This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike License. Your use of this
material constitutes acceptance of that license and the conditions of use of materials on this site.
Copyright 2007, The Johns Hopkins University and Ronald Gray. All rights reserved. Use of these materials
permitted only in accordance with license rights granted. Materials provided “AS IS”; no representations or
warranties provided. User assumes all responsibility for use, and all liability related thereto, and must independently
review all materials for accuracy and efficacy. May contain materials owned by others. User is responsible for
obtaining permissions for use from third parties as needed.
Epidemiology of Domestic Violence
and sexual coercion
Issues covered
1. Magnitude of the problem globally
2. Prevalence and risk factors
3. Health consequences
•
•
•
Physical
Reproductive and sexual health
Birth outcomes/child survival
4. Methodological/ethical research issues
Many forms of violence to
women: beginning early….
•
•
•
•
•
•
selective abortion
female infanticide
neglect, malnutrition
sexual abuse
child prostitution
forced early marriage
Continuing throughout the life
cycle…
• physical partner abuse
• dowry related violence
• non-consensual sex (within marriage or
outside)
• forced prostitution & trafficking
• sexual harassment at work and at school
• violence against women in conflict situations
Domestic (or intimate partner)
violence
• Most common form of violence experienced by
women
• WHO definition: “ the range of sexually,
psychologically and physically coercive acts used
against adult and adolescent women by current or
former male intimate partners.”
• Also female-to-male, female-to-female violence
Prevalence and Risk Factors
Physical assault on women by an intimate male partner:
selected population-based studies, 1982-1999
East er n M edit er r an a
in
Egypt , 19 9 5 - 19 9 6
I sr ael, 19 9 7
West Ban k an d Gaza St r p
i , 19 9 4
16
32
34
52
Eur ope
Net her a
l n ds, 19 8 6
Nor way, 19 8 9
Republci of M oldova, 19 9 7
Swit zer a
l n d, 19 9 4 - 19 9 6
T ur key, 19 9 8
Un ti ed Kin gdom, 19 9 3
21
18
7
14
6
21
58
12
Asia an d West er n Pacif ci
Aust r ala
i , 19 9 6 ( n at o
i n al)
Ban gladesh, 19 9 2
Cambodia, 19 9 6
I n dia, 19 9 9
Papua New Guin ea, 19 8 4
Phip
il pin es, 19 9 8
Republci of Kor ea, 19 8 9
T haia
l n d, 19 9 4
3
30
8
19
14
47
16
40
56
26
38
20
Nor t h Amer ica
Can ada, 19 9 3 ( n at o
i n al)
Un ti ed St at es, 19 9 5 - 19 9 6
3
l as t 12 mont hs
29
1. 3
22
Lat n
i Amer ci a an d t he Car r b
i ean
An t g
i ua, 19 9 0
Bar bados, 19 9 0
Bolvi a
i , 19 9 8
Chie
l , 19 9 7
Colombia, 19 9 5
M exico, 19 9 6
Nicar agua, 19 9 7 ( n at o
i n al)
Par aguay, 19 9 5 - 19 9 6 ( n at o
i n al)
Per u, 19 9 7
Puer t o Rico, 19 9 5 - 19 9 6
Ur uguay, 19 9 7
I n c ur r ent r el at i ons hi p
30
30
17
23
E v er
19
15
27
28
12
10
10
Af r ica
Et hiopia, 19 9 5
Ken ia, 19 8 4 - 19 8 7
Niger a
i , 19 9 3
Sout h Af r ci a ( n at o
i n al) , 19 9 8
Zm
i babwe, 19 9 6
31
13
45
42
31
13
17
0
10
20
Source: World Report on Violence and Health. WHO, Geneva, 2002.
30
40
50
60
WHO Multi-country Study on Women’s Health and Domestic Violence against Women, 2002.
Frequency of lifetime and recent male
against female domestic violence:
Rakai, Uganda, 2000-01
Source: Koenig et al. Bulletin of the WHO, 2003; 81(1):53-60.
Specific types of lifetime and recent male
against female domestic violence:
Rakai, Uganda, 2000-01
Source: Koenig MA et al. Bulletin of the WHO, 2003; 81(1):53-60.
WHO Multi-country Study on Women’s Health and
Domestic Violence against Women, 2002.
