https://orcid.org/0000-0002-3511-7562
Figures
Abstract
Background
Intimate partner violence (IPV) is a widespread public health concern that disproportionately affects women of reproductive age, particularly during pregnancy. Antenatal IPV (AN-IPV) is associated with adverse maternal and infant health outcomes, including poor breastfeeding practices. In Ethiopia, despite the high prevalence of AN-IPV, its association with breastfeeding outcomes remains poorly understood. Therefore, this study aimed to investigate the relationship between AN-IPV and breastfeeding indicators in Ethiopia.
Methods
Data from the Performance Monitoring for Action (PMA) Ethiopia, a nationally representative survey conducted between November 2021 and October 2022, were used. A total weighted sample of 1, 610 postpartum mother-child pairs was included. Data collected at baseline (during enrolment) and six weeks postpartum were used for this analysis. We fitted multilevel binary logistic regression models to estimate the associations between any AN-IPV, as well as physical and sexual AN-IPV separately, and early initiation of breastfeeding (EIBF) and exclusive breastfeeding (EBF), accounting for clustering within enumeration areas. Adjusted Odds Ratios (AORs) with 95% Confidence Intervals (CIs) were calculated to examine associations, and two-sided p-values (<0.05) were used to determine statistical significance.
Results
The prevalence of AN-IPV among mothers was 10.2%, with 6.9% reporting sexual AN-IPV and 4.4% reporting physical AN-IPV. Mothers exposed to any form of AN-IPV had a lower likelihood of initiating breastfeeding within the first hour of birth (AOR = 0.60, 95% CI: 0.36–0.99). Sexual AN-IPV was also associated with a reduced likelihood of EIBF (AOR = 0.50, 95% CI: 0.26–0.95, Model 2); however, this association did not remain in the fully adjusted model (AOR = 0.54, 95% CI: 0.28–1.04, Model 3). We found no evidence of an association between AN-IPV and EBF at six-weeks postpartum.
Conclusions
Our study revealed that maternal exposure to any form of AN-IPV (physical or sexual) was associated with reduced EIBF. However, no association was observed between AN-IPV and EBF at six weeks postpartum. The findings highlight the need for routine identification of women who have experienced or are experiencing AN-IPV, and for the development and provision of targeted, trauma-informed interventions aimed at supporting optimal breastfeeding practices.
Citation: Azene ZN, Dachew B, Reilly N, Damtew SA, MacPhail C (2026) Antenatal intimate partner violence and breastfeeding practices: Evidence from a national longitudinal study in Ethiopia. PLoS One 21(4): e0347323. https://doi.org/10.1371/journal.pone.0347323
Editor: Ari Samaranayaka, University of Otago, NEW ZEALAND
Received: October 26, 2025; Accepted: March 30, 2026; Published: April 24, 2026
Copyright: © 2026 Azene et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data underlying this study are publicly accessible through the Johns Hopkins University (JHU) data repository at https://doi.org/10.34976/8RY9-ZH50 and https://doi.org/10.34976/8R5S-DX31.
Funding: The PMA Ethiopia was funded by the Bill & Melinda Gates Foundation (INV 009466). The funders had no involvement in the study design, data collection, analysis, publication decisions, or manuscript preparation. The first author (ZNA) was supported by an Australian Government Research Training Program (RTP) Scholarship (doi.org/10.82133/C42F-K220).
Competing interests: The authors declare no competing interests.
Abbreviations: AN-IPV, Antenatal Intimate Partner Violence; ANC, Antenatal Care; AOR, Adjusted Odds Ratio; CI, Confidence Interval; CSA, Central Statistical Agency; DAG, Directed Acyclic Graph; DIC, Deviance Information Criterion; EA, Enumeration Area; EBF, Exclusive Breastfeeding; EIBF, Early Initiation of Breastfeeding; HCP, Health Care Provider; HEW, Health Extension Worker; HREC, Human Research Ethics Committee; ICC, Intra-class Correlation Coefficient; IRB, Institutional Review Board; IPV, Intimate Partner Violence; JHSPH, Johns Hopkins Bloomberg School of Public Health; LAM, Lactational Amenorrhea Method; LMICs, Low- and Middle-Income Countries; ORS, Oral Rehydration Salts; PMA, Performance Monitoring for Action; PNC, Postnatal Care; PPFP, Postpartum Family Planning; UOW, University of Wollongong; UNICEF, United Nations Children’s Fund; WHO, World Health Organization
Introduction
Intimate partner violence (IPV) is a significant public health issue involving physical, sexual, or psychological harm, stalking, or reproductive control by a current or former partner [1,2]. It is the most common form of violence against women, affects about one in three women globally, and spans all ethnic, economic, religious, and sexual backgrounds [3]. The prevalence of IPV is particularly high in sub-Saharan Africa, where 46.5% of women have experienced at least one form of violence [4]. In Ethiopia, IPV remains a pervasive issue in both urban and rural communities [5]. A systematic review conducted in the country reported that the lifetime prevalence of physical violence ranged from 31% to 76.5%, sexual violence from 19.2% to 59%, and emotional violence at 51.7% [5].
IPV is more prevalent among women of reproductive age, particularly during pregnancy, which can be a time of increased vulnerability due to various physical, emotional, social, and economic changes and demands [6]. A systematic review conducted globally found that the overall prevalence of antenatal IPV (AN-IPV) ranges widely from 1.5% to 66.9%. In Africa, the prevalence of AN-IPV ranges from 2% to 57%, with an overall prevalence of 15.23% [7]. However, more recent data from 23 sub-Saharan African countries reported a significantly higher pooled prevalence of 41.94% among pregnant women [8].
A multi-country study by the World Health Organization (WHO) on women’s health and domestic violence found that 8% of women in Ethiopia experienced AN-IPV [9]. In contrast, a more recent systematic review reported a higher prevalence of 26.1% among pregnant women [10]. However, despite these differences in estimates, both figures may be underestimates due to the widespread underreporting of IPV and related women’s health issues [11].
