A Mixed Method Study of the Effects of Post-Migration Economic Stressors on the Mental Health of Recently Resettled Refugees

Abstract

After years of emphasis on pre-migration trauma as the major determinant of refugee mental health, researchers have begun to explore the effects of post-migration stressors on refugees’ distress. However, few studies have brought together refugees’ emic understandings of the effects of economic stressors on their mental health with quantitative datasets to further explore the salience of stress processes as an explanatory mechanism. In qualitative interviews, 40% of 290 recently resettled adult refugees noted that economic stressors were a major source of distress and described pathways through which these stressors negatively influenced their mental health by limiting their ability to learn English, obtain meaningful employment, access healthcare, maintain contact with their families, and integrate in their communities. In structural equation modeling of quantitative data, we tested several possible hypotheses that emerged from the qualitative findings. We find that post-migration economic stressors mediated the relationship between migration-related trauma and post-migration emotional distress and PTSD symptoms. These findings provide empirical support for stress proliferation as a mechanism through which trauma exposure contributes to distress.

INTRODUCTION

Life here is very hard. Sometimes I feel like I’m losing my mind. There is so many things to do, I don’t know where to start. When I left Africa, they told us that, “You are going to America. It’s a nice place, safe place,” but I don’t see that. We had so many problems. Before we came, my wife was abused, sexually abused, and she was beaten. Her arms, she was about to lose her arm. Then another thing, right before we came, we lost our kid, one of our children. Then when we arrived here, my wife was pregnant, three months, and then she lost it, too. It’s been so hard for us. Another problem is transport, and I think that’s why my wife lost the baby. It takes us too much time to catch the bus…and then we have to go to different places, so we have to use the bus. It takes so long to get the bus, so she gets so tired. Another thing is this house where we are living now, they told us that next month we are gonna start to pay rent ourselves, and no one is working. We are not sure what we gonna do next month. Maybe we don’t have a place to stay next month. My wife is depressed because she was tortured in Congo and then here, they’re sending us letters saying that our case would be closed. They sent us three letters already saying that our cash assistance would be closed. When we read all these letters and my wife have this problem, then I start thinking about the future. Then, she loses her mind.

Social inequalities influence the experiences of refugees through multiple pathways--from forced displacement as a result of conflict and political violence to structural violence in countries of asylum and resettlement that results in unequal access to resources and opportunities. As this statement by Kalem,¹ a man from the Democratic Republic of Congo who recently resettled in the United States with his wife Brigid and children illustrates, post-migration stressors not only influence the mental health and well-being of newcomers, but also exacerbate the negative effects of pre-migration trauma exposure. For this family, economic stressors influenced their sense of safety and security, transportation, health care, and employment access, and made them very distressed and hopeless. At the time this man was describing their experiences, his wife had been committed to an inpatient mental health hospital for suicidality. This quote not only illustrates the complexity of multiple migration-related traumas and post-migration economic stressors, but also the importance of identifying effective points of intervention to improve the mental health of resettling refugees.

Guided by Pearlin and colleagues’ (2005) explication of stress proliferation as a way of understanding mechanisms through which trauma exposure and social and economic factors contribute to health disparities, this study integrates in-depth qualitative interview data from 290 refugees who had recently resettled in the United States from Afghanistan, Burundi, Democratic Republic of Congo, Iraq, Republic of Congo, Rwanda, and Syria with structural equation modeling of quantitative data from the same participants to build evidence of the importance of addressing the malleable root causes of refugee mental health problems.

BACKGROUND AND LITERATURE REVIEW

The Context of Refugee Resettlement

According to the United Nations High Commissioner for Refugees, over 79 million people were forcibly displaced in 2019, of which 26 million are classified as refugees (UNHCR 2020). By definition, refugees have encountered persecution and been displaced from their land, homes, and countries. Many have experienced the death of family members and friends, harsh living conditions in their home countries and in refugee camps, and a wide range of other traumatic events. Upon resettlement, refugees face additional difficulties and new stressors. Having left family and friends behind, refugees experience culture and language disruptions, making their transition into a different life harder (Knipscheer and Kleber 2006). Adjusting to a host country involves being able to navigate systems in order to access the resources necessary to survive and succeed, such as learning to speak a new language, finding a job, and becoming financially secure.

Refugee Mental Health

When host counties adopt hostile attitudes towards refugees, their symptoms of distress and post-traumatic stress disorder can be worsened (Schweitzer et al. 2006; Pernice and Brook 1996). Even when living in a receptive community, refugees still face challenges, such as unemployment, difficulty accessing education, and low income. This makes resettlement harder than merely engaging with a new and different culture (George and Jettner 2014; Gorst-Unsworth and Goldenberg 1998). Because of their past experiences of trauma, refugees’ abilities to deal with the added stressors of resettlement should be of primary concern. It is important that research focuses on further elucidating refugee resettlement experiences and their effects on mental health, in order to help refugees adapt to their new homes and reduce their distress. People who have experienced serious adversity (which most refugees have, by definition) are at greater risk for mental health problems because the initial trauma(s) and social disruptions they have endured through violence and/or persecution in their home country and displacement from their homes are likely to lead to exposure to additional stressors. This concept of stress proliferation (Pearlin et al. 2005) has not been tested with refugee populations, although it may be an important framework for understanding and addressing refugee mental health disparities.

Refugee mental health has, however, been a focus of study across many disciplines. One of the debates arising from this multidisciplinary research involves the relative utility of trauma-focused and psychosocial approaches in the treatment of mental health problems among war-affected populations. Miller and Rasmussen (2010) identified the limitations of a trauma-focused approach, arguing that the model failed to recognize the daily stressors that result from war and violence, and their effects on mental health. They emphasized that a focus on daily stressors does not intend to ignore the importance of psychiatric treatments, but rather aims to understand how the stress of everyday life may mediate the effects of trauma exposure on mental health. These arguments are consistent with the broader body of work on the adverse consequences of chronic and episodic life stressors, particularly economic stressors, for individual well-being, including both mental and physical health (Pearlin, Lieberman, Menaghan, Mullan 1981; Pearlin, Schieman, Fazio, and Meersman 2005).

