Maternal Intimate Partner Violence and Complementary Feeding Practices in 41 Low- and Middle-Income Countries
Maternal intimate partner violence and complementary feeding practices in 41 low- and middle-income countries
一项基于41个低中收入国家的Demographic and Health Surveys数据的研究发现,母亲遭受亲密伴侣暴力与婴儿辅食摄入不足有关,尤其是在经济和社会地位较低的地区。暴露于亲密伴侣暴力的母亲所生儿童更有可能出现营养不良,尽管没有观察到明显的叠加效应。这些结果强调了在母婴健康策略中考虑暴力相关照顾者限制的重要性。
Introduction
The first 2 years of life are a critical window for growth and development, and adequate feeding during this period is closely associated with health and wellbeing across the life course1,2. Yet suboptimal feeding remains common worldwide. The World Health Organization (WHO) recommends introducing nutritionally adequate complementary foods from around 6 months of age, yet only one in three children aged 6 to 23 months achieves minimum dietary diversity3,4. In low- and middle-income countries (LMICs), inadequate complementary feeding is strongly socially patterned and is associated with socioeconomic disadvantage, lower maternal education, household food insecurity, and limited access to health and nutrition services, underscoring the importance of considering broader constraints on maternal caregiving and feeding practices5,6. At the same time, women in these settings experience a high burden of intimate partner violence (IPV). WHO estimates suggest that approximately one in four ever-partnered women have experienced physical or sexual IPV during their lifetime, with a particularly high burden in sub-Saharan Africa and Southern Asia7.
Maternal health and socioeconomic circumstances are closely linked to child survival and development8. IPV commonly co-occurs with maternal physical and mental health problems, caregiving difficulties, and constrained access to household resources9,10,11. Infant and young child feeding practices are closely associated with child health12. Poor feeding practices are associated with greater vulnerability to malnutrition, particularly stunting and micronutrient deficiencies, as well as greater risks of morbidity and mortality in early life13,14. Previous studies have shown that maternal exposure to IPV has been associated with childhood morbidity, including diarrhoea and acute respiratory infection, in sub-Saharan Africa and with undernutrition, particularly wasting and underweight, across 36 LMICs15,16. However, less is known about whether maternal exposure to IPV is associated with age-appropriate child feeding practices.
Evidence from several settings has reported lower odds of early initiation of breastfeeding and exclusive breastfeeding among women exposed to IPV, and a study from Ethiopia suggests a lower likelihood of achieving a minimum acceptable diet17,18. IPV commonly co-occurs with psychological distress, physical injury and pain, and restrictions on women’s decision-making and access to food, financial resources, and health or nutrition services needed for adequate child feeding19,20. By contrast, other studies have found no clear associations with exclusive breastfeeding, early initiation, or bottle feeding21,22. Most previous work has focused on breastfeeding indicators, with relatively limited attention to child feeding practices during the complementary feeding period. This gap is important because complementary feeding is a nutritionally vulnerable transition that depends on sustained caregiving and access to household resources. IPV is also associated with poorer maternal wellbeing and reduced caregiving capacity and often co-occurs with lower maternal empowerment and socioeconomic disadvantage across LMICs23,24. Whether children exposed to maternal IPV and poor complementary feeding are also at greater risk of early undernutrition remains unclear. The inconsistent evidence and limited multicountry data indicate a need for a multicountry analysis of the associations between maternal IPV and complementary feeding practices, as well as their potential implications for early undernutrition.
This study used Demographic and Health Surveys (DHS) data from 41 LMICs collected between 2008 and 2023 to examine the associations between maternal exposure to IPV and complementary feeding practices among children aged 6 to 23 months. We further examined whether maternal IPV and suboptimal feeding were jointly associated with early undernutrition. Maternal IPV is associated with higher odds of non-attainment of minimum dietary diversity (MDD), minimum meal frequency (MMF), and minimum acceptable diet (MAD), with associations more evident in socioeconomically disadvantaged settings. Children exposed to maternal IPV who do not meet MDD or MAD also have higher odds of wasting and underweight, although no clear additive or multiplicative interaction is observed. These findings highlight the importance of considering violence-related constraints on caregiving in complementary feeding and maternal and child health strategies.
Methods
Data source and study population
We conducted a secondary analysis of 41 country-specific DHS rounds spanning DHS-VI, DHS-7, and DHS-8 (Supplementary Table 1). Countries were included if the selected survey contained data on both maternal IPV and complementary feeding practices were available. The 41 countries included Angola, Armenia, Benin, Burkina Faso, Burundi, Cambodia, Cameroon, Chad, Comoros, Côte d’Ivoire, the Democratic Republic of the Congo, the Dominican Republic, Egypt, Ethiopia, Gabon, The Gambia, Guatemala, Haiti, Honduras, India, Jordan, Kenya, the Kyrgyz Republic, Madagascar, Maldives, Mauritania, Mozambique, Namibia, Nepal, Pakistan, Papua New Guinea, Rwanda, Sierra Leone, South Africa, Tajikistan, Tanzania, Timor-Leste, Togo, Uganda, Zambia, and Zimbabwe. DHS are nationally representative cross-sectional household surveys conducted in more than 90 LMICs using a two-stage cluster sampling design. In the first stage, strata were generated by separating urban and rural areas within each region of the respective countries. Within each stratum, clusters were selected independently with probability proportional to size. In the second stage, a fixed or variable number of households were selected in each selected cluster based on a complete household listing, through equal-probability systematic sampling25,26.
Maternal exposure to IPV was assessed through the DHS domestic violence module, focusing on women who were currently or formerly married or partnered and provided information regarding IPV27. IPV data were collected through private interviews by trained interviewers, with the module administered to one randomly selected eligible woman per household after informed consent had been obtained28.
