BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Harnessing social networks for HIV surveillance - Simon Frost\, De
 partment of Veterinary Medicine\, University of Cambridge
DTSTART:20100511T133000Z
DTEND:20100511T143000Z
UID:TALK23382@talks.cam.ac.uk
CONTACT:Michael Sweeting
DESCRIPTION:The transmission dynamics of HIV are strongly affected by the 
 structure\nof the underlying risk network. However\, collecting data on se
 xual or\ninjection drug use networks is notoriously difficult and costly. 
 I will\ndiscuss how social network data may be harnessed to guide HIV\nsur
 veillance\, and potentially prevention\, by identifying subpopulations\nwi
 th elevated HIV prevalence and/or risk behaviours. I will consider two\nsa
 mpling methods\, respondent-driven sampling (RDS) and time-location\nsampl
 ing (TLS)\, which are widely used to sample at-risk populations. In\nthese
  approaches\, the underlying social network is typically considered\nas a 
 nuisance factor\, and the primary objective is to obtain\ndesign-based unb
 iased estimates of quantities such as HIV prevalence. In\ncontrast\, I wil
 l treat the social network as the primary outcome of\ninterest\, rather th
 an a nuisance.\n\nI will show how in RDS\, individuals infected with HIV o
 r other sexually\ntransmitted infections (STIs) often recruit other infect
 ed individuals.\nUsing Markov tree models\, I will demonstrate that may be
  difficult to\ndistinguish between infected individuals intentionally recr
 uiting other\ninfected individuals and the more likely scenario that indiv
 iduals tend\nto recruit within specific social subpopulations with differe
 nt HIV/STI\nprevalences. In a time-location sample of men who have sex wit
 h men\n(MSM) recruited from bars\, I will demonstrate how specific venues\
 nstructure the population in terms of quantities such as age and HIV\nprev
 alence. To address the severe problems of confounding\, I will show\nhow c
 ausal inference approaches such as marginal structural models can\nbe used
  to consider whether there are venue-specific factors in HIV\nrisk\, beyon
 d easily measured individual factors. Finally\, I will\nconsider how even 
 very crude measures of social networks\, such as the\nnumber of friends on
 e has in a specific subpopulation\, may be used as a\nproxy for belonging 
 to different social subpopulations with different\nrisk behaviors.
LOCATION:Large Seminar Room\, 1st Floor\, Institute of Public Health\, Uni
 versity Forvie Site\, Robinson Way\, Cambridge
END:VEVENT
END:VCALENDAR