Sexual coercion and violence
Sexual victimization by an intimate male partner:
selected population-based studies, 1989-2000
B r a z i l , 2000
2. 8
10. 1
8
C a na da , 1993*
C hi l e , 1997
9. 1
Fi nl a nd, 1997-1998
2. 5
J a pa n, 2000
Indone s i a , 1999-2000
1. 3
5. 9
6. 2
e ve r
13
Me xi c o, 1996*
23
15
Ni c a r a gua , 1997
17. 7
P e r u, 2000
P ue r t o R i c o, 1993-1996*
pa s t 12 mont hs
22. 5
7. 1
5. 7
7. 5
Swe de n, 1991
11. 6
Swi t z e r l a nd, 1994-1995
T ha i l a nd, 2000
T ur ke y, 1998*
29. 9
17. 1
51. 9
Uni t e d Ki ngdom, 1993*
23
6
Uni t e d St a t e s , 1995-1996
We s t B a nk/ Ga z a St r i p,
0. 2
7. 7
27
25
Zi mba bwe , 1996
0
10
20
30
40
Source: World Report on Violence and Health. WHO, Geneva, 2002.
50
60
Source: World Report on Violence and Health. WHO, Geneva, 2002.
Source: World Report on Violence and Health. WHO, Geneva, 2002.
Source: World Report on Violence and Health. WHO, Geneva, 2002.
Risk factors
Risk factors
• Individual level
–
–
–
–
–
Education (--)
Socioeconomic status (-)
Life cycle (age, marital duration, parity) (-)
Extended family structure (-)
Intergenerational transmission of violence (++)
Risk factors
• Individual level risk behaviors
– Alcohol/substance abuse (++)
– HIV risk/risk perceptions (++)
– Infidelity (perception, no. of partners) (+)
• Women’s status
– Savings and credit (-) or (+)
– Women’s autonomy/control of resources (-) or
(+)
Odds ratio for domestic violence
<1 year
Risk Factors for Physical Violence: Rakai,
Uganda
5
4.62***
3.72***
4
3
2
1.84***
1.62***
1.00
1.00
1
0
Never
Sometimes Frequently
Male partner’s consumption of alcohol
Not at
Somewhat Very likely
all/Unlikely
Perception of male partner’s HIV risk
Risk factors: contextual
• Contextual factors
– Neighborhood crime, house ownership levels
(U.S.) (+)
– Community-level women’s status (Bangladesh)
(-) or (+)
– Levels of violent crime (India) (+)
– Norms regarding domestic violence (India)
(++)
Figure 2: Attitudes of men and women toward
domestic violence: Rakai District, Uganda, 2000-2001
Beating of wife/female partner justified for:
100
90
80
70
60
50
40
30
20
10
0
90
87
70
60
42
37 37
th
ea
bo
ve
ty
An
y
of
in
f id
eli
em
ng
ho
us
e
ho
ld
fa
m
ily
du
t
m
ies
io
n
ne
g
lec
tin
g
co
nt
ra
ce
pt
us
in
g
re
fu
s in
g
se
x
16
men
women
27
22
di
so
be
yi
28
50
Health Consequences of Domestic
Violence
Domestic violence associated with adverse:
•
•
•
•
women’s physical health
women’s reproductive health
women’s mental health**
birth outcomes/child survival
Physical
consequences
•
•
•
•
•
•
•
•
Abdominal/thoracic injuries
Bruises, lacerations, abrasions
Disability
Fractures
Gastrointestinal disorders
Ocular damage
Reduced physical functioning
Death
Injuries resulting from domestic violence in
past 12 months: Rakai,Uganda, 2000-01
Source: Koenig MA et al. Bulletin of the WHO, 2003; 81(1):53-60.
