Abstract
Expatriate aid workers (n = 214) representing 19 nongovernmental organizations (NGOs) completed a predeployment survey, including measures of mental health (depression, anxiety, and posttraumatic stress disorder [PTSD]); risk factors (childhood trauma, family risk, and adult trauma exposure); and resilience factors (coping, social support, and healthy lifestyle) to assess their baseline mental health during preparation for deployment. Multiple regression analysis indicated that childhood trauma/family risk was not significantly related to depression, anxiety, or PTSD symptoms when controlling for report of prior mental illness; yet, adult trauma exposure was significantly related to all three. Social support contributed significant variance to depression and PTSD. NGOs can help applicants recognize the effects of recent trauma and the resilience provided by a healthy social network.
Keywords
It is estimated that more than 250,000 individuals are working for humanitarian aid organizations around the world (Stoddard, Harmer, & DiDomenico, 2009), and in the last decade, they have faced increased risk of violence in the settings where they work (Sheik et al., 2000; Stoddard et al., 2009). In 2008, the mortality rate of humanitarian aid workers was higher than that of UN peacekeeping troops; 260 humanitarian aid workers were killed, taken hostage, or injured in a serious assault during their service in that year (Stoddard et al., 2009). Between 2006 and 2009, attacks against expatriate humanitarian aid workers have shown a striking increase, likely due to increased political motives in attacks and high security risks in key areas (e.g., Sudan, Afghanistan, and Somalia; Stoddard et al., 2009).
In the context of this increased risk of violence, aid organizations have increasingly prioritized security and support of staff. Organizations such as People in Aid and the Antares Foundation have offered guidelines of practice for staff-support programs and human resource policies (Antares Foundation, 2006; People in Aid, 2003).
Research in the last decade has highlighted areas of proposed risk and resilience for expatriate aid workers living in a variety of global contexts, and conceptual models support the proposed interrelationship of exposure to traumatic and chronic stressors, individual background variables, resources such as support and healthy coping, and organizational characteristics (Olff, Langeland, & Gersons, 2005). Studies have identified that the frequency of direct exposure to life-threatening traumatic events on the humanitarian field is associated with symptoms of depression (Lopes Cardozo et al., 2005; Lopes Cardozo, & Salama, 2002) and posttraumatic stress disorder (PTSD; Eriksson, Vande Kemp, Gorsuch, Hoke, & Foy, 2001; Jones, Müller, & Maercker, 2006). Indirect or secondary exposure to traumatic events has been associated with negative mental health outcomes in expatriate aid workers (Eriksson et al., 2001; Shah, Garland, & Katz, 2007). Humanitarian service also includes exposure to chronic stressors such as workload and team conflicts that affect aid worker functioning (Curling & Simmons, 2010; Eriksson, Bjorck, & Abernethy, 2003).
In addition to exposure to stressors during humanitarian deployment, personal characteristics of aid workers have been associated with negative mental health. For example, young age, gender, previous psychiatric history (Lopes Cardozo et al., 2005), and exposure to prior traumatic loss (Putman et al., 2009), have all been associated with distress in samples of expatriate and national aid workers.
However, there are resources that have been reported as possible moderators of the effects of chronic and traumatic stress during deployment. For example, social support has been identified as a moderator of the effects of trauma exposure on PTSD for humanitarian workers (Eriksson et al., 2001), and inadequate communication with family and friends was identified as a risk factor for depression (Lopes Cardozo et al., 2005; Lopes Cardozo & Salama, 2002). Lack of support from within the aid organization has also been demonstrated as related to depression (Lopes Cardozo et al., 2005).
As aid workers prepare for deployment to areas of violence and harsh living or working conditions, it is important to consider questions, such as is there preexisting emotional distress? What risk factors for distress do they identify (e.g., earlier adult trauma, childhood abuse, previous mental illness)? What is the current state of their use of resources such as social support, coping, and healthy lifestyle habits? What support services do the NGOs offer? How do these risk and resilience factors interact with each other on their effects on distress outcomes?
