20th International AIDS Conference - Melbourne, Australia


WEPDC01 Counting the Hard to Count
  Oral Poster Discussion Session : Track C
Venue: Room 109-110
Time: 23.07.2014, 13:00 - 14:00
Chair: Basia Zaba, United Kingdom

13:25

Powerpoint
HIV sero-prevalence in Trans population in Mexico
J. Izazola, Switzerland

13:00
WEPDC0101
Abstract
A comparison between respondent-driven sampling and time-location sampling among men who have sex with men in Shenzhen, China
R. Cai1, J. Zhao2, L. Chen2, J.H. Richardus1, S. De Vlas1
1Erasmus MC, University Medical Center Rotterdam, Public Health, Rotterdam, Netherlands, 2Shenzhen Center for Disease Control and Prevention, Shenzhen, China

13:05
WEPDC0102
Abstract
Measuring and accounting for outcome-correlated recruitment and geographic recruitment bias in a respondent-driven sample of people who inject drugs in Tijuana, Mexico
A. Rudolph1, T. Gaines2, R. Lozada3, A. Vera2, K. Brouwer2
1Pacific Institute for Research and Evaluation, Calverton, United States, 2University of California San Diego, San Diego, United States, 3Pro-COMUSIDA, Tijuana Baja California, United States

13:10
WEPDC0103
Abstract
Powerpoint
HIV among older adults in Zimbabwe: ageing with HIV or seroconverting after age 50?
J. Negin1, C. Nyamukapa2,3, J. Eaton3, N. Schur3, A. Takaruza2, P. Mason2,4, S. Gregson2,3
1University of Sydney, Sydney School of Public Health, Sydney, Australia, 2Biomedical Research and Training Institute, Harare, Zimbabwe, 3Imperial College School of Public Health, Department of Infectious Disease Epidemiology, London, United Kingdom, 4University of Zimbabwe, College of Health Sciences, Harare, Zimbabwe

13:15
WEPDC0104
Abstract
Powerpoint
Mapping and size estimation of male-to-female transgender community in 17 states of India: first level findings at the country level
T. Subramanian1, S. Mehendale1, N. Dhingra2, V. Chakrapani3, V. Selvaraj1, E. Noronha4, A. Narang4
1National Institute of Epidemiology, Chennai, India, 2Goverment of India, Department of AIDS Control, Targetted Interventions, New Delhi, India, 3Centre for Sexuality and Health Research and Policy (C-SHaRP), Chennai, India, 4United Nations Development Programme, HIV and Development Unit, New Delhi, India

13:20
WEPDC0105
Abstract
Powerpoint
If you aren't counted, you don't count: estimating the number of female sex workers in Mandalay and Yangon, Myanmar
S.T. Thein1, T. Aung1, H.M. Kyaw1, A. Lancelot1, W. Mcfarland2
1Population Services International/Myanmar, Strategic Information, Yangon, Myanmar, 2San Francisco Department of Public Health, San Francisco, United States

13:25
WEPDC0106
Abstract
HIV sero-prevalence in trans population in 7 urban areas in Mexico
H. Sucilla Pérez1, J.P. Gutierrez Reyes2, A.R. Shiba Matsumoto3, J.A. Izazola Licea4
1CENSIDA, Operational Research, Mexico DF, Mexico, 2National Institute of Public Health, CIEE, Cuernavaca, Mexico, 3National Institute of Public Health, Cuernavaca, Mexico, 4CENSIDA (at the Time of the Study and Analysis), Mexico DF, Mexico

13:30
WEPDC0107
Moderated discussion

Powerpoints presentations
HIV sero-prevalence in Trans population in Mexico - Jose-Antonio Izazola

HIV among older adults in Zimbabwe: ageing with HIV or seroconverting after age 50? - Simon Gregson

Mapping and size estimation of male-to-female transgender community in 17 states of India: first level findings at the country level - Thilakavathi Subramanian

If you aren't counted, you don't count: estimating the number of female sex workers in Mandalay and Yangon, Myanmar - Si Thu Thein



