Effectiveness of HIV Testing through Social Networks among Young, African American, Men who have Sex with Men in St. Louis
1. Effectiveness of HIV Testing through Social
Networks among Young, African American,
Men who have Sex with Men in St. Louis
Nebu Kolenchery
Saint Louis University
APHA Annual Meeting, November 3, 2015
2. PRESENTER DISCLOSURES
(1) The following personal financial relationships with
commercial interests relevant to this presentation
existed during the past 12 months:
Nebu Kolenchery
No relationships to disclose
3.
4.
5. Causes of Disparity
• Poverty
• Access to Clinical Care
• Stigma
• Sexual Behavior
• Innovative Strategies needed
6.
7. Social Network Testing
• First published in 2003 MMWR
• Recruiters were trained
• Found network associates
• Incentives provided to
both parties
• Effective in many populations, no published
results for YAAMSM
12. Total
Gender
Male 750 (47.1%)
Female 842 (52.9%)
Male Sexual Orientation
Not Heterosexual 300 (40.0%)
Heterosexual 450 (60.0%)
Total 1592
Results
Total
Gender
Male 750 (47.1%)
Female 842 (52.9%)
Male Sexual Orientation
Not Heterosexual 300 (40.0%)
Heterosexual 450 (60.0%)
Total 1592
13. Total
Race
African American 1253 (78.7%)
Caucasian/White 225 (14.1%)
Other 114 (7.2%)
Age
Mean (St. Dev) 21.22 (2.98)
Median (Range) 21 (12 – 29)
Total 1592
Results
Total
Race
African American 1253 (78.7%)
Caucasian/White 225 (14.1%)
Other 114 (7.2%)
Age
Mean (St. Dev) 21.22 (2.98)
Median (Range) 21 (12 – 29)
Total 1592
14. Social Network
Testing
Alternate
Venue Testing
In-House
Testing
Gender
Male 38 (76.0%) 236 (43.0%) 476 (47.9%)
Female 12 (24.0%) 313 (57.0%) 517 (52.1%)
Male Sexual Orientation
Not Heterosexual 33 (86.8%) 77 (32.6%) 190 (39.9%)
Heterosexual 5 (13.2%) 159 (67.4%) 286 (60.1%)
Total 50 549 993
Results
Social Network
Testing
Alternate
Venue Testing
In-House
Testing
Gender
Male 38 (76.0%) 236 (43.0%) 476 (47.9%)
Female 12 (24.0%) 313 (57.0%) 517 (52.1%)
Male Sexual Orientation
Not Heterosexual 33 (86.8%) 77 (32.6%) 190 (39.9%)
Heterosexual 5 (13.2%) 159 (67.4%) 286 (60.1%)
Total 50 549 993
15. Social Network
Testing
Alternate
Venue Testing
In-House
Testing
Race
African American 36 (72.0%) 426 (77.6%) 791 (79.7%)
Caucasian/White 12 (24.0%) 95 (17.3%) 118 (11.9%)
Other 2 (4.0%) 28 (5.1%) 84 (8.5%)
Age
Mean (St. Dev) 21.28 (2.96) 21.92 (3.39) 20.83 (2.65)
Median (Range) 21 (15 – 29) 22 (12 – 29) 21 (14 – 29)
Total 50 549 993
Results
Social Network
Testing
Alternate
Venue Testing
In-House
Testing
Race
African American 36 (72.0%) 426 (77.6%) 791 (79.7%)
Caucasian/White 12 (24.0%) 95 (17.3%) 118 (11.9%)
Other 2 (4.0%) 28 (5.1%) 84 (8.5%)
Age
Mean (St. Dev) 21.28 (2.96) 21.92 (3.39) 20.83 (2.65)
Median (Range) 21 (15 – 29) 22 (12 – 29) 21 (14 – 29)
Total 50 549 993
17. • Social Network Testing is an effective
method to reach YAAMSM in St. Louis
• It depends on strong relationships
• Cost-effectiveness
Discussion
18. Limitations
• Not a Randomized Control Trial
• Sample Size
• Missing Data
Future directions could include
• Larger sample sizes
• Randomized Controlled Studies
• Return on Investment Studies
Discussion
19. • Social Network Testing successful for
YAAMSM
• Combination of approaches
Implications
Notas do Editor
In March of 2015, the Centers for Disease Control and Prevention (CDC), in their report, “HIV in the United States: At A Glance,” noted that there were over 1.2 million people in the United States living with HIV. The report mentioned that while the prevalence of infection was steadily increasing in the last decade, the incidence of HIV infection has remained stable with around 50,000 new diagnoses per year. However, the rate of new infections in young (usually defined as younger than 30), African American, Men who have Sex with Men (YAAMSM) are much higher than in any other subset of the population. In 2010, there were 4800 new infections among YAAMSM aged 13-24. This accounted for 45% of new infections among African American MSM and 55% of new infections among young MSM. A study which estimates HIV incidence from surveillance data from 2006-2009 in 16 states and 2 cities confirmed that while the incidence was about 50,000 cases per year, there was a 21% increase in the cases for people aged 13-29 years. A large contributor to this was the 48% increase in HIV infection among YAAMSM who were younger than 30.
