Intimate partner violence is associated with
poorer maternal mental health and
breastfeeding practices in Bangladesh
Lan Mai Tran
*, Phuong Hong Nguyen
, Ruchira Tabassum Naved
Alive &Thrive, FHI360,18 Ly Thuong Kiet Street, Hanoi, Vietnam
Poverty, Health and Nutrition Division, International Food Policy Research Institute, Washington, DC, USA
Health System and Population Studies Division, ICDDR, GPO Box 128, Dhaka 1000, Bangladesh
*Corresponding author. Alive & Thrive, FHI360, 18 Ly Thuong Kiet Street, Hanoi, Vietnam. E-mail:
Accepted on 12 August 2020
Exposure to intimate partner violence (IPV) can have profound adverse consequences on maternal
and child health. This study aimed to: (1) identify factors associated with IPV during pregnancy and
postpartum in Bangladesh; and (2) assess the associations between IPV and maternal mental
health and breastfeeding practices. We used data from a cross-sectional survey of 2000 mothers
with children <6 months in four districts in Bangladesh. We applied multivariable logistic regres-
sion models to examine factors associated with IPV and structural equation modelling to assess
the inter-relationships between IPV, maternal common mental disorders (CMD, measured by Self-
reporting Questionnaire 7) and breastfeeding practices. Overall, 49.7% of mothers experienced
violence during the last 12 months and 28% of mothers had high levels of CMD. Only 54% of
women reported early initiation of breastfeeding and 64% reported exclusive breastfeeding.
Women were more likely to experience IPV if living in food-insecure households, being of low
socio-economic status, having low autonomy or experiencing inequality in education compared
with husbands (OR ranged from 1.6 to 2.8). Women exposed to IPV were 2–2.3 times more likely to
suffer from high levels of CMD and 28–34% less likely to breastfeed their babies exclusively. The in-
direct path (the indirect effects of IPV on breastfeeding through CMD) through maternal CMD
accounted for 14% of the relationship between IPV on breastfeeding practice. In conclusion, IPV is
pervasive in Bangladesh and is linked to increased risks of CMD and poor breastfeeding practices.
Integrating effective interventions to mitigate IPV, along with routine maternal and child health
services and involving men in counselling services, could help both to reduce exposure to IPV
among women and to contribute to better health outcomes for women and children.
Keywords: Intimate partner violence, determinants, mental health, breastfeeding, Bangladesh
Intimate partner violence (IPV)—defined as physical, sexual and
emotional abuse and controlling behaviours by an intimate part-
ner—is a common form of violence against women. The lifetime
prevalence of physical and/or sexual IPV against women is estimated
at 30% globally, ranging from 23% in high-income countries to
38% in the low- and middle-income countries (LMICs) (WHO,
2013a). Bangladesh has one of the highest rates of IPV (Garcia-
Moreno et al., 2005), with two-thirds of ever-married women
reporting at least one form of IPV during their lifetime and half
reporting any form of IPV during the past 12 months (Bangladesh
Bureau of Statistics, 2016). IPV during pregnancy and after
CThe Author(s) 2020. Published by Oxford University Press in association with The London School of Hygiene and Tropical Medicine.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits
unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. i19
Health Policy and Planning, 35, 2020, i19–i29
childbirth is also common, with >30% of women experiencing
physical IPV and 60% emotional IPV (Islam et al., 2017b).
Drivers or triggers of IPV have been identified at multiple levels:
individual, relationship, community and societal (WHO, 2012).
Bangladesh has shared similar determinants of IPV to many other
LMICs (Hindin et al., 2008). At the individual level, women with
low levels of education, who married as children, who are in em-
ployment and whose husbands have extramarital relationships were
significantly more likely to experience IPV (Naved et al., 2017b).
Gambling and substance abuse, increasingly widespread, are more
recent factors that are driving IPV (Naved et al., 2017b). Men’s per-
ceptions of being disadvantaged relative to women also contribute
to IPV (Naved et al., 2017b). At the household level, poverty and
food insecurity are the key factors associated with IPV (VanderEnde
et al., 2015). Other household factors associated with IPV include
women’s decision-making power, the quality of the marital relation-
ship, the number of partners a man has, the number of children in
the household (Naved et al., 2017b) and inequality in education be-
tween women and their husband (Rapp et al., 2012). In terms of
community-level factors, inequitable community gender norms have
been shown to increase the likelihood of IPV (Yount et al., 2018).
Risk factors for IPV during pregnancy are often similar to those for
IPV in general (Jewkes, 2002;WHO, 2011). However, some other
factors are likely to be more important during pregnancy in the
Bangladeshi context, such as unwanted or unplanned pregnancies
(Hindin et al., 2008) and a history of abuse (Naved et al., 2017b).
No clear evidence exists about whether IPV itself increases or
decreases during pregnancy (Jasinski, 2004).
IPV has profound adverse consequences on women’s mental
health and childcaring behaviours. Women who had experienced
IPV had higher levels of emotional distress (Ziaei et al., 2016) and
depressive symptoms (Devries et al., 2013a,Parvin et al., 2018), had
lower mental health and social functioning scores (Dillon et al.,
2013) and were more likely to report suicidal ideation and attempts
(Devries et al., 2011,2013a). In violent domestic environments, the
quality of mothering and the ability of both parents to cope with
children’s needs are impaired (Kong and Lee, 2004;Victora et al.,
2010). Evidence on the influences of IPV on children’s feedings are
still scarce and inconsistent. A population-based study based on
cross-sectional Demographic and Health Surveys from 51 LMICs
showed that all three types of IPV were independently associated
with a decreased likelihood of early initiation of breastfeeding
(EIBF) (Caleyachetty et al., 2019). In Bangladesh, women who
experienced IPV after childbirth were also less likely to exclusively
breastfed their infants (Islam et al., 2017a). However, other studies
did not find an association between IPV and early breastfeeding
practices (Zureick-Brown et al., 2015;Mezzavilla et al., 2018).
Due to widespread violence against women and its importance
for maternal and child health, research has increasingly focused on
this topic, but studies have typically examined each aspect separate-
ly. A large body of literature has described the prevalence of IPV
among women of reproductive age (Devries et al., 2013b) and dur-
ing pregnancy (Devries et al., 2010;Halim et al., 2018), and its asso-
ciated risk factors (WHO, 2012). Another body of literature
provides evidence on the influences of IPV on maternal physical and
psychological health (Devries et al., 2011,2013a; WHO, 2013a).