Poverty is one of the major stressors related to years of conflict experienced by many forcibly displaced people. It can negatively influence family relations, diminish access to education, and ultimately limit one’s ability to earn an income. Taken together, conditions that stem from poverty often cause mental health problems and influence the ability of refugees to cope with other mental health issues (Fernando, Miller, and Berger 2010). Stress proliferation may be one process that explains how an initial trauma causes other stressors or traumas that further erode mental health (Pearlin et al. 2005). For example, the persecution or threat of persecution and attendant violence experienced by refugees in their home countries leads to forced displacement, which results in disruption of social roles and status, and loss of social support, community, and resources, all of which are important predictors of mental health. Thus, for example, the impact of Brigid’s sexual assault on her physical health (miscarriage) and mental health (suicidality) might not have been as dire if she and her family had enough money to pay their rent and access needed health, food, and transportation resources during their resettlement in the United States.

Language proficiency can influence refugees’ economic situations by affecting their ability to access other resources in the community such as public transportation, employment, housing, and social support (Miller et al. 2002). Low paying jobs are often the only jobs available for people with limited language proficiency. These jobs are more likely to be demanding, lacking in employee benefits such as sick days, paid time off, health care coverage, and lacking flexibility in schedules, making it difficult for refugees to invest additional time learning the new language or going to school (Dawood 2011). In turn, lack of language skills often negatively influences access to resources and income earned, making adaptation very difficult and adding to refugees’ distress (Beiser and Hou 2001).

Post-migration social stressors clearly have an effect on the mental health of resettled refugees (Hynie 2018; Kim 2016; Miller et al. 2002; Miller and Rasmussen 2010, 2014, 2017; Rasmussen et al. 2010). A meta-analysis of the effects of current life stressors on the mental health of conflict-affected forced migrants found that daily stressors were related to higher PTSD symptoms and emotional distress and mediated the relationships between previous trauma and anxiety, depression, and PTSD symptoms (Hou et al. 2020). Financial and employment stress have been found to be related to symptoms of anxiety and depression among refugees (Miller et al. 2002; Schweitzer et al. 2006), and unemployment has also been shown to be related to levels of aggression and hostility in exile populations (Lavik et al. 1996). Similarly, Campbell (2012) found that refugees who were dissatisfied with their employment, financial status and housing had worse mental health outcomes. However, we add to this literature by examining the potential role of economic stressors as a mediator and/or moderator of the relationship between refugees’ prior trauma exposure and their PTSD symptoms and emotional distress post-migration. In addition, we include descriptions of refugees’ economic struggles and how they experience these stressors influencing their mental health. We believe that mixed methods approaches are critical for understanding refugees’ perspectives in order to elucidate potential pathways and mechanisms. Understanding the challenges that economic stressors bring to the lives of refugees and how they affect their mental health will allow us to develop effective individual, community, and structural interventions by facilitating the allocation of resources where they can have the greatest positive effects on ameliorating mental health problems among resettled refugees.

METHODS

Data Source

The data analyzed in this study were collected as part of a randomized controlled trial of the Refugee Well-being Project (RWP), a community-based mental health intervention. The study was funded by the National Institute on Minority Health and Health Disparities (R01MD007712) and approved by the University of New Mexico’s Human Research Protections Office. The RWP intervention aimed to promote positive mental health and reduce psychological distress by supporting refugees’ integration into their communities, valuing their knowledge and strengths, connecting them with university students to engage in mutual learning, helping them access resources, and promoting their community involvement and advocacy. The mixed method RCT employed a longitudinal randomized waitlist control group design. Quantitative and qualitative data were collected from a total of 290 adult refugee participants enrolled over four years (2013–2016) in a midsized city in the Southwestern United States. The participants came from three regions: Afghanistan (36%), Iraq and Syria (33%), and the Great Lakes region of Africa (Burundi, Democratic Republic of Congo, Republic of Congo, and Rwanda, 31%). All refugees 18 and older from these regions who had arrived in the United States within the past three years and were living in the city where the study took place were invited to participate; 88.8% agreed. Bilingual/bicultural research team members contacted all refugee households, based on complete lists of refugees resettled each year that were compiled in coordination with the two refugee resettlement organizations in the city and on community networks (see Goodkind et al. 2020 for more details of participant recruitment).

Each participant completed 4 interviews during their participation in the study (initial enrollment/pre-intervention, mid-intervention, post-intervention, and 6-month follow-up), which occurred over 12 months. All interviews included a series of fixed response items that assessed participants’ psychological well-being and distress, social support, environmental mastery, access to resources, and English language proficiency. Quantitative data were collected face-to-face via computer-assisted personal interviews (CAPI) at each participant’s home with a bilingual/ bicultural interviewer in the participant’s native language. Quantitative pre-interview data was used for the analyses in this paper (N=290). Pre-interviews also included a qualitative semi-structured component; at each subsequent time point (mid-, post- and follow-up), a smaller, purposively sampled number of participants (n=66) completed the qualitative interview. The qualitative interviews were conducted in participants’ native languages with interpretation to English by teams of one research assistant and one bilingual, bicultural interpreter.

Participants

At the time of their pre-interviews, participants had been in the United States less than 36 months (M = 28.6, SD = 26.76). Fifty-two percent were women, and their ages ranged from 18–71 years (M = 34.6, SD = 11.58). On average, participants reported low English language proficiency. Forty-nine percent reported not having finished high school, 19% had completed high school, and 31% had completed some post-secondary education. A majority were married (58%), with 33% single, 7% widowed, and 2% divorced. Seventy percent of participants had children (M = 2.3 children, range 0–9).

QUALITATIVE ANALYSIS

The questions in the semi-structured qualitative interviews began with asking about participants’ resettlement experiences, including the best and most difficult things about being in the United States. These were followed by questions about specific domains—how things were going with family in their household, family in other places around the world, their heath, their social network, cultural differences, and access to resources. We used a multi-staged approach to analyze the qualitative pre-interviews. Interviews were professionally transcribed and imported into NVivo10 (QSR International, 2010). Interviews were coded by question automatically using NVivo’s auto-coding feature. Coders then completed a “first cycle” coding using a hierarchical coding tree with descriptive codes (Miles, Huberman & Saldaña 2013). Each interview was coded by at least two trained research assistants and bilingual, bicultural research team members, who worked independently to allow for accuracy and better interpretation of the data. Through querying methods, we were able to gather all responses coded at the “finances” and “mental health” nodes. Subsequently, the second author (with mentorship from the first and fourth authors) reviewed all text at these nodes, using a focused coding technique (Charmaz 2014), to look at the range of responses and patterns within and between nodes, and wrote analytic memos that explored the effects of economic stressors on participants’ mental health.