AN-IPV poses a significant threat to the health and well-being of both women and their infants [12]. Studies have shown that AN-IPV is associated with numerous adverse maternal and infant health outcomes, including low birth weight, premature delivery, miscarriage, abortion, antepartum haemorrhage, intrauterine growth restriction, induced abortion, spontaneous abortion, hypertension, pre-eclampsia, third-trimester bleeding, and sexually transmitted infections [13–19]. Additionally, AN-IPV exposure is strongly associated with poor perinatal mental health outcomes, including depression, anxiety disorders, and post-traumatic stress disorder (PTSD) [18,20]. Furthermore, AN-IPV is believed to negatively affect maternal breastfeeding practices, including early initiation of breastfeeding (EIBF) and exclusive breastfeeding (EBF) [21–26]. However, some studies found no significant association between AN-IPV and breastfeeding indicators (EIBF and EBF) [24,27,28], highlighting inconsistencies in the current body of literature.
AN-IPV may influence breastfeeding practices such as EIBF and EBF through biological, psychological, and social pathways. Biologically, AN-IPV acts as a chronic stressor, triggering the hypothalamic-pituitary-adrenal (HPA) axis and elevating stress hormones such as cortisol and catecholamines. This dysregulation can interfere with hormonal processes vital for lactation, notably by altering prolactin and oxytocin levels—prolactin being essential for milk production and oxytocin for the milk ejection reflex. Such disruptions can impair milk synthesis and let-down, potentially delaying initiation and shortening EBF duration [29–31].
Psychologically, AN-IPV may cause adverse mental health effects such as depression, anxiety, and trauma-related distress [32–36]. These conditions have been associated with reduced breastfeeding self-efficacy and confidence, breastfeeding anxiety, and earlier EBF cessation [37–41]. Trauma symptoms such as fear and hypervigilance may further hinder maternal–infant bonding and responsive feeding [42]. Socially, IPV can restrict maternal autonomy and access to support networks due to controlling behaviours, reduced social and family support, and isolation [43–45]. These factors limit practical support and access to healthcare, weakening the environmental conditions necessary for both EIBF and sustained EBF [46–49].
In Ethiopia, as in many other countries, achieving optimal breastfeeding practices remains a significant public health challenge [50]. The WHO recommends that all newborns be breastfed within the first hour of birth and be exclusively breastfed for the first six months [51]. However, according to the 2016 Ethiopian Demographic and Health Survey (EDHS), only 73% of infants were breastfed within the first hour after birth, and 58% of children under six months were exclusively breastfed [52]. Furthermore, a recent systematic review reported a prevalence of early EBF cessation of 43.31% in Ethiopia [53]. Despite existing challenges, Ethiopia maintains relatively high rates of EBF under six months compared to many developed countries [54].
While several studies in Ethiopia have investigated the association between IPV and maternal morbidity and child nutrition, no study to date has specifically examined the association between AN-IPV and breastfeeding indicators. Therefore, this study aims to investigate the association between AN-IPV and breastfeeding indicators (EIBF and EBF) using nationally representative longitudinal data from the Performance Monitoring for Action (PMA)-Ethiopia.
Methods
Data source and study populations
This study used data from the PMA Ethiopia dataset. PMA-Ethiopia is a survey project that builds on the work of PMA2020 and the PMA Maternal and Newborn Health projects, aiming to provide timely, actionable data on various reproductive, maternal, and newborn health (RMNH) indicators through both cross-sectional and longitudinal data collection methods. The PMA-Ethiopia survey is a nationwide study conducted in partnership with Addis Ababa University (AAU), the Ethiopian Federal Ministry of Health (FMOH), and the Johns Hopkins Bloomberg School of Public Health (JHSPH) [55].
The second cohort of PMA Ethiopia was conducted between 2021 and 2023 and enrolled women at baseline, with follow-up at six weeks, six months, and one year postpartum. We used linked data from the second cohort’s baseline and six-week postpartum surveys for this analysis. These cohort and data collection points were selected for inclusion in this study because they capture women’s exposure to AN-IPV and breastfeeding outcomes. EBF beyond six-weeks postpartum was outside the scope of the current study because breastfeeding practices were not assessed in the subsequent 6-month and 1-year follow-up surveys. The baseline interview collected data on women’s sociodemographic characteristics, whereas information on exposure (AN-IPV), outcomes (EIBF and EBF), and maternal health service-related variables included in this analysis was collected during the six-week postpartum follow-up interview.
The PMA Ethiopia panel’s second cohort six-week follow-up survey included women who had participated in the cohort two baseline survey. Verbal informed consent was obtained from all participants during the screening process and prior to enrolment [55]. Women’s estimated or actual delivery dates were used to schedule a second interview. Among 2,297 eligible women, 2,072 women completed the six-weeks follow-up survey. This analysis included 1,841 women with live births; women with abortions, miscarriages, and stillbirths were excluded. The pathway to recruitment and sample selection for the six-week postpartum follow-up survey of PMA-Ethiopia is summarised in Fig 1.
All data used in this analysis were collected between November 2021 and November 2022. The cohort was drawn from three major regions: Oromia, Amhara, and formerly Southern Nations Nationalities and Peoples’ (SNNP) regions (now subdivided into four regions, namely, South Ethiopia Regional State, Southwest Ethiopia Peoples’ Regional State, Central Ethiopia Regional State, and Sidama Regional State) and one city administration (Addis Ababa). A multistage stratified cluster sampling technique guided data collection. In the first stage, the Amhara, Oromia, and SNNP regions were stratified into urban and rural areas, while Addis Ababa was treated as a single urban stratum. Using the Central Statistical Services (CSS) sampling frame, a total of 162 enumeration areas (EAs) were selected across these strata through probability proportional to size sampling. In the second stage, 35 households were randomly selected within each EA using simple random sampling. Further details regarding the sampling design, selection procedures, and implementation of fieldwork are described elsewhere [55].
