1. Risk Networks, Groups Sex and Other Risk "Nodes," and HIV Transmission: Mixing Patterns and the Limitations of Current HIV and Drug Use Interventions
Samuel R. Friedman

2. I would like to acknowledge
NIDA projects:
R01 DA Social Factors and HIV Risk
R01 DA13128 (Networks, norms & HIV risk among youth)
R01 DA DA A1 Staying Safe: Long-term IDUs who have avoided HIV & HCV
P30 DA (Center for Drug Use and HIV Research; Sherry Deren PI)
Hundreds of participants in these studies
Colleagues and participants who have died of HIV/AIDS and hepatitis C
Many collaborators and co-authors

3. Social and risk networks3

4. Most HIV epidemiology, prevention, and policy has focused on individual knowledge, attitudes, personality and behaviors: 4

5. People also have social and behavioral ties of various types and strengths5

6. Risk (sex, IDU) network ties can carry infections:
Within relationships
To or from an individual
Throughout a community or small group 6

7. Social research insight 1: risk networks carry infections via behaviors And this can help us understand both why HIV can spread quickly in some circumstances and help us develop prevention strategies 7

8. HIV risk is a conditional probability: Risk behaviors with uninfected people do not lead to infection
+, on HAART -
Negative
Unknown, but GC+ and HSV-2+
The probability is socially structured 8

9. What do real risk networks look like?
I will show you one from a study of injection drug users in New York City; one from a study of sex workers in British Columbia (Canada), and then spend some time discussing a New York City mixed-group network. 9

10. SFHR slide of 30-day risk relationships among IDUs in Brooklyn in early 1990s: Note role of shooting gallery managers as nodes
Note the high concurrency

11. Microstructures of large components in SFHR: more risk behavior, more HIV, but more health communication about condoms and bleach.

12. Canada: sex workers are pink circles, clusters of pink circles are brothels (massage parlours), blue squares are commercial clients, blue triangles are boyfriends or husbands, and pink triangles are wives/girlfriends. The 7 circles clusters represent massage parlours where recruitment took place, from 4 cities in the Vancouver area. Source: Valencia Remple.
These massage parlors are quasi-anonymous risk nodes. 12

13. HIV Positive by Gender/Sexuality (MSM=up triangle, WSW=down triangle, other female=circle, other male=square) by Hardest Drug Use Ever (from dark red to light pink: IDU, Crack, Non-injected Heroin or Cocaine; blue=other) by Link Type (sex=yellow line, IDU=red, sex and IDU=blue) (New York) 3-month links

14. HSV2 Positive by Gender/Sexuality (MSM=up triangle, WSW=down triangle, other female=circle, other male=square) by Hardest Drug Use Ever (from dark red to light pink: IDU, Crack, NI Heroin or Cocaine; blue=other) by Link Type (sex=yellow line, IDU=red, sex and IDU=blue)

15. Of 30 HIV-discordant partnerships in the diagram :
HIV discordant couples are where much HIV transmission takes place: Which STIs are likely to be important in facilitating HIV spread?
Of 30 HIV-discordant partnerships in the diagram :
5 were same-sex male partnerships and 25 were opposite-sex partnerships.
HSV-2 was present in 83% of the HIV-discordant couples; CT in 7%; and syphilis & gonorrhea in none.
HSV-2 is probably more important for HIV transmission than bacterial STDs since it is more widespread in NYC—and some African studies.
This may be because of effective syphilis and gonorrhea control in NYC at least. 15

16. Attended group sex event
Hidden risk: Although very little research has been done on this, group sex activities are likely to increase risks
Attended group sex event
Use no drugs or only marijuana
25%
Use non-injected cocaine or heroin
47%
IDU
35%
Female
28%
Male
42%
Non-injecting drug users in rural North Carolina
46%
Lower East Side (New York) youth aged less than 25
34%