And yet another set of studies has focused on associations between
IPV and childcare behaviours, including breastfeeding practices
(Mezzavilla et al., 2018;Caleyachetty et al., 2019). Our study
attempts to bring together an understanding of the linkages across
these domains in the same context and using the same dataset. We
aimed to: (1) examine the magnitude of different forms of IPV dur-
ing pregnancy and postpartum periods; (2) identify factors associ-
ated with IPV; and (3) assess the influences of IPV on maternal
mental health and breastfeeding practices.
Data for this paper were obtained from a baseline household survey
of women who had delivered a baby in the six months prior to the
survey. Data were collected in 2015 as part of an evaluation to test
the feasibility and impacts of integrating intensified maternal nutri-
tion interventions into the existing maternal, newborn and child
health programme platform in Bangladesh (Nguyen et al., 2017b).
The survey was carried out in 20 rural sub-districts (upazilas) from
four districts (Mymensingh, Rangpur, Kurigram and Lalmonirhat).
Within each sub-district, five unions and two villages within each
union were randomly selected to yield a total of 200 villages (each
had an average size of 250 households). Within each village, a
household census was conducted to create a list of mothers with liv-
ing infants <6 months of age. A total of 2000 mother–infant pairs
were selected for the survey using systematic sampling, beginning
with a random seed start point to yield the desired sample size per
•Intimate partner violence (IPV) has profound adverse consequences on maternal and child health.
•Women with low autonomy, living in poor households, experiencing food insecurity and having inequality in education (compared
with their spouses) were more likely to experience IPV.
•Women who experienced IPV were more likely to be depressed and less likely to practise exclusive breastfeeding.
•Health systems play a crucial role in a multisectoral response to violence against women. Integrating effective approaches to screen
for and address IPV along with reproductive, maternal and child health services could be a double duty action to address the high
prevalence of IPV, and improve maternal mental health and suboptimal breastfeeding practices.
•As one of the countries with the largest burden of IPV and also the highest prevalence of maternal and child undernutrition, integrat-
ing screening and support interventions into the health system and across other programmes in Bangladesh has the potential to reduce
IPV, and thus also help achieve several Sustainable Development Goals related to empowering women, promoting gender equality
and improving health and nutrition outcomes.
i20 Health Policy and Planning, 2020, Vol. 35, Suppl. 1
Data were collected via face-to-face interviews using a structured
questionnaire by researchers from Data Analysis and Technical
Assistance Limited (DATA), an experienced and well-qualified sur-
vey firm in Bangladesh. Survey enumerators were trained by senior
researchers using lectures, role play, mock interviews and practice in
a classroom setting, and then field-tested the questionnaire; revisions
were made to the questionnaire based on the results of field testing.
The questionnaire was prepared initially in English and translated
into Bangla, then back translated into English to double check for
accuracy and consistency. All interviews were conducted one-to-one
in a safe and private room at respondents’ homes, and confidential-
ity was ensured for all women.
Ethical approval was obtained from the Institutional Review
Boards of the BRAC University in Bangladesh and the International
Food Policy Research Institute, USA. Written informed consent was
obtained from all women 18 years of age. For women <18 years of
age, we obtained their assent and the permission of their guardians,
i.e. their parents or husbands, to participate in the study.
The survey collected information on IPV experienced by recently
delivered women using questions adapted from the World Health
Organization (WHO) Multi-Country Study on Women’s Health and
Domestic Violence (Garcia-Moreno et al., 2005). The questions
were validated for use in several countries including Bangladesh
(Garcia-Moreno et al., 2005), showing good internal consistency be-
tween each specific measurement (Cronbach’s alpha ranged from
0.75 to 0.90 for different types of IPV) (Islam et al., 2017b).
The respondents were asked about their experience of a range of
controlling behaviours and were asked to report whether they had
experienced physical, emotional or sexual IPV by their intimate
partner (Supplementary Table S1). Seven items were used to meas-
ure physical IPV: (1) pushing, shaking or throwing something at her;
(2) slapping; (3) twisting her arm or pulling her hair; (4) punching
or hitting with a fist or something harmful; (5) kicking or dragging
or physically assaulting her; (6) choking or burning; (7) threatening
or attacking with a knife, gun or any other weapon. Emotional IPV
was measured by at least one affirmative response to questions ask-
ing whether or not the respondent’s husband had insulted her or
made her feel bad about herself; humiliated her in front of others;
done things to scare or intimidate her on purpose; or threatened to
hurt her or someone she cares about. A response was coded as sex-
ual violence by an intimate partner if women reported having been
physically forced to have sexual intercourse; having intercourse out
of fear; or being forced to perform sexual acts that she found
degrading or humiliating. The lifetime prevalence of partner vio-
lence is defined as the proportion of women who report having
experienced one or more acts of violence at any point in their lives.
Current prevalence is the proportion of women reporting that at
least one act of domestic violence had taken place during the
12 months prior to the interview. Some women had been exposed to
all three types of violence (emotional, physical and sexual).
Maternal common mental disorders (CMD), our outcome of
interest, were measured using the 20-item Self-reporting
Questionnaire (SRQ-20) (WHO, 1994). This tool was validated and
adapted for screening mental disorders in developing countries
(Harpham et al., 2005) including Bangladesh (Khan and Flora,
2017). Each item is scored with 0 or 1, depending on responses
related to CMD over the past 30 days. Cronbach’s alpha for the
scale was 0.9 in our sample, indicating good internal consistency.
The scores are added to generate an overall SRQ-20 scale, in which
higher scores denote higher levels of maternal CMD (WHO, 1994).
A cut-off of 7 was considered to ascertain maternal CMD, as recom-
mended by some studies (Stewart et al., 2008;Medhin et al., 2010)
including previous studies in Bangladesh (Khan et al., 2008;Ziaei
et al. 2016;Khan and Flora, 2017).