Study variables and measurements
Outcome variables.
The outcome variables in this study were EIBF and EBF. EIBF was measured by asking mothers, “How long after birth did you first put the baby to the breast?” Responses were recorded in minutes, hours, or days. Based on these responses, a binary variable was generated and categorised as follows: (1) EIBF, if the baby was breastfed within the first hour after birth, and (2) delayed initiation, if breastfeeding began more than one hour after birth [56]. This information was self-reported by mothers who were nearly six weeks postpartum at the time of the interview. EBF was measured as the proportion of infants who had received only breast milk, without any additional liquids or solid foods, on a 24-hour recall at the six-weeks follow-up visit. This excluded medications, immunisations, oral rehydration salts (ORS), or vitamin supplements. For analysis, a binary variable was created: infants who received only breast milk in the past 24 hours were coded as ‘1’ (EBF), and infants who received any other foods or liquids were coded as ‘0’ (non-EBF) [55].
Exposure variable.
The exposure variable in this study was AN-IPV, defined as any form of abuse experienced during pregnancy by individuals in an intimate relationship, including a current or former partner [57]. AN-IPV was assessed at the six-week postpartum follow-up interview using a set of 10 dichotomous (yes/no) items; seven related to physical violence and three to sexual violence (S1 Table). Postpartum women were specifically asked about their experiences of IPV during pregnancy using the question: “At any time during your pregnancy, did your husband/partner do any of the following things to you?” Women who responded “yes” to at least one of these items were classified as having experienced AN-IPV; those who answered “no” to all items were considered not to have experienced AN-IPV [58]. Psychological/emotional IPV data were not collected.
A priori confounding variables.
Covariates included in the analysis were selected based on theoretical considerations and informed by a Directed Acyclic Graph (DAG) developed to identify potential confounders in the relationship between AN-IPV and breastfeeding outcomes (EIBF and EBF). We assumed that the two outcomes, EIBF and EBF, share the same DAG; i.e., the same minimally sufficient set of variables was used in both the EIBF and EBF models (Fig 2). These variables were grouped into five broad categories: sociodemographic characteristics (maternal age, maternal educational status, partner’s educational status, marital status, partner’s age, maternal number of marital unions, family size, place of residence (urban/rural), region, religion, and household wealth index), partner dynamics and relationship factors (partner encouragement for antenatal care (ANC) and postnatal care (PNC), whether the partner had multiple wives, and the belief that using family planning may lead the partner to seek another wife); reproductive and obstetric factors (parity, pregnancy intention status, feelings about the current pregnancy, type of pregnancy (singleton or multiple), number of ANC visits with Health Extension Workers (HEWs) and with other health care providers (HCPs), experience of danger signs during pregnancy, presence of obstetric complications, use of maternity waiting homes, mode of delivery, place of delivery, and skilled attendance at birth); newborn characteristics and birth practices (sex of the child, whether the newborn cried at birth, and whether there was skin-to-skin contact after birth); and PNC and family planning experiences and behaviours (PNC attendance, PNC within the first hour after birth, postpartum contraceptive use at six weeks, intention to use the Lactational Amenorrhea Method (LAM), and whether ANC counselling on postpartum family planning (PPFP) was provided). In total, 15 variables comprised the minimally sufficient adjustment set and were included in the multivariable models.
Cluster variable: Enumeration areas (EAs) were included as the cluster variable in the multilevel models, with mothers nested within clusters (ranging from 1 to 162 EAs).
Stratified variables included AN-IPV types, urban/rural residence, and region.
Statistical analysis
Data preparation was performed using Stata/SE version 17.0.. Data were checked and cleaned before any statistical analysis. The “svyset” command in Stata was used to account for the complex survey design. Sampling weights were applied using the “pweight” command for both descriptive and regression analyses. The PMA Ethiopia data have a hierarchical structure, with women nested within clusters. We assume that women within the same cluster share similar characteristics, which violates the standard assumption of independence of observations in traditional logistic regression. Hence, to account for variability across clusters, a multilevel binary logistic regression model was employed to examine the association between AN-IPV and breastfeeding indicators (EIBF and EBF), and odds ratios (ORs) with 95% confidence intervals (CIs) were calculated to quantify the strength and statistical significance. The Intra-class Correlation Coefficient (ICC) was calculated to assess the variation in breastfeeding outcomes (EIBF and EBF) between clusters. The ICC quantifies the proportion of the total observed variation in EIBF and EBF that is attributable to differences between clusters, reflecting the degree of heterogeneity across clusters. Model fit was assessed using the Deviance Information Criterion (DIC) to compare the sequential models. Because covariates were identified a priori using the DAG, DIC was used solely to evaluate relative model fit, not for variable selection.
A two-level multilevel binary logistic regression model was employed, with mothers at level 1 nested within clusters (enumeration areas) at level 2. Three models were fitted for each type of AN-IPV (physical, sexual, and any AN-IPV) separately. In each model, other forms of violence were not included as covariates. For example, in the model assessing the association between any AN-IPV and EIBF, physical and sexual AN-IPV were not included. The first model was unadjusted. The second model was adjusted for individual-level factors, including sociodemographic characteristics; partner dynamics and relationship factors; reproductive and obstetric factors; newborn characteristics; birth practices; and PNC and family planning experiences and behaviours. The final model was adjusted for all individual-level factors from Model 2 and community-level factors such as place of residence and region. The models were specified sequentially to illustrate the influence of progressive adjustment. Model 3, which included the full minimally sufficient adjustment set, was considered the primary inferential model. Models 1 and 2 are presented to illustrate the crude and partially adjusted associations for comparison.
