The impact of humanitarian emergencies on the prevalence of violence against children: an evidence-based ecological framework

Abstract

Little is known about the patterns and mechanisms by which humanitarian emergencies may exacerbate violence against children. In this article, we propose using the ecological framework to examine the impact of humanitarian emergencies on interpersonal violence against children. We consider the literature that supports this framework and suggest future directions for research to fill identified gaps in the framework. The relationship between humanitarian emergencies and violence against children depends on risk factors at multiple levels, including a breakdown of child protection systems, displacement, threats to livelihoods, changing gender roles, changing household composition, overcrowded living conditions, early marriage, exposure to conflict or other emergency events, and alcohol abuse. The empirical evidence supporting the proposed emergency/violence framework is limited by cross-sectional study designs and a propensity to predominantly examine individual-level determinants of violence, especially exposure to conflict or emergency events. Thus, there is a pressing need to contextualize the relationship between conflict or emergency events and violence against children within the wider ecological and household dynamics that occur during humanitarian emergencies. Ultimately, this will require longitudinal observations of children, families and communities from before the emergency through recovery and improvements to ongoing global surveillance systems. More complete data will enable the humanitarian community to design effective, appropriate and well-targeted interventions.

Keywords:

• Violence against children
• interpersonal violence
• humanitarian emergencies
• child protection
• ecological framework

Introduction

Violence against children is widely recognized as a critical child protection and health concern during humanitarian emergencies, including armed conflict, political unrest and natural disasters (Apfel & Simon, 1996; Machel, 1996; Silverman & La Greca, 2002). Due to a combination of factors (e.g. lack of standardized definitions and tools for measuring violence, a culture of stigma, limited resources and capacity to conduct population-based surveys), it is difficult to quantify the precise magnitude of violence against children in humanitarian emergencies. Still, according to a recent systematic review, numerous studies have documented alarming prevalence rates in settings as diverse as eastern Democratic Republic of the Congo, Afghanistan and Sri Lanka. Both girls and boys are victimized by violence in humanitarian emergencies, albeit in different ways (girls are more likely to be sexually abused and boys are more likely to experience physical violence). Further, in most cases, prevalence rates from the studies in the review did not include infants and young children, suggesting the true prevalence of violence against children is likely even higher than reported (Stark & Landis, 2016). Given the scale of the exposure and the harmful effects of violence against children on individuals’ physical, emotional and social development throughout the life-course, as well as its associated impact on national development potential, preventing and responding to violence against children represents a global public health priority (Felitti et al., 1998; Krug, Mercy, Dahlberg, & Zwi, 2002; Pinheiro, 2006).

Yet, while the importance of addressing violence against children in humanitarian emergencies has become increasingly accepted by programmers and policymakers, most prevention and response work is informed by evidence from outside the humanitarian sphere. The recent systematic review found that there has been a spike in research on violence against children in humanitarian settings in the past five years, but overall, these studies still fail to measure and evaluate variables that adequately capture the rapidly changing contextual dynamics caused by conflict, migration and/or natural disasters. As a result, very little is known about the specific patterns and mechanisms by which humanitarian emergencies may exacerbate violence against children or interact with risk factors for violence against children that precede the emergency. The dearth of information available from the existing literature leaves the humanitarian community unable to identify or to mitigate these risks, or to identify protective factors that may contribute to children’s resilience. In fact, the basic assumption that humanitarian emergencies necessarily increase violence against children is itself unproven and merits critical evaluation (Catani, 2010; Stark & Landis, 2016). The purpose of this article is therefore to propose a framework for thinking about the impact of humanitarian emergencies on interpersonal violence against children, to consider the literature that supports this framework and finally, to suggest future directions for research to fill identified gaps in the framework.

Framework

Definitions

For our purposes, humanitarian emergencies are defined as circumstances brought about by armed conflict, natural disasters or political unrest that cause ‘widespread human, material or environmental losses’ and impair the ability of a society ‘to cope using its own resources’ (Red Cross & Red Crescent, 2012). In keeping with the Sphere Standards, this definition also allows for both slow and rapid onset events, as well as protracted emergencies (Red Cross & Red Crescent, 2011). In addition, a setting is considered an emergency if it occurs within any phase of the humanitarian response cycle, ranging from acute response to early recovery and development (Inter-Agency Standing Committee, 2011).

