Bayesian Diagnostic Reasoning

Bayesian diagnostic reasoning is the clinical application of Bayes' theorem: the probability of a diagnosis given a test result depends not just on the test's accuracy but on the prior probability (pre-test probability) of the disease. A positive mammogram in a 25-year-old with no risk factors means something very different from the same result in a 60-year-old with a family history of breast cancer — because the pre-test probability differs by orders of magnitude. Practitioners use likelihood ratios to update probabilities: a positive result on a test with a likelihood ratio of 10 shifts the probability much more than a test with a likelihood ratio of 2.

When to use it

When interpreting test results, evidence, or signals of any kind; when deciding whether additional information gathering will actually change your decision; when explaining to others why a positive result doesn't necessarily mean what they think it means; or when designing testing or investigation sequences.

How it can help

This model fundamentally changes how you interpret evidence in any domain. Most people treat evidence as binary (the test was positive, so the condition is present), but Bayesian reasoning reveals that the same evidence can be compelling or meaningless depending on prior probability. In business, a positive signal from a market test means more when there's already reason to believe customers want the product (high prior). In security, an alert means more when threat intelligence suggests elevated risk. Bayesian reasoning prevents both under-reaction (ignoring strong evidence because the prior is low) and over-reaction (treating weak evidence as conclusive because the prior is high).