KNOW YOUR EPIDEMIC AND
KNOWING YOUR RESPONSE
MSM AND THEIR NEEDS IN LOW- AND MIDDLE-INCOME
COUNTRIES
Stefan Baral MD MPH CCFP FRCPC
Johns Hopkins School of Public Health, USA
Overview

Know your epidemic

Epidemiology of HIV among MSM



Epidemic Scenarios of HIV among MSM
Burden and Risk Factors for HIV
Know your Response
Combination HIV Prevention Interventions for MSM
 Prevention Expenditures
 Modeling HIV epidemics according to response to MSM
 Human Rights, HIV, and MSM


Moving
g Forward
Introduction

Epidemiology




Responses



Ongoing epidemics among MSM in multiple LMIC
Newlyy identified epidemics
p
in previously
p
y unstudied areas
Resurgent epidemics among MSM in high income countries (HIC)
Inadequate coverage and access for prevention, treatment, and care
I d
Inadequate
““toolkit”
lki ” off prevention
i services
i for
f MSM
Human Rights


Multiple advances in LGBT rights awareness, community empowerment,
activism
Major “pushback” on MSM/LGBT and rights effort
Epidemiology
Systematic search: systematic methodology used to search databases
including PubMed, EMBASE, EBSCO, and the Cochrane Database of
Systematic Reviews through January 30th
30th, 2010
2010.
- MeSH terms:"Homosexual, Men” OR “Homosexual”; “Human
Immunodeficiency Virus”; "Primary Prevention"; "Secondary Prevention";
"Tertiary
Tertiary Prevention
Prevention"
Electronic global consultation: done in October, 2009 to obtain information
on epidemiology
epidemiology, rights contexts
contexts, and programming for MSM
- Used dedicated listserves in Asia, Africa, Latin America and the Caribbean,
and Eastern Europe, including the amfAR MSM Initiative, the MSMGF, UNDP,
and UNAIDS
UNAIDS.
Face to Face consultation (Bangkok, Feb. 2010) with key informants from
28 countries to obtain
b
country specific
f data
d
Search Protocol
Potentially relevant studies
identified and abstracts screened
for retrieval from international
conference searches (n=819)
Potentially relevant studies identified
and abstracts screened for retrieval
from literature Searches (n=1612)
Duplicates Studies
excluded (n=255)
Reports excluded based on
abstract
b t t due
d tto lack
l k off
quantitative data,
geographical context,
sample size, self-reported
HIV status.
((n=1434))
Abstracts excluded based on
abstract due to lack of
quantitative data, geographical
context, sample size, self-reported
HIV status.
(n=503)
Conference abstracts
retrieved for further
analysis (n=61)
(
)
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)
(n
111)
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
q studies retrieved from
US Census Bureau Database for
HIV/AIDS (n=2)
133 prevalence studies
from 130 unique reports:
data from 50 countries
Source: Beyrer, Baral, et al. Epidemiologic Reviews. 2010
Studies retrieved
that were
coordinated by
EuroHIV and
commissioned by
European Union
(n=16)
Epidemic Scenarios Algorithm
HIV prevalence in any high-risk
subgroup >5%
Unavailable Data
HIV prevalence ratio (MSM/gen pop)
Ratio ≥ 10
Ratio < 10
HIV prevalence ratio (IDU/gen pop)
Ratio < 10
HIV prevalence ratio (IDU/gen pop)
Ratio ≥ 10
Ratio ≥ 10
Ratio < 10
% population IDU
SCENARIO 3
< 1%
≥ 1%
% population MSM
< 10%
SCENARIO 1
SCENARIO 4
Source: Beyrer, Baral, et al. Epidemiologic Reviews. 2010
SCENARIO 2
Epidemic Scenarios for MSM
Evidence suggested four epidemic scenarios for LMIC MSM epidemics
-Scenario 5 will come from MENA region:
g
now largely
g y “unavailable data”
Source: Beyrer, Baral, et al. Epidemiologic Reviews. 2010
Scenario 1 - MSM risks are the predominant exposure
mode
d for
f HIV infection
i f i in
i the
h population
l i
40.00%
H
I
V
P
r
e
v
a
l
e
n
c
e
30.00%
20.00%
Aggregate MSM
Prevalence
10.00%
0.00%
General
Population
Prevalence
Scenario 1 - MSM risks are the predominant exposure
mode
d for
f HIV infection
i f i in
i the
h population
l i
Source: Beyrer, Baral, et al. Epidemiologic Reviews. 2010
Scenario 2 - MSM risks occur within established HIV
epidemics
id i d
driven
i
b
by injecting
i j i d
drug use (IDU)
50.00%
H
I
V
P
r
e
v
a
l
e
n
c
e
40.00%
Aggregate IDU
Prevalence
30.00%
Aggregate MSM
P
Prevalence
l
20.00%
General Population
Prevalence
10 00%
10.00%
0.00%
Poland
Serbia
Armenia
Georgia
Moldova
Russia
East Timor
Ukraine
Scenario 2 - MSM risks occur within established HIV
epidemics driven by injecting drug use (IDU)
Source: Beyrer, Baral, et al. Epidemiologic Reviews. 2010
Scenario 3 - MSM risks occur in the context of mature and
widespread
id
d HIV epidemics
id i among hheterosexuals
t
l
30 00%
30.00%
H
I
V
P
r
e
v
a
l
e
n
c
e
20.00%
Aggregate MSM
HIV Prevalence
General Population
Prevalence
10.00%
HIV Prevalence
among Men (15+)
0.00%
Namibia
Botswana
South
Africa
Kenya
Tanzania
Malawi
Nigeria
Sudan
Uganda
Scenario 3 - MSM risks occur in the context of mature and
widespread HIV epidemics among heterosexuals
Source: Beyrer, Baral, et al. Epidemiologic Reviews. 2010
SCENARIO 4 - MSM, heterosexual, and IDU
t
transmission
i i allll contribute
t ib t significantly
i ifi tl tto th
the HIV epidemic
id i
30%
H
I
V
P
r
e
v
a
l
e
n
c
e
20%
Aggregate MSM
Prevalence
10%
General
Population
Prevalence
0%
SCENARIO 4 - MSM, heterosexual, and IDU
t
transmission
i i allll contribute
t ib t significantly
i ifi tl tto th
the HIV epidemic
id i
Beyrer C, et al, Epidemiology Reviews, 2010.
EPIDEMIC SCENARIOS: Unavailable Data
• Kyrgyzstan
• Lebanon

