Prevalence

Prevalence is a measure in epidemiology describing how common a disease, condition, or other health-related characteristic is within a defined population. It includes both newly occurring and previously existing cases that meet the specified definition. Although often called a “prevalence rate,” prevalence is ordinarily a proportion rather than a rate with person-time in its denominator. Some reporting systems also use prevalence to refer to the number of affected people, so counts and proportions must be distinguished. (cdc.gov)

Definition and calculation

For a specified reference date, prevalence is calculated as:

The numerator must come from the population represented by the denominator. Prevalence can be expressed as a fraction, percentage, or number per a stated population size, such as per 1,000 or 100,000 people. The population, reference time, and criteria for identifying cases are essential parts of the measure. (cdc.gov)

For example, in a hypothetical population of 10,000 people, 500 affected individuals correspond to a prevalence of $500/10{,}000=0.05$, or 5%. The same proportion can be expressed as 50 affected people per 1,000 population.

Main types

Point prevalence measures the proportion affected at a particular point in time, such as a survey date.

Period prevalence measures the proportion affected at any time during a specified interval. It includes people already affected at the beginning and those who develop the condition during the interval. A person is counted once even if multiple episodes occur. In populations that change over time, the denominator convention requires specification; average or mid-interval populations are commonly used.

Annual prevalence is period prevalence measured over one year. These measures answer different questions: point prevalence describes a snapshot, whereas period prevalence describes how many people were affected during an interval. (archive.cdc.gov)

Specialized definitions occur in cancer surveillance. Complete prevalence includes people alive on a reference date who have ever received the relevant cancer diagnosis. Limited-duration prevalence includes only those diagnosed within a specified preceding interval who remain alive on the reference date. These measures can include survivors without active cancer; they are not equivalent to the prevalence of currently active disease. (surveillance.cancer.gov)

Relationship to incidence and disease duration

Incidence measures new occurrences, whereas prevalence measures existing cases. Incidence proportion concerns the fraction of an initially disease-free population that develops a condition during an interval; incidence rate relates new cases to time spent at risk. Prevalence therefore does not directly measure the risk of developing disease. (cdc.gov)

Prevalence depends on both the arrival of new cases and how long people remain in the counted disease state. Longer duration can produce substantial prevalence even when incidence is relatively low. Recovery and death remove people from an active-disease prevalence count, while improved survival without recovery can increase it. Consequently, a change in prevalence alone does not establish whether incidence has changed. (archive.cdc.gov)

A commonly used approximation is:

where $I$ is the incidence rate and $D$ is mean disease duration. This relationship requires assumptions, including approximately stable incidence and duration; its interpretation also depends on how the incidence denominator is defined. It is not a universal identity for rapidly changing populations or disease processes. The relationship has been used in statistical methods linking prevalence surveys to incidence estimation. (arxiv.org)

Measurement and comparison

Prevalence can be estimated from population surveys, examinations, disease registries, and health records. These sources do not necessarily measure the same thing. Self-reported diagnosed disease may differ from disease identified through physical measurements, while health-record estimates may represent people receiving care rather than the entire population. (cdc.gov)

Population surveys often require sampling weights to account for unequal selection probabilities and differences between respondents and the target population. Confidence intervals describe sampling uncertainty, and their calculation must account for features such as clustering and stratification. A narrow interval does not by itself eliminate systematic errors in case identification or population coverage. (wwwn.cdc.gov)

Comparisons may use crude prevalence, reflecting the population’s actual composition, or age-standardized prevalence, calculated using a common age distribution. Age standardization helps separate differences in age composition from differences in age-specific prevalence. Changes in survey methods, sampling coverage, or weighting can also alter estimates independently of changes in underlying disease frequency. (cdc.gov)

Uses and interpretive limitations

In public health, prevalence helps describe the burden of existing disease and supports planning within a health system. Cancer prevalence, for example, provides information about the number of survivors who may require continuing services. It does not, by itself, describe disease severity or the amount of care each person needs. (seer.cancer.gov)

Prevalence also affects the interpretation of diagnostic tests. Applying Bayes’ theorem, the positive predictive value of a test is:

where $\mathrm{Se}$ is sensitivity and $\mathrm{Sp}$ is specificity. With sensitivity and specificity held constant, a lower prevalence produces a lower proportion of positive results that are true positives. This is a population-level relationship, not an individualized assessment. (arxiv.org)

Prevalence estimates remain dependent on their operational definitions. Recall error, differing diagnostic thresholds, and imperfect tests can change the measured proportion. Reports therefore require interpretation alongside their case criteria, data source, population coverage, and uncertainty—not solely their numerical value. (cdc.gov)

References

1. Epidemiology Glossarycdc.gov
2. Describing Epidemiologic Datacdc.gov
3. Principles of Epidemiology: Lesson 3, Section 2archive.cdc.gov
4. Cancer Prevalencecdc.gov
5. Measures of Cancer Prevalencesurveillance.cancer.gov
6. Incidence — Health, United Statescdc.gov
7. On the incidence-prevalence relation and length-biased samplingarxiv.org
8. Prevalence to incidencecdn.who.int
9. Comparing Prevalence Estimates From Population-Based Surveys to Inform Surveillance Using Electronic Health Recordscdc.gov
10. NHANES Tutorials — Reliability of Estimates Modulewwwn.cdc.gov
11. NHANES Tutorials — Weighting Modulewwwn.cdc.gov
12. Confidence Intervals for Prevalence Estimates from Complex Surveys with Imperfect Assaysarxiv.org