QUALITATIVE RESULTS

Analysis of the qualitative data revealed that 119 of the 290 participants (41%) described economic concerns that contributed to their emotional distress. The qualitative data revealed that refugee participants’ economic concerns intersected with pre- and post-migration trauma and stressors and influenced their mental health. Participant narratives most clearly underscored the ways in which economic stressors influenced their mental health and how multiple post-migration stressors were interrelated.

Emic Descriptions of Distress Caused by Economic Concerns (Income, Employment, Housing)

Among the different types of emotional distress caused by economic concerns, the most predominant was “worry,” but “stress” and being “afraid” were also frequently used to describe the emotional effects of financial challenges. For example, a 29-year-old Burundian woman said:

“It’s hard because so far I need money for kids, like right now I need money to buy the kids jackets and stuff like that because it’s cold and shoes. I have to look for more money. I need more money to pay rent because it’s not enough. We get like $400 and that’s not even enough to pay the rent. It’s just so hard. I’m really worried.”

Similarly, a 42-year-old Iraqi man referred to his living situation as being so difficult that he did not have any hope of achieving a good life: “We’re not happy here… most importantly [because of] the financial issue. I couldn’t achieve anything and I don’t think I will be able to achieve anything, even for my children, if I stay here. So, I’m not happy.” The three main economic stressors described by refugees were lack of satisfaction with or concerns about their income, their employment, and their housing.

Income and employment.

Not being employed or earning insufficient income were frequently described by participants as having a negative effect on their mental health. Most refugees who were employed had low-wage, entry level jobs with limited or no benefits. Refugees who were unable to work or had not yet found work, and those who received insufficient income from their jobs, relied on government assistance as one of their sources of income. However, an additional economic stressor described by participants was denial, reduction, or discontinuation of government assistance. For example, a 23-year old Congolese woman described how the income from her mother’s recently acquired job went to rent but how subsequent cuts to Supplemental Nutrition Assistance Program (SNAP) benefits that occurred as a result of her mother’s new job influenced her sense of security, stability and her mental health:

Thinking, because first of all, we’re here, and we were thinking once we get to America, they are going to help us. It’s actually not true, because you have to fight for your life…It’s not like they’re going to help you like what we were thinking. And the moment where you start raising up and trying to fight for yourself, the little help they were giving you, they take it away. So, it’s hard. We just have to thank God, but it’s hard. It’s not easy. So, thinking is one thing that’s stressing me out. It’s like I always have something on my mind.

Housing.

Paying rent was one of the biggest concerns for those who talked about economic difficulties affecting their well-being. Some participants mentioned the fear of being homeless as constant in their minds. For example, a 41-year-old Iraqi woman said: “My first worry … I am scared or afraid if I’m going to be homeless one day because a hard or difficult time here… I am worried cause I don’t have anything and I’m scared to be homeless.” Newly arrived refugees often needed help learning how to navigate housing systems, in addition to finding the monetary means to pay for housing. Spending the majority of income on rent with little to nothing left to cover additional necessities was a source of emotional distress.

Pathways through which Economic Stressors Contributed to Mental Health Problems

Many participants said they did not have enough income to cover rent, bills, and other expenses. This situation created difficult living conditions that participants said influenced their mental health.

Foreclosing possibilities for learning English.

With only three months of financial support (the amount refugees in the study received from the federal government through local resettlement agencies), refugees were not able to dedicate sufficient time to learning English before they were expected to find employment. Once employed, they had little time to learn English. Those who did pursue learning English, faced challenges with respect to financial responsibilities. Many participants referred to their financial issues as negatively influencing their ability to focus, as a 30-year-old Afghan woman related: “I think the most important thing about me is paying the rent. It’s very stressful. I can’t even focus on learning English because I’m always stressing about the rent.” Not knowing the language was an impediment for refugees who were seeking better paying jobs, higher education, or re-certification of their previous profession, making it difficult for them to navigate their new communities and socialize with people who did not speak their language.

Lack of opportunity for meaningful work.

Many participants referred to their current job as not paying enough money to cover all expenses, being too hard, or lacking benefits. In addition, beyond the economic aspects of the job, some participants had life goals that were career focused. Others saw having a job as a way to engage in social interactions and meet people, and to be financially independent and feel useful. One of the major problems skilled and educated refugees faced during resettlement was that their certifications and/or degrees were not accepted in the United States. Without certification, refugees’ only option was to accept low wage jobs with no benefits. Accepting a lower status job than one’s previous employment or performing job responsibilities that were not representative of one’s education and skills was very difficult for many participants. Sometimes it represented a loss of one’s own dignity, especially for men, as discussed by a 41-year-old Iraqi man: “We worry especially about financial things because of the job, because I cannot like accept any job…But, I have to do something to support my family…we usually live in our country in decent living and good condition. Yeah, so we will not accept to live like in less economic affairs. It’s difficult for us.” Finding work commensurate with the skill and educational level people arrived with was described as a challenge by many.

Providing for children.

Being able to provide for their children was another primary concern. A 36-year-old Iraqi man voiced fears about becoming homeless, but framed them in the context of providing for his children. Beyond providing basic needs for their children, parents also worried about being able to provide a good quality of life for them, as described by this 32-year-old Congolese woman:

So, what worries me for right now, I hope that my husband will have a better job because on his job now it’s not really good as far as income, so this is the first one. The second one, this apartment, I really have a lot of challenges in it, I don’t have a laundry, it’s very small, so if he got a better job he might get better payments and then we might go to live in a home because for my kids they cannot go and play outside, they don’t have a garden, we don’t have a park or anything, I can’t even buy them a bicycle because where to put it, the apartment is very small so we are not feeling comfortable in this apartment, so I’m worried about us financially because we need something better and if the kids are not feeling comfortable, as a mother I will not feel comfortable. So being in the apartment unable to run, unable to see a garden or play in garden or in a park, that will affect them mentally, and then it will not be good for them. So, that’s what really worries me.