Our study evaluates these questions to assess the mental health status and organizational support for a sample of aid workers as they prepare to deploy for international field assignments. It is hypothesized that the presence of and frequency of exposure to risk factors such as childhood adversity and exposure to trauma will be significantly related to mental health outcomes (depression, anxiety, and PTSD). However, it is also hypothesized that the presence of protective factors such as social support, healthy lifestyle, and healthy coping strategies will account for significant variance in mental health symptoms when controlling for prior trauma. These data represent the first phase of assessment of a larger longitudinal study (Lopes Cardozo et al., 2012).
Method
Participants and Sampling Procedure
Participants were expatriate aid workers who were hired by a humanitarian organization to work outside their native country. They were recruited for participation in the study during the predeployment phase of their current mission. The population included aid workers with and without previous experience. To be recruited, the planned deployment needed to be a minimum duration of 3 months, and a maximum of 12 months. The predeployment assessment was the first part of a longitudinal study.
International, nongovernmental humanitarian aid organizations (NGOs) excluding UN agencies, local aid agencies, or other governmental humanitarian efforts were targeted for recruitment. NGOs meeting the following inclusion criteria were invited to participate: (a) in existence for more than 5 years; (b) record of international funding; (c) operating with a humanitarian imperative; (d) a record of operations in at-risk countries, including low-income countries or those affected by chronic crisis; and (e) deploying a minimum of 20 expatriate staff to the field per year.
Using an archive of organizational information accessed from Relief Web (http://www.reliefweb.int), the researchers developed a preliminary list of 88 NGOs that appeared to meet inclusion criteria. Research team members attempted to contact all 88 NGOs to determine whether they met the inclusion criteria and to inquire about their interest in the study. The size and activity of the organization was taken into consideration for worker recruitment, as the number of annual staff deployments ranged from 20 to 700. The research coordinator sent each NGO contact person the appropriate number of predeployment questionnaires in proportion to the estimated number of annual staff deployments. At the beginning of recruitment (December 2005), 415 packets were distributed to the NGOs. An additional 172 packets were provided to 12 NGOs who had distributed all of the original packets to allow for more participants to participate prior to the December 2007 ending date. Based on the report of the focal persons, 414 survey packets were distributed by the NGO focal persons to deploying aid workers.
Measures
Instruments included in the predeployment self-report survey questionnaire included personal and job-related demographics, exposure to childhood stressors and adult trauma, coping, support, motivation, lifestyle, and mental health outcomes.
Personal and job-related demographics
Items queried the general demographics of the individual (age, gender, marital status, education, previous humanitarian aid experience, and previous treatment for mental health problems), as well as the work type and deployment location.
Coping, support, lifestyle, and motivation
Coping was assessed using an adaptation of the Coping Strategy Indicator (CSI; Amirkhan, 1994). The three items with the highest factor loadings were used for each of the three subscales: Problem Solving, Avoiding, or Social-Support Seeking. In addition, 12 “aid worker–specific” items were developed in consultation with humanitarian aid administrators and staff care consultants (e.g., “overworked,” “wrote letters or email,” “drank more than usual,” etc.), and these items were used descriptively. The personal mean score for the three original CSI subscales were used in analyses. Perceived social support was assessed using the 12-item Social Provisions Scale (Cutrona, 1989; Cutrona & Russell, 1987). This measure assessed agreement in areas such as shared interests, respect, guidance and advice, and support of others. For this study, a 5-point Likert-type scale was used allowing for an (3) unsure response. The negatively phrased items were reverse scored, and a personal mean score of all answered items was used for inferential analyses.
