Millions of children live in nations affected by war, 1 and yet the resources available to ensure their safety, health and psychological recovery from trauma are severely limited. It is well established that children exposed to war are vulnerable to increased rates of post-traumatic stress disorder (PTSD), depression, anxiety and externalising behaviours (i.e. hostility and conduct difficulties) as measured via structured clinical interviews and standardised questionnaires. Reference Betancourt, Newnham, McBain and Brennan2–Reference Okello, Nakimuli-Mpungu, Musisi, Broekaert and Derluyn4 Recent findings suggest a differential impact of war trauma on psychological disorder by gender. Reference Annan, Blattman, Mazurana and Carlson5–Reference Betancourt, Borisova, De la Soudiere and Williamson7 However, the specific intermediate factors through which adversity exerts its effects on psychological health have not been well defined. Reference Brent and Silverstein8 Recent investigations have questioned the hypothesis that distress arises solely from war experiences, and are shifting the focus towards the important role that the post-conflict environment plays in mediating psychological adjustment. Reference Fernando, Miller and Berger9,Reference Miller and Rasmussen10 Daily stressors, the social and material challenges caused or worsened by political instability and poverty, have traditionally been neglected in assessments of psychological outcomes among war-affected populations. An examination of the specific mediating role (the proportion of associations that are explained by indirect pathways) of daily stressors may inform the development and implementation of interventions that have potential to improve a range of modifiable factors in the post-conflict setting. The current study therefore aimed to investigate, first, the degree to which trauma exposure is associated with post-traumatic stress and depression among war-affected youth 6 years following the Sierra Leonean civil war. Second, we aimed to determine the extent to which daily stressors mediate the association between war exposure and symptoms of post-traumatic stress and depression. Building on prior literature investigating mental health outcomes among war-affected youth, Reference Fernando, Miller and Berger9,Reference Jordans, Semrau, Thornicroft and van Ommeren11–Reference Betancourt, McBain, Newnham and Brennan13 it was hypothesised that war trauma would be more closely associated with symptoms of post-traumatic stress than depression. Further, we predicted that in both cases, the relationship between war exposure and psychological outcomes would be mediated by daily stressors.
Method
Participants and study setting
Participants were drawn from a longitudinal cohort study of war-affected youth in Sierra Leone conducted in collaboration with the International Rescue Committee. Children (male and female, n = 529) were interviewed at three time points (2002, 2004 and 2008) following the cessation of conflict. All participants were between 10 and 17 years old at baseline interview. The sample comprised three cohorts: former child soldiers who received non-governmental services for disarmament, demobilisation and reintegration (DDR) following the civil war (n = 264), former child soldiers who did not receive DDR services (n = 128) and a community sample of war-affected youth (n = 137).
In 2008, 387 youth (comprising 73% of the full sample) were traced and interviewed. The current sample comprised youth interviewed at Time 3 who reported at least one war exposure (n = 363). Data for all participants were collected through individual interviews conducted in Krio, the lingua franca, by Sierra Leonean research assistants trained by the study principal investigators. All participants provided verbal informed consent in Krio, and consent and assent procedures were used where participants were younger than 18 years. As a token of appreciation, all participants were provided a small household item and refreshments (given at the end of the interview to avoid possible effects of coercion). Local social workers provided follow-up and referral for any participants who reported potential risk of harm or suicidality (approximately 5% of the original sample). Ethics approval was granted by the Institutional Review Board of the Harvard School of Public Health and the study was conducted in concordance with the Declaration of Helsinki.
Measures
The assessment battery comprised standardised measures forward- and back-translated from English to Sierra Leonean Krio, consistent with a standardised protocol. 14 Cultural adaptation of items was conducted in consultation with Sierra Leonean staff and community members Reference Betancourt, Brennan, Rubin-Smith, Fitzmaurice and Gilman15 and the survey instrument was refined at each wave of data collection. War exposures were assessed at waves 1, 2 and 3, depression and post-traumatic stress symptoms at wave 2 and 3, and daily stressors at wave 3.
War exposure
The Childhood War Trauma Questionnaire Reference Macksoud and Aber16 was adapted to capture the exposures relevant to young people in Sierra Leone. The resulting 55 items were dichotomously coded (yes/no). Items covered violence witnessed, experienced and perpetrated by the participant, as well as bereavement and deprivation experienced during war. We used a total score, summing all war exposures, as we deemed the cumulative effects of war exposure most appropriate to the current analysis. This approach is consistent with prior research examining the effects of war and daily adversity, Reference Fernando, Miller and Berger9,Reference Jordans, Semrau, Thornicroft and van Ommeren11,Reference Mels, Derluyn, Broekaert and Rosseel12 enabling comparison across studies.
