Population size, HIV prevalence, and antiretroviral therapy coverage among key populations in sub-Saharan Africa: collation and synthesis of survey data, 2010-23

Abstract

Background: Key population HIV programmes in sub-Saharan Africa require epidemiological information to ensure equitable and universal access to effective services. We aimed to consolidate and harmonise survey data among female sex workers, men who have sex with men, people who inject drugs, and transgender people to estimate key population size, HIV prevalence, and antiretroviral therapy (ART) coverage for countries in mainland sub-Saharan Africa.

Methods: Key population size estimates, HIV prevalence, and ART coverage data from 39 sub-Saharan Africa countries between 2010 and 2023 were collated from existing databases and verified against source documents. We used Bayesian mixed-effects spatial regression to model urban key population size estimates as a proportion of the gender-matched, year-matched, and area-matched population aged 15-49 years. We modelled subnational key population HIV prevalence and ART coverage with age-matched, gender-matched, year-matched, and province-matched total population estimates as predictors.

Findings: We extracted 2065 key population size data points, 1183 HIV prevalence data points, and 259 ART coverage data points. Across national urban populations, a median of 1·65% (IQR 1·35-1·91) of adult cisgender women were female sex workers, 0·89% (0·77-0·95) were men who have sex with men, 0·32% (0·31-0·34) were men who injected drugs, and 0·10% (0·06-0·12) were women who were transgender. HIV prevalence among key populations was, on average, four to six times higher than matched total population prevalence, and ART coverage was correlated with, but lower than, the total population ART coverage with wide heterogeneity in relative ART coverage across studies. Across sub-Saharan Africa, key populations were estimated as comprising 1·2% (95% credible interval 0·9-1·6) of the total population aged 15-49 years but 6·1% (4·5-8·2) of people living with HIV.

Interpretation: Key populations in sub-Saharan Africa experience higher HIV prevalence and lower ART coverage, underscoring the need for focused prevention and treatment services. In 2024, limited data availability and heterogeneity constrain precise estimates for programming and monitoring trends. Strengthening key population surveys and routine data within national HIV strategic information systems would support more precise estimates.

Funding: UNAIDS, Bill & Melinda Gates Foundation, and US National Institutes of Health.

Figure 1

Identified data for key population size estimates (A), HIV prevalence data (B), and key population antiretroviral therapy coverage (C) Each n represents the number of observations in the associated category, with potential observations for multiple subnational locations from the same study in the same country. CDC=US Centers for Disease Control and Prevention. FSW=female sex worker. MSM=men who have sex with men. PWID=people who inject drugs. TGW=transgender women.

Figure 2

Model estimated urban KPSE proportions for female sex workers, men who have sex with men, people who inject drugs, and transgender women as a proportion of gender-matched adult total population aged 15–49 years (A) Posterior median estimate for each country as a proportion (%). The colour range is different for each key population chloropleth. (B) Posterior median estimate and 95% credible intervals for each country. Points represent observations of subnational key population size estimates proportions that used empirical methods. Countries that had local surveillance data have been shown in blue, and countries informed only by spatial smoothing from neighbouring countries in red. The vertical axis has been shown on log-scale. KPSE=key population size estimate

Figure 3

Key population and total population HIV prevalence on the logit scale (A), natural scale (B), and expressed as a ratio of key population and total population prevalence for female sex workers, men who have sex with men, people who inject drugs, and transgender women (C) Coloured points indicate observed key population prevalence plotted against gender-matched, year-matched, and province-matched total population prevalence. The dotted line represents line of equality. Logit-scale regional correlation coefficients and 95% uncertainty intervals shown in coloured text in panel A. Coloured lines and shading represent the regional estimate and 95% uncertainty results. Points represent country estimates (filled for countries with HIV prevalence data and empty for countries without HIV prevalence data) and 95% uncertainty ranges.

Figure 4

Estimated key population antiretroviral therapy coverage as a function of total population ART coverage on the estimated logit scale (upper) and natural scale (lower) Coloured points indicate observed key population prevalence plotted against gender, year, and province-matched total population ART coverage. Logit-scale correlation coefficients and 95% uncertainty intervals have been shown in part A. The black line and shading represent the estimate for sub-Saharan Africa and 95% uncertainty results. The black points represent country estimates (filled for countries with HIV prevalence data and empty for countries without HIV prevalence data) and 95% uncertainty ranges. The dotted line represents line of equality. ART=antiretroviral therapy.

Figure 5

(A) HIV-positive key population members as a proportion of all people with HIV aged 15–49 years. (B) Estimated number of key populations living with HIV in eastern and southern Africa, and western and central Africa Countries are ordered geographically.