Ethical considerations
Ethical approval was obtained from the University of Wollongong (UOW) Human Research Ethics Committee (HREC) (Ref: 2024/341), and access to the PMA Ethiopia dataset was approved by the data manager after completion of the online registration and data request process. The PMA Ethiopia survey also obtained ethical approval from JHSPH Institutional Review Board (FWA00000287) and from the Addis Ababa University, College of Health Sciences (Ref: AAUMF03–008). Use of this dataset was conducted in accordance with the necessary legal and institutional guidelines.
Results
Maternal and child characteristics
A total of 1,841infants and their mothers were included in the study. After applying sampling weights to account for the survey design and non-response, the weighted sample size was 1,610, which reflects the effective sample size used in the analyses. Among the mothers, 28.4% were aged 25–29 years, nearly half of the women (45.9%) had primary education, and the vast majority (97.4%) were married. Most participants resided in rural areas (73.8%), with more than half from the Oromia region (53.4%). Regarding religion, most women were Muslim (35.6%), followed closely by Orthodox (33.2%). Approximately 43.8% of women had 1–2 children, and 57% reported that their most recent pregnancy was intended. More than half (62.6%) were delivered in a health facility; 62.5% were attended by a skilled healthcare worker; and the majority of births occurred via vaginal delivery (spontaneous and assisted) (94.4%) (Table 1).
The majority of deliveries were singleton (98.3%), and nearly half of the women (45.8%) attended four or more ANC visits. Most women (95.6%) reported that their newborns cried at birth and breathed normally without requiring any resuscitation. More than half of mothers (56.8%) did not attend PNC after delivery (S2 Table).
Prevalence of AN-IPV
Figs 3 and 4 present the prevalence of AN-IPV by type and across regions, respectively. Overall, approximately one in ten women (10.2%) (n = 164) reported experiencing AN-IPV. Specifically, 6.9% (n = 111) experienced sexual AN-IPV, and 4.4% (n = 71) experienced physical. The prevalence of AN-IPV varied across regions, ranging from 3.8% in Addis Ababa to 14.4% in SNNP.
Breastfeeding characteristics
Tables 1 and S2 illustrate the distribution of EBF and EIBF by selected predictors among the cohort of mother-newborn pairs at six weeks postpartum. More than seven in ten (72.4%) initiated breastfeeding within the first hour after birth, and most mothers (75%) exclusively breastfed their infants at six weeks postpartum. EIBF was most common in Oromia (79.6%) and lowest in SNNP (63.6%). The highest EBF prevalence was observed in the Amhara region (83.4%), while the lowest was in Addis Ababa (63.7%). Breastfeeding outcomes varied by mode of delivery, with lower rates observed among caesarean deliveries (EIBF: 42.4%; EBF: 57.5%) compared to vaginal deliveries (EIBF: 74.1%; EBF: 76.1%).
The association between AN-IPV and EIBF
After accounting for a number of sociodemographic characteristics, partner and relationship dynamics, reproductive and obstetric factors, newborn characteristics, birth practices, as well as PNC and family planning experiences and behaviours in a fully adjusted model, we found that exposure to any form of AN-IPV was associated with 40% lower odds of EIBF within the first hour of birth compared to those who were not exposed (AOR = 0.60, 95% CI: 0.36–0.99). When stratified by AN-IPV type, physical AN-IPV was not associated with EIBF, whereas sexual AN-IPV was associated (AOR = 0.50, 95% CI: 0.26–0.95, Model 2), though this association did not remain significant after adjusting for place of residence and region in Model 3 (AOR = 0.54, 95% CI: 0.28–1.04) (Table 2).
The association between AN-IPV and EBF
Compared to mothers not exposed to any AN-IPV, those exposed had lower odds of EBF at six weeks postpartum, although the association was not statistically significant (AOR = 0.69, 95% CI: 0.42–1.13). Similar associations were found for those exposed to physical (AOR = 0.90, 95% CI: 0.49–1.63) and sexual AN-IPV (AOR = 0.50, 95% CI: 0.23–1.05) (Table 3).
Discussion
To the best of the investigators’ knowledge, this is the first study in Ethiopia to examine the association between AN-IPV (any, sexual, and physical) and breastfeeding indicators (EIBF and EBF). After adjusting for key confounders, our findings demonstrated that AN-IPV was significantly associated with lower odds of EIBF. Women exposed to AN-IPV were less likely to initiate breastfeeding within one hour of birth compared to those not exposed. In particular, sexual AN-IPV was associated with delayed initiation of breastfeeding after adjustment for a number of individual-level factors; however, this association was attenuated and became marginally non-significant after adjusting for place of residence and region in Model 3. In contrast, physical AN-IPV was not significantly associated with EIBF. We also found no evidence of an association between AN-IPV and EBF.
In the fully adjusted model, mothers exposed to any type of AN-IPV had 40% lower odds of initiating breastfeeding within the first hour of birth compared to those not exposed. This finding is consistent with earlier studies [22,59]. There are several possible reasons why mothers who experienced AN-IPV may be less likely to initiate breastfeeding within the recommended first hour after birth. One explanation is that IPV-related stress elevates maternal cortisol, which can disrupt oxytocin-mediated pathways, the hormone critical for milk let-down and bonding [60]. Chronic stressful violence and trauma have also been linked with delayed onset of milk production, making EIBF more difficult, due to suppression of the hormones responsible for lactation, such as prolactin [61]. Another possible mechanism is that AN-IPV may create stressful, unstable, and unsafe environments for mothers and interfere with routines essential for EIBF [62]. Moreover, exposure to AN-IPV can reduce a mother’s self-esteem and confidence [63], which may make her less likely to initiate breastfeeding [64].
