Content uploaded by Md Nuruzzaman Khan
Author content
All content in this area was uploaded by Md Nuruzzaman Khan on Mar 19, 2025
Content may be subject to copyright.
Prevalence of co-occurring forms of intimate partner violence
against women aged 15–49 and the role of education-related
inequalities: analysis of Demographic and Health Surveys
across 49 low-income and middle-income countries
Shuangyu Zhao,
a
,
r
Shan Liu,
a
,
r
Jiuxuan Gao,
a
,
r
Ning Ma,
a
Shaoru Chen,
a
Joht Singh Chandan,
b
Rockli Kim,
c
,
d
Peter Karoli,
e
John Lapah Niyi,
f
Jayalakshmi Rajeev,
g
Melkamu Aderajew Zemene,
h
Md Nuruzzaman Khan,
i
Hajirani M. Msuya,
j
Chunling Lu,
k
,
l
S. V. Subramanian,
m
,
n
Feng Cheng,
a
,
o
John S. Ji,
a
Kun Tang,
a
Pascal Geldsetzer,
p
,
q
and Zhihui Li
a
,
o
,
∗
a
Vanke School of Public Health, Tsinghua University, No. 30, Shuangqing Road, Haidian District, Beijing, PR China
b
Department of Applied Health Sciences, University of Birmingham, Birmingham, United Kingdom
c
Interdisciplinary Program in Precision Public Health, Department of Public Health Sciences, Graduate School of Korea University, Seoul,
Republic of Korea
d
Division of Health Policy and Management, College of Health Science, Korea University, Seoul, Republic of Korea
e
National Institute for Medical Research, Dar es Salaam, Tanzania
f
Ghana Health Service, Gushegu Municipal Health Directorate, Gushegu, Ghana
g
Department of Public Health and Community Medicine, Central University of Kerala, Kasaragod, India
h
Department of Public Health, College of Health Sciences, Debre Tabor University, Debre Tabor, Ethiopia
i
Nossal Institute for Global Health, Melbourne School of Population and Global Health, The University of Melbourne, Melbourne,
Victoria, Australia
j
Department of Interventions and Clinical Trials, Ifakara Health Institute, Ifakara, Tanzania
k
Division of Global Health Equity, Brigham and Women’s Hospital, Boston, MA, USA
l
Department of Global Health and Social Medicine, Harvard Medical School, Boston, MA, United States
m
Department of Social and Behavioral Sciences, Harvard T.H. Chan School of Public Health, Boston, United States
n
Harvard Center for Population and Development Studies, Cambridge, MA, United States
o
Institute for Healthy China, Tsinghua University, 100084, Beijing, China
p
Division of Primary Care and Population Health, Department of Medicine, Stanford University School of Medicine, Stanford, CA, USA
q
Chan Zuckerberg Biohub –San Francisco, San Francisco, CA, USA
Summary
Background Women experiencing co-occurring forms of intimate partner violence (IPV; ie, physical, sexual, and/or
psychological) often face more severe psychological and health consequences than those experiencing a single form.
However, research on IPV co-occurrence in low- and middle-income countries (LMICs) remains limited. This study
examines the prevalence of IPV co-occurrence in LMICs and its education-based inequalities.
Methods Data from the most recent Demographic and Health Surveys in 49 LMICs (2011–2023) were used. Our
primary outcome was IPV co-occurrence, defined as a woman aged 15–49 ever experiencing any two or three forms of
physical, sexual, or psychological IPV from her partner within the past year. We categorised IPV co-occurrence into
four subtypes: co-occurrence of (1) physical and sexual IPV, (2) physical and psychological IPV, (3) sexual and
psychological IPV, and (4) all three forms of IPV. We analysed the prevalence of IPV co-occurrence and its
subtypes by women’s education levels, calculating odds ratios to assess inequalities. Nonparametric restricted
cubic splines were used to explore nonlinear relationships between education and IPV.
Findings The study included a total of 344,661 women. The weighted prevalence of IPV co-occurrence varied widely
across countries—from 2.4% in Armenia to 38.9% in Papua New Guinea. Overall, women with no education were
most at risk, experiencing an adjusted prevalence of 14.3% (95% CI: 13.3–15.2), compared to 11.8% (95% CI:
10.8–12.9) among those with primary education, 9.9% (95% CI: 9.3–10.6) for secondary education, and 5.3% (95%
CI: 4.5–6.2) for higher education. The prevalence of IPV co-occurrence involving sexual IPV was highest among
women with primary education, with 4.1% (95% CI: 3.4–4.8) reporting concurrent physical and sexual violence,
compared to 1.5% (95% CI: 1.1–1.9) to 3.7% (95% CI: 3.2–4.1) among other education levels.
*Corresponding author. Vanke School of Public Health, Tsinghua University, 100084, Beijing, China.
E-mail address: zhihuili@tsinghua.edu.cn (Z. Li).
r
Contributed equally.
eClinicalMedicine
2025;82: 103150
Published Online xxx
https://doi.org/10.
1016/j.eclinm.2025.
103150
www.thelancet.com Vol 82 April, 2025 1
Articles
Interpretation IPV co-occurrence remains high, particularly among women with little or no education. Education-
focused interventions are urgently needed to reduce IPV risk and its severe impact. However, the findings may be
influenced by potential reporting biases and cross-country variability in IPV measurement methodologies, which
may limit generalizability.
Funding The China National Natural Science Foundation (Grant numbers 72203119) and The Research Fund, Vanke
School of Public Health, Tsinghua University.
Copyright © 2025 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Keywords: Intimate partner violence co-occurrence; Education-based inequality; Low- and middle-income countries
Introduction
Intimate partner violence (IPV) is widely acknowledged
as a global public health concern leading to both short-
and long-term physical and psychological harm,
including injuries, depression, sexually transmitted in-
fections, and mortality.
1
World Health Organization
(WHO) reported that globally, 27% of ever-partnered
women aged 15–49 years have been subjected to phys-
ical and/or sexual violence from a current or former
male intimate partners at least once in their lifetime,
with 13% experiencing it in the year before being sur-
veyed.
2
This statistic is even higher in LMICs, with a
study suggesting that the prevalence of IPV reached
37.2% within the 12 months preceding the surveys.
