Where does prevalence collide?

In diagnostic medicine, the pretest (or prior) probability is the estimated chance that a patient has a condition before a test result is known, and it is combined with the test's likelihood ratios to give the post-test probability. Fagan's 1975 nomogram made this Bayesian updating a bedside tool. The referent is the baseline probability of the condition before individual evidence, the same kind of quantity that epidemiology calls prevalence and judgment research calls the base rate.[1]

Unit

The probability of the condition before the test result is known

Recovery means

Not applicable

Measured by

Estimated from local prevalence, clinical prediction rules, or clinician judgment, then updated with likelihood ratios

In the psychology of judgment, the base rate is the prior relative frequency of a category in a population, which people tend to underweight when they also have specific, individuating information: the base-rate fallacy. Bar-Hillel's 1980 study of probability judgments documented when and why this neglect occurs. The referent is the unconditional probability of membership in a class before case-specific evidence, which is the same kind of quantity that clinicians call pretest probability and epidemiologists call prevalence, under a different name.[2]

Unit

The unconditional (prior) probability of the category in the population

Recovery means

Not applicable

Measured by

Taken as the marginal frequency of the category, against which conditional judgments are compared

In epidemiology, prevalence is the proportion of a population that has a condition at a given point or over a given period: a snapshot of existing cases, in contrast to incidence, which counts new cases over time. Rothman, Greenland, and Lash's Modern epidemiology defines the measure and its relation to incidence and duration. The referent is the fraction of a defined population currently affected, which is the same kind of probability that decision research names the base rate and clinicians name the pretest probability.[3]

Unit

The proportion of a defined population with the condition at a point or over a period

Recovery means

Not applicable

Measured by

Counted as existing cases divided by the population at risk, from cross-sectional surveys or registries