Algeria
Azerbaijan
j
Djibouti
Iran
Iraq
Jordan
Kazakhstan

94 other
h C
Countries
i






• Libya
• Syria
• Tunisia
• West Bank and Gaza
Assessment of Data Quality

Disease burden among MSM in LMIC
 Data
is predominantly
 Prevalence
 Convenience



Samples
Tells us where epidemic was and not where it is going
May not be generalizable to general population of MSM
 Samples
S
l are among young MSM—so
MSM
lik l very conservative
likely
ti
estimates of disease burden
HIV Incidence has been characterized in cohort studies from:




Kenya
Peru
Brazil
Thailand
Ecological Model for HIV Risk in MSM
Level of Risks
Stage of Epidemic
Public Policy
C
Community
it
Network
Individual
Source: Baral and Beyrer, 2008
HIV among MSM in High Income Countries
Source: Sullivan, et al, 2009. Reemergence of the HIV Epidemic Among Men Who Have Sex With Men in North
America, Western Europe, and Australia, 1996–2005
Number of newly diagnosed HIV infections among men who have sex with men,
Hong
g Kong,
g, Singapore,
g p , Taiwan and Japan,
p , 2002 - 2007
Taiwan
Singapore
180
1200
160
1000
Number of cases
140
120
800
100
600
80
60
400
40
200
20
0
HK
SG
TW
JP
Japan
0
2002
2003
56
38
305
50
54
336
340
2004
67
94
503
449
2005
Year 96
Source: van Griensven, Baral, et al. 2009. Current Opinion in HIV/AIDS
101
584
514
2006
2007
118
108
743
571
168
145
1075
690
Number of cases
Hong Kong
Know Your Response