Daily stressors
An adapted Post-War Adversities Index Reference Layne, Olsen, Baker, Legerski, Isakson and P$ansalić17 was administered to assess current hardship and difficulties. The 30 items examine issues related to housing insecurity, interpersonal adversities, food and economic security and neglect. Items include ‘during the past 6 months, have you (your family) been evicted from your home, or forced to move into worse accommodations than you previously had?’ and ‘during the past 6 months, have people you live with had serious conflicts with each other (arguing, shouting, screaming)’. Items were scored dichotomously (yes or no ratings in response to their occurrence) and a summed total score was used in analyses. The scale has shown strong psychometric properties in prior studies. Reference Betancourt, McBain, Newnham and Brennan13
Post-traumatic stress reactions
A reduced version of the UCLA Post-Traumatic Stress Disorder Reaction Index (PTSD-RI) Reference Pynoos, Rodriguez, Steinberg, Stauber and Frederick18 comprising the nine-item short form plus three items from the original long-form deemed most appropriate to the current setting was used to assess symptoms. Frequency of symptoms experienced in the past 4 weeks was reported on a three-point Likert scale (never, sometimes, often) with higher scores corresponding to higher levels of distress. Continuous scores were used. The PTSD-RI has demonstrated strong psychometric properties in prior studies of war-exposed youth, in a variety of settings including Sierra Leone. Reference Betancourt, Newnham, McBain and Brennan2,Reference Layne, Olsen, Baker, Legerski, Isakson and Pasalic19 The 12-item scale had high internal consistency in this sample (Cronbach’s α = 0.84).
Sample characteristics and correlations with war exposures, daily stressors and mental health outcomes
| n | Mean (s.d.) | Range | Correlations | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Gender | Age | War exposures | Daily stressors | PTSD | Depression | ||||
| Gender, female | 97 | 1 | |||||||
| Age | 20.87 (3.38) | 13–24 | –0.101 | 1 | |||||
| War exposures | 14.91 (7.50) | 1–34 | 0.082 | 0.159Footnote ** | 1 | ||||
| Daily stressors | 8.31 (5.07) | 0–23 | 0.005 | 0.160Footnote ** | 0.198Footnote ** | 1 | |||
| PTSD | 11.22 (7.24) | 0–30 | 0.080 | –0.046 | 0.116Footnote * | 0.439Footnote ** | 1 | ||
| Depression | 29.9 (7.34) | 16–49 | 0.063 | 0.072 | 0.079 | 0.511Footnote ** | 0.542Footnote ** | 1 | |
PTSD, post-traumatic stress disorder.
* Significant at the 0.05 level (2-tailed)
** significant at the 0.01 level (2-tailed).
Depression symptoms
The depression subscale of the Hopkins Symptom Checklist (HSCL-25) Reference Lavik, Hauff, Solberg and Laake20 was used to assess depression symptoms. Items are rated on a four-point scale (not at all, a little, quite a bit, extremely) and continuous scores were used. The scale has been used to examine depression and anxiety in a number of war-affected samples with strong psychometric properties. Reference Lavik, Hauff, Solberg and Laake20,Reference Scholte, Olff, Ventevogel, de Vries, Jansveld and Cardozo21 Internal consistency in the current sample was strong (Cronbach’s α = 0.769).
Demographic characteristics
Participants’ age, gender, education and family socioeconomic resources were assessed using a demographic inventory administered with the interview. The inventory draws from items on the UNICEF Multiple Cluster Indicator Survey 22 used previously in Sierra Leone.
Statistical analysis
For continuous variables, means and standard deviations were calculated, and for categorical data, descriptive statistics were calculated using frequencies. Independent sample t-tests were conducted to determine differences between genders. Correlations and descriptive analyses were conducted in IBM SPSS version 20.0 for Windows.
To investigate whether daily stressors mediated any association between war exposure and mental health outcomes, path analyses were developed. We modelled the associations in separate and then combined models, where correlations between each mental health outcome were taken into account. In the final model, all associations were mutually adjusted for each other and path coefficients represent ‘independent’ associations. Standardised path coefficients and standard errors were estimated using bootstrapping (5000 models) in MPlus for Windows. As all scores were continuous, a maximum likelihood estimator was used.