Breastfeeding practices were assessed using survey questions to
construct standard WHO indicators (Akman et al., 2008) for EIBF
(defined as the proportion of children born in the last 24months
who were put to the breast within 1 h of birth) and exclusive breast-
feeding (EBF) (defined as the proportion of infants 0–5 months of
age who had been fed exclusively with only breast milk in the previ-
ous 24 h). Optimal breastfeeding practice was defined as children
who had been breastfed immediately after birth and had been exclu-
sively breastfed in the 24 h preceding the survey.
Potential factors associated with IPV
We were not able to find any comprehensive frameworks for deter-
minants of IPV or influences of IPV on maternal mental health and
breastfeeding practices. Therefore, we developed a conceptual
framework based on a literature review to guide our analyses
(Figure 1), considering maternal, household levels and the potential
influence of intra-household relationships. IPV may influence breast-
feeding practices directly or indirectly through maternal CMD.
Although we show directionality using arrows in Figure 1 for illus-
trative purposes, we acknowledge the interplay of factors presented
in the framework, and the potential bidirectional nature of the links
Maternal and child factors
Maternal characteristics that were examined include age, education
(categorized as illiterate, elementary school, middle school and high
school or higher) and parity. Age at first marriage was recorded, and
we created a variable for early married (1, 0) using a cut-off of
<19 years of age. Details on mode of delivery (caesarean section),
child age and sex were used as controlled variables in the models of
Women’s autonomy index was measured based on a set of 27
items that covered four different dimensions: women’s economic
decision-making autonomy, familial healthcare and family planning
decision-making autonomy, women’s freedom of movement auton-
omy and women’s attitudes towards gender. Details of these items
were presented in Supplementary Table S2. For each of the questions
in the first three dimensions, the responses were coded as 1 (re-
spondent decided alone or jointly) or 0 (respondent did not make de-
cision). The index of women’s attitudes towards gender dimension
was measured by asking women whether they agreed or disagreed
with statements on women’s roles. Each statement was given a score
of 1 (agree) or 0 (disagree). The sums were used as an overall auton-
omy index (range: 2–27), and this index was divided into tertiles to
obtain high, medium and low categories.
Two key factors were considered at the household level: household
socio-economic status (SES) and food security. An index for house-
hold SES was constructed using a principal components analysis of
variables on housing conditions and asset holdings, and the first
component derived from component scores was used to divide the
SES score into tertiles (Vyas and Kumaranayake, 2006;Gwatkin
et al., 2007). Household food security was measured and calculated
using FANTA/USAID’s Household Food Insecurity Access Scale
(Coates et al., 2007), which provides information related to the
Health Policy and Planning, 2020, Vol. 35, Suppl. 1 i21
household’s experience of food insecurity in the 30days preceding
the survey, including anxiety and uncertainty about access, and in-
sufficient quality and quantity of intake. The households were cate-
gorized into two groups—food secure and food insecure (which
included the mildly, moderately and severely food-insecure groups).
Mild, moderate and severe food insecurity were calculated based on
the responses to more severe conditions and/or the frequency of
experiencing the conditions following steps described in the manual
by Coates et al. (2007).
Educational discrepancy between a woman and her husband was
used as a proxy for potential relationship inequality. Spousal educa-
tion gap was calculated by deducting the wife’s education from the
husband’s education. The couples with an equal educational level
were furthermore divided into two subgroups: couples with second-
ary or higher education and with no spousal education gaps were
termed as ’no gap, high education’, and, equivalently, couples with
no gap and primary or no education were termed as ’no gap, low
education’. Thus, this variable had four categories: (1) no gap, low
education, (2) wives had higher education than their husbands, (3)
husbands had higher education than their wives and (4) no gap, high
We first used descriptive analyses to report the characteristics of the
study sample as well as experiences of different forms of IPV (con-
trolling behaviour, emotional, physical and sexual violence) among
recently delivered women in their lifetime and in the last 12 months.
We then used multivariable logistic regression analyses to examine
factors associated with different forms of IPV and to assess the asso-
ciation between IPV and maternal CMD, as well as between IPV
and breastfeeding practices. Finally, we used structural equation
models, considering all the potential variables in our conceptual
framework, to examine the complex inter-relationship between
these factors and breastfeeding. We presented odds ratios (OR) with
95% confidence intervals (95% CIs) for the multivariable regression
models and coefficients in the figure of the path analysis. All models
were adjusted for geographical clustering at upazila level using a ro-
bust sandwich estimator of the standard errors. Statistical signifi-
cance was defined as Pvalue <0.05. All analysis was done using
Stata version 15.1 software (StataCorp, 2019).
Characteristics of the study sample
The mean age of mothers was 24 years (range 13–44) (Table 1);
>10% of the women were illiterate and only 15% had been edu-
cated at high school or a higher level. Two-thirds of the women
reported their first marriage before 19 years old. The average of par-
ity was two children, and 22% of women had had a C-section dur-
ing their last delivery. Nearly half of women lived in food-insecure
The overall autonomy index was 15.1 64.1 (range 2–25)
(Table 1). Most women had low levels of asset ownership (13–53%)
and purchasing power (40–58%) (Supplementary Table S2).
Although a substantial percentage of women decided on family
healthcare and movement, 52% of women had supportive attitudes
towards wife beating. Regarding educational gaps, there were 47%
of couples in which the wife had a higher educational level than their
husband. The proportion of spouses who were equally poorly
educated was high at 15%, and only 8% of couples were equally
Experiences of IPV
Around three-quarters of women reported one or more controlling
behaviour by their husband (Figure 2). The most common acts of
controlling behaviour were the husband’s expectation that the wife
Figure 1 Conceptual framework for determinants of IPV and inﬂuences of IPV on maternal mental health and child feeding practices
i22 Health Policy and Planning, 2020, Vol. 35, Suppl. 1
ask permission before seeking heathcare (60%), insistence on know-
ing where the wife was at all times (29.2%), and trying to keep the
wife from seeing her friends and getting angry if she spoke to an-
other man (20.3%) (Supplementary Table S1).
The lifetime prevalence of domestic violence was high: 58.3%
for emotional violence, 49.3% for physical violence and 30.3% for
sexual violence; >70% of women had experienced at least one type
of IPV and one-fifth of women had experienced all three types of
violence in their lifetimes. For the 12 months prior to the survey,
half of the women reported experiencing at least one type of IPV
and 11% reported experiencing all three types of IPV (Figure 2).