Although our results indicated that AN-IPV was not significantly associated with EBF, this finding should be interpreted with caution. Several factors may explain the observed null association. First, the short follow-up period: EBF was assessed at six weeks postpartum rather than the WHO-recommended six months. Breastfeeding practices may change over time as maternal and infant factors evolve, and associations with AN-IPV may become more evident later in the postpartum period when sustained breastfeeding challenges arise [47]. Second, breastfeeding is nearly universal in Ethiopia [65]. In contexts characterised by near-universal breastfeeding, social networks, perceived norms, and strong family and community expectations play a central role in shaping mothers’ breastfeeding practices and success, and may buffer potential adverse influences, including AN-IPV [66]. Third, EBF was measured using a 24-hour recall method, which may overestimate the true EBF practices [67,68]. Finally, the relatively small number of women reporting physical and sexual AN-IPV may limit statistical power, which could mask real associations.
In contrast, a large population-based cross-sectional study conducted across 51 low- and middle-income countries reported that mothers exposed to any form of IPV (physical, sexual, and psychological) were less likely to exclusively breastfeed during the first six months [69].
Our study had several limitations. First, both AN-IPV exposure and breastfeeding outcomes (EIBF and EBF) were based on mothers’ self-reports collected via interviewer-administered questionnaires, which may have introduced social desirability bias. This could have resulted in underreporting of AN-IPV experiences or overreporting of adherence to recommended breastfeeding practices, potentially introducing misclassification bias. Such misclassification is likely non-differential and may have affected the observed associations. Second, AN-IPV was assessed retrospectively at the six-week postpartum interview. Although the questions were specifically framed to refer to experiences during pregnancy, self-reported data may be subject to recall bias. Misclassification of AN-IPV exposure due to imperfect recall is possible; it is likely non-differential and may have affected the observed associations with breastfeeding outcomes [70]. Third, women may be reluctant to disclose experiences of AN-IPV due to fear of potential consequences or social stigma, which may also result in underreporting of AN-IPV [71]. If mothers underreport their exposure to AN-IPV, and this misclassification is independent of EIBF and EBF, it could attenuate the observed associations with these outcomes toward the null [72]. Fourth, data on psychological/emotional forms of AN-IPV were not collected in the PMA Ethiopia survey. As a result, the prevalence of AN-IPV in this study likely underestimates the true burden of AN-IPV in the study population. The potential underreporting of IPV is reflected in the overall prevalence of AN-IPV (10.2%) reported by women in this study (ranging from 3.8% to 14.4% across regions), which is substantially lower than the prior pooled prevalence of 26.1% reported among pregnant women in Ethiopia [10]. Excluding psychological/emotional AN-IPV may also have influenced the observed associations, as women exposed solely to psychological/emotional violence would have been misclassified as unexposed. This non-differential misclassification could bias estimates toward the null, potentially attenuating true associations between AN-IPV and breastfeeding outcomes. Fifth, EBF was measured as the proportion of infants who received only breast milk in the 24 hours prior to the interview. While the 24-hour recall method is practical and reliable, it may overestimate the true prevalence of EBF compared to the strict definition of EBF from birth to six months. In our study, 75% of mothers reported EBF at six weeks postpartum, which is higher than the nationally weighted EBF prevalence of 58.97% at six months postpartum [73]. This discrepancy may partly reflect differences in the timing of assessment. Sixth, the relatively low prevalence of sexual and physical AN-IPV in our sample may have limited statistical power to detect associations, particularly in stratified and fully adjusted models examining AN-IPV subtypes. Finally, although we adjusted for several potential confounders identified a priori using a DAG, residual confounding by unmeasured factors such as current or previous maternal mental health disorders not captured in the PMA Ethiopia data should not be overlooked.
Despite these limitations, our study has several strengths. To our knowledge, this is the first study to examine the association between AN-IPV and breastfeeding outcomes (EIBF and EBF) among mothers living in Ethiopia. We used representative longitudinal survey data from PMA-Ethiopia to address this important evidence gap, and our findings provide insights into the intersection of maternal violence exposure and child feeding practices in the early postpartum period in a low-resource setting. Moreover, the application of a DAG, developed a priori, adds methodological rigour by providing a structured approach to identifying and controlling for confounding. This approach reduces the risk of inappropriate adjustment and enhances the validity and credibility of the estimated associations [74]. Finally, the use of survey-adjusted analyses to account for the complex sampling design improves the validity and generalisability of the findings.
Conclusions
This study found that maternal exposure to any form of AN-IPV (physical and sexual) is associated with EIBF in Ethiopia. Specifically, sexual AN-IPV but not physical AN-IPV shows a significant association with EIBF. In contrast, we found no evidence of an association between AN-IPV and EBF at six weeks postpartum. Our findings highlight the need for routine identification of women who have experienced or are experiencing AN-IPV, and for the development and provision of targeted, trauma-informed interventions aimed at supporting optimal breastfeeding practices. For this study, we used a binary approach to categorise AN-IPV exposure. Further research exploring the association between cumulative AN-IPV and breastfeeding outcomes, as well as the relationship between AN-IPV and breastfeeding outcomes beyond the early postpartum period among mothers in Ethiopia, is warranted.
Supporting information
S1 Table. Items used to measure Antenatal (AN) physical and sexual IPV (PMA Ethiopia cohort two, 2021–2023).
https://doi.org/10.1371/journal.pone.0347323.s001
(DOCX)
S2 Table. Distribution of EIBF and EBF by selected predictors among a cohort of mother-newborn pairs at six weeks postpartum in Ethiopia, 2021–2022.
https://doi.org/10.1371/journal.pone.0347323.s002
(DOCX)
Acknowledgments
The authors would like to thank the PMA Ethiopia project team for permission to access the PMA Ethiopia data. We would also like to acknowledge the University of Wollongong for providing a PhD scholarship to the principal investigator of this research.
References
- 1.
Breiding M, Basile KC, Smith SG, Black MC, Mahendra RR. Intimate partner violence surveillance: uniform definitions and recommended data elements. Version 2.0; 2015.
- 2.
Walters ML, Black MC, Basile KC, Breiding MJ, Smith SG, Merrick MT, et al. The national intimate partner and sexual violence survey (NISVS): 2010 summary report; 2011.
