adventures in bayesian medicine: when diagnosing mad cow disease, you need to consider not only the evidence but also your prions