Healthy lifestyle choices were assessed through a series of questions adapted from the Behavioral Risk Factor Surveillance System Questionnaire (BRFSS; Centers for Disease Control and Prevention, 2000). A Healthy Habits Index was created for analysis by summing the positive responses on the following items: (a) healthy eating (defined as eating 2 or more vegetables or fruits a day or 10 or more a week, eating no more than one serving of junk food a day, and eating a normal quantity of food); (b) healthy caffeine consumption (defined as 2 or less servings of caffeine per day); (c) healthy alcohol consumption (defined as 2 or less drinks on average per day, and no reported binge drinking); (d) no tobacco use; (e) no recreational drug use; (f) exercising approximately 3 or more times per week; and (g) sleeping 7 or more hours per night. Each participant received a score ranging from 0 to 7, indicating the sum of healthy habits reported.
In addition, participants were asked a series of questions about their motivation to be involved in humanitarian aid work. The instrument was created by the Antares Foundation Research Team in September 2003 based on the report of humanitarian consultants and aid workers. The instrument included 9 items describing reasons (job related, personal, or existential) why individuals might choose to do humanitarian aid; participants indicated their level of agreement using a 5-point scale.
Childhood risk and trauma exposure prior to deployment
Participants responded to a series of questions regarding traumatic exposure and adverse childhood background characteristics in the predeployment assessment. They were asked to indicate whether they suffered injuries from their parents’ discipline (item from the Assessing Environment III; Knutson, 1988) and whether their parents ever hit or threatened to hit each other (item from the Conflict Tactics Scale; Straus, 1979). Participants were also given a paragraph that described the nature of unwanted sexual advances (which ranged from forced exposure to nudity, physical contact and fondling, to sexual penetration) and were then asked to report occurrence and frequency of these experiences before the age of 18 (Resnick, 1996). Each of these questions regarding childhood trauma exposure were scored dichotomously for any occurrence and summed to create an index of child trauma (ranging from 0 to 3). A list of adverse family factors was developed based on Rutter, Silberg, O’Connor, and Simonoff (1999), including divorce, removal from home, overcrowding in home, mental illness in family, and death of parent or sibling. These items were also scored dichotomously and summed for a 0 through 5 score. For current inferential analyses these two childhood exposure measures were summed to create an overall childhood trauma and risk scale, similar to the Adverse Childhood Experiences Scale (Felitti et al., 1998).
Adult exposure to trauma prior to the current deployment was assessed using queries regarding exposure to romantic partner violence (Conflict Tactics Scale; Straus, 1979) and adult exposure to unwanted sexual contact (Resnick, 1996). Participants also responded to seven additional items describing categories of potentially traumatic events (accidents, natural disaster, life-threatening illness, crime victimization, serious injury or threatened death, traumatic death of a family member, or witnessing threat or serious injury; Widom, Button, Czaja, & DuMont, 2005). Participants indicated whether they thought they were at risk of death or injury, whether they were in fact injured, and whether this event happened during previous aid deployment. For inferential analyses a “yes” or “no” score for each type of adult trauma exposure was summed to create an overall adult predeployment trauma exposure score (ranging 0 to 9).
Mental health outcomes
The Los Angeles Symptoms Checklist assessed severity of symptoms associated with PTSD, including items in the three symptom criteria categories of reexperiencing, avoidance, and hyperarousal (King, King, Leskin, & Foy, 1995). Participants rated each of 17 items with a 5-point scale identifying “how much of a problem” the symptom was for them ranging from 0 (not a problem), to 4 (an extreme problem). A sum of item scores creates an overall measure of severity of PTSD. For the current analysis, each individual was given a mean score of all answered items.
Symptoms of depression and anxiety were measured by the Hopkins Symptom Checklist (HSCL; Derogatis, Lipman, Rickels, Uhlenhuth, & Covi, 1974; Mollica, Wyshak, de Marneffe, Khuon, & Lavelle, 1987). Participants rated the severity of their experience of 10 anxiety items and 15 depression items using the 4-point scale of 1 (not at all) to 4 (extremely). A personal mean score was created for both subscales.