Regarding different acts of violence, 37% of women were insulted
or made to feel bad about themselves, a quarter were slapped and
18% were physically forced to have sexual intercourse
(Supplementary Table S1).
Determinants of IPV
The four key factors found to be associated with IPV were living in
food-insecure households, being of low socio-economic status, hav-
ing low autonomy and there being inequality in education between
the spouses (Table 2). Household food insecurity was significantly
associated with higher odds of being faced with controlling behav-
iour and all types of IPV (OR ranged from 1.6 to 2.3). Women living
in the lowest SES households were more likely to experience physic-
al violence and all types of IPV than those in the highest SES house-
holds (OR ranged from 1.8 to 2.0). Compared to women with high
autonomy, those with low autonomy were 2.5 times (95% CI: 1.6–
4.1) more likely to experience controlling behaviour and 1.6 times
(95% CI: 1.0–2.5) more likely to experience physical IPV.
Inequality in education between husbands and wives increased the
likelihood of domestic violence towards women. Couples with
higher educated wives showed a significantly higher likelihood of
experiencing physical violence than equally highly-educated spouses
(OR: 1.88, 95% CI: 1.17, 3.02). Any educational inequality be-
tween husbands and wives (OR: 2.29 and 2.25), and an equally low
education (OR: 2.79, 95% CI: 1.18–6.55), meant a significantly
higher chance of all types of violence compared with equally highly-
educated couples (Table 2).
Table 1 Sample characteristics
Mental stress scale (range: 0–20) 2000 3.00 (1.00, 7.00)
Common mental disorder 7 559 27.95
EIBF 1071 53.55
EBF 1346 67.30
Age (years) 2000 24.47 65.51
13–19 404 20.20
20–29 1207 60.35
30–44 389 19.45
Illiterate 232 11.60
Elementary school 703 35.15
Middle school 758 37.90
High school or higher 307 15.35
Parity 2000 1.98 60.96
1 772 38.60
2 681 34.05
3–4 547 27.35
Early marriage (age <19) 1708 67.15
Caesarean section 445 22.25
Overall autonomy index 1983 15.09 64.12
Low 734 37.01
Medium 750 37.82
High 499 25.16
Household food insecurity 891 44.55
SES (tertile) 655 33.33
Education gap (range 11 to 14) 1657 0.86 63.27
Wife higher educated 786 47.44
Husband higher educated 481 29.03
No gap, low education 252 15.21
No gap, high education 138 8.33
Child age (months) 2000 3.01 61.70
Child female 2000 47.70
Values are median and interquartile.
Lifetime Last 12 months
Controlling behavior Emotional violence Physical violence
Sexual violence Any violence All types of violence
Figure 2 Experiences of IPV in lifetime and in last 12 months
Health Policy and Planning, 2020, Vol. 35, Suppl. 1 i23
Association between IPV and maternal CMD
The mean score on the CMD scale was 4.7 and the proportion of
mothers with CMD was 28% (Table 1). The three most prominent
symptoms of CMD were headaches (57%), poor appetite (45%),
poor sleep (42%) and difficulty enjoying daily activities (32%)
(Supplementary Figure S1). Women who experienced controlling be-
haviour were more likely to have CMD (OR 1.96; 95% CI 1.36,
2.80) than non-abused women (Figure 3). All types of IPV were sig-
nificantly associated with CMD (OR 1.9–2.3), and the odds of
CMD were highest among women who had experienced all forms of
violence (OR 2.31; 95% CI 1.32, 4.02).
Association between IPV and breastfeeding practices
More than half of the children (54%) had been breastfed within 1 h
of birth, and two-thirds of the children under 6 months of age were
exclusively breastfed. Mothers exposed to IPV were less likely to ini-
tiate breastfeeding early (OR ranged from 0.87 to 0.95); however,
the association was not significant at P<0.05. Both controlling be-
haviour and emotional violence were found to be negatively associ-
ated with EBF in the first 6 months of age (Figure 3). Compared
with women not exposed to IPV, the odds of EBF were lower among
those exposed to controlling behaviour (OR: 0.72, 95% CI: 0.56,
0.92), emotional violence (OR: 0.66, 95% CI: 0.53, 0.82) and all
three forms of violence (OR: 0.71, 95% CI: 0.54, 0.93).
Inter-relationships between determinant factors, IPV,
maternal CMD and breastfeeding practices
In the full path model considering all available factors, we found evi-
dence of strong links between potential determinants, IPV, maternal
CMD and breastfeeding practice (Figure 4). Household food inse-
curity, low SES and spousal education gap showed a direct associ-
ation with experienced IPV (b¼0.19, 0.05 and 0.07, respectively,
all P<0.05). IPV was directly associated with both maternal CMD
(b¼0.11, P<0.001) and optimal breastfeeding practice (b¼
0.05, P<0.05). The indirect path through maternal CMD
accounted for 14% of the relationship between IPV and breastfeed-
ing practice. Household food security also indirectly influenced
breastfeeding practices, though both IPV and maternal CMD, and
low maternal autonomy, had direct associations with breastfeeding
practices (b¼0.08, P<0.05).