Data on complementary feeding practices were derived from the DHS child nutrition module, in which indicators are collected for the youngest eligible child living with the mother using WHO-standardised definitions that enable internationally comparable measurement across countries. This analysis focused on children aged 6–23 months, corresponding to the complementary feeding period and a critical stage of physical and cognitive development encompassing both infancy (6–11 months) and early childhood (12–23 months)29,30,31,32. This period marks a vulnerable transition from exclusive breastfeeding to the introduction of complementary foods alongside continued breastfeeding, during which inadequate feeding practices may significantly elevate the risks of malnutrition, stunting, and micronutrient deficiencies33. The analytic sample therefore comprised the youngest children aged 6–23 months living with their mothers, with no missing observations on maternal exposure to IPV and complementary feeding indicators (Supplementary Fig. 1).
Outcome variables
Primary outcomes were three indicators of complementary feeding practices, constructed as binary indicators from maternal recall of child dietary intake over the preceding 24 h and defined according to WHO standardised criteria for infant and young child feeding practices. These comprised MDD, MMF, and MAD31.
MDD was defined, in accordance with WHO standards, as consumption of at least five of the following eight food groups: (1) breastmilk; (2) grains, roots, tubers, and plantains; (3) legumes and nuts; (4) dairy products; (5) flesh foods (e.g. meat, poultry, organ meats); (6) eggs; (7) vitamin A-rich fruits and vegetables; and (8) other fruits and vegetables.
MMF was defined according to age-specific thresholds, as at least two daily feeds for breastfed infants aged 6–8 months and three or more feeds for those aged 9–23 months. For non-breastfed children aged 6–23 months, the threshold was four or more daily feeds, comprising at least one solid, semi-solid, or soft food.
MAD was operationalised as a composite indicator of both dietary diversity and meal frequency, with criteria contingent on breastfeeding status. For breastfed children, MAD was defined as concurrent fulfilment of both MDD and MMF standards. For non-breastfed children, MAD included modified MDD of at least four of six food groups, excluding dairy, plus a minimum of two milk feeds and MMF fulfilment.
Explanatory variables
The primary explanatory variable was maternal exposure to IPV within the 12 months preceding the survey, comprising physical, sexual, or emotional violence and their co-occurrence. Exposure was assessed using a modified Conflict Tactics Scale (CTS), with responses operationalised into three binary indicators corresponding to each dimension of violence34. Physical IPV included being pushed or shaken, having something thrown at the respondent, being slapped, being punched with a fist or hit with a harmful object, being kicked or dragged, being strangled or burned, being attacked with a knife, gun, or other weapon, or having their arm twisted or hair pulled. Emotional IPV included being humiliated, threatened with harm, or insulted or made to feel bad. Sexual IPV included being physically forced into unwanted sex, being forced into other unwanted sexual acts, or being physically forced to perform sexual acts they did not want to15. Women were classified as exposed (coded as 1; 0 otherwise) upon reporting one or more violent acts within a given category35.
Child nutritional outcomes
Child nutritional outcomes were defined according to the WHO Child Growth Standards. In the DHS, children’s weight and height are measured using standardised anthropometric procedures. Nutritional indicators are expressed as SDs from the median of the WHO reference population for a child’s age. Children were classified as stunted, wasted, or underweight if their height-for-age Z score, weight-for-height Z score, or weight-for-age Z score, respectively, was below -2.0 SD36.
Covariates
To account for potential confounding, we adjusted for prespecified child, maternal, and household characteristics. Child characteristics were age (6–11, 12–18, 19–23 months), sex, birth interval (≤24 versus >24 months), and birth order (1, 2–4, ≥5). Maternal characteristics were age (15–24, 25–34, 35–49 years), educational attainment (none, primary, secondary or higher), marital status (married, divorced, separated), maternal employment (no versus yes), maternal autonomy (no versus yes), and media exposure (no versus yes). Household characteristics included residence (urban versus rural), wealth quintile (poorest, poorer, middle, richer, richest), and household infrastructure indicators, defined by cooking fuel (clean versus not clean), water source (protected versus not protected), and toilet (improved versus unimproved).
Statistics and reproducibility
The final analytic sample comprised 169,644 children. Each child represented one observational unit; as this was a secondary analysis of cross-sectional survey data, experimental replicates were not applicable.
Initially, we merged the domestic violence and complementary feeding records from the DHS datasets of 41 countries, and classified the variables into exposure, outcome, and potential confounder categories. To address substantial disparities in population sizes among included countries, sampling weights were de-normalised by rescaling the standard sampling weights by the ratio of the country-specific population of children aged 0 to 2 years (sourced from the United Nations World Population Prospects) to the respective survey sample size37,38. This ensured that pooled estimates were representative of the aggregate target population across the 41 countries included in the analysis.
Descriptive analyses included calculating unweighted frequencies and percentages for exposure variables (physical, emotional, or sexual violence, and their combinations) and outcome variables (MDD, MMF, MAD). Weighted descriptive estimates were additionally generated using the de-normalised DHS sampling weights, with weighted counts based on these rescaled weights presented in Supplementary Table 2. Country-specific prevalence estimates and corresponding 95% confidence intervals (CIs) for each outcome were calculated using the de-normalised sampling weights and are shown in the figures.
To examine associations between maternal IPV exposure and complementary feeding practices, we used multivariable logistic regression to estimate the adjusted odds ratios (aORs) of non-attainment of MDD, MMF, and MAD, accounting for the DHS complex survey design by incorporating the de-normalised sampling weights, primary sampling units, and survey strata, and adjusting for covariates, survey year fixed effects, and country fixed effects. As co-occurring IPV forms may represent distinct patterns of violence exposure, we examined each pairwise combination of physical, emotional, and sexual IPV, as well as concurrent exposure to all three forms, separately, alongside any IPV and each individual form. Combination-specific estimates provided a more differentiated assessment of multiple-form IPV, indicating whether associations with complementary feeding varied according to the specific forms of violence occurring together. To investigate variation across subpopulations, we conducted subgroup analyses stratified by World Bank income group (low-income, lower-middle-income, and upper-middle-income), child age (6–11, 12–18, and 19–23 months), sex (male and female), birth interval (≤24 and >24 months), birth order (1, 2–4, and ≥5), maternal employment (no and yes), education (none, primary, and secondary or higher), autonomy (no and yes), and media exposure (no and yes), as well as household characteristics, including residence (urban and rural), wealth quintile (poorest, poorer, middle, richer, and richest), cooking fuel (clean and unclean), water source (protected and unprotected), and toilet facilities (improved and unimproved).
