PE0394
An epidemic scenario based approach to assessing global HIV epidemics among
Men who have Sex with Men (MSM) in Low and Middle Income Countries (LMIC)
Chris Beyrer1, Stefan D. Baral1, Frangiscos. Sifakis1, Andrea L. Wirtz1, Benjamin Johns1, Iris Semini2, Robert Oelrichs2, Damian Walker1
1Johns
2The
Hopkins Bloomberg School of Public Health, Center for Public Health and Human Rights, Epidemiology, Baltimore, United States,
World Bank, Global HIV/AIDS Program, Washington D.C., United States
Background:
• HIV epidemic dynamics vary across LMIC, particularly among MSM
• Differences exist in HIV prevalence and incidence, attributable
fraction by risk population, government and non-governmental HIV
prevention, treatment, and care programs, and policy responses.
• Advanced categorization of HIV epidemics among MSM may assist
societies to develop and implement HIV prevention programs and
interventions more fitting to HIV epidemiology and social risk
Algorithm to develop the Epidemic
Scenarios
HIV prevalence in any
high-risk subgroup >5%
Unavailable
Data
HIV prevalence ratio (MSM/gen pop)
Ratio ≥ 10
Methods:
• Systematic review of 133 HIV prevalence studies from 50 LMIC
• Published articles and conference abstracts after January 1, 2000
•Studies required to have a sample size of greater than 50 MSM.
• Review of1679 unpublished reports describing populations or programs
for MSM in LMIC, and data on HIV prevalence among adult MSM, the
general population, sex with women, and those who inject drugs.
• Data were analyzed to identify trends across LMICs.
• Data were used to develop an algorithmic approach to categorize these
epidemics.
ep
de cs
Ratio < 10
HIV prevalence ratio
(IDU/gen pop)
Ratio < 10
HIV prevalence ratio
(IDU/gen pop)
Ratio ≥ 10
Ratio ≥ 10
% population IDU
SCENARIO 3
< 1%
≥ 1%
% population MSM
< 10%
Systematic Review Protocol
Potentially relevant studies
identified and abstracts screened
for retrieval from international
conference searches (n=819)
Duplicates excluded (n=255)
Abstracts excluded based on
abstract
abst
act due to lack
ac o
of
quantitative data, geographical
context, sample size, self-reported
HIV status. (n=503)
Conference abstracts retrieved
for further analysis (n=61)
Abstracts excluded based on
inability to find background data on
specific country HIV prevalence,
inability to find further information
on statistical methods (n=8)
Unique
U
i
studies
t di retrieved
ti
d ffrom
US Census Bureau Database
for HIV/AIDS (n=2)
Potentially relevant studies
identified and abstracts
screened for retrieval from
literature Searches (n=1612)
Reports excluded based on
abstract due to lack of
quantitative data, geographical
context, sample size, selfreported
t d HIV status.
t t
(n=1434)
SCENARIO 1
Studies retrieved that
were coordinated by
y
EuroHIV and
commissioned by
European Union (n=16)
SCENARIO 2
SCENARIO 4
•The algorithmic approach is a dynamic process
• Countries are assigned to a scenario based on current data available;
UNAIDS and UNGASS indicators
•Takes into account prevalence of MSM and IDU in the population, and
HIV prevalence of the general population, MSM, and IDU
• As the epidemic changes or as data become available, scenario
assignment may change
Full texts retrieved for further
analysis with QUOSA (n=178)
Reports excluded based
on lack of HIV prevalence
data, inability to calculate
country population HIV
prevalence (n=111)
Ratio < 10
Results
Five Epidemic Scenarios can be used to classify HIV among MSM:
1) Where MSM are the predominant contributor to HIV;
2) Where MSM transmission occurs within epidemics driven by IDU
3) Where MSM transmission occurs within established heterosexual
epidemics
4) Where sexual and parenteral transmission both contribute significantly
5) Where sociologic data suggest the presence of MSM populations
but epidemiologic data are currently unavailable
133 prevalence studies from
130 unique reports: data
from 50 countries
Conclusions:
•These scenarios allow for more precise understanding of HIV epidemics among MSM and their relationship with the general population but do not
replace the use of generalized or concentrated classifications
•Mapping these scenarios helps predict effective research and intervention strategies by epidemic scenario, facilitates the development of models to
predict costs associated with these strategies, and costs associated with lack of effective and comprehensive prevention and care services for MSM