Network Centrality and Geographical Concentration of Social and Service Venues that Serve Young Men Who Have Sex with Men

Explainable artificial intelligence and domain adaptation for predicting HIV infection with graph neural networks

OBJECTIVE

Investigation of explainable deep learning methods for graph neural networks to predict HIV infections with social network information and performing domain adaptation to evaluate model transferability across different datasets.

METHODS

Network data from two cohorts of younger sexual minority men (SMM) from two U.S. cities (Chicago, IL, and Houston, TX) were collected between 2014 and 2016. Feature importance from graph attention network (GAT) models were determined using GNNExplainer. Domain adaptation was performed to examine model transferability from one city dataset to the other dataset, training with 100% of the source dataset with 30% of the target dataset and prediction on the remaining 70% from the target dataset.

RESULTS

Domain adaptation showed the ability of GAT to improve prediction over training with single city datasets. Feature importance analysis with GAT models in single city training indicated similar features across different cities, reinforcing potential application of GAT models in predicting HIV infections through domain adaptation.

CONCLUSION

GAT models can be used to address the data sparsity issue in HIV study populations. They are powerful tools for predicting individual risk of HIV that can be further explored for better understanding of HIV transmission.

Examination of multidimensional geographic mobility and sexual behaviour among Black cisgender sexually minoritized men in Chicago

Black sexually minoritized men (BSMM) are the most likely to acquire HIV in Chicago- a racially segregated city where their daily travel may confer different HIV-related risks. From survey and GPS data among participants of the Neighbourhoods and Networks Cohort Study, we examined spatial (proportion of total activity space away from home), temporal (proportion of total GPS points away from home), and motivation-specific (discordance between residential and frequented sex or socializing neighbourhoods) dimensions of mobility. To identify potential drivers of BSMM's risk, we then examined associations between mobility and sexual behaviours known to cause HIV transmission: condomless anal sex, condomless anal sex with a casual partner, transactional sex, group sex, and sex-drug use. Multivariable logistic regression models assessed associations. Of 269 cisgender BSMM, most were 20-29 years old, identified as gay, and lowincome. On average, 96.9% (Standard Deviation: 3.7%) of participants' activity space and 53.9% (Standard Deviation: 38.1%) of participants' GPS points occurred outside their 800m home network buffer. After covariate adjustment, those who reported sex away from home were twice as likely to report condomless sex (Odds Ratio: 2.02, [95% Confidence Interval (CI): 1.08, 3.78]). Those who reported socializing away from home were four times more likely to have condomless sex with a casual partner (Odds Ratio: 4.16 [CI: 0.99, 29.0]). BSMM are on the move in Chicago, but only motivation-specific mobility may increase HIV transmission risk. Multidimensional investigations of mobility can inform place-based strategies for HIV service delivery.

Integrated molecular and affiliation network analysis: Core-periphery social clustering is associated with HIV transmission patterns

This study investigates the two-mode core-periphery structures of venue affiliation networks of younger Black men who have sex with men (YBMSM). We examined the association between these structures and HIV phylogenetic clusters, defined as members who share highly similar HIV strains that are regarded as a proxy for sexual affiliation networks. Using data from 114 YBMSM who are living with HIV in two large U.S. cities, we found that HIV phylogenetic clustering patterns were associated with social clustering patterns whose members share affiliation with core venues that overlap with those of YBMSM. Distinct HIV transmission patterns were found in each city, a finding that can help to inform tailored venue-based and network intervention strategies.

Neighborhood and Network Characteristics and the HIV Care Continuum among Gay, Bisexual, and Other Men Who Have Sex with Men