To identify inter-country variations in these associations, country-stratified regression models were fitted. Scatter plots mapping country-level IPV prevalence against the adjusted predicted probabilities of non-attainment of MDD, MMF, and MAD were generated to visually detect potential clustering across included LMICs. We carried out multiple imputation for observations with missing values on one or more covariates in all pooled analyses, using chained equations and logistic regression models to generate five imputed datasets.
We performed several sensitivity analyses. First, regression models were re-estimated without multiple imputation to ensure that results were not influenced by the imputation procedure. Second, given the sampling weights and hierarchical nature of the data, multilevel mixed-effects generalised linear regression models with a random intercept at the cluster level were employed to estimate adjusted ORs and 95% CIs for the association. Third, because non-attainment of complementary feeding indicators was relatively common and ORs may be less interpretable for common binary outcomes, modified Poisson regression models were fitted to estimate adjusted prevalence ratios (PRs) and 95% CIs39.
To assess interactions between maternal IPV and complementary feeding practices in relation to child nutritional outcomes (stunting, underweight, and wasting), we examined both additive and multiplicative interactions. Interaction terms between maternal IPV exposure and each complementary feeding indicator (MDD, MMF, and MAD) were included separately in multivariable logistic regression models for each nutritional outcome. Multiplicative interaction was assessed using the aOR and 95% CI for the interaction term, and additive interaction was assessed using the RERI with corresponding 95% CIs40. To quantify joint associations, we classified children according to combinations of maternal IPV exposure and complementary feeding practices, using those unexposed to IPV who met the respective feeding standard as the reference group, and estimated aORs and 95% CIs for each combination across nutritional outcomes41.
All analyses were performed using Stata (version 17.0) and R (version 4.3.1). A two-sided p-value of less than 0.05 was considered statistically significant.
Ethics approval and consent to participate
The present study was conducted in accordance with the Declaration of Helsinki. The original DHS surveys were reviewed and approved by the ICF Institutional Review Board (IRB) and relevant national review bodies, as detailed in Supplementary Table 3. The ICF IRB operates under Federalwide Assurance No. FWA00000845. Informed consent was obtained from all respondents in accordance with the survey protocols. As this study involved secondary analysis of de-identified DHS data obtained with authorization from the DHS Program, no additional ethical approval or informed consent was required.
Results
Population characteristics, maternal IPV prevalence, and feeding patterns
Table 1 shows the characteristics of the 169,644 children included from 41 LMICs. Of the included children, 87,672 (51.7%) were male, 67,771 (40.0%) were aged 12 to 18 months, 141,473 (83.4%) had birth intervals exceeding 24 months, and 90,370 (53.3%) were of second to fourth birth order. Attainment of recommended feeding standards was observed in 41,257 (24.3%) children for MDD, 71,286 (42.0%) for MMF, and 23,959 (14.1%) for MAD. Stunting was observed in 39,600 (23.3%) children, wasting in 12,198 (7.2%), and underweight in 23,265 (13.7%). The prevalence of meeting MDD was lowest in Chad (9.3%, 95% CI 7.0–11.6) and highest in Maldives (70.4%, 67.1–73.7), the prevalence of meeting MMF was lowest in Mauritania (24.4%, 22.9–25.9) and highest in Nepal (82.1%, 80.2–84.1), and the prevalence of meeting MAD was lowest in Chad (1.0%, 0.0–1.0) and highest in Guatemala (56.7%, 55.0–58.3) (Supplementary Figs. 2–4).
Among maternal characteristics, 82,358 (48.6%) children were born to mothers aged 25–34 years, and 79,792 (47.0%) were born to mothers with secondary education or higher. Overall, 117,456 (69.2%) children had mothers with regular media exposure, 76,316 (45.0%) had unemployed mothers, and 75,968 (44.8%) had mothers with no autonomy. Overall, 26,696 (15.7%) children had mothers who reported exposure to any IPV in the 12 months preceding the survey, including physical IPV in 22,257 (13.1%), emotional IPV in 18,598 (11.0%), and sexual IPV in 10,136 (6.0%). The prevalence of any maternal IPV ranged from 6.3% (95% CI 5.0–7.7) in Chad to 46.0% (43.1–48.9) in Zimbabwe (Supplementary Fig. 5).
IPV and feeding non-attainment
Figure 1 shows that maternal IPV was associated with higher odds of non-attainment of MDD, MMF, and MAD. For MDD, aORs were 1.25 (95% CI 1.16–1.35) for any IPV, 1.38 (1.27–1.49) for physical IPV, 1.22 (1.12–1.34) for emotional IPV, and 1.41 (1.26–1.59) for sexual IPV. Combined exposure to physical and emotional IPV, physical and sexual IPV, emotional and sexual IPV, and all three forms of IPV was also associated with higher odds of MDD non-attainment, with aORs of 1.38 (1.25–1.52), 1.53 (1.35–1.74), 1.31 (1.15–1.49), and 1.35 (1.18–1.55), respectively. For MMF, aORs were 1.08 (1.01–1.14) for any IPV, 1.11 (1.04–1.19) for physical IPV, and 1.69 (1.57–1.82) for emotional IPV, whereas sexual IPV alone was not significantly associated with MMF non-attainment (1.08, 0.98–1.18). Combined exposure to physical and emotional IPV, physical and sexual IPV, emotional and sexual IPV, and all three forms of IPV was associated with higher odds of MMF non-attainment, with aORs of 1.95 (1.79–2.14), 1.17 (1.06–1.29), 2.41 (2.11–2.75), and 2.60 (2.28–2.97), respectively. For MAD, aORs were 1.50 (1.36–1.67) for any IPV, 1.76 (1.56–1.98) for physical IPV, 1.53 (1.35–1.73) for emotional IPV, and 2.27 (1.90–2.73) for sexual IPV. Higher odds of MAD non-attainment were also observed for combined physical and emotional IPV, physical and sexual IPV, emotional and sexual IPV, and all three forms of IPV, with aORs of 1.49 (1.27–1.75), 2.38 (1.65–3.45), 2.44 (1.94–3.07), and 2.68 (2.08–3.47), respectively.