Procedure
A focal person was identified in each NGO to provide initial contact and recruitment for the study. They were not in a supervisory role with participants and were asked to give a 2-year commitment. They were not given financial incentive, but NGOs provided release time of 1 to 2 hr per week. All focal persons took part in training on ethical treatment of human subjects in research and signed a commitment to confidentiality. The focal person explained the research and handed out the first envelope with the questionnaire to all staff about to be deployed, or with larger NGOs sequentially every 2nd or 3rd newly deployed staff. The enrolment process included a standard oral introduction to the study by the focal person. Potential participants received an envelope containing a letter of introduction, consent forms, the initial set of instruments, detailed written instructions about the enrolment process, a website hosted by the Antares organization further explaining the study, and return envelopes for one signed consent form and the completed instruments to be sent back to the research coordinator. The staff who chose to participate sent the completed questionnaire directly to the research coordinator at the Antares Foundation, without knowledge of the focal persons. Compensation was offered to support participation; participants received US$50 for returning predeployment materials.
The survey protocol was reviewed and approved by an institutional review board at Centers for Disease Control and Prevention (CDC), followed by ethical review boards of the other partner research organizations.
Data management and analyses
Data were managed using EPI-INFO 2002. There were cases with missing items in the mental health and predictor scales; the percentage of cases with missing items ranged from 1.9% to 6.1%. Missing data were handled in two ways: (a) to determine cutoff scores for clinical outcome variables, all items were summed and any participant with a score at the cutoff or above was considered a “case,” even if that participant had up to 10% of the items missing and (b) for continuous scale analysis, personal mean scores were created from all answered items up to 10% missing. A conservative approach was used for variables that represented an experience (such as trauma exposure) such that a nonendorsed item was scored as “no exposure.” Data analyses were performed using SPSS 18. Cronbach’s alphas were generated to assess internal consistency of scales, and bivariate Pearson correlations were used to establish relations between variables. Hierarchical multiple regression analyses were used to adjust for potential confounders in the analysis and multiple risk and resilience factors for each outcome. The variables were entered in steps: (a) possible confounding background variables, (b) child trauma/risk exposure and adult trauma exposure, (c) social support and healthy lifestyle, and (d) coping strategies.
Results
From the list of 88 possible NGOs, 18 did not respond to queries, and 22 failed to respond after initial interest and a review of the research study materials. In addition, 29 NGOs declined participation, giving reasons such as not fitting inclusion criteria or a lack of time. A total of 19 NGOs agreed to participate, representing Continental Europe (7), the British Isles (5), and North America (7). This distribution represents regional response rates from the original list of 16% for European NGOs, 33% for NGOs based in the British Isles, and 28% for North American NGOs.
Two hundred fourteen aid workers returned predeployment surveys to the research coordinator, indicating an overall estimated response rate of 51.6%. Two surveys were excluded due to incomplete information. The respondents represented a variety of job functions and backgrounds (see Table 1). They reported 41 different nationalities, with the majority (54.7%) coming from European backgrounds. A total of 40 different countries were identified as the deployment locations, although more than half of the participants (60.4%) deployed to these 6 countries: Sudan (16.5%), Congo (12.2%), Indonesia (10.8%), Pakistan (8.0%), Chad (7.1%), and Afghanistan (5.7%). Thirty-five percent of the participants indicated that their NGO considered their post to be a hardship assignment.
Demographic Characteristics of the Humanitarian Aid Worker Sample
Descriptive statistics for all predictor and outcome scales, including correlations and alpha coefficients are reported in Table 2. Depression, anxiety, and PTSD were significantly correlated. However, the mean scores for each outcome represented levels of either no distress or slight distress in the sample. The scales of motivation and spirituality were not significantly related to the outcomes (Table 2).
Pearson Correlations and Psychometric Properties of the Major Study Variables
Note: M = mean; SD = standard deviation; α = Cronbach’s alpha coefficient; PTSD = posttraumatic stress disorder.
Cronbach’s alpha coefficients were not created for count variables.
p < .05. **p < .01, two-tailed.