Our study confirms that IPV is pervasive in Bangladesh, affecting
more than half of the women in our survey sample during pregnancy
and the lactation period. Women with low autonomy, living in poor
households, experiencing food insecurity and with unequal educa-
tion compared with their husbands were more likely to experience
IPV. All forms of IPV were positively associated with maternal
Table 2 Factors associated with different types of IPV
Controlling behaviour Emotional violence Physical violence Sexual violence All violence
OR [95% CI] OR [95% CI] OR [95% CI] OR [95% CI] OR [95% CI]
13–19 1 1 1 1 1
20–29 1.11 [0.81, 1.53] 0.94 [0.69, 1.27] 0.87 [0.64, 1.18] 0.85 [0.56, 1.30] 0.76 [0.40, 1.43]
30–44 0.9 [0.52, 1.54] 0.82 [0.44, 1.51] 0.73 [0.48, 1.11] 0.93 [0.48, 1.84] 0.64 [0.26, 1.61]
Early married 0.71
[0.56, 0.90] 1.01 [0.63, 1.63] 1.4 [0.94, 2.07] 1.31 [0.86, 2.01] 1.12 [0.57, 2.18]
Parity 1.08 [0.89, 1.32] 1.09 [0.90, 1.33] 1.07 [0.92, 1.25] 0.99 [0.78, 1.24] 1.04 [0.74, 1.44]
[1.58, 4.06] 1.11 [0.71, 1.73] 1.59
[1.01, 2.49] 1.43 [0.78, 2.61] 1.73 [0.88, 3.41]
Medium 1.45 [0.99, 2.11] 1.05 [0.76, 1.45] 1.18 [0.80, 1.75] 1.11 [0.72, 1.71] 1.19 [0.70, 2.02]
High 1 1 1 1 1
Secure 1 1 1 1 1
[1.06, 2.48] 2.01
[1.49, 2.70] 2.29
[1.82, 2.88] 1.78
[1.39, 2.27] 2.10
Low 1.07 [0.78, 1.46] 1.07 [0.78, 1.47] 1.78
[1.30, 2.43] 1.47 [0.94, 2.30] 2.01
Middle 1.09 [0.78, 1.52] 1.16 [0.92, 1.46] 1.26 [0.95, 1.66] 1.04 [0.68, 1.57] 1.47 [0.79, 2.74]
High 1 1 1 1 1
Wife more highly
0.9 [0.60, 1.34] 1.32 [0.91, 1.89] 1.88
[1.17, 3.02] 1.49 [0.92, 2.42] 2.29
0.92 [0.61, 1.41] 1.22 [0.92, 1.63] 1.51 [0.81, 2.82] 1.16 [0.67, 2.01] 2.25
No gap, low
0.76 [0.50, 1.15] 1.34 [0.82, 2.21] 1.8 [0.97, 3.33] 1.31 [0.66, 2.61] 2.79
No gap, high
Observations 1612 1612 1612 1612 1612
CI, conﬁdent intervals; OR, odds ratio; SES, socio-economic status.
i24 Health Policy and Planning, 2020, Vol. 35, Suppl. 1
CMD and negatively associated with EBF practices. The path of in-
fluence from IPV to breastfeeding practice through maternal CMD
was important, explaining 14% of the relationship.
Although a range of studies have explored the determinants of
IPV and its health consequences among married women, our study
contributes to the empirical literature by demonstrating the complex
relationship between IPV, maternal mental health and breastfeeding
practices around the time of pregnancy and lactation—a vulnerable
period for women. Due to a mixture of health behaviours and psy-
chological changes in the pregnancy and postnatal periods, women
need more care and love including more health check-ups, a better
diet, a less heavy workload and more support with household chores
and care of other children. Instead of receiving those forms of spe-
cial care, Bangladeshi women still suffer from very high IPV preva-
lence in this sensitive period, at nearly 50%. Other research suggests
that IPV in this context can contribute to more adverse pregnancy
outcomes such as increased risk of preterm birth, low birth weight
and intrauterine growth restriction (Donovan et al., 2016;Hill
et al., 2016) and poor child feeding practices.
Consistent with other studies (Rahman et al., 2014;VanderEnde
et al., 2015;Islam et al., 2017a;Caleyachetty et al., 2019;
Diamond-Smith et al. 2019), our analysis indicated that women
who live in food-insecure and low SES households were more likely
to have experienced IPV. These results imply that attention must be
paid to opportunities to mainstream IPV prevention and response
through programming in underlying factors such as food security,
livelihoods and economic empowerment. We also found evidence of
an association between a low level of women’s overall autonomy
and a high risk of experiencing controlling behaviour and physical
IPV. In the context of the low economic decision-making power of
Figure 3 Association between IPV and CMD and breastfeeding practices. (A) CMD
. (B) EIBF
. (C) EBF
Model adjusted for maternal age, education, parity,
household SES, food security, child age, gender and geographical clustering.
Model adjusted for maternal common disorder, age, education, parity, BMI, C-sec-
tion, household SES, food security, child age, gender and geographical clustering
Health Policy and Planning, 2020, Vol. 35, Suppl. 1 i25
women and widespread attitudes that support wife beating, as
reported, women are at high risk of IPV. Our results, therefore, indi-
cate that women’s autonomy, particularly economic decision-
making autonomy, and women’s own attitudes towards partner vio-
lence may need to be considered as important socio-cultural deter-
minants for reducing the risk of IPV among women in Bangladesh.
Regarding relationship inequality between spouses, educational in-
equality (either the wife or husband had a higher education than
their spouse) increased the likelihood of IPV. Couples where both
the wife and husband had a low education (less than five years of
schooling) had the highest IPV prevalence. Because of the risk of
educational discrepancy between spouses, it appears that education-
al interventions do not have a purely beneficial effect if either men
or women are the only ones receiving the education. This circum-
stance suggests the need for more directed educational interventions
for both men and women.
Our study demonstrated that all forms of IPV were associated
with maternal CMD, a finding consistent with previous studies in
rural Bangladesh and other settings (Dillon et al., 2013;Ziaei et al.,
2016). Poor mental health lessens the mother’s capability to take ad-
equate care of her child (Stein et al., 2014), which in turn can have
adverse effects on children’s growth (Stewart et al., 2008;Nguyen
et al., 2014) and development (Rahman et al., 2008). While results
from a previous study in LMICs showed a decreased likelihood of
exclusively breastfeeding related to exposure to all forms of IPV
(Caleyachetty et al., 2019), our study only found this association
with exposure to controlling behaviour and emotional violence.
Given the culture of strong influences of the husband and family
members on breastfeeding practices, interventions should be
designed to engage husbands and the broader community to support
mothers to achieve the recommended practices.
Health systems play a crucial role in a multisectoral response to
violence against women, including documenting prevalence, primary
prevention, emphasizing its health burden and advocating for coor-
dinated action with other sectors (Garcia-Moreno et al., 2015).
