Measuring the impact of gender-focused interventions

frequentverseΠολεοδομικά Έργα

16 Νοε 2013 (πριν από 3 χρόνια και 11 μήνες)

77 εμφανίσεις

Measuring the impact of

gender
-
focused interventions



Julie Pulerwitz, ScD

Horizons Program/PATH

At Risk Due to Gender


Gender
-
related dynamics place both women
and men at risk of negative health outcomes
such as HIV/STI and violence


Women: power imbalances resulting in
inability to negotiate condom use and
mutual monogamy


Men: norms that encourage multiple
partners

Addressing Gender in Programs


Increasingly HIV/STI and violence
prevention interventions are seeking to
change gender dynamics in relationships,
posing challenges for evaluators



Few evaluations of programs, and fewer
use evaluated measures to do so

Major Challenges of

Intervention Research


Design Issues
-

How to attribute change


Many activities ongoing so difficult to attribute
change to program


Control/comparison group not possible when
program provides complete coverage



Instruments for Gender Focus


Many agree on importance but no consensus
on definitions and how to operationalize


Gender is complex and multi
-
faceted

Examples

Measuring Gender
-
related Dynamics and
Designing Studies of Program Impact

Development of SRPS

(Pulerwitz et al 2000, Sex Roles)


Sexual Relationship Power Scale (SRPS) to
measure power in sexual relationships



Focus groups with women and theoretical
framework to design items



50 initial items administered to 388 mainly Latina
women at reproductive health clinic in US



Key results with SRPS (23 items)


Highly reliable (alpha > .80)


Negatively associated with sexual and physical violence


Positively associated with education, condom use

Items in SRPS


23 item SRPS with 2 subscales



Relationship Control


My partner always wants to know where I am.


My partner does what he wants, even if I don’t want him
to.


My partner would get angry if I suggested condom use.



Decision
-
making Dominance


Who usually has more say about whether you have sex?


Who usually has more say about important decisions?

SRPS by % Consistent Condom Use*

8
2
13
0
2
4
6
8
10
12
14
16
SRPS
High Power
Moderate Power
Low Power
*p < 0.01
-

Mantel
-
Haenszel chi
-
square test for trend


Logistic Regression Model Predicting
Consistent Condom Use

(*p < 0.05)


OR

95% CI

Sexual Relationship

Power Scale



High

4.95*

(1.20, 20.42)

Moderate

4.12*

(1.02, 16.96)

Low

1.00




Controlling for socio
-
demographic and psycho
-
social variables.
Other significant variables: Peer support for condom use.


Egs of use of the SRPS


School
-
based young men & women in inner
-
city
NY
(Bauman & Silver 2003, Albert Einstein College of Medicine)



Adult men & women at STI clinic in Boston
(Pulerwitz
et al 1999, unpublished manuscript)


Gang
-
affiliated young women in Texas
(Excerpted from
chapter by Amaro, Dai & Arevalo 2003; Study led by Valdez, U of Texas)


ANC clients in South Africa
(Dunkle et al, Lancet May 2004)


Community
-
based youth in Thailand
(ongoing)


OVC in Zimbabwe
(ongoing UCSF study)

SRPS in “Stepping Stones” Study

(Jewkes et al 2002, Medical Research Council, South Africa)


HIV intervention study in South Africa with
Stepping Stones package
-

focused on
gender inequalities and communication



Surveys at baseline and 3 months post
intervention



Sample:

95 women and 107 men (pilot study),
94% response rate for women and 67% for
men

Findings from Stepping Stones


SRPS items relevant and reliable (alpha > .80)
for women, but less so for men



At baseline, less power associated with less
condom use, physical and sexual violence for
women



Post intervention, significant improvement in
relationship power for women, especially
among new condom users



Ongoing large intervention study

Lessons Learned about the SRPS


SRPS appears a useful measure for
relationship power...for adults and youth, for
different ethnicities and cultures.



Relationship Control subscale is sometimes
more useful than Decision
-
making
Dominance subscale.



SRPS is sometimes more useful for women
than men.

Development of GEM Scale

(Pulerwitz, Barker et al 2004, Horizons Report)


Gender Equitable Men (GEM) Scale to measure
attitudes toward gender norms



Qualitative research with young men in Brazil
(Barker 2000) and lit review to design items



35 initial items tested with 749 men aged 15
-
60 in
Rio de Janeiro
(PROMUNDO, NOOS, and TA from Horizons)



At baseline, GEM Scale (24 items) associated
with partner violence, education, and
contraception use

Gender Norms Scale Domains


Home & child
-
care



Sexual relationship



Health & disease prevention



Violence



Homophobia & relations with
other men

Items in GEM Scale


24 item GEM Scale with 2 subscales (alpha
> .80)



‘Traditional’ Norms


Men are always ready to have sex.