Results
Sample characteristics
The analytical sample comprised 363 youth (26.7% female, mean age 20.9 years, s.d. = 3.38) (Table 1). More than half were currently in school (n = 205, 56.5%) and 63.4% had completed grade 7. Most participants (n = 295, 81.3%) were living with a relative and 132 (36.4%) lived with at least one parent at the time of interview.
War exposures and outcomes by gender
War exposures are outlined in online Table DS1. All participants had been exposed to war, and the average number of exposures was high (mean 14.9, s.d. = 7.5). Abduction into armed forces occurred as young as 2 years of age (mean 10.35 years, s.d.= 2.89) and time spent in the fighting forces averaged 2.86 years (s.d. = 2.82). Females reported higher rates of experiencing rape than males in the sample, although sexual assault and exposure to rape was reported by both genders. No significant differences existed by gender for any other war exposure. Participants reported an average of eight daily stressors (s.d. = 5.07). The most commonly reported stressors were having experienced the death of someone close to the participant in the past year (78.2%), lacking money for basic necessities (54.5%), unemployment of the household’s main wage earner (49.6%), personal illness (48.2%) and a member of the household experiencing depression (44.6%). There were no significant differences by gender in reports of post-traumatic stress symptoms (t(1,361) = –1.12, P = 0.263) or depression scores (t(1,360) = –1.19, P = 0.233).
Associations between war exposures, post-war context and mental health
There was evidence that war exposure and daily stressors were significantly associated with post-traumatic stress symptoms, as outlined in Table 1. There was no evidence for an association between the extent of war exposures and depression, although an association between daily stressors and depression was observed. Post-traumatic stress and depression symptoms were correlated to a moderate degree. Daily stressors and war exposure were moderately correlated, suggesting that they are related but to some extent independent.
Mediation models
Direct and indirect pathways from war exposure to post-traumatic stress symptoms and depression were calculated in separate and combined mediation models. The combined model was consistent with the models calculated for each mental health outcome, and is illustrated in Fig. 1. The total association (combination of all direct and indirect paths) from war exposure score to PTSD score was 0.112 (95% CI 0.014–0.21, P<0.05). Of this total, there was evidence that a significant proportion, 0.089 (95% CI 0.04–0.138, P<0.001) was explained by indirect pathways through daily stressors. Thus, 79% of the association between war exposure and
Path model for war exposure, daily stressors and mental health outcomes among war-affected youth in a post-conflict setting.
All path coefficients are standardised and all variables are continuous.
post-traumatic stress was mediated by daily stressors in the post-conflict setting and once daily stressors were accounted for there was no longer evidence for an association between war exposure and PTSD. The total association (combination of all direct and indirect paths) from war exposure score to depression score was 0.083 (95% CI –0.023 to 0.189, P = 0.127). This association and further variance was accounted for by indirect pathways through daily stressors: 0.103 (95% CI 0.048–0.158, P<0.001). Thus, the association between war exposure and depression symptoms was entirely explained by daily stressors, which helps to explain why the indirect pathway association was greater than the total association.
Discussion
Main findings
Six years following war, the high number of war exposures reported by Sierra Leonean youth was significantly associated with symptoms of post-traumatic stress, but not depression. The pathway from war exposure to psychological distress was largely mediated by current daily stressors for post-traumatic stress symptoms, and for depression there was only evidence for a pathway from daily stressors with no impact of war exposure alone. Effects were tested in both individual and combined models that examined the respective impact of trauma and daily stressors on both PTSD and depression symptoms. In all models the mediating role of daily stressors remained strong. Overall, the large effects of daily stressors suggest that despite exposure to severe and chronic violence during childhood, economic and interpersonal stressors experienced in the post-conflict setting appear to play a substantial role in young people’s psychological impairment.