3
Given the constantly high prevalence and far-reaching
consequences to the individual, family and society, the
Sustainable Development Goals (SDGs) calls for the
elimination of all forms of violence against women and
Research in context
Evidence before this study
We searched Google Scholar, Web of Science, and PubMed for
studies with the combination of the following terms:
(“intimate partner violence”OR “domestic violence”OR
“partner violence”OR “spousal violence”OR “marital violence”
OR “multiple forms of intimate partner violence”OR “violence
co-occurrence”) AND (“inequality*”OR “equity”) from Jan 1,
2000, to Jul 1, 2024 with no language restrictions. While
many studies have explored the prevalence of intimate
partner violence (IPV) regarding physical, sexual, or
psychological IPV, few have focused on the co-occurrence of
IPV forms. A study in the United States have observed co-
occurrence of physical violence and psychological aggression,
however cross-cultural research in LMICs remains limited.
Educational attainment is a crucial factor in IPV dynamics, as
increasing women’s education and empowerment in LMICs
challenge entrenched gender norms and power structures
that sustain IPV. Some studies in the US and South Africa has
identified that women at the lowest educational levels are
associated with lower risk of IPV, while another study in
Uganda suggests that higher education is a protective factor
against IPV. This inconsistency indicates the need for
culturally diverse research to fully understand the link
between education and IPV co-occurrence.
Added value of this study
To our knowledge, this is one of the first studies that examine
the prevalence of IPV co-occurrence against women aged
15–49 years in multiple LMICs and explore its relationship
with women’s education. The co-occurrence of IPV was
defined as an ever-partnered woman aged 15–49 ever
experiencing any two or three forms of physical, sexual, or
psychological IPV from her partner within the past year. We
found that the prevalence of IPV co-occurrence is notably
high in LMICs, with 10.9% of women experiencing at least
two forms of IPV within 12 months in the studied countries.
Furthermore, we observed a strong association between
women’s education levels and IPV co-occurrence: overall,
women with higher education levels had a lower prevalence
of IPV co-occurrence, while a higher prevalence of IPV
subtypes involving sexual violence were observed among
women with primary education.
Implications of all the available evidence
Our findings highlight the urgent need for policy and
interventions targeting the co-occurrence of IPV among
women, particularly in LMICs. Educational empowerment is
considered as a central strategy in reducing IPV co-occurrence,
particularly targeting women with primary education as they
experience the highest prevalence of IPV involving sexual
violence. By framing education as not only a social and
economic imperative but also a means of IPV co-occurrence
prevention, policymakers can strengthen the case for
investing in girls’education. However, potential reporting
biases, variations in IPV measurement across countries, and
the inclusion of only surveyed LMICs may limit the
generalizability of findings. Tailored interventions that
address educational inequalities and support IPV prevention
are essential for reducing the burden of IPV and advancing
progress toward achieving the Sustainable Development
Goals (SDGs) related to gender equality and health.
Articles
2 www.thelancet.com Vol 82 April, 2025
girls (Goal 5.2) and prevention of violence-related deaths
(Goal 16.1), which have synergistic effect with other
important SDG targets on health and social de-
terminants of health (Goal 3).
4
The co-occurrence of IPV has become increasingly
prevalent, leading to compounded consequences for
women’s health and social well-being, indicating the
need for integrated and targeted interventions. Research
indicates that physical violence often co-occurs with
psychological abuse, increasing the risks of adverse psy-
chological outcomes and criminal justice involvement.
5–8
A study in Brazil found that 77.4% of women experi-
enced one type of IPV, 20.0% experienced two, and
2.6% reported all three types, while in sub-Saharan
Africa, 7.32% experienced all forms of IPV, high-
lighting its notable prevalence in LMICs.
9,10
Despite
these profound impacts, research on IPV co-occurrence
in LMICs remains limited, with most studies focusing
on developed countries and specificsubtypes.
11
Existing
global research estimates the prevalence of physical
and/or sexual IPV, but ignores the psychological
violence which significantly impacts women’s health.
2
Focusing on co-occurrence provides a more accurate
understanding of how interconnected forms of violence
affect individuals, enabling the development of inte-
grated and effective interventions to address the com-
pounded burden of violence.
12
Education is widely recognised as a direct and im-
pactful driver of structural and cultural change by pro-
moting economic status, participation in household
decisions, expanding media access, and challenging
patriarchal gender roles.
13
In LMICs, rising levels of
women’s education are often considered to be able to
disrupt traditional norms and power structures that
perpetuate IPV.
14
However, the empirical evidence on
the relationship between education and IPV is incon-
sistent across different contexts. For example, studies in
the US and South Africa identified that women with
primary education faced a higher risk of IPV compared to
those with no education,
15
while another study in Uganda
suggests that primary education is a protective factor
against IPV. These divergent findings point to the com-
plex interaction between education and local socioeco-
nomic and cultural norms, where the impact of education
on IPV may vary depending on the prevailing gender
dynamics and societal structures.
16
As an actionable and
scalable intervention, education offers a pathway to
reduce IPV by empowering women and reshaping soci-
etal norms. Yet, most studies focus on specific countries,
limiting generalizability. Broader cross-cultural evidence
is needed to fully understand education’s role in miti-
gating IPV co-occurrence and its subtypes.
Using the latest data from the Demographic and
Health Survey (DHS) between 2011 and 2023, this study
aims to provide the large-scale and country-level preva-
lence of co-occurrence of IPV against women aged
15–49 within past 12 months in 49 LMICs and estimate
the education-related inequalities between and within
countries.
Methods
Data sources and study participants
We obtained the most recent data from DHS con-
ducted after 2010. The DHS are nationally represen-
tative household surveys, selecting countries with
available data on the domestic violence module. The
program uses a two-stage cluster sampling method.
First, each country is divided into multiple geographic
zones, with strata separating urban and rural areas
provided by the national statistical office. Within each
stratum, clusters are sampled independently with the
probability of selection proportional to the cluster’s
population contribution to the total. Second, all
households within the selected cluster are listed, and a
fixed number of households is chosen via equal prob-
ability systematic sampling.
17
The domestic violence
module was administered in a sub-sample of house-
holds in which investigators can carry out the interview
in private. Only one ever-partner woman of reproduc-
tive age per household was randomly selected for an
interview for the domestic violence module. Our study
followed the Strengthening the Reporting of Observa-
tional Studies in Epidemiology (STROBE) reporting
guidelines.