There are times when a woman deserves to be
beaten.


I would be outraged if my partner asked me to use
a condom.


‘Egalitarian’ Norms


A man and a woman should decide together what type
of contraceptive to use.

Egs of use of GEM Scale


Intervention study with young men in India
-

CORO and Horizons/PROMUNDO


National survey with adults in Zambia
-

HCP


Intervention study with young men in Mexico
-

Salud y Genero


Intervention study with adult men in South
Africa
-

MAP and Frontiers/Population Council

‘Program H’ Intervention Study in Brazil

(Instituto PROMUNDO, Horizons, and partners)

Main objectives:


Determine the impact of different combination of
gender
-
focused activities (group education,
education plus community
-
based ‘lifestyle’ social
marketing campaign) on attitudes toward gender
norms, and HIV/STI and violence risk among young
men


Data collection:


Pre and post surveys with 3 groups (2 intervention, 1
control) followed over 1 year (n= 780 at baseline;
over 75% response rate)


In
-
depth interviews with sub
-
sample and partners,
for triangulation and validation from partners

Study Design



Pre
-
test

N = 258

Post
-
test 1

N = 230

Post
-
test 2

N = 217

6 months

6 months

Pre
-
test

N = 250

Post
-
test 1

N = 217

Post
-
test 2

N = 172

6 months

6 months

INTERVENTION 1

INTERVENTION 2

CONTROL



Pre
-
test

N = 272

Post
-
test 1

N = 180

Delayed

intervention

6 months

3 months

Change in GEM Items*

31
52
38
62
21
37
23
43
26
43
29
44
0
20
40
60
80
Baseline (Maré)
6 months (Maré)
1 year (Maré)
Baseline (Bangu)
6 months (Bangu)
1 year (Bangu)
Men need sex more
than women do.

Changing diapers, g
iving
the kids a bath, and
feeding the kids are the
mother's responsibility.

*p < .05 for individual items and full Scale; No significant change in control site

Association Between ‘Traditional’ Norms
and STI Symptoms Over Time



At one year, young men that become
more supportive of equitable norms
report fewer STI symptoms:



4.6* times more likely to report no STI
symptoms in Intervention 1 site


8.3* times more likely to report no STI
symptoms in Intervention 2 site

*p < 0.05


logistic regression for correlated data;

controlling for age, family income, and education

Lessons Learned about Research
with the GEM Scale


Findings suggest GEM Scale a
sensitive and cross
-
culturally relevant
tool



Findings suggest change in attitudes
towards gender norms possible via
interventions, as well as subsequent
changes in HIV/STI risk

‘SDSI’ Intervention Study in Nicaragua

(Puntos de Encuentro, Horizons, PATH, CIDS)


Communication for
social change
program to empower
youth, promote
gender equity, and
reduce violence and
HIV/STI risk


Intervention Activities


Weekly national edutainment
telenovela (Sexto Sentido)


Daily call
-
in radio show to discuss
show themes


Community
-
based activities


Coordination across organizational
networks


Cast visits to schools


Youth training camps



IEC materials


Research Methodology


Longitudinal sample in three representative cities


Esteli, Leon, Juigalpa


Track change of individual people, not just group
overall



Triangulation of results


Pre, mid
-
term and post surveys with 4567 male and
female youth aged 13
-
24 at baseline (over 80%
response rate at mid
-
term), in
-
depth interviews and
focus group discussions, Participatory Action Research



Multivariate and multi
-
level analyses

Emerging Results


At baseline, HIV/STI risk behaviors substantial


40% of sexually active had recent occasional partner


69% did not use condoms consistently with occasional
partners



At mid
-
term:


Exposure to activities associated with better outcomes


More support for gender equity (GEM items) over time


Communication with others about HIV prevention and
sexual behavior increased


Increase in condom use among some groups



Longitudinal analysis ongoing

Conclusions: the Role of
Gender Dynamics


Empirical support that both relationship
power and inequitable gender norms are
key factors in HIV/STI and violence risk.



Evidence that interventions can influence
both relationship power and attitudes
towards gender norms, and subsequent
HIV risk behaviors.


Conclusions:

Evaluation Issues



Evidence of measures and designs
that capture impact of BCC and other
interventions focused on gender
dynamics.



Measuring impact challenging, but
attempts ongoing and growing.