Systematic Review and Global Electronic Consultation


Dearth of data characterizing effective HIV prevention interventions
for MSM in Low and Middle Income Countries
Responding to multiple levels of HIV risk among MSM requires
combination prevention that is multilevel and multimodal
A modified Grade approach for HIV Public Health
I
Interventions
i
for
f MSM

Grading Evidence


3 Primary Parameters

Efficac Data
Efficacy

Biological Plausibility

Community Best Practices
6 Grades

Strong
g

Probable

Possible

Pending

Insufficient

Inappropriate
Combination HIV Prevention Interventions for MSM

Combination HIV Prevention Interventions (CHPI)

Behavioural Interventions



B
Biomedical
d l Interventions
I


Increasing condom and lubricant use during sex
Secondary target is partner reduction
Biomedical interventions aim to decrease transmission and acquisition
risk of but don
don’tt decrease prevalence of risk practices
Structural Interventions

Paucityy of efficacy
y and effectiveness data


Complex study designs required to evaluate these interventions
Challenge to implement and better evaluate
Prevention Expenditures for MSM

Concentrated Epidemics
 MSM
are one of predominant risk groups
 Epidemic Scenarios 1, 2, 4
 3.3%

of total expenditures supporting MSM
Generalized Epidemics
 Epidemic
Scenarios 3,
3 4
 Emerging evidence of risk among MSM
 0.1%
0 1%
of total expenditures
e pendit res ssupporting
pporting MSM
Source: Global HIV Prevention Working Group: Global HIV Prevention: The Access, Funding, and Leadership Gaps. 2009
Prevention Expenditures - Asia
24 000 000 00
24,000,000.00
U
S
D
E
q
u
i
v
a
l
e
n
t
21,000,000.00
18,000,000.00
15,000,000.00
Total HIV
Prevention
Expenditure
12,000,000.00
9,000,000.00
MSM Share
Sh
off
Prevention
Expenditure
6,000,000.00
3,000,000.00
0.00
Thailand
Vietnam
Cambodia
China
Source: USAID, HIV Expenditure on MSM Programming in the Asia-Pacific Region., 2006
Lao
Prevention Expenditures – Latin America
Peru
Source: Medición del gasto en sida: avances y retos de la respuesta latinoamericana al VIH. ONUSIDA. 2010
G
Goals
Model
M
G l
Goals
•
•
Mathematical model
• Inputs:
p
demographics;
g p
; sexual p
practices;; HIV/STI rates
• Outputs: HIV Prevalence and Incidence
g
We have refined the model byy including:
•
Divided “MSM intervention” into separate parameters:
1 outreach
1.
t
h with
ith condom
d promotion
ti and
d distribution
di t ib ti
2. community level behavioral interventions
3. Inclusion of ARV and new findings on efficacy of ARV in
couples (Donnell D, et al, Lancet 2010)
•
Expanded risk categorization for MSM to low
low, medium
medium, and high
risk, and MSM IDU
Source: Bollinger, L. How can we calculate the ‘‘E’’ in ‘‘CEA’’? AIDS 2008, Futures Institute
Scenario 1: Impact of MSM interventions
on allll new HIV infections
i f i
in
i PPeru
Null
Current
100% MSM interventions
Peru

M d li outcomes:
Modeling
t
combined
bi d prevention
ti for
f MSM

Markedly higher coverage of interventions for MSM will
b necessary to change
be
h
trajectory
j
off HIV epidemic
id i iin
Peru
Scenario 2: Impact of MSM interventions on
all
ll new HIV iinfections
f i
in
i Ukraine
Uk i
Null
Current
100% MSM interventions
Scenario 2: Impact of MSM and IDU interventions on all
new HIV infections
i f i
in
i Uk
Ukraine
i
Null
Current
100% MSM interventions 100% MSM + 60% IDU
Ukraine