Note: Estimates are aORs with 95% CIs from multivariable logistic regression models adjusted for child, maternal, and household characteristics and survey year and country fixed effects. Squares indicate aOR point estimates and horizontal error bars indicate 95% CIs; square sizes are proportional to the weights shown, and diamonds indicate summary estimates. The dashed vertical line indicates the null value (aOR = 1). Individual IPV dimensions and co-occurring IPV categories were modelled separately. n = 169,644 independent children. n denotes the unweighted number of children; survey weights were used in estimation. aOR adjusted odds ratio, CI confidence interval, IPV intimate partner violence, MDD minimum dietary diversity, MMF minimum meal frequency, MAD minimum acceptable diet.
Country and subgroup differences
These associations between maternal IPV and complementary feeding practices varied across countries, with the largest statistically significant country-specific aORs observed in the Democratic Republic of the Congo for MDD non-attainment (aOR 4.25, 95% CI 1.79–10.05), Kenya for MMF non-attainment (aOR 2.48, 95% CI 2.00–3.08), and Mauritania for MAD non-attainment (aOR 4.59, 95% CI 1.81–11.67). Cambodia, Guatemala, Haiti, and Kenya showed significant associations across all three feeding indicators (Supplementary Fig. 6). In analyses stratified by World Bank income group, maternal IPV was associated with higher odds of non-attainment of MDD in low-income countries (aOR 1.31, 95% CI 1.15–1.51) and lower-middle-income countries (aOR 1.20, 95% CI 1.08–1.33), with higher odds of non-attainment of MMF in lower-middle-income countries only (aOR 1.11, 95% CI 1.03–1.20), and with higher odds of non-attainment of MAD in low-income countries (aOR 1.69, 95% CI 1.43–1.99) and lower-middle-income countries (aOR 1.50, 95% CI 1.27–1.69), whereas no statistically significant associations were observed in upper-middle-income countries (Fig. 2). No obvious clustering patterns were observed in the scatter plots of country-level IPV prevalence against the adjusted predicted probabilities of non-attainment of MDD, MMF, and MAD (Supplementary Fig. 7 to Supplementary Fig. 9).
Note: Estimates are aORs with 95% CIs for the association between maternal exposure to any IPV and non-attainment of complementary feeding indicators, stratified by World Bank income group. Squares indicate aOR point estimates and horizontal error bars indicate 95% CIs; the dashed vertical line indicates the null value (aOR = 1). Models were adjusted as in the main analysis. n = 41,435, 119,007, and 9202 independent children in low-income, lower-middle-income, and upper-middle-income countries, respectively. n denotes unweighted sample size; survey weights were used in estimation. aOR adjusted odds ratio, CI confidence interval, IPV intimate partner violence, MDD minimum dietary diversity, MMF minimum meal frequency, MAD minimum acceptable diet.
As shown in Fig. 3, subgroup analyses were conducted by child age, sex, birth interval, and birth order; maternal employment, education, autonomy, and media exposure; and household residence, wealth, cooking fuel, water source, and toilet facility. Associations were not consistently significant across strata. For MDD non-attainment, significant associations were observed among children whose mothers had no education, those living in rural areas or in the poorest wealth quintile, and those in households using unclean cooking fuel or unimproved toilet facilities. For MMF non-attainment, significant associations were observed among boys, children aged 19–23 months, those with birth intervals exceeding 24 months or of second to fourth birth order, children whose mothers had no education or regular media exposure, those living in rural areas or in the poorest wealth quintile, and those in households using unprotected water sources or unclean cooking fuel. For MAD non-attainment, significant associations were observed among children whose mothers had no education, those living in rural areas or in the poorest wealth quintile, and those in households using unclean cooking fuel or unimproved toilet facilities.
Note: Estimates are aORs with 95% CIs from multivariable logistic regression models across prespecified subgroups. Squares indicate aOR point estimates and horizontal error bars indicate 95% CIs; the dashed vertical line indicates the null value (aOR = 1). Models were adjusted as in the main analysis. Unweighted subgroup sample sizes were: child age, n = 59,081, 67,771, and 42,792 for 6–11, 12–18, and 19–23 months; sex, n = 87,672 and 81,972 for male and female; birth interval, n = 28,171 and 141,473 for ≤24 and >24 months; birth order, n = 50,964, 90,370, and 28,310 for 1, 2–4, and ≥5; maternal employment, n = 76,316 and 73,145 for no and yes; maternal education, n = 48,806, 41,046, and 79,792 for none, primary, and secondary or higher; maternal autonomy, n = 75,968 and 81,017 for no and yes; media exposure, n = 52,029 and 117,456 for no and yes; residence, n = 51,415 and 118,229 for urban and rural; wealth quintile, n = 43,151, 37,598, 34,411, 29,636, and 24,848 from poorest to richest; cooking fuel, n = 41,201 and 117,222 for clean and not clean; water source, n = 129,942 and 29,538 for protected and not protected; and toilet facility, n = 86,650 and 72,318 for improved and unimproved. For variables with missing values, n denotes observed subgroup membership before multiple imputation. Missing covariate data were handled using multiple imputation as described in the Methods. aOR adjusted odds ratio, CI confidence interval, IPV intimate partner violence, MDD minimum dietary diversity, MMF minimum meal frequency, MAD minimum acceptable diet.
Sensitivity analyses
Sensitivity analyses were broadly consistent with the main findings. In analyses without multiple imputation, any IPV remained associated with non-attainment of MDD (aOR 1.16, 95% CI 1.05–1.27), MMF (aOR 1.05, 95% CI 1.01–1.09), and MAD (aOR 1.28, 95% CI 1.13–1.44) (Supplementary Table 4). In multilevel mixed–effects generalised linear regression models, any IPV remained associated with non–attainment of MDD (aOR 1.38, 95% CI 1.32–1.43), MMF (aOR 1.14, 95% CI 1.10–1.17), and MAD (aOR 1.70, 95% CI 1.62–1.79) (Supplementary Table 5). In analyses using modified Poisson regression, any IPV remained associated with non–attainment of MDD (aPR 1.14, 95% CI 1.13–1.15), MMF (aPR 1.13, 95% CI 1.11–1.16), and MAD (aPR 1.16, 95% CI 1.14–1.18) (Supplementary Table 6).