In March of 2015, the Centers for Disease Control and Prevention (CDC), in their report, “HIV in the United States: At A Glance,” noted that there were over 1.2 million people in the United States living with HIV. The report mentioned that while the prevalence of infection was steadily increasing in the last decade, the incidence of HIV infection has remained stable with around 50,000 new diagnoses per year. However, the rate of new infections in young (usually defined as younger than 30), African American, Men who have Sex with Men (YAAMSM) are much higher than in any other subset of the population. In 2010, there were 4800 new infections among YAAMSM aged 13-24. This accounted for 45% of new infections among African American MSM and 55% of new infections among young MSM. A study which estimates HIV incidence from surveillance data from 2006-2009 in 16 states and 2 cities confirmed that while the incidence was about 50,000 cases per year, there was a 21% increase in the cases for people aged 13-29 years. A large contributor to this was the 48% increase in HIV infection among YAAMSM who were younger than 30.
M from 2006-2009
The disparity in HIV infection among YAAMSM arises from a variety of factors. A systematic review that compared racial disparities in MSM in the United States, Canada and the United Kingdom concluded that African American MSM in the US were more likely than MSM of any other race to not have a high school education, to have a low income, to have a criminal record and to be unemployed. The authors also concluded that disparities in access to HIV clinical care contributed to higher rates of infection.
The stigma surrounding the disease, and the social stress and marginalization that stems from homophobia is associated with delayed HIV diagnosis in MSM population.
African American MSM are also more likely to have an earlier sexual debut, and have sexual partners of their own race (assortative mating), who are often older than them. While African American MSM have lower levels of behaviors that put them at a higher risk for HIV (such as intravenous drug use) when compared with to White MSM, assortative mating with older partners leads to strong, tight-knit social networks that allow the quick spread of disease in a relatively small population. YAAMSM, more than other groups, are more likely to test HIV positive on their first test, and be unaware of their positive serostatus (whether or not they were HIV positive), both of which are risk factors for the rapid spread of disease in a sub-population. This epidemic calls for innovative and targeted prevention strategies that expand testing to YAAMSM and reduce the disparity in the incidence of HIV.
I was born and raised in an Indian community, and for my senior capstone, I wanted to public health in India. Through the Investigative Learning Experience Grant, I spent three weeks in August of 2014 interning with one of the Suraskha Projects, which are targeted HIV prevention interventions for migrant workers in the state of Kerala. The Suraksha Projects target specific sub-populations that have a higher prevalence of HIV (1). One of their strategies involves employing Community Health Workers to educate migrant workers who hail from states that have a much higher rate of HIV infection than Kerala. In this picture, I am accompanying a Community Health Worker to an education session at a plywood factory. The Community Health Workers give the workers classes on HIV and offer them free testing and contraception. However, they also recruit the foreman of the factory, or a worker that is seen as a leader in the community to work for the Suraksha Project. They are given a stipend to host a certain number of classes outside of the work day, and have to meet a target number of workers that they recruit to get tested for HIV. This model takes advantage of the trust and social networks within the target population of migrant workers, and led to a greater level of reach in the community. I wondered if there were similar strategies being employed in the United States. I was particularly interested in those that relied on social ties in vulnerable communities.
To explore this further, I applied for a second internship at Project ARK in St. Louis. Project ARK (AIDS Resources and Knowledge), is a Ryan White Part D funded program of the Department of Pediatrics at the Washington University School of Medicine. Since 1995, Project ARK has been serving the children, youth and women at risk for HIV/AIDS in the St. Louis Region and has found great success in reaching, engaging, and building trust in the community. The SPOT (Supporting Positive Opportunities for Teens) is housed in the same building as Project ARK, was formed in 2009 as a partnership between Project ARK and the Adolescent Center in St. Louis. The SPOT caters to youth between 13 and 24 years of age, and is a free drop-in center that provides comprehensive sexual and reproductive health care, social support, and prevention and case management services.
