Huaqiao University was chosen for its unparalleled demographic diversity, with over 40% of its student body comprising overseas Chinese and individuals from more than 90 countries, creating a natural laboratory to examine how pre-existing cultural norms interact with acculturative influences on sexual behavior. This institution’s role as a global hub for diasporic communities ensures findings transcend localized contexts, offering actionable insights for both Chinese and international public health strategies targeting culturally heterogeneous youth populations. In October 2024, a cross-sectional survey was conducted among all first-year undergraduates enrolled at Huaqiao University’s Quanzhou and Xiamen campuses. Inclusion criteria were: (1) Enrolled as first-year undergraduates in 2024; (2) aged 18–24 years; (3) no visual impairments; (4) absence of severe mental health conditions; (5) capability to independently complete paper-based questionnaires; (6) voluntarily provided informed consent.

Cluster sampling was used, treating academic classes as the cluster units. All first-year undergraduate classes matriculating in September 2023 at Huaqiao University were included, effectively constituting a census of the target population. Because every eligible cluster was retained, no random, stratified, or convenience selection was required. All students within each cluster were invited to participate, and no within-cluster subsampling was performed. This comprehensive inclusion minimized selection bias and maximized sample representativeness.

A self-administered paper questionnaire was distributed and collected on-site. The questionnaire was adapted from the University Students’ HIV Prevention Knowledge and Competency Assessment Questionnaire (12) and updated to align with global advancements in HIV/AIDS and Monkeypox prevention. The questionnaire assessed: (1) Sociodemographic characteristics included sex and place of origin, both treated as nominal categorical variables; age was applied only as an inclusion criterion, was not recorded precisely, and therefore could not be analyzed. (2) Knowledge of HIV/AIDS and Monkeypox prevention (12 items in total), consisting of eight nationally endorsed HIV items and four investigator-developed Monkeypox items; each item was scored 1 for a correct response and 0 for an incorrect or missing response, and the subsection totals (0–8 for HIV, 0–4 for Monkeypox) were analyzed as continuous composite variables. (3) Sexual-behavior variables: sexual experience (binary); stage of sexual debut (ordinal categorical); partner type at debut (nominal categorical); condom use at first intercourse (nominal categorical); commercial-sex engagement (binary); condom use at last commercial encounter (nominal categorical); frequency of condom use in commercial sex (ordinal categorical); attitude toward same-sex relationships (ordinal); awareness of consequences (binary); receipt of HIV-prevention education (binary); participation as a campaign facilitator (binary). The study protocol was approved by the Quanzhou Center for Disease Control and Prevention Ethics Committee (No. 2024003), and all respondents were required to provide written informed consent.

(1) Questionnaire validation: Content validity was ensured through expert consultations (HIV/AIDS, Monkeypox, and sexual behavior specialists) and a pilot survey (July–September 2024) to refine clarity and feasibility. (2) Bilingual adaptation: Professionally translated English questionnaires were reviewed by experts to eliminate ambiguities for participants with limited Chinese proficiency. (3) Data collection rigor: Trained investigators distributed paper questionnaires on-site, provided real-time clarification, and ensured immediate return to minimize non-response and data loss. Anonymous submission safeguarded participant privacy. (4) Data verification: A dedicated quality control team cross-validated all entries.

Data were analyzed in SPSS 26.0 after rigorous cleaning. Continuous variables are summarized as median and inter-quartile range (IQR), and categorical variables as frequencies (%). Univariate analyses of associations between independent variables and sexual behavior were conducted using χ² tests and univariate binary logistic regression, with results reported as crude odds ratios (ORs) and 95% confidence intervals (CIs) calculated as OR = e^β. Variables with p < 0.05 in univariate analyses were entered into a binary logistic regression analysis with sexual activity status (0 = no, 1 = yes) as the dependent variable, using the enter method to estimate adjusted ORs and 95% CIs. Statistical significance was set at α = 0.05.

In addition to the primary analysis, several sensitivity checks were performed to validate model assumptions and robustness. The Box–Tidwell procedure was used to examine the linearity-in-the-logit assumption for the two continuous composite predictors (HIV Knowledge Score and Monkeypox Knowledge Score). For the ordinal predictor Attitude toward same-sex sexual behavior, we conducted a linear-by-linear trend chi-square test and also re-ran the logistic regression treating this three-level attitude variable as a continuous predictor (coded 1 to 2 to 3) to assess the consistency of its effect. Region of origin, a nominal categorical variable with multiple groups, was excluded from these sensitivity analyses because it lacks an inherent ordinal scale for trend or log-linearity testing.

The survey revealed that 4,567 students (93.4%) reported no prior sexual experience, while 325 students (6.6%) indicated having engaged in sexual activity, demonstrating that the vast majority remained sexually inexperienced (Table 1).

Regarding the age of sexual debut, Mainland Chinese and HMT students predominantly initiated sexual activity during high school (50.8% and 49.6%, respectively). In contrast, international students primarily reported first sexual experiences during university (66.7%). When examining partners at sexual debut, most students engaged with heterosexual partners. Notably, international students showed the highest proportion of homosexual partner selection (26.7%). Condom usage during first sexual intercourse varied by region: Mainland Chinese students reported 62.6% usage, HMT students 61.8%, while international students demonstrated significantly lower utilization rates at 46.7%. Analysis of commercial sexual activity within the past year revealed comparable rates between Mainland Chinese (16.2%) and HMT students (16.0%), with international students reporting slightly lower frequency (13.3%) (Table 2).