Lacking networks of support, refugee parents had to deal with their financial struggles on their own. Ensuring the well-being of children was difficult when there was no one to turn to for help in situations of need. These economic stressors resulted in additional stressors in these instances, as parents felt that their social roles as providers were undermined by their inability to access financial resources, which further negatively influenced their mental health. A similar pattern was also evident for those who had left family behind. Unfortunately, reconnecting with their family members abroad was also influenced by refugees’ financial situation.

Family separation and financial support expectations of family members.

Participants talked about two main ways in which the distress of family separation was related to economic stressors: economic barriers to staying in touch with family members and expectations of family members that resettled refugees would send them financial support. For example, a 28-year-old Congolese woman discussed not being in touch with family due to lack of economic resources. She had only had enough money to talk to them three times since arriving in the United States. When asked how this had affected her mental health, she replied, “I’m worried because they are back home.” Being unable to fulfill family members’ expectations for financial support also caused a great deal of distress among those living in the United States, as a 31-year-old Congolese man stated:

My health is good but emotionally, psychological not really because back home everyone has their own issues and come to me. Oh, we have this issue. We have this issue. We need you to send us money. We need your help. Everybody’s asking for money because they need help with the issues they have. Sometimes, I think about it a lot and I get stressed out and maybe depressed and I have medication that I take. But it’s so hard on me. I have all this pressure and everybody’s just counting on me. Everybody is asking for money so we have to keep sending them whatever we have and then we don’t even have.

Being out of touch with family due to financial difficulties and the stress of fielding requests for financial resources from family members in refugees’ home countries were additional salient examples of pathways through which lack of economic resources proliferated additional stressors.

Trauma Impeding Access to Economic Resources

Several participants cited trauma that they suffered prior to their arrival in the United States as a major factor in their ability to work, which in turn, affected their ability to secure financial resources to start a new life. For example, a 36-year old Iraqi man explained:

In my country, in Iraq, I had five surgeries in my head and in my ear. When I went to Lebanon, I did another surgery. At that time, I couldn’t hear at all. But step-by-step, by the time I can hear a little bit. When I first arrived United States, I had some pain. At the same time, I had a trauma. Because of my son. I am blaming myself that I am the reason my son passed away. Until this moment, I cannot forget that. I can see that even in my dreams. All this kind of medication, they don’t help me. Just a little bit.… Because the thing… it is bigger than myself. At the other end, because of my situation, I cannot work. Because I have a lot of reports that I have medical problems, I cannot work. And this affects me also. As you know, I have to pay for my rent, I have to pay for the power, for the utilities, for everything… Since a long time ago, I applied for the housing authority. I still haven’t heard. So my financial issue is very, very weak. I cannot support myself. Because I was living a life, and right now I am living another life… I cannot afford myself. But the problem with my daughter. My daughter asking a lot. She needs a lot. She needs more. She doesn’t understand. At the beginning, she was helping me. Right now, she doesn’t. She wants to go back again to Iraq. She wants to see her mom. This thing is difficult for me.

This participant clearly linked his financial difficulties to his past trauma, which had physical and mental health consequences for him. Economic stressors proliferated as he struggled to support himself and to provide for his child. Ultimately, the participant’s social role as a provider was put into jeopardy by these interrelated and proliferating stressors.

Trauma Exacerbated by Economic Stressors

Some participants’ descriptions suggested that previous traumas were exacerbated by economic stressors. For example, unemployment and the loss of government assistance made a 47-year-old man from Afghanistan concerned that his depression would return:

The problem which we have lately is a financial problem, which is like [what] most of the people have…I worry about, lately we’re concerned about, the financial problem makes me maybe [have] more depression, like I find out a depression problem for me in the future just worrying about that.

Similarly, the experience of Brigid described by her husband Kalem in the opening of this paper powerfully demonstrates how refugees’ past traumas were more likely to lead to ongoing mental health problems when economic stressors were high. These participant perspectives highlighted not only how economic stressors were intertwined with migration-related trauma and subsequent mental health problems, but also how a lack of basic economic security proliferated or led to numerous other stressors that put survival, quality of life, and the future in question.

In summary, participants often identified economic stressors as root causes of their distress. Further, economic stressors were described as precursors to many other stressors that negatively influenced refugees’ mental health, including lack of time to learn English, inability to obtain meaningful work commiserate with qualifications and experience, lack of access to healthcare, inability to fulfill parental roles, and disconnection from separated family. In addition to delineating these pathways, the qualitative data revealed multiple ways in which migration-related trauma, economic stressors, and psychological distress seemed to be interrelated, including the ways economic stressors might exacerbate the mental health effects of past traumas and the ways past traumas might contribute to economic stressors and thus greater distress.

QUANTITATIVE ANALYSIS

Because the current study also included the collection of quantitative data, we were able to test specific hypotheses that emerged from the qualitative data to better understand the relationship between economic stressors and mental health problems. Based on existing literature and our qualitative findings, it seemed that refugees’ migration trauma was related directly to mental health problems but also might be influencing their ability to access economic resources such as housing and employment in the United States, which could be contributing to increased emotional distress and PTSD symptoms (mediating hypothesis). It also seemed plausible that the size of the relationship between migration-related trauma exposure and mental health problems might vary depending on economic stressors post-migration (moderating hypothesis). Therefore, we examined a moderation model, hypothesizing that the relationship between migration-related trauma exposure and mental health problems would be decreased for refugees with fewer economic stressors post-migration (Hypothesis 1), and a mediation model to test the hypothesis that refugees with more migration-related trauma would experience greater economic stressors, and would thus indirectly experience increased emotional distress and PTSD symptoms (Hypothesis 2). The mediation model was tested against an alternative model because the quantitative data we analyzed was cross-sectional.

