خشونت خانگی و تغذیه کودک در لیبریا

Domestic violence against women is endemic globally and is an important social problem in its own right. A compounding concern is the impact of domestic violence against mothers on the nutritional status of their children. Liberia is an apt setting to examine this understudied topic, given the poor nutritional status of young children, high rate of domestic violence against women, and prolonged period of conflict that included systematic sexual violence against women. We expected that maternal exposure to domestic violence would predict lower anthropometric z-scores and higher odds of stunting, wasting, and underweight in children less than five years. Using data from 2467 mother-child dyads in the 2007 Liberia Demographic and Health Survey (LDHS) undertaken between December 24, 2006 and April 19, 2007, we conducted descriptive and multivariate analyses to examine the total, unadjusted and adjusted associations of maternal exposure to domestic violence with these anthropometric measures in children. Maternal reports of sexual domestic violence in the prior year predicted lower adjusted z-scores for height-for-age and weight-for-height as well as higher odds of stunting and underweight. The findings underscore the needs to (1) enhance and enforce conventional and customary laws to prevent the occurrence of domestic violence; (2) treat maternal survivors of domestic violence and screen their children for nutritional deficits; (3) heighten awareness of the intergenerational implications especially of recent sexual domestic violence; and (4) clarify the biological and behavior pathways by which domestic violence may influence child growth, thereby mitigating early growth failure and its adverse implications into adulthood.

Domestic violence refers to “assaultive and coercive behaviors that adults use against their intimate partners” ( Holden, 2003, p. 155), who may include current or former spouses and dating partners ( Saltzman, Fanslow, McMahon, & Shelley, 1999). Domestic violence against women is endemic globally and an important social problem in its own right ( Ellsberg et al., 2008 and Watts and Zimmerman, 2002). A compounding concern is its impact on mothers, who are typically the main caregivers for children ( Davalos and Santos, 2006 and Watts and Zimmerman, 2002). The nutritional effects of domestic violence on young children are probably high; however, this relationship is understudied, and the available findings are mixed, especially in poor settings ( Bair-Merritt et al., 2006, Kishor and Johnson, 2004 and Rico et al., 2010). In a conceptual review of the literature, Yount, DiGirolamo, and Ramakrishnan (2011) specified the biological and behavioral pathways by which domestic violence against mothers may impair the physical growth of children in utero through 36 months of life. Biologically, gestation and infancy are critical periods for the development of stress-responsive biological regulatory systems, which are central to maintain physical and mental health (see also Repetti, Taylor, & Seeman, 2002). A child’s observation of or involvement in domestic violence in these periods may permanently disrupt these regulatory systems. Dysregulation of these systems may cause deficits or delays in metabolic and immune activities; sexual maturation and reproductive function; mental and cognitive functioning; as well as physical growth (see also Mead, Beauchaine, & Shannon, 2010). Chronic deficits or delays in these domains may impair intellectual functioning, behavior, health, and stature into adulthood (Repetti et al., 2002). Behaviorally, a mother’s observation of domestic violence in her own childhood and/or her experiences of such violence in adulthood may impair her health and health behaviors in pregnancy. Such impairments may elevate her risk of a poor pregnancy outcome (see also Shah & Shah, 2010), which would strongly predict her child’s early growth (Saigal & Doyle, 2008). After delivery, poor maternal health and health risk behaviors that are predicted from observing and/or experiencing domestic violence may compromise the mother’s care of her infant. Inadequate psychosocial, nutritional, and health-related care in infancy may also impair the child’s early physical growth. Empirically, the children of mothers who experience domestic violence have elevated levels of biomarkers of stress (see Yount et al., 2011). Mothers experiencing domestic violence also tend to have poorer prenatal health, higher odds of health risk behaviors in pregnancy, less or delayed prenatal care, and higher odds of a range of poor pregnancy outcomes, including fetal death, low birth weight, and preterm birth (Shah & Shah, 2010; see also Yount et al., 2011). Evidence is more mixed, however, regarding the infant care of mothers who have experienced domestic violence or violence in childhood (Yount et al., 2011). On the one hand, the children of these mothers tend to have higher risks of under-two mortality (e.g., Rico et al., 2010), poorer preventive care including lower immunization levels (e.g., Davalos and Santos, 2006 and Kishor and Johnson, 2004), and illness or infection in infancy (see Yount et al., 