Joint exposure and undernutrition
For analyses of interactions, we found no clear additive or multiplicative interaction between maternal IPV and complementary feeding practices for stunting, underweight, or wasting (Supplementary Table 7). As shown in Fig. 4, in pooled joint analyses, compared with children unexposed to IPV who met the respective feeding standard, those exposed to IPV who did not meet MDD had higher odds of wasting (aOR 1.26, 95% CI 1.01–1.57) and underweight (aOR 1.17, 95% CI 1.02–1.36). Similarly, children exposed to IPV who did not meet MAD had higher odds of wasting (aOR 1.27, 95% CI 1.04–1.55) and underweight (aOR 1.17, 95% CI 1.00–1.37). No clear joint associations were observed for stunting or for combinations involving MMF.
Note: Estimates are aORs with 95% CIs from multivariable logistic regression models adjusted for child, maternal, and household characteristics and survey year and country fixed effects. Point estimates indicate aORs and horizontal error bars indicate 95% CIs; the dashed vertical line indicates the null value (aOR = 1). For each complementary feeding indicator, children unexposed to IPV who met the respective feeding standard were the reference group. The analyses included n = 145,275 independent children for stunting, n = 145,254 for wasting, and n = 146,739 for underweight. n denotes unweighted sample size; survey weights were used in estimation. aOR adjusted odds ratio, CI, confidence interval, IPV intimate partner violence, MDD minimum dietary diversity, MMF minimum meal frequency, MAD minimum acceptable diet.
Discussion
In this pooled analysis of DHS data from 41 countries, maternal IPV exposure was associated with higher odds of non-attainment of MDD, MMF, and MAD, although the pattern varied across feeding indicators and contexts. Across World Bank income groups, these associations were observed mainly in low-income and lower-middle-income countries, with significant associations for MDD and MAD in both groups and for MMF in lower-middle-income countries only; no statistically significant associations were observed in upper-middle-income countries. In subgroup analyses, where statistically significant associations were not observed across all strata, significant associations were concentrated among children of mothers with no education, those living in rural areas, and those facing household-level material deprivation, including those in the poorest households and those living in households with unclean cooking fuel or unimproved sanitation; for MMF, significant associations were also observed in several child subgroups, including boys and older children. Country-specific associations were especially notable in the Democratic Republic of the Congo, Kenya, and Mauritania. In joint analyses, compared with children unexposed to IPV who met the respective feeding standard, those exposed to IPV who did not meet MDD or MAD had higher odds of wasting and underweight, whereas no clear joint associations were observed for stunting or for combinations involving MMF.
Existing cross-national evidence has focused primarily on the association between maternal exposure to IPV and breastfeeding outcomes during the first 6 months of life, particularly delayed initiation of breastfeeding. By contrast, evidence relating to the complementary feeding period from 6 to 23 months remains limited, despite this being a critical stage marked by dietary transition, increasing dependence on household food resources, and heightened sensitivity to caregiving conditions17. Unlike previous studies that have mainly examined distal child outcomes such as undernutrition or morbidity, our analysis of specific complementary feeding indicators provides more proximate evidence on the association between maternal IPV and daily feeding practices15. This focus is important because complementary feeding indicators may capture more immediate dimensions of caregiving than outcomes such as stunting or underweight. Several factors documented in previous studies may provide context for interpreting these associations, although the present study did not directly evaluate the underlying mechanisms42,43. First, previous studies have documented the co-occurrence of IPV with maternal physical and psychological morbidity, as well as associations between maternal morbidity and caregiving capacity, responsiveness during feeding, and the consistency of daily feeding routines44,45. Second, IPV frequently coexists with economic control, household deprivation, and food insecurity, all of which are closely tied to access to diverse foods and nutritionally adequate diets for young children46,47. These factors are particularly relevant to the interpretation of MDD and MAD, which depend more directly than feeding frequency alone on dietary quality, food variety, and the sustained availability of adequate foods. Third, fear, psychological distress, and constrained autonomy within violent relationships may likewise coincide with differences in maternal responsiveness, feeding routines, and use of maternal and child health services that support age-appropriate feeding48,49,50.
Our findings suggest that the association between maternal IPV and complementary feeding practices was not uniform across social and contextual settings. Significant associations were observed particularly among children of mothers with no education, those living in rural areas, those in the poorest wealth quintile, and those living in households with unclean cooking fuel or unimproved sanitation. Across World Bank income groups, associations between maternal IPV and complementary feeding non-attainment were observed mainly in low-income and lower-middle-income countries, with no statistically significant associations observed in upper-middle-income countries. The findings suggest that the observed associations may be more evident in settings where women have fewer educational, economic, and household resources to support recommended feeding practices. The subgroup and income-group patterns should therefore be interpreted as reflecting co-occurring social disadvantage. Our analysis did not establish whether constraints in daily caregiving, food access, service infrastructure, or social protection accounted for the observed differences24,51. Associations also varied across countries, with particularly notable estimates in the Democratic Republic of the Congo, Mauritania, and Kenya. Although these cross-country differences should be interpreted cautiously, the analysis was not designed to attribute this heterogeneity to specific country-level characteristics, including food insecurity, gender inequality, service access, or wider social instability51,52. For MMF, the additional associations observed among older children may indicate that associations with feeding frequency vary across stages of complementary feeding.