The participants’ reports of exposure to childhood trauma events, family risk factors, and adult traumas are outlined in Table 3. Two thirds of the participants reported one or more experiences of childhood trauma or risk; the most frequently endorsed type of childhood exposure was unwanted sexual contact. Eighty-two percent of the sample indicated one or more traumatic experiences in adulthood, and the most frequent adult trauma was being a witness to an event where someone was threatened with serious injury or death. The events most often associated with a prior humanitarian deployment were exposure to natural disaster and witnessing life-threatening events.
Report of Child Trauma, Family Risk, and Adult Trauma Exposure
Percentage based on number of reported adult events.
Participants indicated overall healthy levels of social support, coping, and healthy lifestyle habits. The mean score for perceived social support (M = 4.26, SD = .45) highlights that participants felt supported during predeployment. The aid worker participants indicated more frequent use of positive coping strategies, such as seeking social support (M = 2.44, SD = .54) and problem solving (M = 2.57, SD = .45), as compared with avoidance coping (M = 1.61, SD = .49). The most commonly used “humanitarian specific” coping strategies were humor (30.7% reporting using this strategy a lot) and writing letters or email (26.3% reporting a lot). Participants also reported an average of 4 out of 7 healthy lifestyle habits; most commonly reported habits were healthy sleeping (85.8%) and avoiding drugs (85.2%). The least endorsed healthy habit was maintaining a healthy diet (13.2%). A little over half of the sample indicated healthy exercise (56.7%) and nonabusive alcohol consumption (55.1%).
The reported motivation for humanitarian deployment suggests an overarching humanitarian interest. Ninety-six percent of participants indicated that they were motivated to pursue aid work to contribute to a better world [and] help those less fortunate; in fact, half of the sample strongly agreed with this statement. Other strong motivations were to experience other cultures, travel (84.4% strongly agree or agree), to satisfy professional interest in the work and/or professional development (80.1% strongly agree or agree), and to have an adventure or a challenge (72.6% strongly agree or agree).
Bivariate correlations (Table 2) indicated significant relationships between the amount of adult trauma exposure and symptoms of all three mental health outcomes (PTSD, anxiety, and depression). Childhood trauma and risk exposure was not significantly correlated to PTSD. Social support and healthy habits were significantly negatively related to PTSD and depression. The coping strategy of avoidance was significantly associated with all mental health symptoms. The results of the hierarchical multiple regressions are presented in Table 4. Each of the 3 regression models accounted for significant overall variance in the outcome: 23% of the variance in PTSD, 20% of the variance in anxiety, and 28% of the variance in depression. Contrary to the hypothesis, childhood trauma and risk exposure did not contribute significantly to anxiety or depression after controlling for prior mental illness. In addition, healthy lifestyle habits and positive coping strategies were not associated with depression, anxiety, and PTSD after controlling for demographics and trauma exposure. Adult trauma exposure accounted for significant variance in depression, anxiety, and PTSD, and social support was associated with depression and PTSD. Avoidance coping accounted for significant variance in all three mental health outcomes.
Hierarchical Multiple Regression Analyses Predicting Mental Health From Trauma Exposure, Social Support, Healthy Behaviors, and Coping Strategies
Note: n = 203. ΔR2 = change in multiple correlation squared; β = standardized beta; total R2 = total strength of association of all predictors with outcome; PTSD = posttraumatic stress disorder. Standardized betas listed are for variables added in each step.
p < .05. **p < .01. ***p < .001.
Discussion
Overall, this sample of expatriate aid workers preparing for deployment demonstrated low levels of mental health distress and frequent use of proactive coping, social support, and healthy physical habits. They also reported a strong humanitarian motivation, with the desire to contribute to a better world being the most highly endorsed reason to adopt this profession. Although the majority of participants reported a history of trauma exposure, their report of interpersonal and childhood trauma was similar to that reported in an earlier expatriate aid sample (Eriksson et al., 2003) with approximately 2 out of 10 reporting unwanted sexual contact as a child and a similar number reporting violence in a romantic relationship.
