| Abstract |
Background: In the United States, HIV disproportionately affects Black cisgender sexual minority men (BSMM). Although epidemiological and behavioral surveillance are integral to identifying BSMM at risk of HIV infection and transmission, overreliance on self-reported data, inability to observe social contexts, and neglect of populations with limited engagement in health care systems limits their effectiveness. Digital epidemiological approaches drawing on social media data offer an opportunity to overcome these limitations by passively observing in organic settings activities, beliefs, behaviors, and moods that indicate health risks but are otherwise challenging to capture.
Objective: The primary aim of this study was to determine whether features of Facebook communication and networks were associated with biological, behavioral, and psychological indicators of HIV infection and transmission risk.
Methods: Facebook and survey data were collected from BSMM aged 18 to 35 years living in Chicago (N=310). Participants' Facebook posts were characterized using 4 culturally tailored topic dictionaries related to aspects of HIV protection and risk among BSMM (sexual health; substance use; sex behavior; and ballroom culture, a salient subculture in lesbian, gay, bisexual, transgender, and queer communities of color). Social network methods were used to capture structural features of BSMM's Facebook friendships (centrality, brokerage, and local clustering) and Facebook group affiliations. Multivariable regressions revealed relationships between these Facebook features and 5 ground truth indicators of HIV infection and transmission risk (sexually transmitted infection incidence, condomless sex, sex drug use, biomedical prevention, and depression).
Results: Although analysis of participants' Facebook posts revealed that HIV-related topics occupied a small portion of the total messages posted by each participant, significant associations were found between the following HIV risk indicators and Facebook features: Condomless sex, including communication about sexual health (odds ratio [OR] 1.58, 95% CI 1.09-2.29), ballroom culture (OR 0.76, 95% CI 0.63-0.93), and friendship centrality (OR 0.69, 95% CI 0.52-0.92); Sex drug use, including communication about substance use (OR 1.81, 95% CI 1.17-2.79) and friendship centrality (OR 0.73, 95% CI 0.55-0.96) and brokerage (OR 0.71, 95% CI 0.51-0.99); Biomedical prevention, including communication about ballroom culture (OR 0.06, 95% CI 0.01-0.71); and Depression, including communication about sexual health (β=-0.72, 95% CI -1.42 to -0.02), ballroom culture (β=.80, 95% CI 0.27-1.34), friendship centrality (β=-0.90, 95% CI -1.60 to -0.21), and Facebook group affiliations (β=.84, 95% CI 0.25-1.43). Facebook features provided no significant explanatory value for sexually transmitted infection incidence.
Conclusions: Finding innovative strategies to detect BSMM at risk of contracting or transmitting HIV is critical to eliminating HIV disparities in this community. The findings suggest that social media data enable passive observance of social and communicative contexts that would otherwise go undetected using traditional HIV surveillance methods. As such, social media data are promising complements to more traditional data sources.
|