Rapporteur reports

Track C report by Maria Veras


WEPDC01: Counting the Hard to count 

R. Chai compared socio demographic, behavioral characteristics, in addition to HIV status and utilization of HIV services, among MSM, recruited from 2 different methods, respondent-driven sampling and time-location sampling, in Shenzhan, China.  The data from the two samples were adjusted according to the methods requirement.  Men recruited through TLS were older, more likely to be from outside the city, more likely to self-identify as bisexual, less likely to be homosexual/gay, more likely to use HIV services. HIV status was similar. In Shenzhen, TLS was more effective to reach the more marginalized segments of MSM. It was also pointed out that because HIV-related services are available at gay venues in Shenzhen, RDS may be useful to inform further allocation of prevention actions. During discussion, author’s explanation for the differences was attributed to the migrant status. Migrants would be more likely to attend gay identified venues, the locals being more cautious on doing that. It was also emphasized that differences among TLS and RDS are very context dependent.

A Rudolph assessed geographic recruitment bias and outcome-correlated recruitment, among 1048 IDU, using RDS, in Tijuana, Mexico. Participantes were interviewed and a geographic and network clustering of active syphilis were simulated. Unadjusted prevalence was different by gender, but difference disappeared after adjustment. To live, buy or use drugs in the same region and have connection of one or two separation degrees with another active case were correlated. Active cases more likely to live close. From discussion session, audience questioned if some RDS requirements were observed, questioning at the beginning and at the end of the interview about the size of the social network. It was not clear if the author proceeded accordingly.    

S. Gregson presented data from a population based cohort study, carried out in Zimbabwe, from 1998-2011. This analysis focused on HIV infection among those with 50 years or more. The aim was to respond if they were living longer with HIV or seroconverting after the 50s. Participants were interviewed and taken blood sample for HIV testing. The cohort included 58099.9 person- years, 11,883 individuals - average of approximately 5 years of follow-up. Overall incidence rate was 1.22/100 person years; 0.69/100 person-years in 50+.  56,3% of them seroconverted after 50 years, 31,2% in the last 3 years. The recommendation is that the older can no longer be neglected in HIV prevention strategies. Discussion included lost of follow-up, around 50%.

S. Subramarnian presented from a size estimation and mapping of the male-to-female trans population in 17 states of India. The study was conducted between Dec-2012 and Sept-2013. Participatory approaches using key community informants (with Delphi consensus) for site identification and nomination methods for size estimates were followed by a sub-sample of site-visits for verification.  The study estimated 5821 sites and a population size of 62137 Trans male-to-female. 45% of the sites were public and intestate migration was estimated in 10%.  They acknowledge the possibility of duplication but considered the information, although not completed accurate will be very helpful to planning.  

S. Thein presented the results of a population size estimation of FSW in Yangon and Mandalay, Myanmar. Different methods were used: Multiplier methods (Unique object and Unique event), Services (a drop center, a female STI clinic and an HIV testing site) and the Wisdom of Crowd method. The results were then compared with prior official estimates. Multiplier methods were consistent and fell within regional estimates. They estimated Medians of ~5,000 in Yangon and ~3,300 in Mandalay. The Wisdom of the crowd produced different for the 2 cities, close to multiplier methods in Yangon, and very different, lower, in Mandalay. Author concluded that various and different methods must be used. Discussion allowed him to clarify how final estimate was calculated: Median of the 5 multipliers. He also explained that the existing figures were just nomination by stakeholders, not using scientific method.

H Sucilla-Perez conducted a survey in Mexico among transgender, transsexual and travesties, over 18 years. HIV rapid test was performed. Participantes were recruited in social and working places, in seven urban areas in 3 states of Mexico. HIV prevalence: crude 15,83%, adjusted 15,21%. Among the HIV-infected persons 39% knew their serostatus, they were older (33,9 years versus 27) with lower level of education.  The prevalence is higher than between MSM (12,2%).

The discussion session posed questions for all the presenters regarding the balance between RDS and TLS to reach the hidden population. Arguments were made on the need to consider context, if there are connections, RDS better, for more visible population TLS might be preferable. 