The results of the multiple regression analyses highlight the relationship between trauma exposure and distress. Exposure to the combined variable of childhood risk and trauma was significantly correlated with depression and anxiety. Yet when prior mental illness was included in the first step of the regression, childhood risk/trauma was no longer significantly related to any mental health outcomes. This suggests that for this sample the report of prior mental health distress was a better predictor of current distress than simply exposure to childhood risk/trauma. However, the frequency of adult trauma exposure maintained its influence on the outcomes. The summed score of adult trauma exposures contributed significantly to the variance of all three mental health outcomes, even when controlling for prior mental illness.
When controlling for demographic variables and trauma exposure, perceptions of current social support continued to contribute significant variance to the severity of PTSD and depression, demonstrating a negative relationship to both outcomes. Due to the cross-sectional nature of these data, a causal relationship may not be inferred. These results parallel similar studies of social support, trauma exposure, and PTSD in aid workers (Eriksson et al., 2001), as well as unit cohesion, trauma events, and PTSD symptoms in predeployment military units (Brailey, Vasterling, Proctor, Constans, & Friedman, 2007). These results confirm the importance of maintaining a strong social support network. However, healthy lifestyle habits did not retain a significant relationship with depression and PTSD once factors of background and trauma exposure were accounted for.
Finally, neither of the positive coping strategies (problem solving and support seeking) accounted for significant variance in the outcomes. However, report of avoidance coping was significantly positively associated with all three mental health outcomes. This is not surprising, as avoidance and withdrawal are components of both depression and PTSD. Longitudinal analyses can determine whether the use of avoidance coping predicts the development of further symptoms.
One particular limitation of the study is the sampling of NGOs. In spite of several attempts for contact and clarification about the study, most NGOs from the initial list of 88 chose not to participate, or they did not respond to the invitation. This limited response is consonant with other attempts to engage NGOs in a review of staff-support programming, (Ehrenreich & Elliott, 2004). Therefore, the NGOs choosing to be involved in the research study may have a specific interest in the topic of staff support, and/or they may represent NGOs with more resources to invest in the project. These characteristics may have some bearing on how the participating NGOs select their staff. Further research can target a wider sampling of NGOs and assess the specific effects of staff support programming in well-resourced NGOs.
In addition to initial organizational recruitment, there were challenges associated with recruiting subjects for the predeployment assessment. First, each NGO had different logistics regarding how staff were recruited and deployed. Some NGOs had staff travel through the headquarters office, and this allowed an easier contact point with the focal person for recruitment into the study. Other NGOs communicated with prospective staff from a distance, and new staff deployed to the humanitarian field without meeting headquarters’ staff. In these situations, focal persons sent the research packet to the recruits’ homes or locations of deployment. In addition, there was often turnover within the administrative and human resource positions, and this required several shifts of training and preparation for the focal persons.
The implications drawn from this study are particularly relevant for NGOs. Although the general sample of deploying aid workers report low-symptom levels, a small percentage has symptoms suggestive of clinical distress. Previous treatment for a mental disorder was a significant predictor of distress across the three syndromes, suggesting that NGOs can encourage applicants to consider the ways they cope with emotional vulnerabilities and the appropriateness of certain settings (e.g., the threat of political instability, remoteness, or access to supportive services). The significant findings related to adult trauma exposure also suggests that NGOs can help staff to “take an inventory” of their recent exposures. Colluding with a tendency to avoid thinking about those experiences, and to move quickly to another deployment, may be a disservice to valued staff. NGOs should consider the timing of redeploying staff that have experienced recent traumas, and when assessing the impact of these traumas symptoms of depression should be highlighted in addition to PTSD. Finally, organizations can contribute to the development of social support, through peer support programs, social activities, and access to email and phone calls to friends and family (Curling & Simmons, 2010; Lopes Cardozo et al., 2005). Further longitudinal analyses of these factors can contribute to the understanding of risk and resilience in aid deployments and appropriate organizational response.
Footnotes
Authors’ Note
The findings and conclusions reported in this article are those of the authors and do not necessarily represent the views of the US Centers for Disease Control and Prevention.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research for this article was funded by the US Centers for Disease Control and Prevention and the Antares Foundation through a cooperative agreement.