The study initially identified data from 428,755
women aged 15–49 who participated in the domestic
violence survey conducted from 2013 to 2023. After
excluding women whose domestic violence sample
weight was zero, indicating no representation in the
sample (n = 1162), those who were never in union
(n = 82,892), those whose age not in 15–49 (n = 28), and
who did not have complete data on IPV (n = 12), the final
dataset consisted of 344,661 women samples (Fig. 1)
from 49 countries (Eastern and Southern Africa n = 16;
Western and Central Africa n = 13; East Asia and Pacific
n = 5; South Asia n = 5; Latin America and Caribbean
n = 4; Europe and Central Asia n = 3; Middle East and
North Africa n = 2; South America n = 1). Information on
the selected surveys for each country is provided in
Appendix Table S1.
Written informed consent was obtained from all in-
dividuals conducted by the DHS team. A parent or
guardian must provide consent prior to participation by
a child or adolescent under 18. The data used in the
study consisted of a publicly available de-identified
dataset, which was retrieved from the DHS website
with permission (https://dhsprogram.com/). This study
was approved by the Tsinghua Institutional Review
Board (Project No. 20220005).
Outcomes
The primary outcome identified in this study was IPV
co-occurrence against women, defined as the presence
Articles
www.thelancet.com Vol 82 April, 2025 3
of any two or three forms of IPV experienced by a
woman from her husband or partner within the past 12
months. The three forms of IPV studied were physical,
sexual, and psychological IPV within 12 months. Phys-
ical IPV was defined as women aged 15–49 who ever
been “pushed, shook or had something thrown”,“slap-
ped”,“punched with fist or hit by something harmful”,
“kicked or dragged”,“strangled or burnt”,“attacked with
knife/gun or other weapon”,or“had arm twisted or hair
pulled”by husband/partner or a combination of these
acts within 12 months. Sexual IPV was defined as
women aged 15–49 who have ever been “physically
forced into unwanted sex”,“forced into other unwanted
sexual acts”,or“physically forced to perform sexual acts
respondent didn’t want to”or a combination of these
acts within 12 months. Psychological IPV was defined as
women aged 15–49 who have ever been “humiliated”,
“threatened with harm”,or“insulted or made to feel
bad”by their husband/partner or a combination of these
acts within 12 months.
Additionally, we defined four subtypes of IPV co-
occurrence, including 1) the co-occurrence of physical
and sexual IPV, 2) the co-occurrence of physical and
psychological IPV, 3) the co-occurrence of sexual and
psychological IPV, and 4) the co-occurrence of all three
forms of IPV.
Exposures and covariates
We identified women’s education using the highest
educational attainment and the number of years of ed-
ucation completed. The woman’s educational attain-
ment was categorised into four groups based on DHS’s
definition: no education, primary education, secondary
education, and higher education. Covariates adjusted in
models included woman’s age (15–19 years, 20–34
years, or 35–49 years), place of residence (urban or ru-
ral), child marriage (aged at first cohabitation <18, or
aged at first cohabitation ≥18), marital status (married,
living with a partner, or others), the parity number (0, 1,
2, 3, 4, 5 or above), contraception use (yes or no),
household wealth index (poorest, poorer, middle, richer,
or richest), and women’s employment (currently
employed or not employed). Definitions of the cova-
riates are listed in Appendix Table S2.
Statistical analysis
We first estimated the co-occurrence of IPV along with
95% confidence intervals (CIs) at pooled, income, and
country levels. For subgroup analyses, countries were
classified into income groups based on World Bank
classifications to examine differences in IPV prevalence
across income groups. The weighted national preva-
lence was calculated according to the original sampling
Women agreed to participate
in domestic violence module
from 49 selected countries
(N = 428,755)
Excluded women who were ne ver in u nion
(N =82,892 )
Excluded women whose dome stic violence
sample weight was zero
(N = 1,162)
Excluded women who did not have
complete data on intimate partner violence
(N = 12)
Demographic analysis:
eligible women with data on
sele cted fac tors a nd outcom es
(N = 344,661)
Exclud ed wome n whose ma ternal age was
not in 1 5–49
(N = 28)
Fig. 1: Flow chart of sample exclusion.
Articles
4 www.thelancet.com Vol 82 April, 2025
weights of each country provided by DHS; the pooled
and income-level prevalence was calculated using sam-
pling weights rescaled by the population size of each
country in the survey year.
Second, in order to examine education-related in-
equalities after controlling for the impact of potential
confounders, we employed logistic regression models
adjusted to maternal age, child marriage, place of resi-
dence, parity number, women’s employment, marital
status, household wealth index, and contraception use,
and measured the adjusted prevalence of IPV co-
occurrence in each income group by education level.
To further strengthen the results on inequality, we also
reported the adjusted odds ratios (aORs) between the co-
occurrence of IPV and women’s education estimated
from the above models. The outcomes were categorised
as IPV co-occurrence or subtypes and not experiencing
any form of IPV: IPV co-occurrence was defined as
experiencing any two or three forms of IPV compared to
no IPV; co-occurrence of physical and sexual IPV
referred to experiencing both physical and sexual IPV
compared to no IPV; co-occurrence of physical and
psychological IPV referred to experiencing both physical
and psychological IPV compared to no IPV; co-
occurrence of psychological and sexual IPV referred to
experiencing both psychological and sexual IPV
compared to no IPV; and co-occurrence of all three
forms of IPV referred to experiencing physical, psy-
chological, and sexual IPV compared to no IPV.
Finally, we examined the potential nonlinear rela-
tionship between women’s education and IPV co-
occurrence using nonparametrically restricted cubic
splines. Women’s education was measured by the
number of years of education completed. Statistical
significance was set at p< 0.05. We dropped samples
with missing IPV responses, exposure and covariates.
The percentages of missing values of all variables were
less than 0.5%. The ‘svy’command was used in all an-
alyses to account for the sample design of the surveys
including weight, sampling strata, and cluster. All sta-
tistical analyses were done in Stata (version 17.0).
Role of the funding source
The funder of the study had no role in study design, data
collection, data analysis, data interpretation, writing of
the report, or submission for publication. All authors
had full access to the data in the study, and ZL had final
responsibility for the decision to submit for publication.