- 3. Devries KM, Mak JYT, García-Moreno C, Petzold M, Child JC, Falder G, et al. Global health. The global prevalence of intimate partner violence against women. Science. 2013;340(6140):1527–8. pmid:23788730
- 4. McCloskey LA, Boonzaier F, Steinbrenner SY, Hunter T. Determinants of intimate partner violence in Sub-Saharan Africa: a review of prevention and intervention programs. Partner Abuse. 2016;7(3):277–315.
- 5. Semahegn A, Mengistie B. Domestic violence against women and associated factors in Ethiopia; systematic review. Reprod Health. 2015;12:78. pmid:26319026
- 6. Mojahed A, Alaidarous N, Kopp M, Pogarell A, Thiel F, Garthus-Niegel S. Prevalence of intimate partner violence among intimate partners during the perinatal period: a narrative literature review. Front Psychiatry. 2021;12:601236. pmid:33633606
- 7. Shamu S, Abrahams N, Temmerman M, Musekiwa A, Zarowsky C. A systematic review of African studies on intimate partner violence against pregnant women: prevalence and risk factors. PLoS One. 2011;6(3):e17591. pmid:21408120
- 8. Melkam M, Fentahun S, Rtbey G, Andualem F, Nakie G, Tinsae T, et al. Domestic violence and associated factors among pregnant women in sub-Saharan African countries from the recent demographic and health survey data: a multilevel analysis. Front Public Health. 2024;12:1386524. pmid:39257957
- 9.
Organization WH. World health organization multi-country study on women’s health and domestic violence against women; 2004.
- 10. Alebel A, Kibret GD, Wagnew F, Tesema C, Ferede A, Petrucka P, et al. Intimate partner violence and associated factors among pregnant women in Ethiopia: a systematic review and meta-analysis. Reprod Health. 2018;15(1):196. pmid:30514311
- 11. Shanmugam D, Hou K, Pierson E. Quantifying disparities in intimate partner violence: a machine learning method to correct for underreporting. npj Womens Health. 2024;2(1).
- 12. Belay AS, Kassie Gidafie A, Abera Gudeta T. Perinatal intimate partner violence during COVID-19 and its associated factors among postpartum mothers attending newborn immunization in southern Ethiopia, 2021: a cross-sectional study. SAGE Open Med. 2022;10:20503121221116671. pmid:35983083
- 13. Rodrigues T, Rocha L, Barros H. Physical abuse during pregnancy and preterm delivery. Am J Obstet Gynecol. 2008;198(2):171.e1-6. pmid:17905171
- 14. Martin SL, Li Y, Casanueva C, Harris-Britt A, Kupper LL, Cloutier S. Intimate partner violence and women’s depression before and during pregnancy. Violence Against Women. 2006;12(3):221–39. pmid:16456149
- 15. Alhusen JL, Ray E, Sharps P, Bullock L. Intimate partner violence during pregnancy: maternal and neonatal outcomes. J Womens Health (Larchmt). 2015;24(1):100–6. pmid:25265285
- 16. Stöckl H, Filippi V, Watts C, Mbwambo JKK. Induced abortion, pregnancy loss and intimate partner violence in Tanzania: a population based study. BMC Pregnancy Childbirth. 2012;12:12. pmid:22390254
- 17. Donovan BM, Spracklen CN, Schweizer ML, Ryckman KK, Saftlas AF. Intimate partner violence during pregnancy and the risk for adverse infant outcomes: a systematic review and meta-analysis. BJOG. 2016;123(8):1289–99. pmid:26956568
- 18. Howard LM, Oram S, Galley H, Trevillion K, Feder G. Domestic violence and perinatal mental disorders: a systematic review and meta-analysis. PLoS Med. 2013;10(5):e1001452. pmid:23723741
- 19. Steele-Baser M. Intimate partner violence and pregnancy and infant health outcomes—pregnancy risk assessment monitoring system, nine US jurisdictions, 2016–2022. MMWR Morb Mortal Wkly Rep. 2024;73.
- 20. Paulson JL. Intimate partner violence and perinatal post-traumatic stress and depression symptoms: a systematic review of findings in longitudinal studies. Trauma Violence Abuse. 2022;23(3):733–47. pmid:33252020
- 21.
United Nations Children’s Fund. Intimate partner violence and breastfeeding: a summary of the evidence base; 2023 [cited 2025 Jun 2]. Available from: https://www.unicef.org/media/128256/file/UNICEF_GBViE_Breastfeeding_Brief_September_2022.pdf
- 22. Martin‐de‐las‐Heras S, Velasco C, Luna‐del‐Castillo JdD, Khan KS. Breastfeeding avoidance following psychological intimate partner violence during pregnancy: a cohort study and multivariate analysis. BJOG. 2019;126(6):778–83.
- 23. Khalid N, Zhou Z, Nawaz R. Exclusive breastfeeding and its association with intimate partner violence during pregnancy: analysis from Pakistan demographic and health survey. BMC Womens Health. 2024;24(1):186. pmid:38509533
- 24. Olubodun T, Asefa A, Banke-Thomas A, Balogun M, Okafor I, Odukoya O, et al. Maternal exposure to intimate partner violence and breastfeeding practices of children 0–23 months: findings from the 2018 Nigeria Demographic and Health Survey. J Glob Health Rep. 2023;7.
- 25. Ariyo T, Jiang Q. Intimate partner violence and exclusive breastfeeding of infants: analysis of the 2013 Nigeria demographic and health survey. Int Breastfeed J. 2021;16(1):15. pmid:33485361
- 26. Kjerulff Madsen F, Holm-Larsen CE, Wu C, Rogathi J, Manongi R, Mushi D, et al. Intimate partner violence and subsequent premature termination of exclusive breastfeeding: a cohort study. PLoS One. 2019;14(6):e0217479. pmid:31181090
- 27. James JP, Taft A, Amir LH, Agius P. Does intimate partner violence impact on women’s initiation and duration of breastfeeding? Breastfeed Rev. 2014;22(2):11–9. pmid:25109096
- 28. Wallenborn JT, Cha S, Masho SW. Association between intimate partner violence and breastfeeding duration: results from the 2004-2014 pregnancy risk assessment monitoring system. J Hum Lact. 2018;34(2):233–41. pmid:29596755
- 29.
