R package and its accompanying web application, allowing biomedical researchers to compute Bayes factors for a selection of designs common in biomedicine in a user-friendly way. A comparison is made between frequentist and Bayesian approaches to statistical inference in biomedicine, and advantages of Bayes factors are highlighted. Further, the procedure for computing Bayes factors with baymedr is explained step by step, so that not only statistical experts can compute Bayes factors. Lastly, default settings are developed, providing biomedical researchers with solid starting points for their analyses. In conclusion, this dissertation provides biomedical researchers with tools to compute Bayes factors for biomedical designs and calls for an increased adoption of Bayesian statistics in biomedicine, coupled with the development of user-friendly software for Bayesian computation.