Incidence vs. Prevalence

Incidence counts new cases that develop in a previously unaffected population over a defined time — it captures the speed at which disease appears. It comes in two flavors: cumulative incidence (risk), a proportion over a fixed period, and the incidence rate, cases per person-time. Prevalence instead counts all existing cases — new and old — at a given moment, expressed as the fraction of the population currently affected. Roughly, prevalence ≈ incidence × average duration of disease.

The distinction matters because the two answer different questions. Incidence is the right measure for studying causes and the effect of prevention, since it reflects new disease arising. Prevalence describes the current burden on a health system and is useful for planning services. Confusing them leads to wrong conclusions.

A classic pitfall is using prevalent cases to study etiology. Prevalence is inflated by anything that lengthens survival and depleted by rapid recovery or death, so a factor that prolongs life with disease will look like a "risk factor" in prevalence data even if it never caused a single case (prevalence–incidence, or Neyman, bias). Prevalence studies are therefore vulnerable to selection effects related to survival.

For example, a fast-fatal cancer can have high incidence but low prevalence, while a chronic, manageable condition can have modest incidence yet high prevalence.