2011). On the other hand, the breast-feeding practices and parenting of mothers who experience domestic violence or violence in childhood range from deficient to what researchers have called compensatory (see Yount et al., 2011). Still, these findings overall are troubling, as 11%–71% of women globally report any prior domestic violence ( Ackerson and Subramanian, 2008, Garcia-Moreno et al., 2006 and Kishor and Johnson, 2004), and millions of children in the U.S. alone are exposed annually ( McDonald, Jouriles, Ramisetty-Mikler, Caetano, & Green, 2006). Also, rates of domestic violence may be higher in conflict and post-conflict settings ( World Health Organization (WHO), 2002), where sexual violence is a common tactic of war and exposure to conflict-related violence predicts men’s perpetration of domestic violence ( Bracken and Petty, 1998 and Ward and Marsh, 2006). Despite evidence of the effects of domestic violence on predictors of child growth, evidence of its total (unmediated) effects on child growth is limited and mixed ( Yount et al., 2011). In an analyses of five national Demographic and Health Surveys (DHS), maternal exposure to any physical or sexual violence since age 15 years predicted higher odds of stunting in children 6–59 months in only one country ( Rico et al., 2010). In contrast, a few longitudinal studies mostly in wealthier settings have shown negative effects of domestic violence against mothers on children’s growth ( Asling-Monemi et al., 2009, Montgomery et al., 1997 and Peck and Lundberg, 1995). Yet, the applicability of findings from wealthier settings to poorer countries with different socio-political and epidemiologic environments is unclear. Also, a prior review has exposed a remarkable dearth of research on the intergenerational effects of domestic violence in African countries ( Yount et al., 2011). The above discussion motivates the present study. We examine the total, unadjusted and adjusted associations of maternal exposure to various forms of domestic violence with available anthropometric measures in children less than five years, using data from the 2007 Liberia Demographic Health Survey (LDHS). Liberia is an appropriate setting to examine this association, given its poor nutritional status of young children, high rate of domestic violence against women, and prolonged civil conflict in which sexual violence against women was a systematic tactic of war. Setting

Only recently have researchers begun to appreciate the myriad potential adverse effects on young children of maternal experiences of domestic violence. The gaps in this field are substantial and persistent; yet, young children are disproportionately exposed to domestic violence because of their high dependence on maternal care (Yount et al., 2011). These gaps in research are especially large in African countries, even though domestic violence and child malnutrition are highly prevalent in this region (Yount et al., 2011). This study, thus, contributes new knowledge on the association of domestic violence against mothers with child growth in Liberia. The analysis leveraged representative data from Liberia in which maternal reports of domestic violence, objective measures of child anthropometry, and important potential confounders of household, maternal, and child attributes were considered. The results revealed significant adjusted associations between anthropometric measures in children less than five years and maternal reports of especially sexual domestic violence in the prior year. The persistence and consistency of these relationships, even in an impoverished, post-conflict setting such as Liberia, has several implications. First, these associations were apparent even with controls for multiple other stressors, such as another family death and poverty. Second, these associations were most evident for nutritional indicators that relate to the overall (underweight) and long-term (stunting) nutritional status of children, which strongly predict adverse outcomes into adulthood. Third, although physical and emotional domestic violence are documented elsewhere as having multiple adverse effects, only the associations of sexual domestic violence with child nutrition were evident in this setting. This finding corroborates those of prior studies (Reza et al., 2009 and Ward and Marsh, 2006) and may indicate the severity and embeddedness of sexual domestic violence in post-conflict Liberia, where sexual violence has been used as a systematic tactic of war. Further research across different countries would help to corroborate or refute this interpretation. Also consistent with unpublished studies (Kishor & Johnson, 2004), we found persistent associations of lower mean waz and haz scores in the children of mothers who reported witnessing father-to-mother physical violence as a child. These findings suggest that children’s exposure to domestic violence may have long-term effects on the next generation. Finally, this study revealed the needs for more theorizing and/or more empirical research in several domains. Much needed are longitudinal studies of the implications of domestic violence for children’s well-being into adulthood. Also needed is better