Violence is defined in terms of four forms of intentional, interpersonal harm: physical violence, sexual violence, mental violence and neglect. Definitions of these forms of violence are drawn from the 2014 UNICEF report on violence against children (UNICEF, 2014). Physical violence is defined as ‘all corporal punishment and all other forms of torture, cruel, inhuman or degrading treatment or punishment as well as physical bullying and hazing by adults or by other children.’ Sexual violence is defined as ‘any sexual activities imposed by an adult on a child against which the child is entitled to protection by criminal law’ or ‘committed against a child by another child if the offender is significantly older than the victim or uses power, threat or other means of pressure.’ Mental violence is defined as ‘psychological maltreatment, mental abuse, verbal abuse and emotional abuse or neglect.’ Neglect is defined as ‘the failure to meet children’s physical and psychological needs, protect them from danger or obtain medical, birth registration or other services when those responsible for their care have the means, knowledge and access to services to do so.’ Finally, in accordance with the United Nations Convention on the Rights of the Child, children are defined as those under the age of 18 years (UNICEF, 1989).

Relationship between humanitarian emergencies and violence against children

The relationship between humanitarian emergencies and violence against children depends on ecological, household and individual risk factors and this relationship may fluctuate at different stages of the humanitarian response cycle (e.g. conflict and post-conflict). Further, any theoretical model must recognize that the relationship between humanitarian emergencies and violence against children will differ across contexts due to variability in the underlying population characteristics, gender roles and features of the emergency itself (e.g. scale, duration, involvement of civilians, levels of morbidity and mortality, levels of displacement). Brofenbrenner’s ecological framework has been influential in drawing attention to the interplay between environmental and social forces and individual behaviors and traits (Bronfenbrenner, 1979). The framework has previously been applied to understanding issues ranging from early childhood education to public health promotion and may also be usefully applied to understand how emergencies affect violence against children.

At the ecological level, humanitarian emergencies are associated with a breakdown of systems, including legal, medical and social services, compared to prior to the emergency (Haj-Yahia & Abdo-Kaloti, 2003; Wexler, Branski, & Kerem, 2006). Informal community groups such as religious associations, school clubs and women’s groups and networks of extended family also tend to be disrupted by emergencies and related displacement (Wessells & Monteiro, 2004). These factors lead to weakened community coherence, impaired social support and isolation, all of which make it more difficult for violence against children to be recognized, reported and prevented. In addition, displacement creates situations where populations may resettle to locations where they are perceived as a stigmatized minority, resulting in repeated exposure to discrimination (Stark, Plosky, Horn, & Canavera, 2015). Subsequent ‘minority stress’ is a risk factor for poor mental health, which is known to exacerbate violence (Charles & Denman, 2013; Pascoe & Smart Richman, 2009). To some extent, the arrival of UN agencies and NGOs may introduce new services and structures to reduce violence against children and ‘build back better,’ especially in the post-conflict period. However, truly embedding these programs within the social fabric of a destabilized community is a tremendous challenge (Canavera, Lanning, Polin, & Stark, in press; Training Resources Group & Play Therapy Africa, 2012).

Humanitarian emergencies also have profound economic consequences which originate at the ecological level, but ultimately manifest in households and individuals. Emergencies threaten basic livelihoods and increase poverty and food insecurity (Cliffe, 1994; Drapcho & Mock, 2000). Often these extra financial pressures generate anxiety and tensions amongst parents and caregivers struggling to provide for their children and such feelings can fuel violence. Further, changing economies frequently lead to changing gender roles. For example, men may be unable to maintain their traditional role as ‘breadwinner,’ and women may take up income-generating activities that would otherwise not have been within their domain. More women and girls may engage in transactional sex or be forced into early marriage, making them vulnerable to violence. In some cases, men may resent these changes to social norms and attempt to reassert their power through violence against their families and others. Men may also be prone to drink alcohol to cope with their shame and boredom, which in turn may contribute to further violence. Alternately, women may become empowered by their new economic roles and be more able to assert their independence from men (Biswas, Rahman, Mashreky, Rahman, & Dalal, 2010; Boutron, 2012; Charles & Denman, 2013; Horn, 2010; Horn, Puffer, Roesch, & Lehmann, 2014; Payne, 1998).