17. Attended Group Sex Party (Pluses) and Had Unsafe Sex at Group Sex Party (Circles) by Gender/Sexuality (MSM=up triangle, WSW=down triangle, other female=circle, other male=square) by Hardest Drug Use Ever (from dark red to light pink: IDU, Crack, NI Heroin or Cocaine; blue=other) by Link Type (sex=yellow line, IDU=red, sex and IDU=blue) 17

18. Two views of sexual networks, where lines represent partnerships between people (circles)
a. Several unconnected sexual networks
b. Several people attend a group sex event and some have sex with unknown strangers under the influence of drugs and the excitement of the event
GSE

20. The epidemiologic impact of GSEs may be increased by the rapidity of partner change
Many participants have sex with several others in an hour or less.
Penises, fingers and/or sex toys move rapidly among different people’s vaginas and/or anuses.
These frequent sex acts may bruise or abrade vaginal, anal and penile tissue—which may facilitate transmission of HIV or other infectious agents.
Women’s risk may be increased by a widespread practice of men keeping the same condom on while having vaginal or anal sex with many partners.
Thus, studies and interventions that focus on “using condoms for every sex act” may miss sexual risk for the many community members—both drug users and others–who attend GSEs.

21. The epidemiologic implications of GSEs are further increased by:
Many people attend GSEs quite frequently—several times a week. For HIV, this means that GSEs can be (and have been) central to acute HIV infection outbreaks.
GSEs seem to mix people from different “risk groups.” People who use non-injected drugs, IDUs, non-users, straight men, gays, and women (including lesbians) may have sex together at these events
At GSEs in the Bushwick network study , approximately a quarter of attendees injected drugs, and about a fifth of men had sex with another man there (Friedman et al., 2008).
11% tested positive for HIV, 51% for HSV-2, and 10% for Chlamydia.
Thus, infections are present as are the uninfected.

22. Epidemiologic implications (cont)
Even a person with only one (uninfected) partner at a given event might be at high risk if any of several people to whose fluids he or she is directly or indirectly exposed was infectious.
This in turn puts her or his non-attending partners at risk.

23. Impact on epidemiologic interpretation of sexual network and genetic data
Epidemiologically, to the extent that indirect transmission takes place at GSEs, standard sexual network measures like in-degree, out-degree, centrality, and concurrency are greatly increased beyond what would be apparent from partnership data alone
We conceptualize this in a new term— “effective event concurrency” based on numbers of persons to whom one could transfer an infection during a given GSE.
An important research issue is to understand the factors that affect the probability of indirect transmission.
Parameters from such research, together with further data on group sex attendance and activities, should become part of mathematical models of transmission dynamics.

24. Interaction of infectiousness with networks opens paths for interventions

25. Role of AHI in Sustaining an Epidemic
Hollingsworth et al suggest that the relative infectiousness of AHI may vary during the history of a local epidemic:
Early in an epidemic: The number of infected individuals grows exponentially, most are in the early stages, most transmission is caused by individuals with AHI
Epidemic progresses: Proportion of transmission due to AHI decreases, while proportion due to asymptomatic or AIDS increases
Established epidemic: Transmissions due to AHI = 11%, Asymptomatic= 68%, AIDS = 21%
Finally, these points are from an oral presentation and a poster presented at CROI in Denver. The group from the Imperial College of London did some math. Modeling using data from Wawer et al, from the Rakai study on seroconversion among discordant couples, to estimate the impact of AHI in transmission in a community. They concluded that the impact of AHI would vary depending, among other things, on the stage of the epidemic in a given country. But in an established epidemic, like the one in the US, the new cases due to chronic HIV inf. Would by far surpass those due to AHI. The debate continues. More research is needed.
Modified from Hollingsworth, TD et al. Has the role of PHI being overstated? 13th CROI, Denver, Feb 06 as presented by Birkhead at Acute HIV Infection: A Multidisciplinary Symposium, NYC, June, 2006

26. Calculated probabilities of transmission per coital act
Pilcher CD, Tien HC, Eron JJ, et al. Brief but efficient: acuteHIV infection and the sexual transmission of HIV. J Infect Dis 2004; 189:1785–92.