In order for treatment as prevention to work as a national strategy to contain the HIV/AIDS epidemic in the United States (US), the HIV care continuum must become more robust, retaining more individuals at each step. The majority of people living with HIV/AIDS (PLWHA) in the US are gay, bisexual, and other men who have sex with men (MSM). Within this population, there are distinct race- and ethnicity-based disparities in rates of HIV infection, engagement, and retention in HIV care, and viral suppression. Compared with White MSM, HIV-infected Black MSM are less likely to be on anti-retroviral therapy (ART), adhere to ART, and achieve viral suppression. Among MSM living in urban areas, falling off the continuum may be influenced by factors beyond the individual level, with new research identifying key roles for network- and neighborhood-level characteristics. To inform multi-level and multi-component interventions, particularly to support Black MSM living in urban areas, a clearer understanding of the pathways of influence among factors at various levels of the social ecology is required. Here, we review and apply the empirical literature and relevant theoretical perspectives to develop a series of potential pathways of influence that may be further evaluated. Results of research based on these pathways may provide insights into the design of interventions, urban planning efforts, and assessments of program implementation, resulting in increased retention in care, ART adherence, and viral suppression among urban-dwelling, HIV-infected MSM.

Socio-Structural and Neighborhood Predictors of Incident Criminal Justice Involvement in a Population-Based Cohort of Young Black MSM and Transgender Women

Black men who have sex with men (MSM) and transgender women are disproportionately affected by criminal justice involvement (CJI) and HIV. This study recruited 618 young Black MSM and transgender women in Chicago, IL, using respondent-driven sampling between 2013 and 2014. Random effects logistic regression evaluated predictors of incident CJI over 18 months of follow-up. Controlling for respondent age, gender and sexual identity, spirituality (aOR 0.56, 95% CI 0.33-0.96), and presence of a mother figure (aOR 0.41, 95% CI 0.19-0.89) were protective against CJI. Economic hardship (financial or residential instability vs. neither aOR 2.23, 95% CI 1.10-4.51), two or more past episodes of CJI vs. none (aOR 2.66, 95% CI 1.40-5.66), and substance use (marijuana use vs. none aOR 2.79, 95% CI 1.23-6.34; other drug use vs. none aOR 4.49, 95% CI 1.66-12.16) were associated with CJI during follow-up. Research to identify and leverage resilience factors that can buffer the effects of socioeconomic marginalization may increase the effectiveness of interventions to address the socio-structural factors that increase the risk for CJI among Black MSM and transgender women. Given the intersection of incarceration, HIV and other STIs, and socio-structural stressors, criminal justice settings are important venues for interventions to reduce health inequities in these populations.

The Social Context of HIV Prevention and Care among Black Men Who Have Sex with Men in Three U.S. Cities: The Neighborhoods and Networks (N2) Cohort Study

Background: In many parts of the world, stark racial disparities in human immunodeficiency virus (HIV) prevalence, incidence, prevention, and care outcomes persist among gay, bisexual, and other men who have sex with men (MSM), with Black MSM significantly impacted in the United States (U.S.). Individual-level characteristics, including sexual behaviors and socioeconomic status, do not fully account for racial/ethnic disparities in HIV among MSM. We hypothesize that neighborhood contexts and network characteristics influence risk for HIV infection as well as HIV-related prevention and care behaviors. As such, the study design includes the use of real-time geospatial methods and in-depth assessments of multiple network typologies to investigate the impact of neighborhood and network-level factors on HIV prevention and treatment among Black MSM residing in longstanding priority HIV elimination areas in the U.S., namely Chicago, Illinois and in the Deep South (Jackson, Mississippi and New Orleans, Louisiana) (n = 450, n = 50, and n = 100, respectively). We describe the design, sampling methods, data collection, data management methods, and preliminary findings of the ongoing 'Neighborhoods and Networks (N2) Cohort Study'. Methods/Design: N2 employs a prospective longitudinal design. The sample includes Black MSM participants in Chicago recruited via respondent-driven sampling and assessed every six months over two years of follow-up. Participants enrolled in Jackson and New Orleans are being recruited through existing health and community services and assessed every six months over one year of follow-up. Mobility within and between neighborhoods is being assessed using global positioning system (GPS) technology. Social and sexual networks among Black MSM are being studied through egocentric network inventories as well as newer methods of creating meso-level networks that involve social media (Facebook) and mobile phone contacts. Key HIV prevention outcomes such as pre-exposure prophylaxis (PrEP) care engagement, and HIV/STI (sexually transmitted infections) biomarkers will be examined at baseline and follow-up. Results: As of 31 December 2018, a total of 361 men were enrolled across all study sites: 259 in Chicago and 102 in the Deep South (75 in New Orleans and 27 in Jackson). At baseline, participants ranged in age from 17 to 65 years old (mean = 34.3, standard deviation = 5.1) with 123 men (34.1%) self-reported as HIV positive. While HIV treatment levels were similar between sites, men in the Deep South reported higher rates of adherence than men in Chicago (63.3% versus 49.4%, p = 0.03). Sexual risk profiles were mainly the same between men from different study sites, with 22.9% of men in Chicago and 28.9% in the Deep South reporting consistent condom use during vaginal and anal sex (p = 0.26). Regarding their home neighborhoods, men in the Deep South were more likely than those in Chicago to characterize theirs as having a good reputation (43.1% versus 24.7%, p < 0.001) and as being safe (37.3% versus 21.2%, p = 0.002). Conclusions: The focus on Black MSM in the N2 Study will allow for a nuanced exploration of the attitudes, beliefs, behaviors, and practices of a diverse group of Black MSM. The study is also positioned to provide novel insight about neighborhood and network characteristics that influence HIV-related behaviors. A health equity framework ensures that Black MSM are not explicitly or implicitly deemed as deviant, disordered, or the non-reference group. Findings from N2 will provide guidance for the implementation of more impactful HIV prevention interventions that engage a diverse population of Black MSM as we work toward HIV elimination in the U.S.