Modeling outcomes: combined prevention for MSM


Higher
Hi
h coverage off iinterventions
t
ti
for
f MSM has
h iimpactt on
the HIV epidemic in Ukraine
C bi ti
Combinations
with
ith IDU iinterventions
t
ti
hhave th
the greatest
t t
impact on the HIV epidemic
Scenario 3: Impact of MSM interventions
on all new HIV infections in Kenya
Ken a
Null
Current
100% MSM interventions
Scenario 3: Impact of MSM interventions
on all new HIV infections in Kenya
12
125
120
115
Thousands
110
105
100
95
90
85
80
75
2008
2010
Null
Current
2012
100% MSM interventions
2014
Kenya

Modeling outcomes: combined prevention for MSM

EEven where
h
th
the HIV prevalence
l
is
i hi
highh and
d mature
t
among the heterosexual population, MSM specific
interventions positively impact the general population
Scenario 4: Impact of MSM interventions
on all new HIV infections in Thailand
Null
Current
100% MSM interventions
Scenario 4: Impact of MSM and IDU interventions
on allll new HIV infections
i f i
in
i Th
Thailand
il d
Null
Current
100% MSM interventions 100% MSM + 60% IDU
Thailand

Modeling outcomes: combined prevention for
MSM



Thai epidemic is stable overall
Higher levels of coverage for MSM with existing
interventions would lead to overall declines in HIV
p
epidemic
Substantial increases in coverage of IDU interventions
j impact
p on the HIV epidemic
p
have major
Modeling the impact of MSM interventions and ARVs
Key messages
K

MSM specific interventions impact new infections among MSM
and general population but must include access to ARVs for MSM
 Community-based behavioral interventions
 Distribution of condoms and lubricants


Where IDU plays a role in the epidemic, the greatest
p among
g the general
g
population
p p
can be seen when
impact
coverage of needle exchange and substitution therapy
are increased as well
Benefits
f is seen in 4/4 epidemic
d
scenarios
Human Rights Contexts

Methods



Health Impact Assessment used to assess criminalization of samesex practices a risk factor for HIV among MSM
Participatory methods in selected case study countries including
Results

Increased enforcement of laws criminalizing same sex practices
resulted
l d in



widespread fear and hiding
interfered with the ability to provide HIV prevention, care, and treatment
services limiting coverage
interfered with the ability of MSM to seek evidence-based HIV services
li iti uptake
limiting
t k
Conclusion and Moving Forward


HIV continues to disproportionately affect MSM in high and low
income settings with both individual and structural drivers of HIV
infection
To improve the health outcomes of MSM in low and middle income
settings, a comprehensive effort is needed

Know your Epidemic by

Generating high quality epidemiologic data to




characterize populations
demonstrate need
inform p
prevention strategies
g
Know your Response by


Adopting combination prevention strategies that address multiple levels of risk
A
Appropriately
i t l resourcing
i prevention
ti programs ffor MSM iin response tto attributable
tt ib t bl
fraction of HIV disease
Acknowledgements
Center for Public Health and Human Rights, Johns
Hopkins

Chris Beyrer MD, MPH

Frangiscos Sifakis, PhD, MPH

Andrea Wirtz, MHS

Damian Walker, PhD

Benjamin Johns, MHA
UNDP

Jeffrey O’Malley

Mandeep Dhaliwal, MD LLB
UNAIDS

WHO

The Futures Institute


John Stover, PhD
Lori Bollinger, PhD
Global HIV/AIDS Program of The World Bank

Robert Oelrichs, MD, PhD, MPH

Iris Semini, PhD

Laith Abu-Raddad, PhD

David Wilson, PhD
Michael Bartos, Phd
Ying Ro Lu, MD
Know your Epidemic and Response Session Team

Shiv Khan, OBE

Zoryan Kis

Andy
dy Sea
Sealee

Rob Carr

Joel Nana

Nyambura Njoroge

Maria Prins

Allison Talan