QUANTITATIVE MEASURES

Dependent Variables

Emotional distress was measured with the Hopkins Symptoms Checklist-25 (HSC-25; Mollica et al. 2004; Derogatis 1974). Responses are based on a 4-point Likert-type scale with categories ranging from 1 (not at all) to 4 (extremely). Ten questions measure symptoms of anxiety (M = 1.45, SD = .59, Cronbach’s α = .92), and 15 measure symptoms of depression (M = 1.54, SD = .59, Cronbach’s α = .93). The total score is the mean of all items. Participants’ scores ranged from 1.00 to 3.67, with M = 1.50, SD = .58. Cronbach’s α = .96. PTSD symptoms were measured by the PTSD Symptom Checklist-Civilian version (PCL-C). The PCL-C is a 17-item self-reported measure on a 1–5 rating scale from “not at all” to “extremely.” A total score is calculated and higher scores indicate higher levels of PTSD symptoms. Participant scores ranged from 17 to 81, M = 29.95, SD = 15.02, Cronbach’s α = .95.

Migration-Related Trauma Exposure

Migration-related trauma exposure was measured using 27 population appropriate dichotomous items adapted from existing trauma scales (Weine et al. 1995; Foa et al. 1993; and the Harvard Trauma Questionnaire: Mollica et al. 1992). These included experiencing the destruction of one’s home, being tortured or witnessing torture, and being separated from family and friends. Scores ranged from 0 to 27 (M = 5.92, SD = 4.91, Cronbach’s α = .92).

Economic Stressors

We created a measure from three questions assessing how participants felt about their financial situation, all on a scale of 0 (very dissatisfied) to 6 (very satisfied). In the analyses, these were reverse coded so that a higher score would indicate higher stress (0 = very satisfied and 6 = very dissatisfied). The questions asked about satisfaction with income (M = 3.81; SD = 1.80); employment situation (M = 3.62; SD = 1.93), and housing (M = 2.57 SD = 1.78). Questions were combined to form a scale with an overall mean score of 3.32 (SD = 1.22, Cronbach’s α = .70).

Covariates

Covariates were included in fully specified multivariate models and retained based on performance in model (see Aneshensel, 2013). These controls included gender, age, national origin, time in the United States, household size, monthly income, and physical limitations. Gender was a dichotomous variable (1 = female, 0 = male; M = .48, SD = .50). National origin was categorized using three item responses, with “Iraqi” as the reference group (0), “Afghan” (1), and “Great Lakes African” (2). Time in the United States was measured in weeks (M = 29.91, SD = 27.96, Range = 0 to 166.57). Household size was measured using a single continuous variable with the number of people living in the household (including the participant) ranging from 1.00 to 11.00 (M = 5.01, SD = 2.18). Monthly income was measured using a single continuous variable with income range from 0.00 to $3000.00 (M = 706.79, SD = 667.77). Physical limitations were measured using two questions that asked participants to respond about how their physical limitations influenced their daily activities. Items were measured on a 5-point Likert scale from not at all (1) to an extreme amount (5). Total scores ranged from 2.00 to 10.00 (M = 4.27, SD = 2.42; Cronbach’s α = 85).

QUANTITATIVE RESULTS

Main Analyses for Path Models

Path analyses were conducted to assess the conceptual models using maximum likelihood (ML) estimations in AMOS structural equation modeling (SEM) software (Arbuckle 2013). Using ML estimations sidesteps issues associated with influential outliers that would influence model fit, as well as issues of normality that influence parameter estimates (Hancock and Liu 2012). Normality assumptions were met, and no issues of collinearity were present. All constructs were examined as manifest indicators. This was done to conserve power given our relatively small sample size (Cole and Preacher 2014).

Comparative Fit Index (CFI) and Goodness of Fit Index (GFI) were ≥.95 (adequate if ≥.90), and the Root Mean Square Error of Approximation (RMSEA) was ≤.06 (adequate if ≤.08). The Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) were used to compare model fit between models (West, Taylor, and Wei, 2012). These fit indices were assessed as path models were generated. Bollen-Stine bootstrap procedures with 6000 bootstrap resamples were also used to assess the consistency of the proposed model to the sample data. Bollen-Stine bootstrap results with a p value greater than .05 indicate that the proposed model is consistent with the sample data (Walker and Smith 2016).

Hypothesis 1: Moderation Model Results

The moderation model included the direct relations between both migration-related trauma exposure and economic stressors and emotional distress and PTSD symptoms (see Figure 1, Model 1). Additionally, we examined the interaction between migration-related trauma exposure and economic stressors on emotional distress and PTSD symptoms. Covariates were examined for inclusion in the fully specified model. Gender, national origin, age, and physical limitations were retained for subsequent analyses based on performance in the fully specified model (see Table 1).

Figure 1. Standardized Moderating and Mediating Path Models (N = 290).

Note. Model 1 fit: χ² = 13.48 (15), p = .56, GFI = .98; AGFI = .97; RMSEA = .02 (95% CI: .01, .05); AIC =55.48; BIC = 132.55 (Saturated BIC = 204.11). Covariates in the model were gender, national origin, age, and physical limitations. Post-traumatic Stress Disorder (PTSD). Model 2 fit: χ² (14) = 15.29, p = .37; GFI = .99; AGFI = .97; CFI = .99; RMSEA = .02 (95% CI = .01 – .06); AIC = 77.29 (Saturated Model = 90.00); BIC = 191.06. Covariates in the model were gender, national origin, age, income and physical limitations. Post-traumatic Stress Disorder = PTSD. *p < .05, ** p < .01, ***p < .001

Table 1.

Unstandardized and Standardized Coefficients, and Significance Levels for Moderation Model (Standard Errors in Parentheses; N = 288)

Parameter Estimates	Unstandardized	Standardized
Structural Model
Migration Trauma Exposure → Emotional Distress	.10 (.02)	.31***
Migration Trauma Exposure → PTSD Symptom Severity	.20 (.02)	.43***
Economic Stressors → Emotional Distress	.10 (.02)	.27***
Economic Stressors → PTSD Symptom Severity	.11 (.01)	.43***
Migration Trauma Exposure × Economic Stressors → Emotional Distress	NS	NS
Migration Trauma Exposure × Economic Stressors → PTSD Symptoms	NS	NS
Controls
Gendera → Emotional Distress	.13 (.04)	.22***
Gendera → PTSD Symptom Severity	.14 (.03)	.15***
Nationalityb → Emotional Distress	−.14 (.02)	−.35***
Nationalityb → PTSD Symptom Severity	−.21 (.02)	−.38***
Age → PTSD Symptom Severity	.03 (.001)	.11**
Physical Limitations → Emotional Distress	.04 (.01)	.31***
Physical Limitations → PTSD Symptom Severity	.04 (.01)	.23***

Notes. NS = Not Significant; Model Fit: χ² (21) = 24.62, p = .26; GFI = .98; AGFI = .97; CFI = .99; RMSEA = .02 (95% CI = .01 – .06); AIC = 92.05 (Saturated Model = 110.00); BIC = 217.40.