Survey results indicated that a total of 52 first-year university students engaged in commercial sexual activity. Analysis of protective practices during commercial sexual encounters revealed significant regional variations. During their most recent commercial sexual activity, condom usage rates were 20.7% among Mainland Chinese students, 23.8% among HMT students, and 50.0% among international students.

More concerning were the patterns observed over the past year. The proportion of students reporting never using condoms was alarmingly high: 57.1% in HMT students, 37.9% in Mainland Chinese students, and 50.0% among international students. Conversely, consistent condom use (defined as use during every encounter) remained critically low: 6.9% (Mainland Chinese), 4.8% (HMT), and 0.0% (international students) (Table 3).

Significant gender-based disparities were observed, with male students demonstrating a higher frequency of sexual activity (8.3%) compared to their female counterparts (4.7%) (χ² = 62.344, p < 0.05).

Regional analysis revealed pronounced differences: HMT students exhibited the highest sexual activity rate at 17.4%, significantly exceeding rates among Mainland Chinese (4.5%) and international students (9.0%) (χ² = 127.456, p < 0.05) (Table 1).

Both composite scores were non-normally distributed; results are therefore summarized as median (IQR). Among 4,892 respondents, the HIV Knowledge Score had a median of 7.0 with an IQR of 6.0–8.0 on the 0–8 scale, whereas the Monkeypox Knowledge Score had a median of 3.0 with an IQR of 1.0–3.0 on the 0–4 scale. In univariate logistic regression, higher HIV knowledge was inversely associated with sexual activity (OR = 0.852, 95% CI: 0.794–0.913), while higher Monkeypox knowledge was positively associated (OR = 1.122, 95% CI: 1.025–1.229) (Table 4).

Students who lacked awareness of same-sex sexual behavior exhibited a significantly higher frequency of sexual activity (12.4%) compared to those with prior awareness (5.1%) (χ² = 71.673, p < 0.001).

Attitudinal differences further revealed distinct patterns: Students holding opposing views toward same-sex sexual behavior demonstrated the highest sexual activity rate (12.6%), surpassing rates among those with neutral (4.0%) or supportive attitudes (4.3%) (χ² = 121.027, p < 0.001) (Table 5).

Survey results indicated that first-year students who did not receive HIV-prevention education in the past year reported a significantly higher frequency of sexual activity (10.5%) than those who did receive such education (5.3%) (χ² = 39.762, p < 0.001). In addition, students who served as campaign facilitators for HIV-prevention activities showed a higher frequency of sexual activity (9.9%) compared with non-participants (4.0%) (χ² = 66.396, p < 0.001) (Table 6).

The results demonstrated several significant associations. Compared to Mainland Chinese students, those from Hong Kong, Macau, and Taiwan (HMT) showed 3.737 times higher odds of sexual activity (95% CI: 2.846–4.908). Higher knowledge levels about Monkeypox were positively associated with sexual activity (OR = 1.174, 95% CI: 1.060–1.301), while awareness of potential consequences of same-sex sexual behavior was protective (OR = 0.453, 95% CI: 0.346–0.594). Students with neutral (OR = 0.377, 95% CI: 0.287–0.496) or supportive (OR = 0.371, 95% CI: 0.218–0.629) attitudes toward same-sex behavior had lower odds of sexual activity compared to those with opposing views. Notably, while receiving HIV prevention education was associated with reduced odds of sexual activity (OR = 0.294, 95% CI: 0.213–0.405), students who participated as campaign facilitators showed substantially higher odds (OR = 4.280, 95% CI: 3.140–5.834) (Table 7).

In the supplementary sensitivity analyses, all key non-binary predictors deviated significantly from normality (Shapiro–Wilk tests, p < 0.001 for HIV Knowledge Score, Attitude toward same-sex sexual behavior, and Monkeypox Knowledge Score; Supplementary Table S1). The Box–Tidwell tests confirmed that the logit-linearity assumption was satisfied for both continuous knowledge scores: the interaction terms for HIV Knowledge Score and Monkeypox Knowledge Score were non-significant (Wald χ² = 0.496, p = 0.481; and 2.229, p = 0.135, respectively; Supplementary Table S2). Likewise, treating Attitude toward same-sex sexual behavior as an ordinal or continuous predictor yielded results consistent with the main analysis: a linear-by-linear trend test showed a highly significant association (χ² = 91.817, p < 0.001), and coding this attitude variable as a continuous covariate in the logistic model produced an adjusted OR of 0.467 (95% CI: 0.370–0.590, p < 0.001), indicating a robust protective effect (Supplementary Table S3).

Taken together, the normality checks, Box–Tidwell tests, and trend analyses demonstrate that the chosen logistic-regression specification is statistically appropriate: the two knowledge scores perform well as linear continuous terms, and the attitude variable retains a stable effect when treated either categorically or as an ordinal predictor, confirming that our main findings are not sensitive to alternative model parameterisations.