Track C report by Mark Stoove


Title: A comparison between respondent-driven sampling and time-location sampling among men who have sex with men in Shenzhen, China

Author: Rui Cai

Overview:

R Cai compared HIV-related characteristics of MSM recruited through respondent-driven (n=621) and time-location (recruited through bars, saunas, gyms and public spaces) (n=533) sampling in Shenzhen, China. Based on RDS-adjusted estimates, RDS-recruited MSM were more likely to be younger, originate from locally Shenzhen or other parts of Guangdong province, to self-identify as gay (less likely to self-identify as bi-sexual), and report exclusively receptive anal sex in the past six months. RDS-recruited MSM were less likely to report multiple anal sex partners in the past six months and less likely to have recently accessed HIV-related services. There was no significant difference in HIV prevalence by recruitment strategy. The authors concluded that, in Shenzhen, time-location sampling is more effective to reach the more marginalized segments of MSM.

 

Title: Measuring and accounting for outcome-correlated recruitment and geographic recruitment bias in a respondent-driven sample of people who inject drugs in Tijuana, Mexico

Author: Abby Rudolph

Overview:

In the context of controversies in relation to the validity of respondent-driven sampling (RDS) to produce random samples, A Rudoph presented data from 1048 RDS-recruited people who inject drugs (PWID) in Tijuana, Mexico, to examine the presence of non-random recruitment bias. Using active syphilis infection as an outcome, geographic recruitment bias and outcome-correlated recruitment were assessed by comparing gender-specific estimates from RDSAT with predicted probabilities estimated via logistic regression accounting for network/spatial correlates of syphilis and the shared social environment among those in the same recruitment chain. Syphilis prevalence varied greatly between participant clusters characterised by geography and first and second degree social network connections, and varied from overall unadjusted and adjusted RDS-estimates. The authors concluded that because active syphilis is not randomly distributed, preferential recruitment of peers with consistent outcome status or social environments could lead to inaccurate prevalence estimates if clustering is not accounted for.

 

Title: HIV among older adults in Zimbabwe: ageing with HIV or seroconverting after age 50?

Author: Simon Gregson

Overview:

Using data from a prospective cohort in Eastern Zimbabwe, S Gregson reported on HIV-related outcomes among older PLWH across five survey waves between 1998 and 2011, aiming to determine the proportion of adults aged 50 years or over who seroconverted at aged 50 years or over. HIV incidence across the cohort was 1.22 per 100PY and did not vary by gender. Among those aged 50 years or over HIV incidence was 0.69 per 100 PY and significantly higher among males (0.99/100PY) compared to females (0.54/100PY). Among participants aged 50 years and over who were living with HIV at last interview, 56% seroconverted when aged ≥50 years rather than aging with HIV. The authors concluded that older individuals should not be excluded from targeted HIV prevention programmes and services.

 

Title: Mapping and size estimation of male-to-female transgender community in 17 states of India: first level findings at the country level

Author: Thilakavathi Subramanian

Overview:

T Subramanian reported on a 2012/13 mapping and population size estimation of the male-to-female transgender (TG) community in 17 states of India using participatory approaches with key community informants and site-visit verification. In the target states, there were an estimated 62,157 TGs, with most (61%) residing in the states Maharashtra, Uttar Pradesh, Odisha, West Bengal and Andhra Pradesh. TGs commonly (47%) reported being involved in a gharānā (a system of social organization linking musicians or dancers) and most (62%) were involved in commercial sex work (62%) and resided in urban areas (71%).

Title: If you aren't counted, you don't count: estimating the number of female sex workers in Mandalay and Yangon, Myanmar

Author: Si Thu Thein

Overview:

S T Thein reported on sample size estimation for female sex workers in Mandalay and Yangon, Myanmar. Multiplier methods - unique object, unique event and three service multipliers (individual visits to drop-in centres, STI treatment, and HIV testing) were used, and compared with official and other published estimates. The median estimate of the female sex worker population in Yangon and Mandalay were 5000 and 3000, respectively. These estimates are higher than prior official and published estimates and data supported the use of median estimates generated from multiple approaches to sample size estimation.

 

Title: HIV sero-prevalence in trans population in 7 urban areas in Mexico

Author: Héctor Sucilla Pérez

Overview:

H S Pérez presented data estimating the HIV sero-prevalence in transgender populations in 7 urban areas of Mexico using cross sectional survey and HIV serology data. Recruited via gay community gathering places and locations frequented by sex workers (73% were recruited in public spaces) in urban centres across three states and 7 metro areas, HIV prevalence was estimated at 15.8%. Only 39% of serologically positive participants self-reported being HIV positive (61% undiagnosed HIV). HIV prevalence among transgender populations was approximately equivalent with previous estimates of HIV prevalence among MSM in Mexico.




