Population size estimates (PSE) for gay, bisexual, and other cisgender men who have sex with men (GBMSM) are critical for HIV resource allocation but remain challenging where stigma and criminalization limit field-based data collection. This preprint examines whether internet-derived signals—specifically Google Trends search volumes and aggregated metrics from a gay social networking app—can provide standardized, low-cost PSEs for GBMSM in four countries: Senegal, Ghana, Mozambique, and Indonesia.
The authors analyzed country-level and local Google Trends relative search volumes for the search terms “porn” and “gay porn” over a one-year period (September 2024 to September 2025). They used Pornhub global review data to translate relative search volumes into proportional PSEs (GTPSE). Aggregated, non-public data from a gay social networking app provided monthly active user counts, which the authors extrapolated to annual user estimates to derive app-based proportional PSEs. Both proportional approaches were multiplied by the country male population aged 15–49 years to produce absolute PSEs. Estimates were compared to existing values in the UNAIDS Key Population Atlas.
Google Trends approach: country-level and local relative search volumes for “porn” and “gay porn” were extracted and normalized; Pornhub review data were applied to produce proportional GBMSM PSEs. The resulting proportional estimates are referred to as GTPSE.
App-based approach: the study used aggregated monthly active user counts from a gay social networking app, extrapolated to annual users to estimate the proportion of GBMSM represented by the platform in each country.
Conversion to absolute PSE: proportional estimates from each method were multiplied by the male population aged 15–49 to generate absolute population size estimates for GBMSM.
Comparisons and sensitivity analyses: final estimates were compared with UNAIDS Key Population Atlas estimates. Sensitivity analyses included using survey-based multipliers and alternative combinations of search terms to assess robustness.
The Google Trends–derived proportional estimates (GTPSE) and app-derived proportional estimates varied by country. Reported GTPSE ranged from 0.37% in Mozambique to 1.92% in Indonesia. App-based proportional estimates ranged from 0.29% in Mozambique to 1.34% in Ghana. When compared with UNAIDS Key Population Atlas estimates, GTPSE exceeded UNAIDS estimates in Ghana, Mozambique, and Indonesia. App-based estimates exceeded UNAIDS estimates in Ghana and Mozambique. At a subnational level, the study found that Maputo’s estimated GBMSM proportion was substantially higher than Mozambique’s national estimate, illustrating geographic concentration differences.
The authors conducted sensitivity analyses using survey-based multipliers and alternative search-term combinations; these were intended to test the stability of internet-derived proportional estimates. The preprint reports that GTPSE generally exceeded UNAIDS estimates across several settings, while app-based estimates were more variable and in some instances lower than UNAIDS estimates, suggesting platform access and geographic distribution differences can substantially affect app-derived PSEs.
The authors suggest that the higher PSEs from Google Trends may reflect underestimation by traditional field-based methods in settings with stigma or criminalization, where GBMSM are less likely to be reached through household surveys or mainstream health services. Conversely, the lower or variable app-derived estimates in some countries highlight differences in access to or use of specific social platforms and the uneven geographic concentration of social networks. The Maputo versus national Mozambique discrepancy underscores the need to consider subnational heterogeneity when planning HIV services.
The preprint emphasizes that internet-derived estimates are subject to biases related to search behavior, platform access, and the representativeness of app users. Aggregated social networking app data are not publicly available due to data-use and confidentiality restrictions; the authors note these data may be obtained from the corresponding author upon reasonable request. The manuscript states that no IRB approval was required for the analyses and that the authors declared no competing interests. The study is a preprint and has not been peer reviewed; the authors caution it should not be used to guide clinical practice without further validation.
Internet-derived data streams, including Google Trends and aggregated social networking app metrics, may offer a feasible, low-cost approach to estimating GBMSM population sizes in low- and middle-income countries. The differing results between GTPSE and app-based estimates highlight that multiple digital sources capture different segments of GBMSM populations. The authors recommend further validation through digital data linkages to improve the reliability of these approaches and to support HIV prevention and treatment planning in contexts where traditional surveillance is constrained.