Neighbourhoods, networks and pre-exposure prophylaxis awareness: a multilevel analysis of a sample of young black men who have sex with men

OBJECTIVE

Young black men who have sex with men (YBMSM) in the USA represent a subgroup that has the highest HIV incidence among the overall population. In the USA, pre-exposure prophylaxis (PrEP) is an effective prevention intervention to prevent HIV acquisition when taken regularly. Neighbourhood and network factors may relate to PrEP awareness, but have not been studied in YBMSM. This study aimed to examine the relationship of neighbourhood and network characteristics with PrEP awareness among YBMSM.

METHODS

We used data collected from a sample of 618 YBMSM in Chicago (2013-2014). Home addresses were collected for participants and enumerated network members. Administrative data (eg, 2014 American Community Survey, Chicago Department of Public Health) were used to describe residence characteristics. Network member characteristics were also collected (eg, sexual partners' sex-drug use, confidant network members who were also MSM). Multilevel analysis was performed to examine the relationships of neighbourhood and network characteristics to PrEP awareness.

RESULTS

Higher neighbourhood-level educational attainment (adjusted odds ratio (aOR) 1.02, p=0.03) and greater primary care density (aOR 1.38, p=0.01) were associated with greater PrEP awareness; greater neighbourhood alcohol outlet density (aOR 0.52, p=0.004) was associated with less PrEP awareness. Sexual network members residing in the same neighbourhood as the participants (aOR 2.58, p=0.03) and discussions around avoiding HIV acquisition with confidants (aOR 2.26, p=0.04) were associated with greater PrEP awareness.

CONCLUSIONS

The results suggest that neighbourhood and network characteristics can influence PrEP awareness in YBMSM. Additional studies are needed to understand the influences of neighbourhood (eg, MSM serving venues) and network (eg, peer to peer communication) characteristics on dissemination of PrEP information, uptake and adherence and the related mechanisms behind the associations.

Statistical adjustment of network degree in respondent-driven sampling estimators: venue attendance as a proxy for network size among young MSM

We introduce a new venue-informed network degree measure, which we applied to respondent-driven sampling (RDS) estimators. Using data collected from 746 young MSM in 2014-2016 in Chicago, IL, and Houston, TX, we estimated the population seroprevalence of HIV and syphilis and risk/protective behaviors, using RDS estimates with self-reported network size as a standard degree measure as well as our proposed venue-informed degree measure. The results indicate that the venue-informed degree measure tended to be more efficient (smaller variance) and less biased than the other measure in both cities sampled. Venue attendance-adjusted network size may provide a more reliable and accurate degree measure for RDS estimates of the outcomes of interest.