Respectfully, what is actually interesting is NOT that critical-care physicians use Bayesian reasoning while questioning its application in critical-care trials. The interesting thing is that…after decades of RCT reversals, many for harm, trialists and statisticians expect more trust and mathematical discretion over priors, shrinkage, borrowing, covariate adjustment, and complex estimands without engaging in substantive failure-mode analysis or a corresponding increase in causal discipline. Indeed, we have already experienced such “strapped-on” Bayesian discretion in the REMAP-CAP cause agnostic RCT, (corticosteroid domain) which apparently no one defends.
Statisticians should embrace the healthy critical views of critical care physicians and engage in the failure mode analysis required to regain their trust.