During pregnancy, antenatal care (ANC) visits have been identified
as an important platform for IPV screening and prevention because
ANC provides an opportunity to enquire about IPV, and also allows
for the possibility of follow-up during ANC with appropriate sup-
portive interventions, such as counselling and empowerment inter-
ventions (WHO, 2016). The current WHO recommendation on
ANC for a positive pregnancy experience strongly encourages coun-
tries to include IPV components at ANC visits when assessing condi-
tions that may be caused or complicated by IPV, to improve clinical
diagnosis and subsequent care (WHO, 2016). In addition, providers
must be trained to ask questions in the correct way and to respond
appropriately to women who disclose violence (WHO, 2013b,
2019). A previous systematic review of the existing literature found
that evidence of effective interventions for IPV during the perinatal
period is lacking, but home visitation programmes and some multi-
faceted counselling interventions did produce promising results (OR
ranged from 0.47 to 0.92) (Van Parys et al., 2014). Some countries
(Spain, India, Lebanon, Brazil and South Africa) have guidelines or
protocols to incorporate care for IPV into their healthcare systems,
but overall system development and implementation have been slow
to progress (Garcia-Moreno et al., 2015).
To address the high burden of IPV, the Bangladesh policy frame-
work has included a number of conventions, policies and acts on
violence against women, among them the Suppression of Violence
against Women and Children Act in 2000, the Domestic Violence
(Prevention and Protection) Act in 2010 and the Domestic Violence
(Prevention and Protection) Rules in 2013 (Naved et al., 2017a). In
response to the global recommendations, the health sector also
included support for IPV in the essential health service package
(Gov of Bangladesh, 2016). However, there are substantial gaps in
the implementation of laws and policies where the IPV screening
guideline was not mandatory in the national strategy for maternal
and neonatal health (Gov of Bangladesh, 2009). Given that health
platforms play a critical role in reaching pregnant women and their
families, IPV screening and support services, including couple coun-
selling sessions, should be included as part of routine antenatal care
to ensure that all pregnant women receive appropriate support.
Figure 4 Inter-relationships between determinants factors, IPV, maternal CMD and breastfeeding practices
Optimal breastfeeding practice includes EIBF and
The indirect path through maternal CMD accounted for 14% of the relationship between IPV and breastfeeding practice
i26 Health Policy and Planning, 2020, Vol. 35, Suppl. 1
Some governments and professional organizations recommend
screening all women for IPV rather than asking only women with
symptoms (WHO, 2013b), and it is imperative that this be priori-
tized in Bangladesh, given the very high IPV burden and the range of
poor health outcomes that are associated with IPV.
Previous studies in Bangladesh have shown positive effects of
breastfeeding counselling on mitigation of the negative impact of
IPV on the duration of EBF (Frith et al., 2017). Combining cash
transfers with group-based nutrition behaviour change communica-
tion had an impact not only on nutrition behaviours (Hoddinott
et al., 2018) but also on IPV (Roy et al., 2017) in Bangladesh. This
suggests that combined intervention may have positive impacts on
IPV and on nutrition behaviours. Given the effectiveness of large-
scale maternal nutrition (Nguyen et al., 2017a) and child feeding
programmes (Menon et al., 2016) that include interpersonal coun-
selling and community mobilization, adding an IPV component to
the counselling content and involving men in the counselling ses-
sions can bring additional benefits to reduce the prevalence of IPV,
improving maternal mental health and amplifying effects on child
breastfeeding practices. As noted above, mandating and supporting
the integration of IPV screening and support as part of routine care
during pregnancy could go a long way in recognizing the challenge
and creating supportive structures within the health system.
Our study has some limitations. To collect information on moth-
ers and on feeding for children born to those mothers, we only
included women with living children <6 months, thus excluding
women who died or whose babies died. We acknowledge that those
women may be the most vulnerable to IPV and its consequences.
The assessment of breastfeeding practices was based on mothers’
reports of all foods and liquids given to children in the first few days
after birth (for EIBF) and in the 24 h prior to the survey (for EBF),
which may be subject to recall bias. Given that all mothers reported
on feeding practices that were ongoing or within the last six months,
we believe, however, that mothers would have had good recall.
Data for CMD were also collected based on maternal recall of their
symptoms of mental disorders in the last four weeks. Maternal recall
has its limitations and CMD may influence the recall, either through
under- or over-reporting of information. Due to the study’s cross-
sectional design, we are not able to establish causal relationships bet-
ween variables, but we can infer associations. Finally, our data were
collected from rural areas, thus they have limited generalizability at
the national level.
Despite these limitations, the study findings have important
implications for policy and practice in Bangladesh based on insights
into the complex relationship between factors influencing IPV, as
well as the adverse consequences of IPV on maternal mental health
and child feeding practices. Given the high IPV prevalence in
Bangladesh, integrating actions both to reduce IPV and to mitigate
the impacts of IPV into existing prenatal and postnatal care services,
can serve as multiple duty actions to simultaneously address the
high prevalence of IPV, poor maternal mental health and suboptimal
breastfeeding practices. Several options are available to policy-
makers, at least some of which have been tested in rural Bangladesh,
but further action is urgently needed given the high prevalence of
IPV as well as the huge burden that exposure to IPV places on the
lives of women and their children in this context.
Supplementary data are available at Health Policy and Planning online.
This work was supported by the HSG Women Mentorship Program, which
supports individual capacity for early-career women in low- and middle-
income countries. The programme paired the author with an experienced
mentor, Ligia Paina, who provided guidance during the paper preparation
process. We thank Ligia Paina for valuable recommendations, suggestions
and language editing. We thank Data Analysis and Technical Assistance
Limited—the survey ﬁrm that collected the data in Bangladesh.
The Bill & Melinda Gates Foundation and the Canadian Department of
Foreign Affairs, Trade and Development through Alive & Thrive, managed
by FHI 360; and the CGIAR Research Program on Agriculture for Nutrition
and Health (A4NH), led by the International Food Policy Research Institute.
Conﬂict of interest statement. None declared.
Ethical approval: Ethical approval was obtained from the Institutional
Review Boards of the BRAC University in Bangladesh and the International
Food Policy Research Institute, USA. Written informed consent was obtained
from all women >18 years. For women <18 years of age, we obtained their
assent and the permission of their guardians, i.e. their parents or husbands, to
participate in the study.
Akman I, Kuscu MK, Yurdakul Z et al. 2008. Breastfeeding duration and post-
partum psychological adjustment: role of maternal attachment styles.
