Data from a randomized controlled trial in 2015 were used to estimate the growth trajectories of peer norms, self-efficacy, and condom use behavior, and to identify associated sociodemographic and behavioral factors among a sample of 804 Chinese men who have sex with men (MSM). Latent class analysis and growth mixture modeling were conducted using Mplus. Two growth trajectories were estimated for each outcome variable with good model fit. The growth trajectories of peer norms were related to age (β = - 0.066, p < 0.05). The growth trajectories of self-efficacy were related to age (β = 0.057, p < 0.01) and using a condom during first sexual encounter with another man (β = 0.777, p < 0.001). The growth trajectories of condom use behavior were related to income (β = 0.366, p < 0.01) and having casual male partners (β = - 1.016, p < 0.001). Predictors for the growth factors within each latent class were also estimated. For subsets of MSM who are older, richer, used a condom during their first sexual encounter with another man, and do not have a casual male partner, condom videos may not have sufficient efficacy and other interventions may be necessary.
Keywords: Chinese MSM; Condom use; Growth mixture modeling; Growth trajectory; Latent class analysis.