Cusp Catastrophe Modeling of Suicide Behaviors among People Living with HIV in China.

The number of people living with HIV (PLWH) increases rapidly with advancement in antiretroviral therapy. Suicide prevention is challenging, particularly for PLWH because of the negative impact of HIV-related stigma and social capital reduction. However, only a small proportion of the variance in suicide risk can be explained by these variables if a linear paradigm is used as guidance. In this study, we tested a nonlinear cusp catastrophe modeling. Participants (N = 523) were PLWH selected through a risk venue-based method in Wuhan, a provincial capital city in China. Suicidal ideation post HIV+ diagnosis and in the past 30 days and suicide plan and attempt were assessed. Data were collected using self-report questionnaire and were analyzed using both the direct and multivariate stochastic cusp catastrophe modeling methods with social capital as asymmetry variable and HIV-related stigma as bifurcation variable. The analysis was executed using R, including nls() function for the direct method and 'cusp' package for the stochastic modeling. Results from stochastic cusp modeling analysis indicated that social capital was significantly associated with risk of suicide after controlling for key covariates; the association was significantly bifurcated by HIV-related stigma. The data fit the cusp model better than the alternative linear model (R2 =.483 vs. .127). Findings of this study indicate suicide behaviors among Chinese PLWH follow a nonlinear dynamic system. In addition to enhancing our understanding of suicide risk, findings of this study underscore the significance in social capital enhancement and stigma reduction for suicide prevention among PLWH in China.