MeSH Consortium logo

Estimating Population Sizes - Demystifying Key Words Campaign

April 29, 2022

Written by Thomas Hartney

The Demystifying Key Words campaign’s first post discussed key populations – those populations who are at increased risk of HIV – and described why it’s important to understand their HIV prevention needs. A question that researchers and policy makers may need to consider is the size of each key population. Having a good estimate of this can help to allocate resources effectively and assess how well HIV prevention measures, such as HIV testing or pre-exposure prophylaxis, have reached those who need them. HIV surveillance systems can provide estimates of the number of new cases (i.e. the numerator), but to calculate the rate of new cases you also need to know how many people are at risk of infection (i.e. the denominator).

While common approaches such as censuses and birth and death registers can be used to estimate the size of the general population, there are numerous challenges around identifying the size of key populations, often relating to stigma and criminalisation. These include lower visibility of the status of these individuals, higher rates of mobility and migration compared to the general population, and lower rates of engagement with healthcare and government services.

Over the last few decades researchers have developed different methods of estimating key population size, and population size estimation (PSE) studies have become a vital part of the toolkit for evidence-based HIV responses. There is no one gold standard method of estimating populations sizes, and so often multiple data sources and methods are combined to gain a reasonable grasp of the lower and upper bounds that are likely to exist. These methods can be broadly categorised into two types: empirical methods (based on formalised study designs and statistical concepts), and non-empirical methods (based on observation or opinion).1 While empirical methods might be expected to provide more accurate estimates than the non-empirical methods, often the latter is sufficient for rapid planning purposes.

Empirical methods include:
• Service Multiplier Method – researchers compare the number of members of key populations identified by program data to self-identified clinic attendees.
• Capture-Recapture – researchers identify individuals randomly such as by distributing cards in areas/venues, and compare the number of cards distributed with the number of recaptures (number of key population members reporting to have received cards) over a given time period.
• Handcock’s Respondent Driven Sampling (RDS) based Method – this method, based on the respondent-driven sampling approach of accessing members of hard-to-reach populations, uses individuals recruited via personal networks of purposively selected individuals2.
• Census – researchers count members of a particular key population (e.g. female sex workers) on one high activity night or other specified time period.
• Network Scale-up Method - this relies on asking participants about the risk behaviours of people they know, rather than asking people directly about their own behaviour. This is based on the assumption that you can extrapolate from the prevalence of risk behaviours in the social networks of a representative sample of the general population to estimate the prevalence in the whole population3.

Non-empirical methods include:
• Delphi method – this is a structured communication technique that uses a panel of experts to assess evidence from multiple sources4.
• Wisdom of the crowds – this uses a non-expert panel, such as participants in a survey, to contribute best guesses on the number and plausible range of key population sizes, assuming that when responses are combined this will provide a good overall estimate.
• Hotspot mapping – this can take place as part of a programme mapping exercise, by extrapolating total numbers of key population members based on those identified at venues where HIV prevention activities could be targeted.

PSE studies are becoming increasingly common, however there are still many countries where estimates of key population sizes are not available or not current. A recent review found that fewer than half of African countries have published size estimation data for any key population.5 There’s evidence that PSE studies are not always incorporated into national operational plans and strategy documents, with results often being only published in surveillance reports, and there is room to use these further to guide HIV responses for those most at need.6 Future research challenges will include estimating population sizes at smaller than sub-national levels, improving methods of synthesising and triangulating the results of different methods, and representing key population sizes with measurements of statistical variability, such as confidence intervals, rather than as point estimates.

 

1. Neal JJ, Prybylski D, Sanchez T, Hladik W. Population size estimation methods: Searching for the holy grail. JMIR Public Heal Surveill. 2020;6(4):1-7. doi:10.2196/25076
2. Sulaberidze L, Mirzazadeh A, Chikovani I, Shengelia N, Tsereteli N, Gotsadze G. Population Size Estimation of Men Who Have Sex with Men in Tbilisi, Georgia; Multiple Methods and Triangulation of Findings. PLoS One. 2016;11(2):1-14. doi:10.1371/journal.pone.0147413
3. Sharifi H, Karamouzian M, Baneshi MR, et al. Population size estimation of female sex workers in Iran: Synthesis of methods and results. PLoS One. 2017;12(8):1-12. doi:10.1371/journal.pone.0182755
4. Abdul-quader AS, Baughman AL, Hladik W. Estimating the size of key populations: current status and future possibilities. 2019;9(2):107-114. doi:10.1097/COH.0000000000000041.Estimating
5. Sabin K, Zhao J, Garcia Calleja JM, et al. Availability and quality of size estimations of female sex workers, men who have sex with men, people who inject drugs and transgender women in low- and middle-income countries. PLoS One. 2016;11(5):1-11. doi:10.1371/journal.pone.0155150
6. Viswasam N, Lyons CE, MacAllister J, et al. The uptake of population size estimation studies for key populations in guiding HIV responses on the African continent. PLoS One. 2020;15(2):1-16. doi:10.1371/journal.pone.0228634

Back to Updates

Follow Us

@MeSHConsortium

Bill & Melinda Gates Foundation logo
crossmenuchevron-down-circle