Pramod DA. Link between cortisol levels and breastfeeding - connection and importance; 2024 [cited 2026 Feb 19]. Available from: https://www.icliniq.com/articles/womens-health/link-between-cortisol-levels-and-breastfeeding?utm_
- 30. Ziomkiewicz A, Babiszewska M, Apanasewicz A, Piosek M, Wychowaniec P, Cierniak A, et al. Psychosocial stress and cortisol stress reactivity predict breast milk composition. Sci Rep. 2021;11(1):11576. pmid:34078999
- 31. Boutet C, Vercueil L, Schelstraete C, Buffin A, Legros JJ. Oxytocin and maternal stress during the post-partum period. Ann Endocrinol (Paris). 2006;67(3):214–23. pmid:16840912
- 32. Chambliss LR. Intimate partner violence and its implication for pregnancy. Clin Obstet Gynecol. 2008;51(2):385–97. pmid:18463468
- 33. Kiely M, Gantz MG, El-Khorazaty MN, El-Mohandes AAE. Sequential screening for psychosocial and behavioural risk during pregnancy in a population of urban African Americans. BJOG. 2013;120(11):1395–402. pmid:23906260
- 34. Connelly CD, Hazen AL, Baker-Ericzén MJ, Landsverk J, Horwitz SM. Is screening for depression in the perinatal period enough? The co-occurrence of depression, substance abuse, and intimate partner violence in culturally diverse pregnant women. J Womens Health (Larchmt). 2013;22(10):844–52. pmid:23931153
- 35. Huth-Bocks AC, Krause K, Ahlfs-Dunn S, Gallagher E, Scott S. Relational trauma and posttraumatic stress symptoms among pregnant women. Psychodyn Psychiatry. 2013;41(2):277–301. pmid:23713621
- 36. Lawrence E, Yoon J, Langer A, Ro E. Is psychological aggression as detrimental as physical aggression? The independent effects of psychological aggression on depression and anxiety symptoms. Violence Vict. 2009;24(1):20–35. pmid:19297883
- 37. Chowdhury AN, Ramakrishna J, Chakraborty AK, Weiss MG. Cultural context and impact of alcohol use in the Sundarban Delta, West Bengal, India. Soc Sci Med. 2006;63(3):722–31. pmid:16584824
- 38. Hasselmann MH, Werneck GL, Silva CVC d. Symptoms of postpartum depression and early interruption of exclusive breastfeeding in the first two months of life. Cad Saude Publica. 2008;24:s341–52.
- 39. Kendall-Tackett K. Breastfeeding and trauma: what lactation specialists need to know. Breastfeed Med. 2020;15(8):495–500.
- 40. Tran LM, Nguyen PH, Naved RT, Menon P. Intimate partner violence is associated with poorer maternal mental health and breastfeeding practices in Bangladesh. Health Policy Plan. 2020;35(Supplement_1):i19–29. pmid:33165581
- 41. Jackson KT, Mantler T, Davidson CA, Siwik E. Intimate partner violence and breastfeeding outcomes among canadian women: an exploratory correlational study. Matern Child Health J. 2024;28(10):1737–48. pmid:39115628
- 42. Cavallé-Abasolo E, Dikmen-Yildiz P, Gómez-Gómez I, Barros-Martins L, Motrico E. Impact of traumatic childbirth and birth-related posttraumatic stress disorder on breastfeeding outcomes: a systematic review of longitudinal and cohort studies. Birth. 2025;52(4):553–71. pmid:41054020
- 43. Shroff MR, Griffiths PL, Suchindran C, Nagalla B, Vazir S, Bentley ME. Does maternal autonomy influence feeding practices and infant growth in rural India? Soc Sci Med. 2011;73(3):447–55. pmid:21742425
- 44. Sulaiman LA-R, Ojogiwa OT, Ajayi CE. Intimate partner controlling behaviour and intimate partner violence among married women in rural areas in South Africa. BMC Womens Health. 2025;25(1):199. pmid:40264113
- 45. Wessells MG, Kostelny K. The psychosocial impacts of intimate partner violence against women in LMIC contexts: toward a holistic approach. Int J Environ Res Public Health. 2022;19(21):14488. pmid:36361364
- 46. Walters CN, Rakotomanana H, Komakech JJ, Stoecker BJ. Maternal experience of intimate partner violence is associated with suboptimal breastfeeding practices in Malawi, Tanzania, and Zambia: insights from a DHS analysis. Int Breastfeed J. 2021;16(1):20. pmid:33602285
- 47. Normann AK, Bakiewicz A, Kjerulff Madsen F, Khan KS, Rasch V, Linde DS. Intimate partner violence and breastfeeding: a systematic review. BMJ Open. 2020;10(10):e034153. pmid:33130559
- 48. Leight J, Wilson N. Intimate partner violence and maternal health services utilization: evidence from 36 National Household Surveys. BMC Public Health. 2021;21(1):405. pmid:33632170
- 49. Jackson KT, Larose S, Mantler T. Accessing trauma- and violence-informed breastfeeding support from primary care providers among women with histories of intimate partner violence: an exploratory interpretive description study. Can J Nurs Res. 2025;57(2):177–87. pmid:39359241
- 50. Temesgen K, Andarge E, Fikadu T, Bekele M, Chisha Y, Esubalew H, et al. Early cessation of breastfeeding and the associated factors among mothers with children aged 2 to 3 years in rural Southern Ethiopia: a community-based cross-sectional study. BMC Nutr. 2023;9(1):22. pmid:36721185
- 51.