theorizing and research on the effects of domestic violence against mothers on children of different ages. Finally, with growing evidence of the total effects of domestic violence on child growth, research is needed to disentangle the biological and behavioral pathways by which this relationship operates ( Yount et al., 2011). This study does have some limitations. First, the analysis was subject to sample selection bias resulting from the high percentage of non-response on children’s anthropometric measures (25%) and maternal exposure to domestic violence (17%). Fortunately, the distributions of measures for domestic violence did not differ across children with and without anthropometric measures, and the estimated coefficients for sexual domestic violence were robust to the inclusion of observed controls. Second, although important confounders were included in the analysis, others were omitted that preclude making causal inferences. Future research might include, for example, information on women’s exposure to sexual violence by non-intimate partners, which may co-occur with other measured forms of violence. Also missing was information related to the complex circumstances of a post-conflict society, which may elevate the risk of violence against women. Such contextual measures may include, for example, limited work opportunities, social alienation, and conflict-related trauma among men. The DHS does not systematically collect such data, and so the results here are interpreted as associational. We recommend that the DHS collect more specific conflict-related outcomes in the post-conflict settings in which it operates. A third limitation is that women may have misreported experiences of domestic violence because of problems with recall, misinterpretations of the questions, or the social stigma associated with disclosing violence (Ellsberg et al., 2008 and Ward and Marsh, 2006). To enhance comprehension and disclosure, however, questions on domestic violence in the Liberian and other DHS focus on specific acts of aggression. Similar questions have been validated cross-culturally (Strauss, 1990) and typically generate higher rates than do single-question formats (Kishor & Johnson, 2004). Still, estimates of domestic violence may reflect minimum bounds of such violence and its associations with child nutrition. A fourth limitation is that some measures of child nutrition had inconsistent or unexpected associations with some control variables, especially the death of a brother, maternal age at first marriage, marital status, and the age and gender of the child. Some of these associations, however, were not significant at the level of 0.05 and should be interpreted with caution. Also, age-related variables are difficult to measure precisely in such contexts, warranting further caution in interpretations. Finally, residual unobserved heterogeneity may have affected some of these estimates. A final limitation is that the DHS is cross-sectional and does not collect comprehensive data on the psychosocial and biological mediators of domestic violence and child nutrition. As a result, longitudinal data collections are needed to disentangle the pathways of effect. Despite these limitations, our findings suggest the need for further action in four areas. First, there is a need to enhance and enforce conventional and customary laws to prevent the occurrence of domestic violence and ensure the rights and protection of survivors. The current Liberian government is trying to do so through the formulation of a Gender-Based Violence National Action Plan and the passage of a Gender and Sexually Based Violence Act in 2008. According to the law, rape and domestic violence are prohibited, but spousal rape is not recognized as a crime ( U.S. Department of State, 2010). Also, the law often is not enforced because of a weak legal system and taboos against reporting sexual violence. As a result, rape and domestic violence remain threats in Liberia ( Rehn and Sirleaf, 2002 and Warner, 2007) and continue to be tolerated ( Casella et al., 2005). Second, attention should be paid to the treatment of mothers who are identified survivors of domestic violence and to screen their children for nutritional deficits. Third, efforts are needed to promote normative change by disseminating information on the likely intergenerational effects of especially sexual domestic violence. Finally, additional research is needed to clarify the biological and behavior pathways by which domestic violence may influence child growth. This research would help to prioritize modifiable points of intervention, thereby mitigating early growth failure and its adverse implications into adulthood. In sum, preventing domestic violence appeals to the ethics of improving public health and ensuring basic human rights (Heise, 1998 and World Health Organization [WHO], 2002). This study has important implications in the context of ongoing post-conflict reconstruction in Liberia. It sheds light on the association of domestic violence against mothers with child nutrition and heightens the need to prevent and remediate violence against women, including that perpetrated in the supposed safety of the home.