Domestic violence and reproductive
health
•
•
•
•
•
Gynecological morbidity/ STIs
HIV
Contraceptive use
Unintended pregnancy
Abortion
Gynecological morbidity associated with
intimate partner violence
•
•
•
•
Gynecological disorders
Infertility
Pelvic pain
Sexual dysfunction
Domestic violence and gynecological morbidity
Pelvic pain
STD history
% reporting
Odds Ratios
48%
21%
2.25*
1.54
1.00
Physical abuse
No abuse
Norway (Schei, et al. 1990)
Physical/sexual
abuse
Physical abuse
No Abuse
U.S. (Martin, et al. 1998)
Violence and HIV/AIDS
Recent studies in the U.S. and Africa suggest a complex dynamic:
– Violence as a risk factor for HIV
– Violence as a consequence of disclosure of HIV status
•
•
Epidemics overlap in SS Africa
Many similar risk factors
•
Direct link:
– Coerced sex with infected partner
•
Æ
HIV
Indirect link
– Violence limits women’s ability to negotiate safe sex with or by partner
(monogamy, condom use, non-risky sexual practices)
– Women’s early experience with sexual abuse leads to later high risk
sexual behavior (~ U.S.-based evidence)
Associations between physical/
sexual violence and HIV risks
2.5
2.10
***
Odds Ratio
2
1.5
**
1.48
1.30
1.00
1.00
1
0.5
0
Broad violence
Limited/ no
violence
South Africa
Dunkle, et al. (2004)
** p<.01
***p<.001
Coercion at
first sex
Coercion not at
first sex but
subsequently
No coercion
Uganda
Koenig, et al. (unpublished)
Violence and Non-use of Contraception
• Violence or threat of abuse makes birth control
negotiation difficult
• Women may be deterred from independent
contraceptive behavior
• Barrier methods may not be feasible
Evidence on domestic violence and
non-use of contraception
• Several U.S. studies show small but
significant differences in current use
• Rakai, Uganda (class reading)
• North India (Stephenson, et al.2005)
• Qualitative evidence (Ghana, Uganda,
Bolivia)
Evidence on domestic violence
and non-use of contraception
Outcome
Adjusted OR
(95% CI)
Current use of contraception
0.46 (0.24-0.87)
Condom use at last sex
0.27 (0.13-0.57)
Consistent condom use < 6 mos
0.21 (0.08-0.53)
Adjusted for age at first sex, education, religion and current marital status
Source: Koenig MA et al. (2004)
Unintended pregnancy and violence:
U.S. and international studies
• Women whose pregnancy was unintended had 2-4
times increased risk of physical violence versus
women whose pregnancy was intended
- both population- and clinic-based studies
• Unintended pregnancy results from partner’s
violence through partner’s control of
contraception, unprotected forced sex, and/or
coercing a woman to have a child
- qualitative research
Source: Gazmararian et al. Maternal and Child Health Journal, 2000; 4(2):79-84
Violence and abortion
• Little research: two studies among women
seeking abortion services
– found physical violence related to likelihood of
reporting previous pregnancy termination or
miscarriage
Sources: Amaro H et al. Am J Public Health. 1990 May;80(5):575-9.
Webster et al. Am J Obstet Gynecol. 1996 Feb;174(2):760-7.
Domestic Violence and Birth
Outcomes
• Direct effects
– Physical trauma
– Resulting chronic maternal condition
• Indirect effects
–
–
–
–
Elevated stress levels
Delay in seeking prenatal care
Poor nutrition
Substance abuse
Violence during pregnancy
US:
• 1%-20% - any time during pregnancy
- majority of the studies report <10%
Developing countries:
- Significantly higher, poor birth outcomes also much
higher (4%-28% in WHO study)
- 10% -18% during most recent pregnancy in India
study
• More common than many key risk factors for poor birth
outcomes (pre-eclampsia, gestational diabetes, placenta
previa)
Domestic violence and pre-natal care
Odds ratio
1.8
Delay in entering
pre-natal care
1.0
Physical violence: <12 months
No violence: <12 months
Dietz et al. (1997): U.S.
Domestic violence and pregnancy outcomes
LBW
Infant
death
Fetal
death
3.0
Fetal
death
2.02**
1.5*
1.47**
ea
te
n
rb
N
ev
e
er
b
Ev
Ev
1.00
N
o
ab
us
e
du
r in
g
er
b
eg
n
pr
pr
g
ur
in
ed
A
bu
s
ea
te
n
an
cy
y
eg
na
nc
le
n
vi
o
N
o
1.00
ea
te
n
1.0
ce
1.00
V
io
le
nc
e
ce
le
n
vi
o
N
o
V
io
le
nc
e
1.00
Infant
death
ea
te
n
LBW
rb
4.0
N
ev
e
Complications
during
pregnancy
Mexico (Valdez- Santiago, et al. 1996)
U.S. (McFairlane, et al. 1996)
India (Jejeebhoy, 1998)
Logistic Regression Results for Maternal Care
Utilization by Violence Status
OR
Crude
95% CI
OR
Adjusted
95% CI
Antenatal care 0.48
0.36-0.63
0.66
0.49-0.88
Tetanus toxoid 0.57
0.44-0.74
0.81
0.63-1.04
Delivery care
0.49
0.30-0.79
1.05
0.64-1.73
Postpartum
care
0.41
0.26-0.66
0.61
0.37-0.99
Adjusted for: women’s age, parity, women’s education, husband’s education,
socio-economic status, caste, gender and wantedness of the index child.
Hazards Regression Results for Early child
Mortality by Violence Status
HR
Crude
95% CI
Adjusted
95% CI
HR
Perinatal
2.05
1.20-3.49
2.59
1.48-4.51
Neonatal
1.91
1.13-3.26
2.37
1.36-4.15
Post-neonatal
1.03
0.53-2.02
0.997 0.50-1.98
Child
1.38
0.17-11.54
0.998 0.07-14.37
Adjusted for: women’s age, parity, women’s education, husband’s education,
socio-economic status, caste, gender and wantedness of the index child.