Journal of Paediatrics and Child Health 44: 369–73.
Bangladesh Bureau of Statistics. 2016. Bangladesh Violence against Women
(VAW) Survey 2015. Bangladesh Bureau of Statistics (BBS), Bangladesh
Bureau of Statistics (BBS), Dhaka, Bangladesh.
Caleyachetty R, Uthman OA, Bekele HN et al. 2019. Maternal exposure to in-
timate partner violence and breastfeeding practices in 51 low-income and
middle-income countries: a population-based cross-sectional study. PLoS
Medicine 16: e1002921.
Coates J, Swindale A, Bilinsky P. 2007. Household Food Insecurity Access
Scale (HFIAS) for Measurement of Household Food Access: Indicator
Guide (v. 3). Washington, DC: Food and Nutrition Technical Assistance
Project, Academy for Educational Development.
Devries KM, Kishor S, Johnson H et al. 2010. Intimate partner violence during
pregnancy: analysis of prevalence data from 19 countries. Reproductive
Health Matters 18: 158–70.
Devries KM, Mak JY, Bacchus LJ et al. 2013a. Intimate partner violence and
incident depressive symptoms and suicide attempts: a systematic review of
longitudinal studies. PLoS Medicine 10: e1001439.
Devries KM, Mak JY, Garcia-Moreno C et al. 2013b. Global health. The glo-
bal prevalence of intimate partner violence against women. Science 340:
Devries K, Watts C, Yoshihama M et al. 2011. Violence against women is
strongly associated with suicide attempts: evidence from the WHO
multi-country study on women’s health and domestic violence against
women. Social Science & Medicine 73: 79–86.
Diamond-Smith N, Conroy AA, Tsai AC, Nekkanti M, Weiser SD. 2019.
Food insecurity and intimate partner violence among married women in
Nepal. Journal of Global Health 9: 010412.
Dillon G, Hussain R, Loxton D, Rahman S. 2013. Mental and physical health
and intimate partner violence against women: a review of the literature.
International Journal of Family Medicine 2013: 1–15.
Donovan BM, Spracklen CN, Schweizer ML, Ryckman KK, Saftlas AF. 2016.
Intimate partner violence during pregnancy and the risk for adverse infant
outcomes: a systematic review and meta-analysis. BJOG: An International
Journal of Obstetrics & Gynaecology 123: 1289–99.
Frith AL, Ziaei S, Naved RT et al. 2017. Breast-feeding counselling mitigates
the negative association of domestic violence on exclusive breast-feeding
Health Policy and Planning, 2020, Vol. 35, Suppl. 1 i27
duration in rural Bangladesh. The MINIMat randomized trial. Public
Health Nutrition 20: 2810–8.
Garcia-Moreno C, Hegarty K, d’Oliveira AF et al. 2015. The health-systems
response to violence against women. The Lancet 385: 1567–79.
Garcia-Moreno C, Jansen HAFMEllsberg M, Heise L, Watts C. 2005. WHO
Multi-country Study on Women’s Health and Domestic Violence against
Women: Initial Results on Prevalence, Health Outcomes and Women’s
Responses. Geneva, Switzerland: World Health Organization.
Government of Bangladesh. 2009. National Neonatal Health Strategy and
Guidelines for Bangladesh. Dhaka, Bangladesh: Ministry of Health and
Family Welfare. Government of the People’s Republic of Bangladesh.
Government of Bangladesh. 2016. Bangladesh Essential Health Service
Package (ESP). Dhaka, Bangladesh: Ministry of Health and Family
Welfare. Government of the People’s Republic of Bangladesh.
Gwatkin DR, Rutstein S, Johnson K et al. 2007. Socio-economic differences in
health, nutrition, and population within developing countries: an overview.
Nigerian Journal of Clinical Practice 10: 272–82.
Halim N, Beard J, Mesic A et al. 2018. Intimate partner violence during preg-
nancy and perinatal mental disorders in low and lower middle income coun-
tries: a systematic review of literature, 1990–2017. Clinical Psychology
Review 66: 117–35.
Harpham T, Huttly S, De Silva MJ, Abramsky T. 2005. Maternal mental
health and child nutritional status in four developing countries. Journal of
Epidemiology & Community Health 59: 1060–4.
Hill A, Pallitto C, McCleary-Sills J, Garcia-Moreno C. 2016. A systematic re-
view and meta-analysis of intimate partner violence during pregnancy and
selected birth outcomes. International Journal of Gynecology & Obstetrics
Hindin MJ, Kishor S, Ansara DL. 2008. Intimate Partner Violence among
Couples in 10 DHS Countries: Predictors and Health Outcomes. DHS
Analytical Studies No. 18. Calverton, MD: Macro International, Inc.
Hoddinott J, Ahmed A, Karachiwalla NI, Roy S. 2018. Nutrition behaviour
change communication causes sustained effects on IYCN knowledge in two
cluster-randomised trials in Bangladesh. Maternal & Child Nutrition 14:
Islam MJ, Baird K, Mazerolle P, Broidy L. 2017a. Exploring the inﬂuence of
psychosocial factors on exclusive breastfeeding in Bangladesh. Archives of
Women’s Mental Health 20: 173–88.
Islam MJ, Broidy L, Baird K, Mazerolle P. 2017b. Intimate partner violence
around the time of pregnancy and postpartum depression: the experience of
women of Bangladesh. PLoS One 12: e0176211.
Jasinski JL. 2004. Pregnancy and domestic violence: a review of the literature.
Trauma Violence Abuse 5: 47–64.
Jewkes R. 2002. Intimate partner violence: causes and prevention. The Lancet
Khan AM, Flora MS. 2017. Maternal common mental disorders and associ-
ated factors: a cross-sectional study in an urban slum area of Dhaka,
Bangladesh. International Journal of Mental Health Systems 11: 23.
Khan NZ, Muslima H, Bhattacharya M et al. 2008. Stress in mothers of pre-
term infants in Bangladesh: associations with family, child and maternal fac-
tors and children’s neuro-development. Child: Care, Health and
Development 34: 657–64.
Kong SK, Lee DT. 2004. Factors inﬂuencing decision to breastfeed. Journal of
Advanced Nursing 46: 369–79.