Results
Our analysis included a total of 344,661 women aged
15–49 from 49 LMICs who were currently or formerly in
an intimate partnership. There were 99,981 women
(29.0%) with no education, 96,174 (27.9%) with primary
education 113,268 (32.9%) with secondary education,
and 35,151 (10.2%) with higher education (Table 1).
Women with no education and primary school living at
rural were 82.7% and 75.4% respectively while, only
63.2% with secondary school and 39.4% with higher
education were living at rural area. Furthermore, 36.6%
and 25.4% participant with no education and primary
school respectively had lower wealth index as compared
to 11.0% and 1.9% for those with secondary or higher
education. Same trends seen for not having a child,
being not married and for early marriage. Thus, women
with primary or lower education levels were more likely
to live in rural areas, with worse wealth status, not
having a child, not married, and married before 18.
We presented the pooled prevalence of IPV co-
occurrence and the prevalence by country income
groups in Fig. 2. Overall, we found that 10.9% (95% CI:
10.4–11.5) of the women experienced at least two forms
of IPV within the 12 months before the surveys. The
prevalence of IPV co-occurrence was the highest in low-
income countries (LICs) (18.3%, 95% CI: 17.7–19.0) and
the lowest in lower-middle-income countries (LMCs)
(9.0%, 95% CI: 8.4–9.7); while the upper-middle-income
countries (UMICs) was in the middle at 10.8% (95% CI:
10.2–11.4). For subtypes of co-occurrence of IPV, we
consistently found LICs to be at higher risk compared to
LMCs and UMICs. For example, the co-occurrence of
physical and psychological IPV was 15.9% (95% CI:
15.3–16.5) in LICs, compared to 8.0% (95% CI: 7.4–8.6)
in LMCs and 9.1 (95% CI: 8.6–9.7) in UMICs.
The country-level prevalence of IPV co-occurrence
and its subtypes were presented in Fig. 3. Remarkably,
nearly one third of the countries exhibited notably high
prevalence of IPV co-occurrence, with 17 out of the 49
countries surpassing a prevalence of 15%, nearly all of
which were LICs. It is worth noting that four countries
showed exceptionally high prevalence exceeding 25%,
including Papua New Guinea (38.9%, 95% CI:
35.6–42.4), Afghanistan (30.1%, 95% CI: 27.5–32.8),
Sierra Leone (28.7%, 95% CI: 26.4–31.2), and Congo
Democratic Republic (25.2%, 95% CI: 22.7–27.8).
Similarly, the co-occurrence of physical and psycholog-
ical IPV was very high in Papua New Guinea (35.1%,
95% CI: 31.7–38.7), Afghanistan (28.5%, 95% CI:
25.9–31.2), Sierra Leone (27.3%, 95% CI: 25.0–29.8),
and Congo Democratic Republic (20.1%, 95% CI:
17.8–22.5), all exceeding 20%. Papua New Guinea re-
ported the highest prevalence of IPV co-occurrence and
all its subtypes, with the co-occurrence of physical and
sexual IPV at 20.9% (95% CI: 20.9–22.9), psychological
and sexual IPV at 20.3% (95% CI: 18.3–22.5), and all
three forms of violence at 18.7% (95% CI: 16.8–20.7).
Fig. 4 showed the adjusted prevalence of IPV co-
occurrence by women’s education levels at the pooled
level and within each income group. Overall, women
with no education faced the highest risk of experiencing
IPV co-occurrence, with a prevalence of 14.3% (95% CI:
13.3–15.2), compared to 11.8% (95% CI: 10.8–12.9)
among those with primary education, 9.9% (95% CI:
Articles
www.thelancet.com Vol 82 April, 2025 5
9.3–10.6) among those with secondary education, and
5.3% (95% CI: 4.5–6.2) among those with higher edu-
cation. However, when examining subtypes involving
sexual IPV, such as psychological and sexual violence,
women with primary education had the highest preva-
lence, showing a prevalence of 4.1% (95% CI: 3.4–4.8).
This percentage surpasses that of women with other
education levels, which ranged from 1.5% (95% CI:
1.1–1.9) to 3.7% (95% CI: 3.2–4.1). The other two sub-
types of IPV co-occurrence involving sexual IPV–
physical and sexual violence and all three violence–
showed similar patterns. When stratified by income
levels, this relationship appeared to be particularly
notable in LICs, where the prevalence of co-occurrence
of psychological and sexual IPV among women with
primary education was 7.5% (95% CI: 7.1–7.9),
compared with other educational levels ranging from
3.3% (95% CI: 2.2–4.4) to 6.0% (95% CI: 5.5–6.6).