Finally, households and individuals are directly impacted by humanitarian emergencies in a myriad of ways that affect violence. Household composition is regularly altered, especially as able-bodied men join the armed forces, migrate or die, leaving behind disproportionate numbers of women, children and elderly (Brück & Schindler, 2009; Ezeoha, 2015; Hill, 2004). Households may also absorb extended relatives or neighbors, including children who have been separated from their primary caregiver as a result of the emergency (Stark et al., 2016). These changes increase the household dependency ratio and may cause further emotional and financial stress and violence, especially for single parents or caregivers and unaccompanied children (Ezeoha, 2015; Hadley, Belachew, Lindstrom, & Tessema, 2011). At the same time, people are often living in overcrowded housing conditions and may have limited freedom of movement outside the home due to security concerns and/or curfews (Charles & Denman, 2013; Haj-Yahia & Abdo-Kaloti, 2003). Again, these conditions likely increase stress and violence, though it has also been posited that close, multi-family living arrangements may also be protective of violence against children because perpetrators are embarrassed to display violence in the presence of strangers (Usta & Masterson, 2012).

In terms of the direct effects of humanitarian emergencies on individuals, it is important to recognize that the emergency shapes the experiences and behaviors of both adults and children. Both age groups should be considered in tandem to fully understand the dynamics of violence. Adults and children are more likely to witness traumatic events during humanitarian emergencies, including shelling, gunfire, abduction, torture, destruction of property, death of loved ones and domestic abuse between adults in their own household (Haj-Yahia & Abdo-Kaloti, 2003; Qouta, Punamäki, Miller, & El-Sarraj, 2008). Children who witness conflict or other emergency events are also more likely to show signs of post-traumatic stress and displays of anger and aggressiveness, which may provoke violent reactions from parents, caregivers and peers (Qouta et al., 2008). Adults and older children with post-traumatic stress are prone to perpetrate interpersonal violence themselves (Catani, 2010; Catani, Jacob, Schauer, Kohila, & Neuner, 2008; Catani et al., 2009). The effects of post-traumatic stress on violence perpetration have been documented to persist over the lifecourse and across generations (Devakumar, Birch, Osrin, Sondorp, & Wells, 2014; Saile, Ertl, Neuner, & Catani, 2014). These long-term effects are often compounded by lack of educational opportunities for children growing up in emergencies and permanent disabilities caused by military victimization or other injuries (Charles & Denman, 2013; Qouta et al., 2008).

Evidence

The empirical evidence supporting the proposed emergency/violence framework is limited by cross-sectional study designs and a propensity to predominantly examine individual-level determinants of violence. Despite these limitations, a growing body of quantitative research has consistently found that parental exposure to conflict or other emergency events predicts the perpetration of violence against children, and that this association is highly correlated with children’s exposure to conflict or emergency events, post-traumatic stress in parents and children, paternal alcohol use, children’s aggressive behavior and parental socio-demographics characteristics like low income and low education (Catani et al., 2008, 2009; Haj-Yahia & Abdo-Kaloti, 2003; Qouta et al., 2008; Saile et al., 2014; Sriskandarajah, Neuner, & Catani, 2015). The directionality between parental exposure to conflict or emergency events and the correlated variables cannot be established due to the cross-sectional nature of the data. This means, for example, it is unknown whether parental low income increases the risk of parental exposure to conflict or emergency events, or parental exposure to conflict or emergency events increases the risk of parental low income. However, the relationship between exposure to conflict or emergency events and violence against children does appear to exhibit a dose-response pattern, with more frequent and severe exposure predicting greater risk of violence against children (Catani et al., 2008, 2009; Haj-Yahia & Abdo-Kaloti, 2003). Further, correlations between exposure to conflict or emergency events and violence against children have been detected decades after the emergency has ended, though measurement of conflict or emergency events may be particularly subject to recall bias in studies with long time lags (Devakumar et al., 2014; Gupta et al., 2009; Saile et al., 2014).

Of note, the literature on the trauma and violence effects associated with acute natural disasters is much more limited than the literature from protracted conflict settings. One study from Bangladesh did find a significant association between exposure to floods and violence against children (Biswas et al., 2010). However, a study from Sri Lanka that measured both Tsunami exposure and war exposure, only war exposure predicted violence against children and in fact, Tsunami exposure was protective. The author hypothesized that the short duration of the Tsunami might have led to a period of family cohesion and that the longer duration of the war hindered similar household dynamics, but this theory is purely speculative (Catani et al., 2008).