^**p < .01,

^***p < .001

^aReference group for gender is male.

^bReference group for nationality is Iraqi.

For the moderation model, Figure 1, Model 1 presents the over-identified path model, with standardized regression weights reported. The unconstrained model demonstrated good model fit: χ² (21) = 24.62, p = .26; GFI = .98; AGFI = .97; CFI = .99; RMSEA = .02 (95% CI = .01 – .06); AIC = 92.05 (Saturated Model = 110.00); BIC = 217.40. Bollen-Stine bootstrapping results showed that the p-value was greater than .05 (p < = .67), indicating that the proposed model was consistent with the sample data. This model accounted for 45% of the variance in emotional distress and 48% of the variance in PTSD symptom severity. Migration-related trauma exposure was positively associated with emotional distress (p < .001) and PTSD symptoms (p < .001), and economic stressors were also positively associated with emotional distress (p < .001) and PTSD symptoms (p < .001). However, the interaction effect of migration-related trauma exposure and economic stressors had no direct association with outcome variables. Thus, the best-fitting model showed the direct and independent associations of migration-related trauma exposure and economic stressors on both emotional distress and PTSD symptoms but did not support the moderating hypothesis that the relationship between migration-related trauma exposure and mental health problems would be decreased for refugees experiencing fewer economic stressors post-migration. Women reported higher emotional distress and PTSD symptoms (p < .001) than men. Afghans and Great Lakes Africans were less likely to report emotional distress and PTSD symptoms (p < .001) than Iraqi participants. Increased age was associated with higher PTSD symptoms (p < .001), and more physical limitations were associated with increases in emotional distress and PTSD symptoms (p < .001).

Hypothesis 2: Mediation Model Results

Figure 1, Model 2 presents the over-identified path model, with standardized regression weights reported (see Table 2). This model tested the mediating influence of economic stressors between migration-related trauma and emotional distress and PTSD symptom severity. Gender, national origin, age, income and physical limitations were retained as covariates based on performance in the fully specified model. The unconstrained model demonstrated good fit χ² (14) = 15.29, p = .37; GFI = .99; AGFI = .97; CFI = .99; RMSEA = .02 (95% CI = .01 – .06); AIC = 77.29 (Saturated Model = 90.00); BIC = 191.06. Bollen-Stine bootstrapping results showed that the p-value was greater than .05 (p < = .58), indicating that the model was consistent with the sample data. The model accounted for 12% of the variance in economic stressors, 46% of the variance in emotional distress, and 42% of the variance in PTSD symptoms. A positive direct association was present between migration-related trauma exposure and economic stressors (p < .001). A positive association was identified between migration-related trauma exposure and both emotional distress (p < .001) and PTSD symptoms (p < .001). Economic stressors also had a positive direct relationship with both emotional distress (p < .001) and PTSD symptoms (p < .001). Using bias-corrected bootstrap confidence intervals, the following partial indirect associations from migration-related trauma through economic stressor were significant (p < .05): migration-related trauma to emotional distress (indirect effect: .08, 95% CI = .03, .09); and migration-related trauma exposure to PTSD symptom severity (indirect effect: .06, 95% CI = .03, .07). The indirect effect proportions indicated that economic stressors mediated 15.7% of the association between migration-related trauma exposure and emotional distress, and 11% of the effect between migration-related trauma exposure and PTSD symptom severity. Results suggest that while direct associations were present between migration-related trauma exposure and both emotional distress and PTSD symptom severity, economic stressors partially mediated these effects, indicating that economic stressors may exacerbate the examined negative outcomes. Thus, hypothesis 2 was supported - refugees with more migration-related trauma had higher emotional distress and PTSD symptoms in part because they experienced greater economic stressors related to housing, employment, and income². Women reported higher emotional distress and PTSD severity (p < .001) when compared to men. Afghans and Great Lakes Africans were less likely to report emotional distress and PTSD symptoms (p < .001), when compared to Iraqi participants. Higher monthly income was associated with lower economic stressors (p < .01). Older age was associated with higher PTSD symptom severity (p < .001). Similarly, increased physical limitations were associated with higher emotional distress and PTSD symptom severity (p < .001), and higher economic stressors (see Table 2).

Table 2.

Unstandardized and Standardized Coefficients, and Significance Levels for Main Analytic Mediation Model and Alternative Model (Standard Errors in Parentheses; N = 288)

	Unstandardized	Standardized
Main Analytic Mediation Structural Model
Migration Trauma Exposure → Economic Stressors	.27 (.05)	.27***
Migration Trauma Exposure → Emotional Distress	.10 (.02)	.32***
Migration Trauma Exposure → PTSD Symptom Severity	.19 (.02)	.44***
Economic Stressors → Emotional Distress	.11 (.01)	.29***
Economic Stressors → PTSD Symptom Severity	.10 (.02)	.22***
Controls
Gendera → Emotional Distress	.15 (.03)	.22***
Gendera → PTSD Symptom Severity	.13 (.04)	.15***
Nationalityb → Emotional Distress	−.14 (.03)	−.35***
Nationalityb → PTSD Symptom Severity	−.21 (.02)	−.38***
Income → Economic Stressors	−.10 (.01)	−.16**
Age → PTSD Symptom Severity	.03 (.001)	.10*
Physical Limitations → Economic Stressors	.06 (.02)	.15*
Physical Limitations → Emotional Distress	.04 (.01)	.31***
Physical Limitations → PTSD Symptom Severity	.04 (.01)	.23***
Alternative Structural Model
Emotional Distress → Economic Stressors	.79 (.19)	.38***
Economic Stressors → Migration Trauma Exposure	.11 (.05)	.11*
PTSD Symptoms → Migration Trauma Exposure	.65 (.15)	.45***
Controls
Nationalityb → Migration Trauma Exposure	.34 (.08)	.24***
Nationalityb → Economic Stressors	−.15 (.07)	−.12*
Age → Migration Trauma Exposure	.02 (.004)	.20***
Income → Economic Stressors	−.12 (.01)	−.18***

Notes. Main Model Fit: χ² (14) = 15.29, p = .37; GFI = .99; AGFI = .97; CFI = .99; RMSEA = .02 (95% CI = .01 – .06); AIC = 77.29 (Saturated Model = 90.00); BIC = 191.06. Alternative Model Fit: χ² (14) = 29.96 p =.008; GFI = .97; AGFI = .93; CFI = .97; RMSEA = .06 (95% CI = .03 – .09); AIC = 130.22 (Saturated Model = 56.00); BIC = 199.85

^*p < .05,

^**p < .01,

^***p < .001

^aReference group for gender is male.

