[Latent class analysis on characteristics of self-reported sexual behavior and associated factors in college students in Tianjin].

Objective

To identify latent categories of HIV infection risk in college students with self-reported sexual behaviors in Tianjin, analyze the associated factors, and provide evidence for targeted HIV/AIDS prevention interventions in universities.

Methods

Based on cross-sectional study design, an anonymous online questionnaire survey was conducted among college students selected by stratified cluster sampling from 54 colleges or universities in 16 districts of Tianjin from September to December 2024. In principle, the survey was required to cover no less than 10% of the total enrolled students at each university (or campus). The sample size of 5 034 was estimated based on previous survey data. College students who self-reported sexual behaviors were included, and information on demographic characteristics, self-reported sexual behaviors, sexual attitudes and relationship status, and HIV/AIDS-related knowledge and behaviors was collected. Latent class analysis (LCA) was performed on sexual behavior characteristics, and χ² test and multilevel multinomial logistic regression model were used to analyze the influencing factors of different latent classes.

Results

The effective questionnaire response rate was 98.6% (101 593/102 996), and 6 325 students who reported their sexual behaviors were included in the study. LCA identified three latent classes: the low sexual active group (25.4%,1 605/6 325), consisting of those who reported no sexual behaviors in the past year; the high-exposure group (5.2%, 329/6 325), consisting of those who reported having multiple types of sexual partners, having both homosexual and heterosexual partners, and having no consistent condom use at sex; and the steady partner group (69.4%,4 391/6 325), consisting of those who reported having only steady romantic partners in the past year and consistently using condoms during sexual activity. Taking the low sexual activity group as a reference, multivariate analysis showed that positive factors associated for the high-exposure group of self-reported sexual behavior latent classes among college students included being in a relationship (aOR=1.59, 95%CI: 1.22-2.05), being willing to accept sexual partners with prior sexual experience (aOR=1.93, 95%CI: 1.40-2.62), and never having seen AIDS prevention posters (aOR=2.50, 95%CI: 1.81-3.38); negative factors in the high-exposure group latent classes included being from rural area (aOR=0.66, 95%CI: 0.51-0.85), being junior (third-year) students (aOR=0.66, 95%CI: 0.44-0.95) or graduate students (aOR=0.55, 95%CI: 0.31-0.90), and making friends with HIV-infected individuals only under certain conditions (aOR=0.60, 95%CI: 0.41-0.86). The positive associated factors for the steady partner group latent class included being female (aOR=1.52, 95%CI: 1.30-1.78), majoring in humanities and social sciences (aOR=1.20, 95%CI: 1.03-1.39), being in a relationship (aOR=6.17, 95%CI: 5.39-7.05), being accepting of sexual activity during college (aOR=2.77, 95%CI: 1.86-3.93), awareness of HIV/AIDS prevention knowledge (aOR=1.24, 95%CI: 1.02-1.49), and supporting the rights of people living with HIV with condition (aOR=1.21, 95%CI: 1.02-1.41); the negative factor in the steady partner group latent classes was being graduate students (aOR=0.74, 95%CI: 0.56-0.95).

Conclusions

Significant heterogeneity exists in self-reported sexual behavior characteristics among college students in Tianjin, which can be classified into a low sexual activity group, a high-exposure group, and a steady partner group. HIV/AIDS prevention efforts in universities should shift from mere knowledge dissemination to an integrated "knowledge-attitude-practice" intervention, and stratified, targeted intervention strategies should be implemented.