In contrast to the relatively robust research agenda to identify individual-level determinants of violence against children in emergencies, knowledge about ecological and household-level determinants of violence against children in emergencies is based on a combination of indirect quantitative evidence and non-representative qualitative narratives. For instance, there are studies that show that emergencies cause changes to household employment status, and there are studies that show changes in employment status increase the risk of violence, but there are no studies that have evaluated the full causal pathway from emergency to change in employment status to violence against children (Brück & Schindler, 2009; Krishnan et al., 2010). Without this type of longitudinal follow-up in two exchangeable populations with varying levels of emergency exposure, it is impossible to estimate the isolated effects of the humanitarian emergency, rather than the general effects of other economic fluctuations in society. Simplistic comparisons between rates of violence in conflict-affected areas versus stable areas or between rates of violence during times of conflict and times of peace are deeply flawed because such comparisons are obscured by a plethora of residual confounders (e.g. governance, culture) (Catani, 2010; Peterman, Palermo, & Bredenkamp, 2011; Usta & Farver, 2010). Also, without accounting for the ecological and household-level determinants of violence against children in study designs, the aggregate effects of humanitarian emergencies on violence remain unknown. In other words, while exposure to conflict or emergency events seems to increase violence against children, if multi-family living arrangements or greater economic independence for women simultaneously decrease violence against children, capturing the interaction between these variables is crucial to measuring the cumulative impact of emergencies on violence against children.

It is worth noting that qualitative narratives can and have introduced useful insights regarding the complex dynamics driving violence against children in humanitarian emergencies. Qualitative methods have revealed potential mechanisms and depths of understanding about violence that would otherwise not have been evident to researchers, such as awareness of local gender norms, details about living conditions and ideas about how both of these elements have changed as a result of the emergency (Charles & Denman, 2013; Horn, 2010; Usta & Masterson, 2012). Because this type of data is not representative by nature, qualitative studies are not appropriate for definitively describing population-wide trends. Still, qualitative approaches are an essential formative tool in generating hypotheses, developing surveys, explaining perceptions and beliefs and triangulating findings (Bolton, Tol, & Bass, 2009).

Future directions

In synthesizing the theories and evidence on the impact of humanitarian emergencies on violence against children, several research gaps have emerged. First, there is a pressing need to contextualize the relationship between conflict or emergency events and violence against children within the wider ecological and household dynamics that occur during humanitarian emergencies. In what ways do various features of emergencies predict the degree and type of violence against children? Which features are most significant? Do protracted conflicts have fundamentally different effects than natural disasters, and why might this be the case? Does violence against children increase and decrease at different stages of the humanitarian response cycle? If so, how and why do these changes happen and what are the optimal periods in which to intervene to prevent violence? Are there any factors that tend to emerge during emergencies which may be protective of violence against children and can these factors be harnessed as part of interventions?

Ultimately, the answers to these questions depend on causal inference, which will require longitudinal observations of children, families and communities from before the emergency through recovery. Improvements to ongoing global surveillance systems at multiple levels will also facilitate the availability of baseline data when emergencies inevitably, though unexpectedly, arise. The national Violence Against Children Surveys (VACS) that have been conducted in 11 countries so far are an excellent example of the movement towards better surveillance (Chiang et al., 2016). VAC-type studies should continue to be supported and expanded, with consideration of creating specialized modules for emergency-prone locations and inclusion of community-level data.

For many years, the humanitarian community mistakenly operated under the assumption that HIV prevalence was heightened during conflict. Comprehensive surveillance data from seven countries allowed a rigorous appraisal of this assumption and eventually led to a paradigm shift in understanding and mitigating HIV risk amongst displaced populations (Spiegel et al., 2007). Once again, the humanitarian community must carefully examine its basic assumptions, this time with regards to violence against children. More complete data on violence against children will enable the humanitarian community to design effective, appropriate and well-targeted interventions. Children in emergencies deserve an evidence-based approach to their protection and well-being.

Funding

This work was supported by Know Violence in Childhood.

Disclosure statement

No potential conflict of interest was reported by the authors.