Equivalence of prospective and retrospective models in the Bayesian analysis of case-control studies