Improving physicians’ understanding of statistical tests

To correctly interpret claims made in the medical literature, the reader must understand the statistical significance testing that underpins them. There are concerns, however, that many physicians misunderstand the probabilistic logic underlying these tests. As well as having a potentially negative effect on the quality of medical research, this lack of understanding could ultimately have adverse implications for public health.Westover and colleagues highlight and attempt to address this problem in their debate published today in BMC Medicine.

By administration of a simple quiz to 246 physicians, Westover et al. found that a startling 93% answered incorrectly. In light of these results, the authors provide an accessible explanation of the underlying probabilistic concepts that need to be followed if intelligent interpretation of the medical literature is going to be achieved.

The debate focuses in particular on Bayes’ rule and its ability to protect against many common errors in statistical reasoning. Bayes’ rule refers to the interpretation of probabilities as degrees of belief in order to answer the question “what way should one’s confidence in a particular claim change in response to new data?”.

This article highlights the importance of physicians’ understanding of this rule and provides a step-by-step approach to interpreting the concepts it expresses. They conclude that “a deeper appreciation of Bayes’ rule may go a long way toward making physicians less vulnerable to the fallacies inherent in conventional applications of statistical significance testing.

View the latest posts on the On Medicine homepage

Comments