Finally, our joint analyses examined the co-occurrence of maternal IPV, suboptimal complementary feeding, and early undernutrition. Compared with children whose mothers were unexposed to IPV and who met the respective complementary feeding standard, children whose mothers were exposed to IPV and who did not meet MDD or MAD had higher odds of wasting and underweight, whereas no clear joint associations were observed for stunting or for combinations involving MMF. One possible interpretation is that, in the context of maternal IPV, failure to meet MDD or MAD may serve as a marker of co-occurring dietary and socioeconomic vulnerability53,54. MDD and MAD capture dimensions of dietary diversity and overall dietary adequacy that are not reflected by feeding frequency alone, although the present analysis could not determine why the joint associations differed across feeding indicators55,56. The absence of a corresponding pattern for stunting should be interpreted in light of the longer-term and cumulative nature of linear growth faltering, which may not be captured by recent feeding practices and IPV exposure measured in cross-sectional data57,58. At the same time, because we did not observe clear additive or multiplicative interaction, these findings are better interpreted as evidence of co-occurring vulnerability than of synergistic effect amplification.
These findings have several implications for policy and programme design. First, maternal IPV could be considered in complementary feeding and nutrition programmes as a marker of vulnerability to suboptimal feeding. In resource-constrained settings, IPV-sensitive feeding support could be embedded within existing maternal and child health and nutrition platforms, including immunisation, growth monitoring, and nutrition contacts. In settings where access to facility-based care is constrained, community-based infant and young child feeding services delivered by trained community health workers may provide an additional avenue for delivering support beyond health facilities, with referral pathways linking women experiencing IPV to appropriate health and psychosocial care59,60,61. Second, the subgroup patterns may help inform how programme support is prioritised across populations. Given that the observed associations were particularly evident among children of mothers with no education, those living in rural areas, and those facing household-level material deprivation, these populations may warrant particular attention in efforts to improve complementary feeding, including prioritisation for home-based counselling, more frequent follow-up, and linkage to food, cash, or social support where available5. Third, the joint findings for MDD and MAD suggest that programme design in IPV-affected settings could consider factors beyond behaviour-change messaging alone. The pattern observed here does not identify the underlying barriers to dietary quality and overall dietary adequacy. Programme responses could nevertheless consider material barriers to adequate diets alongside efforts to improve feeding frequency62. Taken together, these findings suggest that complementary feeding programmes could consider individual feeding behaviours alongside the broader contexts of violence, poverty, and women’s capacity to provide care.
Several limitations should be noted. First, the analysis included only 41 LMICs with available DHS data on both IPV and complementary feeding, which may limit generalisability to other settings. Second, the cross-sectional design precludes establishing temporal ordering and causal inference. Third, self-reported IPV and complementary feeding practices may be affected by recall and social desirability bias; underreporting of IPV could attenuate associations, although direction and magnitude of bias cannot be quantified. Fourth, despite adjustment for a broad range of child, maternal, and household characteristics, residual confounding remains possible. Fifth, the study team did not include researchers based in the countries represented in this analysis. Given the context-specific nature of IPV and infant and young child feeding, this may have limited the interpretation of country-specific findings and the local relevance of the policy and programme implications. Future multicountry studies would benefit from involving researchers from participating countries in study design, interpretation, and knowledge translation. Finally, the joint and interaction analyses require cautious interpretation because IPV referred to the previous 12 months, feeding indicators to the preceding 24 h, and nutritional outcomes were measured at survey. Their temporal ordering therefore could not be established, and the joint associations should not be interpreted as evidence of a defined causal sequence or mutually reinforcing effects.
This multicountry analysis across 41 LMICs indicates that maternal IPV was associated with higher odds of non-attainment of MDD, MMF, and MAD among children aged 6–23 months. Children whose mothers were exposed to IPV and who did not meet MDD or MAD also had higher odds of wasting and underweight. Associations were concentrated in socioeconomically disadvantaged settings, suggesting that IPV may be a relevant consideration in the design of complementary feeding and maternal and child health services in these settings.
Data Availability
The survey data used in this study are available from the Demographic and Health Surveys website (https://dhsprogram.com). Access to datasets requires user registration and approval through the Demographic and Health Surveys Program’s data access system. Administrative permission was obtained to access and use these datasets for this research. The datasets are not redistributed by the authors and are currently stored in secure, password-protected storage accessible only to the study team. Source data underlying Figures 1–4 are provided in Supplementary Data.
Code availability
The statistical analysis code used in this study is publicly available at https://github.com/wangxinfeng112800/maternal-ipv-complementary-feeding.
References
Black, M. M. et al. Early childhood development coming of age: science through the life course. Lancet 389, 77–90 (2017).
White, J. M., Bégin, F., Kumapley, R., Murray, C. & Krasevec, J. Complementary feeding practices: current global and regional estimates. Maternal Child Nutr. 13, e12505 (2017).
WHO Guideline for Complementary Feeding of Infants and Young Children 6-23 Months of Age. World Health Organization: Geneva (WHO, 2023).
United Nations Children’s Fund (UNICEF) Fed to Fail? The Crisis of Children’s Diets in Early Life. 2021 Child Nutrition Report. (UNICEF, 2021)
Gatica-Domínguez, G., Neves, P. A., Barros, A. J. & Victora, C. G. Complementary feeding practices in 80 low-and middle-income countries: prevalence of and socioeconomic inequalities in dietary diversity, meal frequency, and dietary adequacy. J. Nutr. 151, 1956–1964 (2021).
Shagaro, S. S., Mulugeta, B.eT. & Kale T.D. Complementary feeding practices and associated factors among mothers of children aged 6-23 months in Ethiopia: Secondary data analysis of Ethiopian mini demographic and health survey 2019. Arch. Public Health 79, 205 (2021).
World Health Organization Violence Against Women Prevalence Estimates, Global, Regional and National Prevalence Estimates for Intimate Partner Violence Against Women and Non-partner Sexual Violence Against Women. (WHO, 2025)
Clark, H. et al. A future for the world’s children? A WHO–UNICEF–Lancet commission. Lancet 395, 605–658 (2020).
de Souza Mezzavilla, R., de Figueiredo Ferreira, M., Curioni, C. C., Lindsay, A. C. & Hasselmann, M. H. Intimate partner violence and breastfeeding practices: a systematic review of observational studies. J. de. Pediatr. 94, 226–237 (2018).
Rivara, F. P. et al. Healthcare utilization and costs for women with a history of intimate partner violence. Am. J. Prevent. Med. 32, 89–96 (2007).