The rising incidence of HIV and other Sexually Transmitted Infections is a huge public health issue among African American youth in St. Louis. From 2008 to 2010, the new cases of HIV/AIDS among male youth in St. Louis increased by 33% compared to only 2% in adult males. Both the SPOT and Project ARK have sound relationships with the African American MSM community in St. Louis and have employed multiple recruitment strategies to reach the local HIV positive population.
My project looked at three different testing models: Social Network Testing (SNT), Alternate Venue Testing (AVT), and In-House Testing (IHT).
This study will make a case for the use of a Social Network Testing model to better reach the young, African-American, Men who have Sex with Men population in urban settings in the United States.
The concept of the social network was first explored by Barnes in 1954 when he described the relationships in a Norwegian village that extended beyond family and work, and found that members of close-knit networks have influence on one another. Community-based interventions to reduce risky behavior among MSM in urban areas have shown that popular opinion leaders in the MSM community are able to have a significant influence on the sexual practices of their peers in their social network. Within the context of HIV testing, Social Network Testing (SNT) was first given national exposure in a 2005 Morbidity and Mortality Weekly Report from the CDC and was found to be a successful method to reach MSM that were at high-risk for HIV.
The Social Network Testing method involves enlisting individuals that are at high-risk for HIV to recruit their peers in their respective social network to be tested. The former are called Recruiters and the latter are called Network Associates. At Project ARK, the target population was non-white MSM clients, and clients of demographics were welcome to get recruited and tested. The staff acted as Gatekeepers, and enlisted Recruiters. These Recruiters were past clients, or were recruited through other testing methods. The staff selected Recruiters based on their whether they exhibited qualities that showed that they were a leader in their social network and have access to individuals that was not previously tested, as well as a demonstrated ability to be comfortable talking about HIV and testing to their peers. The Recruiters were chosen irrespective of their serostatus, and the staff then gave them an informal training on approaching their peers, and offered advice on how to answer questions about HIV, testing, and their own serostatus. Each Recruiter was given a set of ten business cards, and the Recruiters were asked to find Network Associates (NA) to be tested. Recruiters had to talk to their NAs about HIV and testing at Project ARK. If they presented for testing with a business card, the NAs were given a $10 gift card, and their respective Recruiter, identified by virtue of his or her business card, was given $15 gift card. Thus both parties were incentivized to participate.
Past studies have shown that Social Network Testing is extremely effective with many different populations including high-risk African American women, Intravenous Drug Users (particularly African American IDUs), African American MSM in Baltimore, San Francisco and New York. This shows the potential for SNT to be effective in urban areas similar to St. Louis. There have been recommendations that SNT be used among YAAMSM population. To my knowledge, there has not been a published study that uses SNT in this sub-population.
For AVT, the staff at the SPOT and Project ARK went to locations where they were more likely to encounter members of a sub-population that are at a higher risk for contracting HIV. The testing was conducted at local libraries, bars/nightclubs, community centers, and churches, among other locations. This method of testing would ideally reach a larger population of racial and ethnic minorities and MSM clients. The IHT method relied on the strong reputation of Project ARK and the SPOT as an agency that offered HIV testing services in the St. Louis area. Most clients that were recruited by this method were walk-in clients that heard that free testing was offered through their healthcare provider, online, or from their peers. When being tested, clients were offered HIV Counseling and a Risk Assessment questionnaire. However, some clients were also tested through Express Testing, which meant that they did not complete the entire Counseling and Risk Assessment Process.
HIV testing data from September 2013 through April 2014 was analyzed through SPSS and R
All clients 30 years of age and older were excluded from the analysis to limit the data to young clients. 30 years was chosen as the cut-off because most studies classify young MSM as those that are 13-29. All clients with an invalid test result, or those that were missing data for their test result were excluded as well. If the remaining individuals identified as African American and MSM, they were coded to be a YAAMSM client.
Chi-squared analyses were performed to examine differences in gender, race, and sexual orientation across the three different test types. Age distribution was compared using a One-Way ANOVA. The primary outcome that was analyzed was HIV positivity in the three different testing strategies. The second hypothesis examined whether or not SNT would recruit a higher proportion of YAAMSM compared to AVT and IHT. These hypotheses were tested using a Fisher’s Exact Test.
From September 2013 to April 2014, Project ARK and the SPOT serviced 2263 clients. With the administration of the aforementioned exclusion criteria, this number was reduced to 1592 clients below 30 with conclusively Positive or Negative test results. 18 of these 1592 were positive, which makes for an overall prevalence rate of 1.13%.