Organization WH. Indicators for assessing infant and young child feeding practices: definitions and measurement methods; 2021.
- 52.
Central Statistical Agency (CSA) [Ethiopia] and ICF. The DHS program: demographic and health surveys. Rockville: ICF Macro; 2016.
- 53. Emagneneh T, Mulugeta C, Alamrew A, Ejigu B, Abebe W. Early cessation of exclusive breastfeeding and associated factors in Ethiopia: a systematic review and meta-analysis. Front Nutr. 2025;12:1500077. pmid:40395551
- 54.
UNICEF. UNICEF calls for Strengthening Efforts to Expand Breastfeeding Support throughout the Nation 2024 [cited 2025 Jun 11]. Available from: https://www.unicef.org/ethiopia/press-releases/unicef-calls-strengthening-efforts-expand-breastfeeding-support-throughout-nation
- 55. Zimmerman L, Desta S, Yihdego M, Rogers A, Amogne A, Karp C, et al. Protocol for PMA-Ethiopia: a new data source for cross-sectional and longitudinal data of reproductive, maternal, and newborn health. Gates Open Res. 2020;4:126. pmid:33150302
- 56.
World Health Organization. Indicators for assessing infant and young child feeding practices. Part 2: Measurement. Geneva: World Health Organization; 2010 [cited 2025 Aug 07]. Available from: https://iris.who.int/bitstream/handle/10665/44306/9789241599290_eng.pdf
- 57. Deshpande NA, Lewis-O’Connor A. Screening for intimate partner violence during pregnancy. Rev Obstet Gynecol. 2013;6(3–4):141–8. pmid:24920977
- 58. Abramsky T, Watts CH, Garcia-Moreno C, Devries K, Kiss L, Ellsberg M, et al. What factors are associated with recent intimate partner violence? Findings from the WHO multi-country study on women’s health and domestic violence. BMC Public Health. 2011;11:109. pmid:21324186
- 59. Boyce SC, McDougal L, Silverman JG, Atmavilas Y, Dhar D, Hay K, et al. Associations of intimate partner violence with postnatal health practices in Bihar, India. BMC Pregnancy Childbirth. 2017;17(1):398. pmid:29187158
- 60. Jarnecke AM, Barden E, Back SE, Brady KT, Flanagan JC. Intimate partner violence moderates the association between oxytocin and reactivity to dyadic conflict among couples. Psychiatry Res. 2018;270:404–11. pmid:30308464
- 61. Nagel EM, Howland MA, Pando C, Stang J, Mason SM, Fields DA, et al. Maternal psychological distress and lactation and breastfeeding outcomes: a narrative review. Clin Ther. 2022;44(2):215–27. pmid:34937662
- 62. Cerulli C, Chin N, Talbot N, Chaudron L. Exploring the impact of intimate partner violence on breastfeeding initiation: does it matter? Breastfeed Med. 2010;5(5):225.
- 63. Mechanic MB, Weaver TL, Resick PA. Mental health consequences of intimate partner abuse: a multidimensional assessment of four different forms of abuse. Violence Against Women. 2008;14(6):634–54. pmid:18535306
- 64. Scott JA, Binns CW, Graham KI, Oddy WH. Temporal changes in the determinants of breastfeeding initiation. Birth. 2006;33(1):37–45. pmid:16499530
- 65. Hassen HM. Trends and disparities in ever-breastfeeding practice and early breastfeeding initiation in Ethiopia: a 20-year trend analysis from EDHS datasets. BMC Public Health. 2024;24(1):2558. pmid:39300468
- 66. Carlin RF, Mathews A, Oden R, Moon RY. The influence of social networks and norms on breastfeeding in African American and Caucasian mothers: a qualitative study. Breastfeed Med. 2019;14(9):640–7. pmid:31433206
- 67. Aarts C, Kylberg E, Hörnell A, Hofvander Y, Gebre-Medhin M, Greiner T. How exclusive is exclusive breastfeeding? A comparison of data since birth with current status data. Int J Epidemiol. 2000;29(6):1041–6. pmid:11101545
- 68. Fenta EH, Yirgu R, Shikur B, Gebreyesus SH. A single 24 h recall overestimates exclusive breastfeeding practices among infants aged less than six months in rural Ethiopia. Int Breastfeed J. 2017;12(1):36.
- 69. Caleyachetty R, Uthman OA, Bekele HN, Martín-Cañavate R, Marais D, Coles J, et al. Maternal exposure to intimate partner violence and breastfeeding practices in 51 low-income and middle-income countries: a population-based cross-sectional study. PLoS Med. 2019;16(10):e1002921. pmid:31574100
- 70.
De González AB, Richardson DB, Schubauer-Berigan MK. Statistical methods in cancer research volume V: bias assessment in case–control and cohort studies for hazard identification; 2024.
- 71. Ellsberg M, Heise L. Bearing witness: ethics in domestic violence research. Lancet. 2002;359(9317):1599–604. pmid:12047984
- 72. Gracia E. Unreported cases of domestic violence against women: towards an epidemiology of social silence, tolerance, and inhibition. BMJ. 2004:536–7.
- 73. Yismaw GA, Abuhay HW, Alemayehu MA, Derseh NM, Agimas MC, Tesfie TK. Geospatial variation of exclusive breastfeeding and its determinants among mothers of infants under 6 months in Ethiopia: spatial and geographical weighted regression analysis. BMJ Paediatr Open. 2024;8(Suppl 2):e002573. pmid:38684333
- 74. Tennant PWG, Murray EJ, Arnold KF, Berrie L, Fox MP, Gadd SC, et al. Use of directed acyclic graphs (DAGs) to identify confounders in applied health research: review and recommendations. Int J Epidemiol. 2021;50(2):620–32. pmid:33330936