MeSH Consortium logo

Comparing population size estimate methods for female sex workers in Zimbabwe - a first of its kind study

May 28, 2021

Image Source: https://sisters-zimbabwe.org/

Population size estimates (PSE) help inform HIV programming and modelling, and delivery of services for female sex workers (FSW). Yet, different methods of estimating population sizes exist and it is unclear how to obtain national-level estimates. A 2020 MeSH study by Fearon et al., looked into estimating the FSW population size in Zimbabwe, and through this compared different estimation methods. They estimated that there were 40,491 (plausibility bounds 28,177–59,797) FSWs in Zimbabwe, the first ever national-level estimate, now adopted into national policy-making.

Four PSE methods were tested across 20 sites in Zimbabwe using respondent-driven sampling (RDS) survey data; social and geographic mapping data; sex worker program data; and population census data.

The four PSE methods tested on individual sites were:

  • Service Multiplier Method (SMM) – researchers compared the number of women identified by sex worker program data to self-identified clinic attendees via the RDS survey data.
  • Unique Object Multiplier Method (UOMM) – the epidemiologists compared a known number of wristbands distributed among the target population (equivalent to the number of women identified by sex worker program data) to the number of individuals reporting having received a wristband in the RDS survey.
  • Capture-Recapture – teams distributed cards in areas/venues identified through mapping and comparing the number of cards distributed with the number of recaptures (number of women reporting to have received cards the night before) over 2 nights.
  • Census – survey assistants counted FSWs on one high activity night.

Development of a national level estimate was done through discussion with key stakeholders and experts in the country. They defined additional hotspots and matched these into strata with the study sites to attribute PSEs for each hotspot. They then estimated that 95% of FSWs in the country could be found in these hotspots and used this assumption to calculate the national PSE.

High variability was found within sites and across methods, where multiplier methods (UOMM and SMM) produced the highest PSEs. The study supports the recommendation that multiple PSE methods are used to obtain site-level PSEs, preferably those using different data sources and making different assumptions

The researchers also determined that most FSWs could be found in a few large sites such as urban areas like Harare (2.55% of FSWs among women aged 15-49 years) and Bulawayo (4.09%). This highlights a need for national program and service provision in such urban areas. The full text of the paper is available open access from the Journal of Acquired Immune Deficiency Syndrome.

Back to Updates

Follow Us

@MeSHConsortium

Bill & Melinda Gates Foundation logo
crossmenuchevron-down-circle