Estimates of the size of key populations (A–E)

Export Indicator

What it measures

Number of people engaging in the specific behaviour that put the given population at risk for HIV transmission or a proxy for those types of behaviour:

A. Sex workers.
B. Gay men and other men who have sex with men.
C. People who inject drugs.
D. Transgender people.
E. People in prisons and other closed settings.

Programme planning for key populations can be more efficient if the size of these populations can be accurately estimated. The figures enable national AIDS programmes, health ministries, donors and not-for-profit and multilateral organizations to efficiently allocate resources to adequately meet the prevention needs of specific populations at higher risk. Size estimates are also important for modelling the HIV epidemic.


Not applicable


Not applicable


Not applicable

Method of measurement

Several methods for estimation are available, including capture-recapture, service multipliers and network scale-up. See the Further Information section below for specific details.

Measurement frequency

Population size should be estimated every five years. However, any time an integrated biobehavioural survey is implemented, size estimates should be incorporated, if only to add to the database to confirm or refine estimates.

  • Estimating population sizes by age or sex is generally impractical. However, if a survey measures women who inject drugs or male sex workers, for example, a size estimate should be included.
Additional information requested

To better understand the size estimates submitted, we request that the following additional information be included in the comment box:

  • Definition used for the population, and inclusion criteria used in the study/survey, as applicable.
  • Method to derive the size estimate.
  • Site-specific estimates for all available estimates.

In keeping with efforts to provide more granular data presentations, the latter will offer the opportunity for mapping denominator data with programme data if they are collected in the same survey areas.

If there are subnational data available, please provide the disaggregation by administrative area, city, or site in the space provided. You may also upload an Excel spreadsheet of these data instead of entering them in the online tool. Submit the digital version of any available size estimation reports using the upload tool.


Strengths and weaknesses

The quality of population size estimates varies according to the methods used and the fidelity with which the methods are implemented. Every effort to assess bias and adjust the estimates accordingly should be attempted and explained. Size estimates for small areas should not be presented as national estimates: either a rational approach to extrapolation should be used and explained or the small area estimates should explicitly be submitted for the relevant areas explicitly. Please indicate in the comment field whether a multi-stakeholder consensus has been reached for the reported size estimates.

Please note that new guidance from the World Health Organization and UNAIDS suggests that size estimates for gay men and other men who have sex with men should not represent less than 1% of the adult male population. If the size estimate is calculated as less than 1%, then the results should be reviewed, as per the guidance.

Technical brief: reasonable population size estimates for men who have sex with men. Geneva: World Health Organization and UNAIDS; 2020 (

Further information

UNAIDS/WHO Working Group on Global HIV/AIDS and STI Surveillance. Guidelines on estimating the size of populations most at risk to HIV. Geneva: World Health Organization, and UNAIDS; 2010 (