Men who have sex with men (MSM) are a recognized high‑risk population for sexually transmitted diseases (STDs), including HIV and syphilis. In China, published estimates report elevated HIV prevalence among MSM and among student MSM specifically. In Hainan Province, young students account for a notable proportion of HIV cases, with male student infections largely attributable to homosexual contact. Reliable estimates of the size and geographic distribution of MSM within colleges are essential to design targeted prevention, testing, and surveillance interventions. This study used data from the geosocial networking app Blued together with a questionnaire and the multiplier method to estimate the population size and spatial distribution of college MSM across Hainan Province.
Hainan Province had 21 colleges and universities with 25 campuses enrolling students in 2023. The institutions included undergraduate and junior colleges spanning multiple disciplines (financial, science and engineering, comprehensive, medical, normal, language and literature, political science and law, and sports). Campus distribution by city was: 15 campuses in Haikou, six in Sanya, and one each in Danzhou, Wuzhishan, Qionghai, and Wenchang.
To explore geographic patterns, the authors summed Blued campus users by city and computed MSM population density by dividing estimated MSM counts by the number of male students enrolled in all colleges within the same city. ArcMap 10.8.1 was used to map campus‑level MSM density and identify hotspots across Hainan.
The study applied the multiplier method to estimate total student MSM. In this implementation, r represented the number of Blued users observed within campus geofences and p was the proportion of student MSM reporting Blued use from a questionnaire. The basic multiplier logic equates the observed institutional count to the estimated population multiplied by the proportion who attend or are observed at the institution. Confidence intervals were calculated using standard formulas with N substituted where necessary; the source provides the method but not step‑by‑step computations beyond reported results.
Blued data collection occurred nightly between 10:30 and 11:30 from May 9 to May 18, 2023. The research team geofenced central campus locations with radii ranging from 0.20 to 0.90 km depending on campus size and recorded publicly visible fields including nickname, age, height, weight, gender role, and online status. Only aggregated, anonymized public data were used; no personally identifiable information was collected. From an initial 3,362 online annotations, 234 duplicates/abnormal records were removed, leaving 3,128 valid observations aged 18–24 for analysis. Definitions used in analysis included: Registered user (observed within campus geofence at any time during the 10‑day period), Active user (online on five or more days during the 10‑day period), Online user (status online at time of capture), Offline user (status offline but listed), and Secret user (private mode; 37 identified, 2.34% of registered users). Sensitivity analyses using four‑ and six‑day thresholds for active user status returned similar patterns.
Binary logistic regression was used to examine determinants of active user status (dependent variable: active vs non‑active as defined above). Independent variables included age, BMI, college level, college location, college classification, and campus. Adjusted odds ratios (aOR) with 95% confidence intervals were reported. All tests were two‑sided with α = 0.05 and analyses were performed with SPSS 27.0. The study received ethical approval from the Ethical Committee of Hainan Medical University (HYLL-2024–838). The authors state data were anonymized and protected.
A questionnaire administered April 6–10, 2025 surveyed 181 MSM in Hainan about Blued usage. Among these, 164 (90.61%) had used Blued; 28 respondents were college students and 25 of those students (89.28%) reported Blued registration. Among the 25 student users, 76.00% had used the app for over two years and 76.00% reported using Blued primarily to seek sexual partners; 56.00% used it for social interaction. The survey results informed the multiplier parameter (p) used to estimate total college MSM via Blued capture counts.
During the ten‑day capture window, the average number of college students online on Blued was 939. The total number of registered users observed across the period was 1,579 (after de‑duplication). Using the proportion of student MSM reporting Blued registration from the questionnaire (89.28%), the multiplier estimate for MSM in colleges across Hainan Province was 3,504 with a 95% confidence interval of 3,464–3,544.
Of the observed Blued users, 2,596 were included in analyses of online patterns; 860 (33.13%) met the active user definition (online ≥5 days). Multivariate logistic regression identified higher odds of active use among younger age groups (18–19 and 20–21 years) compared with 22–24 years, and a higher proportion of active users among undergraduates versus junior college students (aOR = 1.90, 95% CI: 1.38–2.61). The analysis also found differences by college classification: normal colleges (aOR = 0.55, 95% CI: 0.36–0.84) and sports colleges (aOR = 0.27, 95% CI: 0.10–0.72) differed relative to comprehensive universities. City‑level patterns included higher active‑user proportions in Danzhou and Wenchang (42.22% and 41.30%, respectively) despite smaller absolute numbers of registrants.
Geographically, the majority of observed college Blued users were located in Haikou (72.22%, 2,259/3,128), followed by Sanya (20.75%, 649/3,128). Other cities showed more dispersed distributions without clear aggregation. Online activity rate (active users normalized by city male student enrollment) was highest in Wenchang (1.21%), then Sanya (0.82%) and Haikou (0.74%). Mapping with ArcMap highlighted Haikou and Sanya as population concentration hotspots.
The authors conclude that combining Blued geosocial data with a questionnaire and multiplier method yielded an estimated 3,504 MSM among college students in Hainan, and identified urban hotspots (Haikou, Sanya) and cities with high active‑user proportions (Danzhou, Wenchang). They recommend prioritizing sexual health education and enhanced STD surveillance in specific colleges and regional hotspots identified by the analysis. The study documents ethical approval, funding sources, and absence of competing interests.
Limitations and additional specifics that were not fully reported in the source text include the full questionnaire instrument content, detailed stepwise calculations for the multiplier CI beyond reported estimates, and the truncated portion of the discussion that ended mid‑sentence in the source. These items were not provided in the article text excerpt used as the source.