Medhin G, Hanlon C, Dewey M et al. 2010. The effect of maternal common
mental disorders on infant undernutrition in Butajira, Ethiopia: the
P-MaMiE study. BMC Psychiatry 10: 32.
Menon P, Nguyen PH, Saha KK et al. 2016. Impacts on breastfeeding practices
of at-scale strategies that combine intensive interpersonal counseling, mass
media, and community mobilization: results of cluster-randomized program
evaluations in Bangladesh and Viet Nam. PLoS Medicine 13: e1002159.
Mezzavilla RS, Ferreira MF, Curioni CC, Lindsay AC, Hasselmann MH.
2018. Intimate partner violence and breastfeeding practices: a systematic re-
view of observational studies. Jornal de Pediatria 94: 226–37.
Naved RT, Samuels F, Gupta T et al. 2017a. Understanding Intimate Partner
Violence in Bangladesh through a Male Lens. London: Overseas
Naved RT, Samuels F, Masson VL et al. 2017b. Understanding Intimate
Partner Violence in Rural Bangladesh: Prevention and Response. Overseas
Development Institute, London.
Nguyen PH, Kim SS, Sanghvi T et al. 2017a. Integrating nutrition interven-
tions into an existing maternal, neonatal, and child health program
increased maternal dietary diversity, micronutrient intake, and exclusive
breastfeeding practices in Bangladesh: results of a cluster-randomized pro-
gram evaluation. The Journal of Nutrition 147: 2326–37.
Nguyen PH, Saha KK, Ali D et al. 2014. Maternal mental health is associated
with child undernutrition and illness in Bangladesh, Vietnam and Ethiopia.
Public Health Nutrition 17: 1318–27.
Nguyen PH, Sanghvi T, Kim SS et al. 2017b. Factors inﬂuencing maternal nu-
trition practices in a large scale maternal, newborn and child health program
in Bangladesh. PLoS One 12: e0179873.
Parvin K, Mamun MA, Gibbs A, Jewkes R, Naved RT. 2018. The pathways
between female garment workers’ experience of violence and development
of depressive symptoms. PLoS One 13: e0207485.
Rahman A, Patel V, Maselko J, Kirkwood B. 2008. The neglected ‘m’ in MCH
programmes—why mental health of mothers is important for child nutri-
tion. Tropical Medicine & International Health 13: 579–83.
Rahman M, Hoque MA, Mostofa MG, Makinoda S. 2014. Association be-
tween adolescent marriage and intimate partner violence: a study of young
adult women in Bangladesh. Asia Paciﬁc Journal of Public Health 26:
Rapp D, Zoch B, Khan MMH, Pollmann T, Kra¨mer A. 2012. Association be-
tween gap in spousal education and domestic violence in India and
Bangladesh. BMC Public Health 12: 467.
Roy S, Hidrobo M, Hoddinott J, Ahmed A. 2017. Transfers, Behavior Change
Communication, and Intimate Partner Violence. Postprogram Evidence
from Rural Bangladesh. IFPRI Discussion Paper 01676. Washington, DC:
International Food Policy Research Institute.
StataCorp. 2019. Stata Statistical Software. College Station, Texas 77845
USA. Copyright 1985-2019 StataCorp LP.
Stein A, Pearson RM, Goodman SH et al. 2014. Effects of perinatal mental dis-
orders on the fetus and child. The Lancet 384: 1800–19.
Stewart RC, Umar E, Kauye F et al. 2008. Maternal common mental disorder
and infant growth—a cross-sectional study from Malawi. Maternal & Child
Nutrition 4: 209–19.
VanderEnde KE, Sibley LM, Cheong YF, Naved RT, Yount KM. 2015.
Community economic status and intimate partner violence against women
in Bangladesh: compositional or contextual effects? Violence against
Women 21: 679–99.
Van Parys AS, Verhamme A, Temmerman M, Verstraelen H. 2014. Intimate
partner violence and pregnancy: a systematic review of interventions. PLoS
One 9: e85084.
Victora CG, de Onis M, Hallal PC, Blossner M, Shrimpton R. 2010.
Worldwide timing of growth faltering: revisiting implications for interven-
tions. Pediatrics 125: e473–80.
Vyas S, Kumaranayake L. 2006. Constructing socio-economic status indices: how
to use principal components analysis. Health Policy and Planning 21: 459–68.
World Health Organization (WHO). 1994. A User’s Guide to the Self
Reported Questionnaire (SRQ). Geneva: World Health Organization.
World Health Organization (WHO). 2011. Intimate Partner Violence during
Pregnancy. Geneva: Department of Reproductive Health and Research.
World Health Organization.
World Health Organization (WHO). 2012. Understanding and addressing vio-
lence against women. Geneva: WorldHealth Organization.
World Health Organization (WHO). 2013a. Global and Regional Estimates
of Violence against Women: Prevalence and Health Effects of Intimate
Partner Violence and Non-partner Sexual Violence. Geneva: World Health
World Health Organization (WHO). 2013b. Responding to Intimate Partner
Violence and Sexual Violence against Women. WHO Clinical and Policy
Guidelines. Geneva: World Health Organization.
World Health Organization (WHO). 2016. WHO Recommendations on
Antenatal Care for a Positive Pregnancy Experience. Geneva: World Health
i28 Health Policy and Planning, 2020, Vol. 35, Suppl. 1
World Health Organization (WHO). 2019. Health Care for Women Subjected
to Intimate Partner Violence or Sexual Violence. A Clinical Handbook.
Geneva: WHO Press, World Health Organization.
Yount KM, James-Hawkins L, Cheong YF, Naved RT. 2018. Men’s
perpetration of partner violence in Bangladesh: community gender
norms and violence in childhood. Psychology of Men & Masculinity 19:
Ziaei S, Frith AL, Ekstrom EC, Naved RT. 2016. Experiencing lifetime
domestic violence: associations with mental health and stress among pregnant
women in Rural Bangladesh: the MINIMat randomized trial. PLoS One 11:
Zureick-Brown S, Lavilla K, Yount KM. 2015. Intimate partner violence and
infant feeding practices in India: a cross-sectional study. Maternal & Child
Nutrition 11: 792–802.
Health Policy and Planning, 2020, Vol. 35, Suppl. 1 i29