27. HIV infection chains
The concepts below will be shown visually over the next few slides
Infection chains show how transmission events move through injection or sexual networks over time, and provide cues for locating and recruiting people with highly-infectious acute HIV infection (AHI).
Venues are locations where people take risks together (shooting galleries; group sex events; brothels) or meet new partners (bars, strolls)

28. Stages of HIV infection and CURRENT testing technologies
At Risk for HIV Infection
Negative on all tests
VL positive, antibody negative; duration ~ 2 months
Non-acute infection detected within about 6 months of HIV acquisition; duration ~ 4 months
Duration ~ 10 years before AIDS
Acute HIV Infection (AHI)
Recent HIV Infection (RHI)
Chronic HIV Infection

29. + Infection Chains + + - - - - - -
Original Infection - - - - -
Solid arrows are actual infection paths
Dotted lines are other IDU or sexual networks
Chronic HIV Infection
Acute HIV Infection
Recent HIV Infection
At Risk for HIV Infection 29

30. + Infection Chains + + + + - - - - - -
Source Infection - - - - -
Solid arrows are actual infection paths
Dotted lines are other IDU or sexual networks
Chronic HIV Infection
Acute HIV Infection
Recent HIV Infection
At Risk for HIV Infection 30

31. + Infection Chains + + + + + + - - - - -
Original Infection + - - - -
Solid arrows are actual infection paths
Dotted lines are other IDU or sexual networks
Chronic HIV Infection
Acute HIV Infection
Recent HIV Infection
At Risk for HIV Infection 31

32. Social network and venue tracing save time+ + + + + + + + - + + + - -
Original Infection - -
Venues (shooting gallery, crack house, sex party, singles bar) + + - + - +

33. Social research insight 2: social influence networks carry messages, norms, and other influence that can shape risk behaviors, the extent to which people seek medical treatment or counseling, and the extent to which people adhere to medications. 33

34. To or from an individual
This involves the activity of SOCIAL network ties, which can carry influence:
Within relationships
To or from an individual
Throughout a community or small group 34

35. Drug users and others provide risk reduction and medical adherence education and persuasion to others
We invented the word “intravention” to describe people in a community acting to help others to take actions that will protect them or improve their health.
When they urge others to behave in safer ways, or to adhere to the rules on when to take their medicines, this is how norms are enacted and maintained. 35

36. Alters’ urging shapes Ego’s external normative environment

37. Community organization in relationship to networks: Even the most “rooted” CBO has the problem of trying to reach and influence non-members
CBO members
CBO members

38. Schematic diagram of intravention as a context for intervention efforts (1) (NOTE: Behaviors here are INTERACTIVE behaviors) 38

39. Schematic diagram of intravention as a context for intervention efforts that can diffuse through networks (2) (Note: Behaviors here are INTERACTIVE behaviors) 39

40. Those who are at risk, whether sex workers, drug users, men who have sex with men, or other people, are those who REALLY do HIV prevention 40

41. They do this as individuals, small groups, and formal organizations
This has been shown, for example, in:
Gay men in many cities
New York IDUs knew of the new disease long before CDC—and began to reduce risk behaviors
International Network of People Who Use Drugs
Sex workers in Calcutta
Youth groups in many parts of Africa 41

42. Summary People are people, not just behaviors.
Sexual and drug injection networks are key to HIV spread
Social networks (including often sexual or drug relationships) are key to adherence and prevention.
The people are the ones whose action spreads or stops epidemics. They are often ahead of public health agencies in this.
Community, medical and counseling agencies must work with the people in this—or be of little help. 42

43. Additional thoughts
Overlapping sexual (and injection, where relevant) networks of heterosexuals, MSM, and WSW; and the implications of this for transmission patterns.
Venues like massage parlours in Vancouver, shooting galleries in SFHR and group sex parties in NNAHRAY are important.
Infection chain logic is important: It provides the basis, together with what we know about high rates of HIV infection during the first weeks of infection, network recruitment and social intervention models, to develop new HIV prevention approaches