Fig. 5 presented the odds ratio for the association
between IPV co-occurrence and women’s education,
adjusted for maternal age, child marriage, place of resi-
dence, parity, employment status, marital status, house-
hold wealth index, and contraception use. Women with
no education, compared to those with primary education,
have a significantly higher risk of experiencing IPV co-
occurrence overall (OR: 1.3, 95% CI: 1.1–1.5) and the
co-occurrence of physical and psychological violence (OR:
1.3, 95% CI: 1.1–1.5). In contrast, women with secondary
or higher education have significantly lower odds of
experiencing these forms of IPV. Regarding sexual IPV,
women with no education show a similar risk to those
with primary education, while the other two education
Overall (N = 344,661) Women’s education level
No education
(N = 99,960)
Primary education
(N = 96,229)
Secondary education
(N = 113,310)
Higher education
(N = 35,162)
Women’s age, years
15–19 9993 (2.9) 1400 (1.4) 2501 (2.6) 5097 (4.5) 281 (0.8)
20–34 174,354 (50.6) 34,993 (35.0) 45,586 (47.4) 65,469 (57.8) 22,673 (64.5)
35–49 160,227 (46.5) 63,588 (63.6) 48,087 (50.0) 42,702 (37.7) 12,197 (34.7)
Place of residence
Urban 110,608 (32.1) 17,297 (17.3) 23,659 (24.6) 41,683 (36.8) 21,302 (60.6)
Rural 233,966 (67.9) 82,684 (82.7) 72,515 (75.4) 71,585 (63.2) 13,849 (39.4)
Wealth index
Poorest 66,847 (19.4) 36,593 (36.6) 24,428 (25.4) 12,459 (11.0) 668 (1.9)
Poorer 72,706 (21.1) 27,395 (27.4) 26,544 (27.6) 21,181 (18.7) 2285 (6.5)
Middle 73,050 (21.2) 20,496 (20.5) 21,928 (22.8) 26,618 (23.5) 4183 (11.9)
Richer 70,982 (20.6) 11,298 (11.3) 16,253 (16.9) 30,243 (26.7) 8753 (24.9)
Richest 60,990 (17.7) 4199 (4.2) 7021 (7.3) 22,767 (20.1) 19,262 (54.8)
Parity number
0 28,944 (8.4) 4999 (5.0) 5097 (5.3) 10,534 (9.3) 6011 (17.1)
1 62,713 (18.2) 8098 (8.1) 10,579 (11.0) 25,259 (22.3) 12,690 (36.1)
2 117,155 (34.0) 22,896 (22.9) 30,006 (31.2) 47,006 (41.5) 12,865 (36.6)
3 67,537 (19.6) 24,595 (24.6) 24,621 (25.6) 19,822 (17.5) 2706 (7.7)
4 33,768 (9.8) 17,097 (17.1) 13,080 (13.6) 6683 (5.9) 598 (1.7)
≥5 34,457 (10.0) 22,296 (22.3) 12,791 (13.3) 3964 (3.5) 281 (0.8)
Marital status
Married 315,630 (91.6) 90,883 (90.9) 84,441 (87.8) 105,339 (93.0) 32,620 (92.8)
Living with a partner 6547 (1.9) 700 (0.7) 3751 (3.9) 2039 (1.8) 949 (2.7)
Others 22,397 (6.5) 8398 (8.4) 7982 (8.3) 5890 (5.2) 1582 (4.5)
Child marriage
No 195,373 (56.7) 40,292 (40.3) 41,836 (43.5) 70,566 (62.3) 32,620 (92.8)
Yes 149,201 (43.3) 59,689 (59.7) 54,338 (56.5) 42,702 (37.7) 2531 (7.2)
Contraception use
No 122,668 (35.6) 35,893 (35.9) 32,988 (34.3) 39,417 (34.8) 13,990 (39.8)
Yes 221,906 (64.4) 64,088 (64.1) 63,186 (65.7) 73,851 (65.2) 21,161 (60.2)
Women’s employment
Not employed 234,999 (68.2) 63,188 (63.2) 60,782 (63.2) 82,346 (72.7) 24,430 (69.5)
Currently employed 109,575 (31.8) 36,793 (36.8) 35,392 (36.8) 30,922 (27.3) 10,721 (30.5)
Note: The numbers and percentages are unweighted.
Table 1: Sample characteristics by education levels, using the most recent Demographic Health Surveys pooled across 49 countries, n (%).
Articles
6 www.thelancet.com Vol 82 April, 2025
categories have significantly lower risks. The results from
unadjusted models were consistent (Appendix Fig. S1),
and the nonparametric restricted cubic splines showed
consistent findings (Appendix Fig. S2).”
Discussion
To our knowledge, this is one of the first studies
examining the prevalence of IPV co-occurrence against
women aged 15–49 years in multiple LMICs and
exploring its relationship with women’s education.
Three key findings have emerged: firstly, the prevalence
of IPV co-occurrence is notably high in LIMCs, with
10.9% of women experiencing at least two forms of IPV
within 12 months. Secondly, we observed that women’s
increased level of education were associated with a lower
prevalence of IPV co-occurrence, both on a pooled level
and in LICs and UMICs. Thirdly, for subtypes of IPV co-
Fig. 2: Prevalence of co-occurrence of multiple forms of intimate partner violence against women aged 15–49 years in the past 12
months according to income group, % (95% confidence interval). Note: Physical & Sexual IPV: co-occurrence of physical and sexual intimate
partner violence; Physical & Psychological IPV: co-occurrence of physical and psychological intimate partner violence; Psychological & Sexual IPV:
co-occurrence of psychological and sexual intimate partner violence; All three forms of IPV: co-occurrence of all three forms of intimate partner
violence. The y-axis range is set from 0 to 10 for co-occurrence of physical and sexual IPV, co-occurrence of psychological and sexual IPV, and
co-occurrence of all three forms of IPV.
Articles
www.thelancet.com Vol 82 April, 2025 7
occurrence involving sexual violence, women with pri-
mary education had the highest prevalence.
Our findings indicate a pervasive pattern of IPV co-
occurrence in LMICs, which poses a significant barrier
to achieving SDG 5.2 on eliminating all forms of
violence. These findings are consistent with previous
research in Sub-Saharan Africa, which reported simi-
larly high rates of IPV co-occurrence in countries such
as Congo Democratic Republic, Uganda, and Burundi.
10
This challenge extends globally, affecting not only
LMICs but also high-income countries like the US and
Europe.
18,19
Indeed, to combat this challenge, some high-
income countries have implemented intervention pro-
grams. For example, educational programs such as
‘Shifting Boundaries’in the US have demonstrated
effectiveness in reducing sexual violence victimization
and perpetration among middle school students,
particularly through cost-efficient building-based in-
terventions.
20
Meanwhile, in LMICs, some programs in
sub-Saharan Africa like “Stepping Stones”in South
Africa, have utilised manuals to reduce conflicts within
intimate relationships and foster mutual understanding,
achieving notable success in reducing physical IPV and
forced sex.
21
Despite these efforts, LMICs are in urgent
need to effectively address co-occurrence of IPV. The
complexity and persistence of IPV co-occurrence require
more targeted and comprehensive strategies to mitigate
its impact across diverse contexts.
We observed a notably high prevalence of IPV co-
occurrence in several countries, such as Papua New
Guinea, Afghanistan, Sierra Leone, and the Congo
Democratic Republic. One country that stands out is
Papua New Guinea, where 38.9% of the women were
subjected to IPV co-occurrence. This could be attributed
to women marginalised social status.
22
There is an
enduring and dominating silence about women’s expe-
riences of IPV in Papua New Guinea due to traditional
gender and societal norms, belief systems, economic
Fig. 3: Prevalence of co-occurrence of multiple forms of intimate partner violence against women aged 15–49 years in the past 12
months across 49 LMICs, % (95% confidence interval). Note: LMICs: Low- and middle-income countries; Physical & Sexual IPV: co-occurrence
of physical and sexual intimate partner violence; Physical & Psychological IPV: co-occurrence of physical and psychological intimate partner
violence; Psychological & Sexual IPV: co-occurrence of psychological and sexual intimate partner violence; All three forms of IPV: co-occurrence
of all three forms of intimate partner violence.