^bReference group for nationality is Iraqi.

When testing the alternative model specification (emotional distress and PTSD symptom severity directly predicting migration-related trauma exposure and indirectly increasing migration-related trauma exposure through economic stressors), the overall model fit to the data had less than adequate fit: χ² (14) = 29.96 p =.008; GFI = .97; AGFI = .93; CFI = .97; RMSEA = .06 (95% CI = .03 – .09); AIC = 130.22 (Saturated Model = 56.00); BIC = 199.85. See Table 2 for unstandardized and standardized regression coefficients, and standard errors. Fit indices were outside the range of acceptable model-to-data fit. Further, the AIC value for the hypothesized model (AIC = 77.29) was closest to the saturated model of 90.00, when compared to the alternative model (AIC = 148.39; Saturated Model = 72.00). The BIC for the hypothesized model was 191.06 and 229.14 for the alternative model, with a difference greater than 10, further indicating support for the hypothesized model. Bollen-Stine bootstrapping results showed that the p-value was less than .05 (p = .005), indicating that the alternative model was inconsistent with the sample data and that there were issues with model specification (Enders 2002). Consequently, bias‐corrected bootstrap confidence intervals to test indirect paths yielded non-significant results. This suggests that mediation in the alternative direction was not present, and no results are provided for this model. Therefore, we conclude that the most probable order of association is as we hypothesized (economic stressors mediated the association between migration-related trauma exposure and both emotional distress and PTSD symptoms).

DISCUSSION

This mixed method study brought together in-depth qualitative interview data from 290 refugees who had recently resettled in the United States from Afghanistan, Burundi, Democratic Republic of Congo, Iraq, Republic of Congo, Rwanda, and Syria and structural equation modeling of quantitative data from the same participants to provide important insight into the ways in which economic stressors negatively influence refugee mental health. Results from open-ended qualitative questions revealed that economic stressors were salient to more than 40% of respondents in terms of identifying the primary causes of their emotional distress. In addition to having a direct effect on refugees’ mental health, economic stressors appeared to contribute to the presence of other daily stressors in the lives of resettled refugees.

Economic difficulties influenced the extent to which refugees were able to access education and healthcare. Without the means to afford basic necessities such as rent, food, and clothing, necessary healthcare was often put on hold. While many people in the United States face these same dilemmas, it is critical to recognize and address these issues for refugees because most arrive with significant physical health (and sometimes mental health) problems due to their forced displacement and flight experiences. If these issues are not immediately addressed, our findings show that they lead to multiple other intersecting difficulties that impede integration and well-being. Moreover, lack of language acquisition was also related to financial difficulties. Learning English was often put aside to address economic struggles, just as healthcare was. Delays in gaining English proficiency impede refugees’ ability to find better paying jobs, having long-term effects on educational attainment and economic security. A related important finding was that the extent to which refugees were able to communicate with loved ones abroad depended on whether they had the economic means to do so. Research has shown that family separation negatively influences refugee mental health more strongly than multiple other traumatic experiences (Miller, Hess, Bybee, and Goodkind 2017). These findings reinforce the importance of addressing economic stressors in efforts to reduce refugee health disparities.

Furthermore, these results support Pearlin and colleagues’ stress process model (Pearlin et al. 1981; 2005), particularly in terms of illustrating mechanisms of stress proliferation through which migration-related traumas lead to economic stressors, which in turn lead to other daily stressors that contribute to poor mental health. In addition to the focus on major trauma as a primary stressor that results in numerous additional stressors, Pearlin and colleagues propose another pathway through which a stressor may proliferate further hardships—disruption of roles/status—which is evident in our qualitative data. Participants frequently described status loss in terms of prior career attainment and current employment (or lack of) and their inability to adequately fulfill their roles as parents. Thus, the pathways participants described through which trauma exposure and economic stressors contributed to poor mental health reveal that displacement and economic strain are especially debilitating when they result in social role disruption. However, our research also suggests that there is the possibility—through reducing economic stressors by increasing access to financial resources—of interrupting these cascades. These findings also align with research demonstrating the salience of changed economic status on refugee mental health after resettlement (Porter and Haslam 2005). Social status and valued social roles should be considered in efforts to support refugees’ mental health (Rousseau 2018).

Another significant contribution of these findings is additional empirical support for the growing evidence that daily post-migration stressors mediate the relationship between pre-migration trauma (e.g., war-related violence exposure) and emotional distress (Fernando, Miller and Berger 2010; Miller and Rasmussen 2010). The supported mediation path model demonstrates that economic stressors significantly mediate the effect of displacement-related trauma on both emotional distress and PTSD symptoms. These findings highlight the critical contribution of economic stressors to participants’ mental health, above and beyond the influence of past traumas and provide important empirical support for stress proliferation as an explanatory mechanism that connects trauma exposure, economic stressors, and mental health.

Importantly, this study also builds empirical support for Miller and Rasmussen’s (2017) ecological model for understanding and addressing the mental health of refugees and other forcibly displaced populations, which suggests first attending to and reducing post-migration daily stressors for all refugees and then providing specialized mental health treatments for those who need them. However, our findings also extend and build upon this work because of our specific focus on economic stressors (Miller and Rasmussen’s model incorporates but does not disentangle a broad range of post-migration stressors). Although our study and many others have demonstrated that past trauma significantly influences refugees’ mental health problems, our findings also highlight that the post-migration context, particularly in terms of economic resources and stressors, is critical for facilitating healing and reducing the negative effects of pre-migration trauma. Economic stressors can be diminished by providing resettled refugees the necessary tools to succeed when they arrive. Thus, it is particularly important to consider policies and programs that ensure adequate economic resources for refugees.