Methodological and Ethical Issues in
Domestic Violence Research
Methodological issues
• Definition of violence
• Study population
• Facilitating the disclosure of violence
Definition of violence
• Who defines violence: the researcher or the
respondent?
• Types and severity of violence included in the
questions
• Lifetime vs. recent (<12 mos) experience
• From current partner vs. any male partner? Male
partner vs. any family member?
• Inclusion of questions on sexual violence
• Sexual abuse/rape prior to/outside of primary
relationship
Defining the study population
• Age range
• Marital status
• Sex of respondent
Male vs. Female Reports of Domestic
Violence: Rakai, Uganda 2000-2001
30.4
% reporting
24.6
24.1
18.5
19.9
8.1
Physical violence: ever Physical violence: <12
months
Male
Female
Coercive sex: Ever
Factors that affect disclosure of
violence
• How questions are phrased
• Number of opportunities to disclose
• Context in which questions asked
• Interviewer characteristics and skill
• Stigma attached to issue (may vary across settings)
Ethical guidelines for violence research
(WHO recommendations)
•
Measures included to protect safety of respondents and interviewers
– Total privacy/confidentiality
– Interview only one member per household
– Presented as a woman’s health survey
– Referral mechanisms for services/counseling
•
Special training and emotional follow-up for interviewers
•
Incorporation into multi-purpose surveys only when
ethical/methodological requirements can be met
Interventions to prevent violence
Structural intervention to prevent IPV and HIV in
South Africa
Pronyk et al Lancet 2006;368:1973
• Community randomized trial in Limpopo province,
South Africa
• Pair-matched communities randomized to either:
– Structural intervention of microfinance and
“learning/action” education (N = 4)
– Control communities, no intervention for 3 years,
then provided with the intervention
• Endpoints:
– Intimate partner violence past 12 months
– Unprotected sex with a non-spouse past 12
months
– HIV incidence
Exposure groups
• Cohort 1: Women who applied for a loan in
intervention villages, age and sex matched
with controls from control villages (n = 860).
Follow up 2 years
• Cohort 2: Woman aged 15-35 co-resident
with a loan acceptor in intervention arm
compared with co-resident with a control arm
subject (n = 1835). Follow up 2 yrs
• Cohort 3: 14-35 year old randomly selected
men and women in intervention arm,
compared similarly selected with control arm
residents (n = 3881). Follow up 3 years
Interventions
• Poverty focused microfinance
– Field workers identify poorest households,
form groups of 5 and offer microfinance
loans, repaid over 10-20 weeks
– Meet every 2 weeks for business
assessment, 10 session “life, gender and
HIV” training program
– Community mobilization with leaders
Effects in cohort 1. Matched loan
applicants vs controls
Selected outcomes
Intervent Control
%
%
Household assets > 2000
58
49
rand
Communication about
sexual matters
86
55
Intimate partner violence
< 12 months
6
12
Adj RR
(955CI)
1.15
(1.04-1.28)
1.58
(1.21-2.07)
0.45
(0.23-0.91)
Are women who accept loans self-selected for greater autonomy,
or does loan acceptance increase a woman’s status
and deter violence?
Effects in cohort 2. Matched co-resident
women in intervention vs controls
Selected outcomes
Communication about
sexual matters
Intervent Control
%
%
66
50
> 1 sexual partner past 12
mths
18
16
Unprotected sex with nonspousal partner
48
48
Household loan acceptance does not affect
behaviors in co-residents
Adj RR
(955CI)
1.32
(0.90-1.95)
1.16
(0.85-3.32)
1.02
(0.85-1.23)
Effects in cohort 3. Matched males and
females in intervention vs controls
Selected outcomes
Intervent Control
%
%
> 1 sexual partner past 12
15
19
mths
Unprotected sex with nonspousal partner
43
48
HIV incidence
11
11
Adj RR
(955CI)
0.64
(0.19-2.16)
0.89
(0.66-1.19)
1.06
(0.66-1.69)
Intervention did not affect population-level behaviors
Or HIV incidence
Authors conclusions
• Combined microfincance and training
interventions can reduce intimatepartner violence in program participants
• Social and economic development
interventions have the potential to alter
risk environments for HIV and intimate
partner violence
Caveats
• This was a community randomized trial analyzed
as an individual-level trial
• The main effects in loan applicants vs controls is
an “as-treated”, not intent-to-treat analysis (i.e.,
selected on most compliant intervention arm
participants.)
• No “spill over” effects to household co-residents
• No impact on HIV incidence