Overall Low-income countries Lower-middle-income countries Upper-middle-income countries
IPV co-occurrence IPV co-occurrence IPV co-occurrence IPV co-occurrence
Physical & Sexual IPV Physical & Sexual IPV Physical & Sexual IPV Physical & Sexual IPV
Physical & Psychological IPV Physical & Psychological IPV Physical & Psychological IPV Physical & Psychological IPV
Psychological & Sexual IPV Psychological & Sexual IPV Psychological & Sexual IPV Psychological & Sexual IPV
All three forms of IPV All three forms of IPV All three forms of IPV All three forms of IPV
Prevalence, %
25
20
15
10
5
0
10
8
5
2
0
Education of Women No education Primary education Secondary education Higher education
25
20
15
10
5
0
25
20
15
10
5
0
25
20
15
10
5
0
25
20
15
10
5
0
25
20
15
10
5
0
25
20
15
10
5
0
25
20
15
10
5
0
25
20
15
10
5
0
25
20
15
10
5
0
25
20
15
10
5
0
25
20
15
10
5
0
10
8
5
2
0
10
8
5
2
0
10
8
5
2
0
10
8
5
2
0
10
8
5
2
0
10
8
5
2
0
10
8
5
2
0
Fig. 4: Adjusted prevalence of co-occurrence of multiple forms of intimate partner violence against women aged 15–49 years in past 12
months according to income group and education of women, % (95% confidence interval). Note: Physical & Sexual IPV: co-occurrence of
physical and sexual intimate partner violence; Physical & Psychological IPV: co-occurrence of physical and psychological intimate partner
violence; Psychological & Sexual IPV: co-occurrence of psychological and sexual intimate partner violence; All three forms of IPV: co-occurrence
of all three forms of intimate partner violence. The y-axis range is set from 0 to 10 for co-occurrence of psychological and sexual IPV, and co-
occurrence of all three forms of IPV. The prevalence is adjusted to maternal age, child marriage, place of residence, parity number, women’s
employment, marital status, household wealth index, and contraception use.
Articles
www.thelancet.com Vol 82 April, 2025 9
All three forms of IPV
Physical & Psychological IPV Psychological & Sexual IPV
IPV co−occurrence Physical & Sexual IPV
0.0 0.5 1.0 1.5 2.0
0.0 0.5 1.0 1.5 2.0 0.0 0.5 1.0 1.5 2.0
0.0 0.5 1.0 1.5 2.0 0.0 0.5 1.0 1.5 2.0
OR (95% CI)
Education of Women No Education Primary Education Secondary Education Higher Education
Fig. 5: Odds ratios for association between co-occurrence of multiple forms of intimate partner violence against women aged 15–49
years in past 12 months and education of women in adjusted models, odds ratio (95% confidence interval). Note: IPV: Intimate partner
violence; Physical & Sexual IPV: co-occurrence of physical and sexual intimate partner violence; Physical & Psychological IPV: co-occurrence of
physical and psychological intimate partner violence; Psychological & Sexual IPV: co-occurrence of psychological and sexual intimate partner
violence; All three forms of IPV: co-occurrence of all three forms of intimate partner violence. The reference group is women with primary
education. The logistic regression models are adjusted to maternal age, child marriage, place of residence, parity number, women’s employment,
marital status, household wealth index, and contraception use.
Articles
10 www.thelancet.com Vol 82 April, 2025
resources, and more.
23
Shockingly, 68.9% of women in
Papua New Guinea justified physical IPV.
24
This wide-
spread acceptance of violence is a risk factor for all
forms of IPV—physical, sexual, and psychological—
thereby increasing the likelihood of IPV co-occurrence.
Despite criminalizing marital rape through the Sexual
Offences and Crimes against Children Act 2002 and
ratifying the Convention on the Elimination of All
Forms of Discrimination against Women, IPV remains
neglected in Papua New Guinea and other LMICs due to
local understandings and perceptions of marriage.
25
This neglect is largely driven by deeply entrenched
local understandings and perceptions of marriage that
often justify or minimize violence.
26,27
Our findings
highlight the importance of developing and imple-
menting more effective and tailored interventions to
empower women and address IPV in Papua New
Guinea.
We also found that women with higher education
consistently experience a 60%–70% lower risk of IPV co-
occurrence and its subtypes compared to women with
only primary education. This aligns with substantial
evidence showing that higher education reduces IPV
risk.
14,16,28,29
For example, a multi-national study of 44
countries identified education as a protective factor,
16
and another study found that increasing a woman’s
schooling by one year decreased her probability of
recent poly-victimization by 1 percentage point and
lifetime poly-victimization by 2 percentage points.
30
The
consistency of our findings highlights that the protective
effects of education extend beyond single form of IPV to
the co-occurrence of multiple types, underscoring the
broader relevance of education in mitigating complex
patterns of violence. Evidence suggests that education
lowers IPV risk through three pathways: enhancing
women’s personal resources, delaying family formation,
and influencing partner selection.
30
It strengthens
cognitive skills and job opportunities, reducing eco-
nomic dependence and empowering women to leave
abusive relationships.
31
It also delays marriage and
childbirth, allowing women to accumulate resources
and negotiate better within relationships.
31
Additionally,
higher education increases the likelihood of choosing
partners with similar education and financial stability,
reducing economic strain and related stressors that can
contribute to IPV.
32,33
However, unequal access to edu-
cation globally may limit these findings’applicability
and exacerbate IPV risks. In Afghanistan, where over
1.4 million girls are banned from secondary education,
the lack of schooling exacerbates gender inequalities,
34
limiting women’s ability to recognise, report, or escape
IPV. These disparities highlight the critical role of ed-
ucation in preventing IPV and underscore the need for
targeted efforts to expand educational access, particu-
larly in low-resource settings.
Furthermore, we found that women with primary
education had the highest prevalence of subtypes of IPV
co-occurrence involving sexual violence. Some previous
studies also reported similar findings. For example, a
study in Kenya found that women living in rural areas
with less than a primary education were 35% less likely
to experience sexual violence compared to those with a
primary education.
35
Research conducted in Malawi
revealed that women exposed to the Universal Primary
Education reform are more likely to justify intimate
partner violence and experience sexual violence.
