“…The latter is what Bittl et al do in their Bayesian metaanalysis of the evidence supporting the relative benefits of treatments for unprotected left main coronary artery disease. 3 Although the notion of epistemic probability might seem far afield from practice guidelines, in fact, such guidelines have their own language of epistemic uncertainty. Instead of stating that the proposition that PCI is comparable to CABG, has, say, an 80% chance of being true, the level of evidence for this statement is classified as B.…”