For SNT, there were 5 staff that served as Gatekeepers to enlist Recruiters. In total, the staff found 44 recruiters (each staff found anywhere from 3-19 Recruiters). Each Recruiter was given 10 business cards. Some recruited no Network Associates, while others successfully recruited 10 Network Associates. 50 total Network Associates were recruited as part of SNT. Through AVT, the 549 clients were tested, and 993 clients were tested through IHT.
The total sample was 47.1% male (40% of which were MSM).
They were 78.7% African American, and had a mean age of 21.22, with a median age of 21 (ranging from 12-29).
This table shows the demographic variables stratified by type of testing strategy. 76% of those tested with SNT were male, and 86.8% of those were MSM. This is much higher compared to the other two methods.
With respect to Race, SNT recruited 72% African Americans which was slightly lower percentage than AVT and IHT. The age distribution among the three types of testing are relatively similar, with all having mean ages that are between 20 and 22, and median ages of 21 and 22.
This table show the results of the two hypotheses that were tested. SNT showed a positivity rate of 6.0%, while AVT and IHT were 0.9% and 1.0% respectively (p-value = 0.032). 12.5% of all clients were YAAMSM. 44% of the SNT sample was YAAMSM, compared to 7.8% in AVT and 13.5% in IHT (p-value < 0.001). Both Fisher’s Exact Tests yielded p-values that were both less than 0.05 and statistically significant. Therefore, SNT yielded a significantly higher positivity rate and recruited a significantly higher number of YAAMSM clients.
The results suggest both that SNT is an effective method to reach YAAMSM in St. Louis, and that it is more effective at finding positives through targeted intervention. 16 out of the 18 positives that were found over the course of eight months were YAAMSM, and all 3 positives that were found from SNT were YAAMSM as well. While 3 positives is a rather small number, Project ARK and the SPOT are reputable public health entities in the St. Louis area, and 3 out of 18 positives, or 16.7% of their HIV positive clients out of the 1592 tested (2263 total) were found through SNT. SNT had about 6 times the positivity rate of both AVT and IHT. The success of SNT within the context of a real service provider is evidence that the strategy is suitable for YAAMSM in urban communities.
In addition to identifying HIV positive clients, there is evidence to show that YAAMSM suffer from significant social isolation, and that HIV prevention interventions based on community building through virtual and non-virtual social networks would be effective in this population. Social Network Testing relies on the strength of relationships between individuals and that the entity carrying out the intervention is in good standing and can maintain relationships with the target population.
Among African American MSM in cities, HIV prevalence is often further complicated by stigma and urban poverty, and necessitates interventions that address both the social factors that affect HIV and increase the number of YAAMSM being tested. While successful implementation of SNT is resource-intensive (25), a 2006 study found that identifying one HIV positive person through an SNT-type model would cost $4,929, while location-based and CBO-based testing cost $8,250 and $11,481 respectively (26). This is due to the relatively small number of individuals that need to be recruited and tested to find positives when compared to other strategies.
This study did have its limitations. This is not a randomized controlled trial, and since it is a secondary analysis of existing data, there is potential for selection and information bias in the design.
The sample size of 50 for SNT was much smaller than the 549 for AVT and the 993 for IHT, and this does limit the statistical power to draw conclusions from the data. However, previous published studies of SNT have had similarly small sample sizes. In addition, 3 out 18, or 16.7% of the HIV-clients came from 3.14% of all clients serviced during the time that data was collected. Within the context of a clinical service provider, this does establish that SNT is still very effective in identifying positivity.
Since the data was collected in a non-controlled setting, there is a lot of missing data that had to be excluded from the study. For example, there were many tests that were interpreted to be invalid or left blank for the result. These were left out of the analysis, and might have an effect on the results. The data also did not include any variables that measured the socioeconomic status (SES) of the clients. SES could potentially be a confounder in this study due to the previously established relationship between poverty and prevalence of HIV, particularly in MSM.
Given these limitations, future directions for this type of study could involve larger sample sizes and randomizing targeted study participants to different types of recruiting methods. In addition, the potential cost-effectiveness could be measured with a younger population since this could save much more through early detection and management of HIV.
SNT shows great promise as a potential solution to the rising incidence of HIV in YAAMSM. Social network based approaches have also found success in reducing risky sexual behaviors in MSM like Unprotected Anal Intercourse. SNT Recruiters could have a similar effect in reducing risky behaviors in YAAMSM.
It is also important to note that the complex and multi-faceted nature of HIV transmission in YAAMSM calls for multiple testing approaches. Community-based organizations like Project ARK and the SPOT should use a combination of approaches in order best serve their clients. Nonetheless, Social Network Testing can be a powerful, economical addition to the arsenal of a service provider that caters to YAAMSM in cities.