Daily stressors in post-conflict settings
In recent decades, wars have exerted a disproportionate burden on low- and middle-income countries. Reference Song and Shaheen23 These nations have significantly fewer resources available to revive the health, education, legal and logistical systems necessary to support post-conflict recovery. As such, government and non-governmental agencies must make difficult decisions to direct the few resources available to programmes that will provide the greatest benefit in recovery and rebuilding after conflict. Reference Annan, Green and Brier24 The current findings support a growing evidence base on the importance of daily stressors. For example, among adult Darfuri refugees living in camps in Chad, levels of distress were more closely correlated with basic needs and concerns about safety than with war traumas. Reference Rasmussen, Nguyen, Wilkinson, Vundla, Raghavan and Miller25 Although many of the daily stressors arose from the experience of war, most salient to the participants were the impact of ongoing needs and hardships. Similar findings for Afghan, Reference Miller, Omidian, Rasmussen, Yaquibi, Daudzai and Nasiri26,Reference Panter-Brick, Eggerman, Mojadidi and McDade27 Bosnian, Reference Miller, Weine, Ramic, Brkic, Bjedic and Smajkic28 Iraqi and Bhutanese refugee adults Reference Jordans, Semrau, Thornicroft and van Ommeren11 have been documented, where a range of daily hardships and current stressors were significantly associated with poorer mental health outcomes in situations of ongoing instability. In studies of war-affected adults, variations in gender effects exist. Some studies have reported higher rates of stress among males Reference Rasmussen and Annan29 and others in females, Reference Jordans, Semrau, Thornicroft and van Ommeren11 perhaps reflecting perceptions of belonging, achievement and contribution to the household, as well as the burden of responsibility. In contrast, there were no significant differences in post-traumatic stress reactions or depression symptoms in the current sample.
Among children and adolescents in adverse settings, a combination of traumatic events and daily stressors have been associated with internalising and externalising symptoms. Reference Fernando, Miller and Berger9,Reference Mels, Derluyn, Broekaert and Rosseel12 A study of Sinhalese, Tamil and Muslim students who had experienced trauma related to war and/or the 2004 tsunami in Sri Lanka demonstrated that despite the dual burden of war and disaster, daily stressors of material deprivation, abuse and interparental conflict convincingly mediated the relationship between trauma exposure and PTSD. Reference Fernando, Miller and Berger9 Similarly, internally displaced Eastern Congolese adolescents reported a significantly higher rate of traumatic exposure, higher vulnerability to insecurity and more daily stressors than adolescents who had returned home or those in a non-displaced cohort. Reference Mels, Derluyn, Broekaert and Rosseel12 Interestingly, those who had returned home did not differ significantly on daily stressors from the non-displaced group, suggesting that displacement and ongoing insecurity were particularly influential. Mels and colleagues Reference Mels, Derluyn, Broekaert and Rosseel12 found a stronger association between traumatic events, in combination with daily stressors, and post-traumatic stress reactions and internalising symptoms than that reported in the current study, perhaps because of higher levels of ongoing insecurity in the displaced cohort.
The current findings complement an established evidence base highlighting the important role of resettlement stressors on psychological outcomes among refugee children and adolescents in high-income nations. Reference Lustig, Kia-Keating, Knight, Geltman, Ellis and Kinzie30 It is well recognised that vulnerability to hardship, isolation and adversity together with barriers to inclusion at school and community settings (including language barriers, cultural attitudes and stigma) place refugee-origin children and youth at risk of psychological difficulties. Reference Howard and Hodes31 Socio-economic circumstances, family health and cohesion, community integration and access to education following resettlement are important predictors of psychological adjustment among children resettled in both high- and low-income countries. Reference Reed, Fazel, Jones, Panter-Brick and Stein3,Reference Lustig, Kia-Keating, Knight, Geltman, Ellis and Kinzie30,Reference Fazel, Reed, Panter-Brick and Stein32 Building on our understanding of the challenges facing children who are displaced or affected by disaster, the current study extends a more sophisticated picture of mental health for populations that remain in the post-conflict environment.
Possible targets for intervention
Wars destroy the societal infrastructure crucial to recovery. Health, education, food and water security, legal and protection systems are often damaged in prolonged conflicts, and this has significant implications for survivors. Youth in the current sample reported a high number of stressors 6 years following war, including economic, family and community stressors. Although common to post-conflict settings, these risk factors have the benefit of being potentially modifiable, and are increasingly the target of recovery and rehabilitation efforts. Reference Annan, Green and Brier24 Ensuring child protection, increasing security in camps and improving access to medical care and education are already key targets of humanitarian programming, Reference Neuner33 yet the current findings suggest that more could be done to address the long-term impact of war on young people’s structural and interpersonal environments. Economic opportunity relates strongly to depression in both high- and low-resource settings. Reference Brent and Silverstein8,Reference Lund, Breen, Flisher, Kakuma, Corrigall and Joska34 For youth returning from war, a lack of opportunities in educational and vocational sectors is a significant challenge, especially for females whose prospects are particularly poor. Reference Annan, Green and Brier24 It is important that interventions for youth reporting distress in the post-conflict setting address psychological symptoms in the context of poverty, insecurity and interpersonal stability. Reference Betancourt, Newnham, Hann, McBain, Akinsulure-Smith, Weisz, Raynaud, Gau and Hodes35 Thus, psychological programmes that link with educational, vocational and industry partners may enable greater gains for youth seeking intervention after war. Reference Betancourt, Newnham, Hann, McBain, Akinsulure-Smith, Weisz, Raynaud, Gau and Hodes35
Similarly, social support and interpersonal relationships are well-established protective factors for depression and post-traumatic stress disorder. Reference Betancourt, Newnham, McBain and Brennan2,Reference Betancourt, McBain, Newnham and Brennan13 For children affected by war, including those who were actively involved in the violence, safe and reliable caregiver relationships are vital to psychosocial recovery. Reference Fazel, Reed, Panter-Brick and Stein32 Many children face family loss and disintegration in the aftermath of war, and some return to situations of domestic violence, child abuse and neglect. Reference Annan and Brier36 Sadly, a landscape of loss often cultivates ongoing hardship. For those who were able to return to caregivers, the burden of illness and disability among family members may have increased stress and reduced opportunities to access school and work. Given the relative emphasis on the quality and well-being of caregivers in the Post-War Adversities Index used in the current study, this provides some indication of possible intervention targets. Further exploration of the child’s interaction with their social ecology through adolescence and early adulthood will inform future modes of intervention.