Limitations

Although our findings contribute to understanding the economic and social complexities that affect the mental health of refugee populations, one limitation is the variability in length of time that participants had been living in the United States. Among the participants were people who had been in the United States for a week and some who had been here for two years, a difficult period during which things can still be unsettled. A longitudinal study would establish the relationship between the variables over time, and thus provide a better understanding of how financial difficulties are related to mental health as time after resettlement increases. In addition, although the sample for this study included 89% of all refugees resettled within a three-year period in one city, it cannot be assumed that these results are generalizable to other geographic contexts within the United States and around the world or to other refugee populations not included in the sample. For example, local attitudes towards newcomers, availability and accessibility of social services, local economic conditions, and political climate vary substantially across communities, which could possibly influence experiences of economic stressors and their effects on valued social roles, social status, and mental health.

It is also important to note that our measure of economic stressors involves participants’ subjective assessments of their income, employment, and housing conditions. Thus, this variable may in part reflect participants’ ability to cope with their economic conditions. Therefore, the mediating relationships we detected may indicate not only that previous trauma results in poorer economic outcomes, but also that previous trauma may make one less able to cope with economic stressors. We took this approach because we wanted to link qualitative data on people’s lived experiences and their subjective assessment of their economic stressors with similar quantitative data. However, we examined the relationship between one objective measure of economic conditions (monthly income) and our subjective measure of economic stressors (r= −.69, p<.001), and we included monthly income as a covariate in our models. It would be useful in future research to include objective measures of economic conditions as main analytic variables.

Finally, while there is a general agreement that medium size samples (N = 200, 300) are reasonable to obtain accurate model-to-data fit in SEM (West, Taylor, and Wei 2012), some more recent investigations suggest that model fit indices tend to approximate a closer estimation to the sample at N ≥ 500 (Shi, Lee, and Maydeu-Olivares 2019). It is possible that additional significant relationships (e.g., moderation) would have been detected with a larger sample size. Thus, future studies are encouraged to draw on larger sample sizes to ensure greater accuracy in model-to-data fit and findings, although this can be challenging with relatively small populations, such as refugees. Despite these limitations, the corroboration of our findings using both qualitative and quantitative data, as well as our ability to explore mechanisms through which trauma exposure and economic stressors contribute to poor mental health outcomes in the qualitative data and to test specific hypotheses about how economic stressors influence the relationship between prior trauma exposure and mental health problems, provide meaningful insight into understanding and addressing mental health disparities experienced by refugees.

Directions for Future Research

In addition to the importance of longitudinal research that can examine and potentially replicate and extend these findings over longer periods of the resettlement process, our qualitative findings suggest several important pathways that could be tested quantitatively with larger samples of refugees. In particular, complex models to more fully test stress proliferation processes could include not only migration-related traumas and subjective economic stressors, but also the many other stressors that refugees described as influencing their mental health.

Conclusion

A growing body of research has shifted the emphasis of research on refugee mental health from a sole focus on pre-migration traumas to include the relevance of post-migration daily stressors, but studies that specifically examine economic stressors as mediating and/or moderating the relationship between refugees’ prior trauma exposure and their PTSD symptoms and emotional distress post-migration have been lacking. In addition, few studies have included emic descriptions of refugees’ economic stressors and pathways through which these stressors influenced their mental health. Mixed method data from this study revealed that financial difficulties lead to mental health problems among refugees through myriad other stressors that proliferate from forced displacement and economic stressors, including inability to learn English and further education and professional recertification, disruption of parental and other social roles, and barriers to healthcare. Our results provide important policy guidance for reducing refugee mental health disparities. Hosting nations should allocate resettlement resources where they can most effectively make a positive change in the lives of refugees. Providing additional time before refugees have to be economically self-sufficient would maximize their opportunities to learn English, gain understanding of the ways in which new health, employment and other systems work, and rebuild their lives in a new country. It is important that refugees are empowered to play significant roles in society, which requires making sure that our systems are providing the resources necessary for them to have control over their own lives and situations. Returning to the experiences of the Congolese family described at the opening of this paper, the United States should strive to be the “nice place, safe place” that refugees are expecting, a place where they would not be subjected once more to the fear or reality of economic precarity and losing their homes, but this time from structural instead of political violence.

Table 3.

Unstandardized and Standardized Coefficients, and Significance Levels for Supplemental Model Testing the Association between Depression, Anxiety, and PTSD Symptoms through Economic Stressors (Standard Errors in Parentheses; N = 288)

Structural Model	Unstandardized	Standardized
Migration Trauma Exposure → Economic Stressors	.23 (.06)	.23***
Migration Trauma Exposure → PTSD Symptoms	.17 (.02)	.39***
Migration Trauma Exposure → Depression Symptoms	.08 (.02)	.23***
Migration Trauma Exposure → Anxiety Symptoms	.08 (.02)	.24***
Economic Stressors → PTSD Symptoms	.07 (.02)	.15***
Economic Stressors → Depression Symptoms	.08 (.02)	.21***
Economic Stressors → Anxiety Symptoms	.07 (.02)	.24***
Controls
Nationalitya → PTSD Symptoms	−.18 (.02)	−.34***
Nationalitya → Depression Symptoms	−.12 (.02)	−.28***
Nationalitya → Anxiety Symptoms	−.11 (.02)	−.26***
Income → Economic Stressors	−.10 (.06)	−.16***
Income → Anxiety	−.10 (.01)	−.10***
Age → Economic Stressors	.02 (.01)	.26***
Physical Limitations → Economic Stressors	.06 (.02)	.15**
Physical Limitations → PTSD Symptoms	.05 (.01)	.25***
Physical Limitations → Depression Symptoms	.05 (.01)	.35***
Physical Limitations → Anxiety Symptoms	.04 (.01)	.34***

Notes. Model Fit: χ²(10) = 7.05 p = .72; GFI = .99; AGFI = .98; CFI = . 99; RMSEA = .01 (95% CI = .01 −.05); AIC = 77.05 (Saturated Model = 90.00); BIC = 205.50.

^**p < .01,

^***p < .001

^aReference group for nationality is Iraqi.