Prevalence

Prevalence is the share of a defined population carrying a characteristic — a pattern of use, a condition, an exposure — at a given moment or across a stated time window. In harm reduction contexts, it most often appears as the fraction of a group that uses a particular substance, carries a particular health risk, or has experienced a particular outcome.

It is distinct from incidence, which counts how many new cases arise in a period. Prevalence counts existing ones. A condition that lasts a long time accumulates high prevalence even when incidence is low; one with high mortality or rapid resolution may show low prevalence regardless of how frequently it appears.

How it is done

Prevalence is estimated through surveys, administrative records, and indirect measurement. Population surveys — household interviews, school studies, treatment registries — ask respondents whether they have used a substance within a defined window: past month, past year, or lifetime. The resulting count divided by the total sampled gives the estimate for that population and that window.

The window definition shapes the number substantially. Past-month prevalence and lifetime prevalence of the same substance may differ by an order of magnitude, and the two figures answer different questions. Estimates drawn from treatment populations reflect people in contact with services, not the broader group that uses.

Indirect methods — wastewater epidemiology, naloxone distribution records, overdose surveillance — can reach populations surveys miss. Each carries its own measurement assumptions and a denominator that is, at best, estimated.

What it cannot tell you

A prevalence figure alone cannot distinguish frequent use from occasional, dependent use from casual, or harmful patterns from non-harmful ones. Two populations with identical prevalence of a substance can have very different harm profiles depending on route, setting, and what else is used alongside it.

The most consequential gap is systematic undercount. Stigmatized and hidden populations — people experiencing homelessness, those in carceral settings, people who use highly criminalized substances — are consistently underrepresented in surveys. Recorded prevalence in these groups is almost certainly lower than true prevalence.

The wrong conclusion a low figure invites is that the problem is small because the number is small. Low survey prevalence may instead reflect reluctance to report, inaccessible sampling frames, or high mortality removing cases from the population before they can be counted. A number that looks reassuring may be measuring the difficulty of measurement.