Track C report by Maria Veras


1. The first study compared socio demographic, behavioral characteristics, in addition to HIV status and utilization of HIV services, among MSM, recruited from 2 different methods, respondent-driven sampling and time-location sampling, in Shenzhan, China.  The data from the two samples were adjusted according to the methods requirement.  Men recruited through TLS were older, more likely to be from outside the city, more likely to self-identify as bisexual, less likely to be homosexual/gay, more likely to use HIV services. HIV status was similar. In Shenzhen, TLS was more effective to reach the more marginalized segments of MSM. It was also pointed out that because HIV-related services are available at gay venues in Shenzhen, RDS may be useful to inform further allocation of prevention actions. During discussion, author’s explanation for the differences was attributed to the migrant status. Migrants would be more likely to attend gay identified venues, the locals being more cautious on doing that. It was also emphasized that differences among TLS and RDS are very context dependent.

2. This study assessed geographic recruitment bias and outcome-correlated recruitment, among 1048 IDU, using RDS, in Tijuana, Mexico. Participantes were interviewed and a geographic and network clustering of active syphilis were simulated. Unadjusted prevalence was different by gender, but difference disappeared after adjustment. To live, buy or use drugs in the same region and have connection of one or two separation degrees with another active case were correlated. Active cases more likely to live close. From discussion session, audience questioned if some RDS requirements were observed, questioning at the beginning and at the end of the interview about the size of the social network. It was not clear if the author proceeded accordingly.    

3. 3rd author, presented data from a population based cohort study, carried out in Zimbabwe, from 1998-2011. This analysis focused on HIV infection among those with 50 years or more. The aim was to respond if they were living longer with HIV or seroconverting after the 50s. Participants were interviewed and taken blood sample for HIV testing. The cohort included 58099.9 person- years, 11,883 individuals - average of approximately 5 years of follow-up. Overall incidence rate was 1.22/100 person years; 0.69/100 person-years in 50+.  56,3% of them seroconverted after 50 years, 31,2% in the last 3 years. The recommendation is that the older can no longer be neglected in HIV prevention strategies. Discussion included lost of follow-up, around 50%.

 4. Data were presented from a size estimation and mapping of the male-to-female trans population in 17 states of India. The study was conducted between Dec-2012 and Sept-2013. Participatory approaches using key community informants (with Delphi consensus) for site identification and nomination methods for size estimates were followed by a sub-sample of site-visits for verification.  The study estimated 5821 sites and a population size of 62137 Trans male-to-female. 45% of the sites were public and intestate migration was estimated in 10%.  They acknowledge the possibility of duplication but considered the information, although not completed accurate will be very helpful to planning.  

5. The 5th author presented the results of a population size estimation of FSW in Yangon and Mandalay, Myanmar. Different methods were used: Multiplier methods (Unique object and Unique event), Services (a drop center, a female STI clinic and an HIV testing site) and the Wisdom of Crowd method. The results were then compared with prior official estimates. Multiplier methods were consistent and fell within regional estimates. They estimated Medians of ~5,000 in Yangon and ~3,300 in Mandalay. The Wisdom of the crowd produced different for the 2 cities, close to multiplier methods in Yangon, and very different, lower, in Mandalay. Author concluded that various and different methods must be used. Discussion allowed him to clarify how final estimate was calculated: Median of the 5 multipliers. He also explained that the existing figures were just nomination by stakeholders, not using scientific method.

6. The last presentation was  a survey conducted in Mexico among transgender, transsexual and travesties, over 18 years. HIV rapid test was performed. Participantes were recruited in social and working places, in seven urban areas in 3 states of Mexico. HIV prevalence: crude 15,83%, adjusted 15,21%. Among the HIV-infected persons 39% knew their serostatus, they were older (33,9 years versus 27) with lower level of education.  The prevalence is higher than between MSM (12,2%).

 The discussion session posed questions for all the presenters regarding the balance between RDS and TLS to reach the hidden population. Arguments were made on the need to consider context, if there are connections, RDS better, for more visible population TLS might be preferable.

 

 




   

    The organizers reserve the right to amend the programme.