Dadi, A. F. et al. Association between maternal mental health-related hospitalisation in the 5 years prior to or during pregnancy and adverse birth outcomes: a population-based retrospective cohort data linkage study in the Northern Territory of Australia. The Lancet Regional Health–Western Pacific 46, 101063 (2024).
Wilson, A. C. et al. Relation of infant diet to childhood health: seven year follow up of cohort of children in Dundee infant feeding study. Bmj 316, 21–25 (1998).
Black, R. E. et al. Maternal and child undernutrition and overweight in low-income and middle-income countries. Lancet 382, 427–451 (2013).
Bhutta, Z. A. et al. What works? Interventions for maternal and child undernutrition and survival. Lancet 371, 417–440 (2008).
Dadi, A. F. et al. Intimate partner violence and childhood health outcomes in 37 sub-Saharan African countries: an analysis of demographic health survey data from 2011 to 2022. Lancet Glob. Health 12, e1785–e1793 (2024).
Gao, Y., Zhao, S., Zhou, Y., Zhou, J. & Tang, K. Maternal experiences of multiple forms of intimate partner violence and associations with undernutrition among children under 5: evidence from 36 low-and middle-income countries based on demographic and health surveys. BMJ Global Health 10, 101063 (2025).
Caleyachetty, R. 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. 16, e1002921 (2019).
Tsedal, D. M., Yitayal, M., Abebe, Z. & Tsegaye, A. T. Effect of intimate partner violence of women on minimum acceptable diet of children aged 6–23 months in Ethiopia: evidence from 2016 Ethiopian demographic and health survey. BMC Nutr. 6, 28 (2020).
Tran, L. M., Nguyen, P. H., Naved, R. T. & Menon, P. Intimate partner violence is associated with poorer maternal mental health and breastfeeding practices in Bangladesh. Health Policy Plan. 35, i19–i29 (2020).
Bhatt Carreno, S. et al. Linkages between maternal experience of intimate partner violence and child nutrition outcomes: a rapid evidence assessment. PLoS ONE 19, e0298364 (2024).
Tchakounte, D., Kamga, B. F. & Mbassa, M. R. Z. The association between intimate partner violence and breastfeeding practices in Cameroon: an analysis from the Demographic and Health Survey of 2018. Heliyon 10, e32062 (2024).
Zureick-Brown, S., Lavilla, K. & Yount, K. M. Intimate partner violence and infant feeding practices in India: a cross-sectional study. Matern. Child Nutr. 11, 792–802 (2015).
Coll, C. V., Ewerling, F., García-Moreno, C., Hellwig, F. & Barros, A. J. Intimate partner violence in 46 low-income and middle-income countries: an appraisal of the most vulnerable groups of women using national health surveys. BMJ Glob. health 5, e002208 (2020).
Coll, C. V. et al. Identifying the women most vulnerable to intimate partner violence: a decision tree analysis from 48 low and middle-income countries. EClinicalMedicine 42, (2021).
Chen, S. et al. Prevalence, disparities, and trends in intimate partner violence against women living in urban slums in 34 low-income and middle-income countries: a multi-country cross-sectional study. EClinicalMedicine 81, (2025).
Corsi, D. J., Neuman, M., Finlay, J. E. & Subramanian, S. Demographic and health surveys: a profile. Int. J. Epidemiol. 41, 1602–1613 (2012).
Simon, D. J. et al. Intimate partner violence during pregnancy against 601,534 women aged 15 to 49 years in 57 LMICs: prevalence, disparities, trends and associated factors using Demographic and Health Survey data. EClinicalMedicine 86, (2025).
Ma, N. 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 secondary analysis of population-based surveys. Lancet Glob. Health 11, e1863–e1873 (2023).
World Health Organization Essential Nutrition Actions: Improving Maternal, Newborn, Infant and Young Child Health and Nutrition. (WHO, 2013).
Keeley, B., Little, C. & Zuehlke, E. The State of the World’s Children 2019: Children, Food and Nutrition—Growing Well in a Changing World. (UNICEF, 2019).
World Health Organization & United Nations Children’s Fund (UNICEF) Indicators for Assessing Infant and Young Child Feeding Practices: Definitions and Measurement Methods. (WHO and UNICEF, 2021).
He, C. et al. Assessment of heatwave impacts on child feeding practices across 36 low-income and middle-income countries: a cross-sectional analysis. Lancet Planet. Health 9, 101281 (2025).
Lutter, C. K., Grummer-Strawn, L. & Rogers, L. Complementary feeding of infants and young children 6 to 23 months of age. Nutr. Rev. 79, 825–846 (2021).
Kishor, S., Johnson, K. Profiling Domestic Violence: A Multi-Country Study. ORC Macro: Calverton, Maryland, 2004.
Croft, T. N. et al. Guide to DHS Statistics. ICF: Rockville, Maryland, 2018.
Mei, Z. & Grummer-Strawn, L. M. Standard deviation of anthropometric Z-scores as a data quality assessment tool using the 2006 WHO growth standards: a cross country analysis. Bull. World Health Organ. 85, 441–448 (2007).
ICF International Demographic and Health Survey Sampling and Household Listing Manual. (ICF International, 2012).
United Nations Department of Economic and Social Affairs, Population Division World Population Prospects 2024: Summary of Results. (United Nations, 2024).
Zou, G. A modified Poisson regression approach to prospective studies with binary data. Am. J. Epidemiol. 159, 702–706 (2004).
Andersson, T., Alfredsson, L., Källberg, H., Zdravkovic, S. & Ahlbom, A. Calculating measures of biological interaction. Eur. J. Epidemiol. 20, 575–579 (2005).
Knol, M. J. & VanderWeele, T. J. Recommendations for presenting analyses of effect modification and interaction. Int. J. Epidemiol. 41, 514–520 (2012).
Campbell, J. C. Health consequences of intimate partner violence. Lancet 359, 1331–1336 (2002).
Ellsberg, M., Jansen, H. A., Heise, L., Watts, C. H. & Garcia-Moreno, C. Intimate partner violence and women’s physical and mental health in the WHO multi-country study on women’s health and domestic violence: an observational study. Lancet 371, 1165–1172 (2008).