36
These
findings suggest that while higher education empowers
women with greater awareness of their rights and ca-
pacity to protect themselves, primary education may
offer limited protection against sexual violence. One
possible explanation is that primary education may
provide women with some degree of autonomy and
awareness but not enough to fully protect them against
IPV, particularly in contexts with entrenched gender
norms.
36
In patriarchal societies, increased education
levels may challenge traditional power dynamics,
potentially provoking violent responses from male
partners.
37
Furthermore, women with primary educa-
tion may lack access to resources or support systems to
navigate or escape abusive relationships.
38
Thus, efforts
of extending and amending compulsory schooling pol-
icies according to specific countries would empower
women and reshape gender norms, helping to alleviate
multiple forms of IPV by ensuring that women at all
educational levels are aware of their rights and mea-
sures to protect themselves from violence.
This study had several limitations. First, all IPV es-
timates in this study relied on women’s self-reported
experiences, which are prone to recall and reporting
bias. This is particularly true for sexual violence, which
is highly sensitive and often underreported due to
embarrassment, stigma, or fear of retaliation, with
willingness to disclose potentially varying by educational
level.
39,40
Additionally, the large teams and long in-
struments used in DHS fieldwork may influence
reporting, as non-DHS surveys often indicate higher
IPV prevalence. However, the inclusion of follow-up
questions during the survey helped to significantly
reduced these potential biases. Second, this study
included only countries that were surveyed for IPV,
potentially skewing representation toward nations
where governments prioritise these issues and choose to
include the domestic violence module. Third, there are
variations in definitions and measurements of psycho-
logical IPV across surveys and countries. However, the
DHS has been incorporating the best approaches to
researching psychological violence against women with
act-based standardised questions based on constructs
that have been validated. Furthermore, this study fo-
cuses on education due to its unique capacity to drive
structural and cultural change and its relevance as a
scalable intervention for reducing IPV. Future research
could explore the combined effects of wealth and other
socioeconomic characteristics. Finally, because the
Articles
www.thelancet.com Vol 82 April, 2025 11
study included only 49 LMICs, the findings may not be
representative of all LMICs.
Despite its limitations, our study provides important
insights into the pervasive issue of IPV co-occurrence in
LMICs. We conducted comprehensive multi-country
analyses with substantial sample sizes, which en-
hances the generalizability of our findings. One of the
key strengths of our study is the focus on the complex
relationship between education levels and the risk of
IPV co-occurrence, offering a nuanced understanding of
how educational inequality influences these outcomes
across different contexts.
Our findings highlight the critical association be-
tween education levels and IPV co-occurrence, sug-
gesting that various forms of intimate violence often
arise from similar individual, relationship, community,
and societal risk factors. This co-occurrence presents
significant opportunities for prevention.
41
While educa-
tion programs are typically driven by social and eco-
nomic goals, their role in preventing IPV is rarely
acknowledged. This study underscores girls’education
as an actionable strategy to reduce violence against
women, framing IPV prevention as an added value to
strengthen education investment cases. We advocate for
the urgent implementation of integrated programs in
LMICs that specifically target the co-occurrence of IPV.
These programs should address shared risk factors and
prioritise educational empowerment as a central strat-
egy to reduce IPV co-occurrence.
Contributors
SZ, SL, and JG had full access to all the data in the study, accessed and
verified the underlying data, and take responsibility for the integrity of the
data and the accuracy of the data analysis. SZ, SL, JG and ZL conceptualised
and designed the study, and SZ, SL, and JG led the data analysis and
interpretation. SZ, SL, and JG wrote the initial manuscript. NM, SC, JC, RK,
PK, JN, JR, MA, MNK, HM, CL, SVS, FC, JSJ, KT, and PG contributed to
the study design, data analysis, interpretation of the results, and writing. All
authors contributed to the critical revision of the manuscript for important
intellectual content. ZL provided overall supervision of the study. All the
authors approved the final submission of the study and accept responsibility
for the decision to submit for publication.
Data sharing statement
The data used in the analyses is publicly available upon request from the
DHS. More information on DHS and access to the survey datasets can
be found at https://dhsprogram.com/.
Declaration of interests
All authors declare no competing interests.
Acknowledgements
This research is supported by the China National Natural Science
Foundation (Grant numbers 72203119) and The Research Fund, Vanke
School of Public Health, Tsinghua University, China.
Appendix A. Supplementary data
Supplementary data related to this article can be found at https://doi.
org/10.1016/j.eclinm.2025.103150.
References
1World Health Organization. Global and regional estimates of violence
against women: prevalence and health effects of intimate partner violence
and non-partner sexual violence. Geneva: World Health Organiza-
tion; 2013.
2Sardinha L, Maheu-Giroux M, Stöckl H, Meyer SR, García-
Moreno C. Global, regional, and national prevalence estimates of
physical or sexual, or both, intimate partner violence against
women in 2018. Lancet. 2022;399:803–813.
3Ma N, Chen S, Kong Y, et al. Prevalence and changes of intimate
partner violence against women aged 15 to 49 years in 53 low-
income and middle-income countries from 2000 to 2021: a sec-
ondary analysis of population-based surveys. Lancet Global Health.
2023;11:e1863–e1873.
4 The sustainable development goals (SDGs) goal 5.2 & goal 16.1.
https://sdgs.un.org/goals. Accessed May 7, 2024.
5Ouellet F, Hetroy E, Patard G, Gauthier-Davies C, Leclerc C. Co-
occurrence of violence on the severity of abuse in intimate re-
lationships. J Interpers Violence. 2022;37:NP21927–N21952.
6Winstok Z. Conflict escalation to violence and escalation of violent
conflicts. Child Youth Serv Rev. 2008;30:297–310.
7Simmons J, Wijma B, Swahnberg K. Lifetime co-occurrence of
violence victimisation and symptoms of psychological ill health: a
cross-sectional study of Swedish male and female clinical and
population samples. BMC Public Health. 2015;15:1–14.
8Gottlieb A, Mahabir M. The effect of multiple types of intimate
partner violence on maternal criminal justice involvement.
J Interpers Violence. 2021;36:6797–6820.