Limitations
Several limitations should be considered. First, although efforts were made to sample from multiple districts in Sierra Leone and across former child soldier and community populations; the sample comprised youth who were all exposed to a particularly brutal and prolonged civil war, and the purposive recruitment methods mean that care should be taken in generalising to other populations. Second, self-reported measures of psychological symptoms were used. We have made substantial efforts to reliably translate, culturally adapt and pilot all measures; however the lack of psychiatric diagnostic assessment available in Sierra Leone means that the symptom scores do not reflect valid diagnoses. Third, self-reported psychological symptoms and stressors may be vulnerable to reporting bias, yet the stressor events covered by the Post-War Adversities Index are largely objective, and the bias in retrospective reports linked to psychopathology is likely to be minimal. Reference Brewin, Andrews and Gotlib37 Fourth, the model was built with cross-sectional data and thus it is difficult to determine the directional nature of the pathways. It is possible that mental health difficulties play a role in generating hardships and ongoing stressors. Yet, the importance of these stressors for mental health, particularly depression, suggests that irrespective of war exposure, daily hardships contribute to mental health problems. Longitudinal research is required to examine the directional nature of the relationship between daily stressors and psychological symptoms among youth living in post-conflict settings. It will also be important to determine the most potent stressors and clusters of adversities that will inform psychotherapeutic treatment mechanisms and broader public health interventions.
Implications
Severe daily stressors in the post-conflict environment were found to mediate the relationship between the extent of war exposure and psychological difficulties among youth in Sierra Leone. These findings suggest potential avenues for intervention. Children and adolescents affected by war are not reacting solely to traumatic exposure; rather, it appears that their engagement with economic opportunities and healthy interpersonal relationships in the aftermath of war are vital to their psychological recovery. Although a small number of interventions for war-affected children have demonstrated effectiveness, Reference Bolton, Bass, Betancourt, Speelman, Onyango and Clougherty38,Reference Ruf, Schauer, Neuner, Catani, Schauer and Elbert39 a large number of treatments are implemented in the absence of empirical support. Reference Tol, Barbui, Galappatti, Silove, Betancourt and Souza40 The model illustrated in Fig. 1 builds on an existing literature that draws attention to the proximal environmental context crucial to recovery, Reference Fernando, Miller and Berger9,Reference Miller and Rasmussen10 and has significant implications for clinical and public health interventions. Despite ongoing debate regarding the timing and focus of psychological interventions in post-conflict settings, Reference Miller and Rasmussen10,Reference Neuner33 it is clear that 6 years on from war, daily stressors are a critical factor in adolescent and youth health. Clinical programmes implemented in post-conflict settings have significant opportunity to address these modifiable factors, and should work to integrate with educational, vocational and social support programmes. Opportunities to engage with families and communities to embrace stability and healthy functioning, in addition to trauma-focused care, are vital for sustainable improvement. Expanding treatment approaches by incorporating a focus on daily stressors therefore reflects an important shift in addressing the needs of youth in post-conflict settings.
Acknowledgements
We wish to thank our research assistants and coordinators in Sierra Leone, and colleagues at the International Rescue Committee for their collaboration in this work. We are most grateful to the participants and their families who have donated their time and thoughts, and enhanced our understanding of mental health and resilience in the post-conflict environment.
eLetters
No eLetters have been published for this article.