Din, M. A. C., Teng, N. I. M. F. & Manaf, Z. A. Maternal depression and child feeding practices: determinants to malnutrition among young children in Malaysian rural area. Women’s. Health 19, 17455057221147800 (2023).
Helle, C., Hillesund, E. R. & Øverby, N. C. Maternal mental health is associated with children’s frequency of family meals at 12 and 24 months of age. Matern. Child Nutr. 20, e13552 (2024).
Chai, J. et al. Association between intimate partner violence and poor child growth: results from 42 demographic and health surveys. Bull. World Health Organ. 94, 331 (2016).
Vahedi, L., Orjuela-Grimm, M., Chan, P. M. P. & Meyer, S. R. Intimate partner violence and women’s dietary diversity: a population-based investigation in 8 low-and middle-income countries. J. Nutr. 155, 1236–1245 (2025).
Mallan, K. M., Daniels, L. A., Wilson, J. L., Jansen, E. & Nicholson, J. M. Association between maternal depressive symptoms in the early postnatal period and responsiveness in feeding at child age 2 years. Matern. Child Nutr. 11, 926–935 (2015).
Fitzpatrick, S. & Whitfield, K. C. Maternal postpartum depression and responsive feeding in the first 2 years: a review. Infant Behav. Dev. 80, 102073 (2025).
Gautam, S. & Jeong, H.-S. The role of women’s autonomy and experience of intimate partner violence as a predictor of maternal healthcare service utilization in Nepal. Int. J. Environ. Res. Public Health 16, 895 (2019).
Heise, L. L. & Kotsadam, A. Cross-national and multilevel correlates of partner violence: an analysis of data from population-based surveys. Lancet Glob. Health 3, e332–e340 (2015).
Sardinha, L., Maheu-Giroux, M., Stöckl, H., Meyer, S. R. & García-Moreno, C. Global, regional, and national prevalence estimates of physical or sexual, or both, intimate partner violence against women in 2018. Lancet 399, 803–813 (2022).
Yount, K. M., DiGirolamo, A. M. & Ramakrishnan, U. Impacts of domestic violence on child growth and nutrition: a conceptual review of the pathways of influence. Soc. Sci. Med. 72, 1534–1554 (2011).
Waterman, E. A., McLain, M., Zulfiqar, H., Ahmar Qadeer, T. & Ciavoi, S.-M. The link between intimate partner violence and food insecurity: a review of quantitative and qualitative studies. Trauma Violence Abus. 25, 1511–1530 (2024).
Moursi, M. M. et al. Dietary diversity is a good predictor of the micronutrient density of the diet of 6-to 23-month-old children in Madagascar3. J. Nutr. 138, 2448–2453 (2008).
Bhutta, Z. A. et al. Evidence-based interventions for improvement of maternal and child nutrition: what can be done and at what cost? Lancet 382, 452–477 (2013).
Victora, C. G., De Onis, M., Hallal, P. C., Blössner, M. & Shrimpton, R. Worldwide timing of growth faltering: revisiting implications for interventions. Pediatrics 125, e473–e480 (2010).
Mertens, A. et al. Causes and consequences of child growth faltering in low-resource settings. Nature 621, 568–576 (2023).
World Health Organization. Infant and young child feeding. https://www.who.int/news-room/fact-sheets/detail/infant-and-young-child-feeding (2023).
Janmohamed, A., Sohani, N., Lassi, Z. S. & Bhutta, Z. A. The effects of community home visit and peer group nutrition intervention delivery platforms on nutrition outcomes in low and middle-income countries: a systematic review and meta-analysis. Nutrients 12, 440 (2020).
World Health Organization Strengthening Health Systems to Respond to Women Subjected to Intimate Partner Violence or Sexual Violence: A Manual for Health Managers. (World Health Organization, 2017).
Puspitasari, M. D. & Gayatri, M. Indonesia infant and young child feeding practice: the role of women’s empowerment in household domain. Glob. J. Health Sci. 12, 129–144 (2020).
Acknowledgements
We thank the Demographic and Health Surveys Program for granting access to the data used in this study.
Funding
No external funding was received.
Author information
Author notes
These authors contributed equally: Xinfeng Wang, Jingfei Lyu.
Authors and Affiliations
Institute for Global Public Policy, Fudan University, Shanghai, China
Xinfeng Wang
Saw Swee Hock School of Public Health, National University of Singapore, Singapore, Singapore
Xinfeng Wang & Natasha Howard
School of Public Health, Fudan University, Shanghai, China
Jingfei Lyu
National Health Commission Key Laboratory of Health Technology Assessment, Fudan University, Shanghai, China
Jingfei Lyu
Department of Global Health and Development, London School of Hygiene and Tropical Medicine, London, UK
Natasha Howard
Guangdong Provincial People’s Hospital, Southern Medical University, Guangzhou, China
Biaochuan He
Guangdong Provincial Key Laboratory of South China Structural Heart Disease, Guangdong Provincial People’s Hospital, Guangzhou, China
Biaochuan He
Authors
- Xinfeng Wang
- Jingfei Lyu
- Natasha Howard
- Biaochuan He
Contributions
X.W. had access to the data in the study. B.H. and N.H. conceptualized and designed the study. X.W. and J.L. led data analysis and interpretation. X.W. and J.L. wrote the initial manuscript. B.H. and N.H. provided overall supervision. All authors contributed to critical revision of the manuscript for intellectual content. All authors approved the version submitted.
Corresponding authors
Correspondence to Natasha Howard or Biaochuan He.
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Wang, X., Lyu, J., Howard, N. et al. Maternal intimate partner violence and complementary feeding practices in 41 low- and middle-income countries. Commun. Health 1, 45 (2026). https://doi.org/10.1038/s44528-026-00049-8
Received: 26 May 2026
Accepted: 30 September 2026
Published: 11 November 2026
Version of record: 11 November 2026
DOI: https://doi.org/10.1038/s44528-026-00049-8
来源:Nature 健康护理主题 · nature.com