9 Valério ID, Soares ALG, Leite de Moraes C, Gonçalves H. SciELO
Brazil - prevalence, co-occurrence, and associated factors of inti-
mate partner violence among Brazilian university students. https://
www.scielo.br/j/csc/a/MLFDmycGk5n3qJFdPpmrzMt/?lang=en.
Accessed January 19, 2025.
10 Mossie TB, Mekonnen FH, Tadesse M, Tadele A. Mapping the
disparities in intimate partner violence prevalence and de-
terminants across Sub-Saharan Africa. Front Public Health. 2023;11:
1188718.
11 Blasdell R. The co-occurrence of physical and sexual intimate
partner violence among us college females. Int J Crim Justice Sci.
2021;16:97–109.
12 Hamby S, Grych J, Hamby S, Grych J. Tracing the threads of the web:
the epidemiology of interconnections among forms of violence and victim-
ization. The web of violence: exploring connections among different forms
of interpersonal violence and abuse. 2013:9–27.
13 Wang L. Factors influencing attitude toward intimate partner
violence. Aggress Violent Behav. 2016;29:72–78.
14 Oluwagbemiga A, Johnson A, Olaniyi M. Education and intimate
partner violence among married women in Nigeria: a multilevel
analysis of individual and community-level factors. J Interpers
Violence. 2023;38:3831–3863.
15 Peterman A, Behrman J, Palermo T. Primary education and adult
women’s experience of intimate partner violence: quasi-experimental
evidence from government education policies. Sexual Violence Research
Initiative. 2014.
16 Heise LL, Kotsadam A. Cross-national and multilevel correlates of
partner violence: an analysis of data from population-based surveys.
Lancet Global Health. 2015;3:e332–e340.
17 USAID. The DHS program: available datasets. https://dhsprogram.
com/data/available-datasets.cfm. Accessed May 7, 2024.
18 Hacıaliefendio˘
glu A, Yılmaz S, Koyutürk M, Karakurt G. Co-
Occurrence patterns of intimate partner violence. In: Bio-
computing 2 021. Kohala Coast, Hawaii, USA: World Scientific;
2020:79–90.
19 Sullivan TP, McPartland TS, Armeli S, Jaquier V, Tennen H. Is it
the exception or the rule? Daily co-occurrence of physical, sexual,
and psychological partner violence in a 90-day study of substance-
using, community women. Psychol Violence. 2012;2:154.
20 Taylor BG, Stein ND, Mumford EA, Woods D. Shifting boundaries:
an experimental evaluation of a dating violence prevention program
in middle schools. Prev Sci. 2013;14:64–76.
21 Jewkes R, Nduna M, Levin J, et al. Impact of stepping stones on
incidence of HIV and HSV-2 and sexual behaviour in rural South
Africa: cluster randomised controlled trial. BMJ. 2008;337. https://
doi.org/10.1136/bmj.a506.
22 Gender-based Violence in Papua New Guinea: Addressing the
Issue. United nations population Fund. UNFPA; 2021. https://www.
unfpa.org/publications. Accessed September 2, 2024.
23 Kelly-Hanku A, Aeno H, Wilson L, et al. Transgressive women
don’t deserve protection: young men’s narratives of sexual violence
against women in rural Papua New Guinea. Cult Health Sex.
2016;18:1207–1220.
Articles
12 www.thelancet.com Vol 82 April, 2025
24 Adu C, Asare BY-A, Agyemang-Duah W, Adomako EB,
Agyekum AK, Peprah P. Impact of socio-demographic and economic
factors on intimate partner violence justification among women in
union in Papua New Guinea. Arch Public Health. 2022;80:136.
25 Zorn J. Engendering violence in the Papua New Guinea courts:
sentencing in rape trials. Engendering Violence in Papua New Guinea.
2012:163.
26 Heise L. What works to prevent partner violence? An evidence overview.
2011.
27 Jewkes R, Abrahams N. The epidemiology of rape and sexual
coercion in South Africa: an overview. Soc Sci Med. 2002;55:1231–
1244.
28 Nabaggala MS, Reddy T, Manda S. Effects of rural–urban residence
and education on intimate partner violence among women in Sub-
Saharan Africa: a meta-analysis of health survey data. BMC Womens
Health. 2021;21:1–23.
29 Osorio E. The effect of education on intimate partner violence: evidence
from an education reform in Tanzania. 2023. Available at: SSRN
4526328.
30 Weitzman A. Does increasing women’s education reduce their risk
of intimate partner violence? Evidence from an education policy
reform. Criminology. 2018;56:574–607.
31 Kalmuss DS, Straus MA. Wife’s marital dependency and wife
abuse. In: Physical violence in American families. Routledge;
2017:369–382.
32 Felson RB, Messner SF. The control motive in intimate partner
violence. Soc Psychol Q. 2000:86–94.
33 Fox GL, Benson ML, DeMaris AA, Van Wyk J. Economic distress
and intimate violence: testing family stress and resources theories.
J Marriage Fam. 2002;64:793–807.
34 Afghanistan Humanitarian Needs and Response Plan 2024. United
nations office for the coordination of humanitarian affairs (OCHA).
https://www.unocha.org/publications/report/afghanistan/afghanistan-
humanitarian-needs-and-response-plan-2024-december-2023-endarips;
2023. Accessed May 15, 2024.
35 Abuya BA, Onsomu EO, Moore D, Piper CN. Association between
education and domestic violence among women being offered an
HIV test in urban and rural areas in Kenya. J Interpers Violence.
2012;27:2022–2038.
36 Abrahamsson S. Does primary education affect intimate partner
violence against women? Evidence from Malawi; 2016.
37 True J. From domestic violence to war crimes: the political economy of
violence against women. The Political Economy of Violence against
Women. 2012:3–16.
38 Bloch F, Rao V. Terror as a bargaining instrument: a case study of
dowry violence in rural India. Am Econ Rev. 2002;92:1029–1043.
39 Scurich N. Introduction to this special issue: underreporting of
sexual abuse. Behav Sci Law. 2020;38:537–542.
40 Spohn R, Bjornsen A, Wright EM. Factors associated with reporting
of sexual assault among college and non-college women. J Aggress
ConflPeace Res. 2017;9:279–289.
41 CDC. Intimate partner violence prevention: resource for action.
https://www.cdc.gov/violenceprevention/intimatepartnerviolence/
index.html. Accessed May 7, 2024.
Articles
www.thelancet.com Vol 82 April, 2025 13
