Ask Mallard

Prevalence, incidence and person-time

Prevalence counts existing cases at a moment; incidence counts new cases arising over a period.

The relationship between them

In a steady state, prevalence is approximately incidence multiplied by average duration. That single relationship explains most confusion in this area: prevalence rises if more people get the disease, if they survive with it longer, or if they recover more slowly, and the number alone cannot distinguish the three.

Incidence, duration and prevalence

Prevalence as a stock and incidence as a flow. New cases flow into the pool of people living with the condition; people leave the pool by recovering or leaving the population. The size of the pool depends on both the inflow and how long people stay.Incidence fills the pool; duration decides how full it getsNew cases: incidencePeople living with thecondition: prevalenceRecovery, or leaving thepopulationIn a steady state, prevalence is approximately incidence multiplied by average duration, so prevalence canrise because people live longer with the condition.
Conceptual diagram. It shows the structure of the idea, not data from any study.

Two kinds of incidence

  • CUMULATIVE INCIDENCE, or risk: new cases divided by the number of people at risk at the start. A probability between 0 and 1, requiring a stated period and complete follow-up.
  • INCIDENCE RATE: new cases divided by the total PERSON-TIME at risk. Handles people entering and leaving at different times, and has units of one over time.
  • Person-time is accumulated only while someone is genuinely at risk: not before they enter, not after the outcome, not while they are immune or already have the condition.
  • A rate of 5 per 100 person-years is not 5 percent per year, though it is close when the rate is low.
  • Odds is a third thing again, and is what a case-control study estimates directly.

Where the denominator goes wrong

Counting time before someone could have had the outcome creates immortal time and biases the rate downward. Leaving prevalent cases in the denominator of an incidence calculation dilutes it. Using the whole population rather than the susceptible population overstates the denominator for conditions people can only get once.

Which to use

Use incidence for questions about causation and about what a treatment changes. Use prevalence for questions about burden: how many people need this service, how many beds, how many prescriptions. A study measuring prevalence cannot answer an incidence question, whatever the analysis.

What to watch out for

Prevalence depends on how long people stay in the diseased state, so a condition can rise in prevalence purely because treatment lets people live longer with it. Reading a prevalence change as a change in occurrence is the standard error. And a RATE is not a RISK: a rate has time in its denominator and can exceed 1, while a risk is a probability bounded by 1 and is meaningless without the period it refers to.

Clinical example

Diabetes prevalence rises in a region. That is consistent with more new cases, with better survival among people who have it, with earlier diagnosis, or with any combination, and only incidence data separate them.

Code for this method
Loading…
Recent studies that used this
Searching PubMed…
Related Learn pages
Sources & further reading
  1. Measures of disease frequency: prevalence and incidence (opens in a new tab) · Nephron Clin Pract · doi 10.1159/000286345
  2. Overview of clinical study designs (opens in a new tab) · Clin Exp Emerg Med · doi 10.15441/ceem.23.036
Last source review: 2026-09-06
AI-generated educational content
Mallard uses reputable methodological and reporting sources, but this material may contain errors or omit important context. Verify important methodological decisions with the cited sources and, when appropriate, a qualified research methodologist or statistician.