The quality and safety of clinical trials depend upon the ability of experts to make accurate predictions about the risks and likely benefits. However, we know very little about whether or not experts are actually good at making these risk/benefit predictions.
In this talk, Jonathan Kimmelman will describe a series of studies that he and his team at McGill University have conducted, which measured whether experts could accurately forecast the events, outcomes, and risks in trials. His findings suggest that although a subset of clinical experts show a high level of skill in predicting trial outcomes, most clinical experts under-perform simple prediction algorithms.
Please join us for a discussion of the moral, policy, and scientific implications of these findings.