Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Aug 1.
Published in final edited form as: J Trauma Stress. 2023 Jun 20;36(4):796–807. doi: 10.1002/jts.22948

Persistence of the association between mental health and resource access: A longitudinal reciprocal model in a diverse refugee sample

Meredith A Blackwell 1, David Lardier 2, Ryeora Choe 3, Jessica R Goodkind 3
PMCID: PMC11214801  NIHMSID: NIHMS2000073  PMID: 37339147

Abstract

Stress associated with resource deprivation is an active social determinant of mental health. However, mixed findings around the strength of this association and its persistence over time obscure optimal interventions to improve mental health in forcibly displaced populations. A reciprocal model was analyzed between resource access and measures of depression and anxiety at three different assessments conducted 6 months apart (Time [T] 1, T2, and T3). Participants included resettled refugees (N = 290) from three geocultural regions (i.e., Afghanistan, the Great Lakes Region of Africa, and Iraq/Syria). The results showed that although limited resource access at T1 was related to depressive and anxiety symptoms, B = 0.26, SE = 0.16, p = .023, r2 = 0.55; posttraumatic stress disorder (PTSD) symptoms, B = 0.20, SE = 0.10, p < .001, r2 = .56; and culturally specific depression and anxiety at T2, B = 0.22, SE = 0.16, p < .001, r2 = 0.65, these were not reciprocally related to resource access at T3. The results help clarify the strength and direction of effects between resource deprivation and measures of depression and anxiety over time. Although resource deprivation is predictive of depression and anxiety among recently resettled refugees, the effect may not persist in the long term. These findings have critical implications, including the urgency of ensuring initial access to resources for resettled refugees to stave off the development of depression and anxiety, as delaying immediate resource access may result in the development of chronic, hard-to-treat mental health disorders.

Most refugees are forced from their homes due to destabilizing and violent conflicts in their country of origin; thus, refugees have a much higher risk of experiencing potentially traumatic events (PTEs) than citizens of stable nations (Aragona et al., 2013; Onyut et al., 2009). Most civilians in conflict-affected regions will experience at least one PTE during their lifetime, such as witnessing or experiencing violent attacks, bombings, torture, sexual violence, separation from family, and forced migration (Fasfous et al., 2013; Johnson & Thompson, 2008; Khamis, 2005; Qouta & El Sarraj, 2004). This increased exposure puts refugees at a much higher risk of developing emotional disorders, such as anxiety, depression, and posttraumatic stress disorder (PTSD) compared to citizens of stable nations (Blackmore et al., 2020). Additionally, being female and being older, especially in conjunction with social isolation, can increase the risk of developing symptoms of mental health disorders in refugee populations (Alemi et al., 2015; Bogic et al., 2015; Naal et al., 2021).

High levels of stress related to both past trauma exposure and daily stressors can affect refugee mental health. The World Health Organization (WHO) reports that 1 in 5 people living in settings affected by conflict has a mental health disorder (WHO, 2022), and in specific settings, these rates may be closer to 33%–55% (Blackmore et al., 2020; Giacco et al., 2018; Patanè et al., 2022). Although widely heterogeneous methodologies and discrepancies in rigor may partially explain these large differences in rates (Patanè et al., 2022), individuals’ environments also influence the development of mental health disorders (Gleeson et al., 2020; Miller & Rasmussen, 2010; Wu et al., 2021). In addition to PTEs experienced during periods of violent conflict, prolonged exposure to adversity, such as detention, insecure residency status, low access to material services, an inability to find work, and other postmigration stressors, can also cause compounding mental strain (Goodkind et al., 2021; Li et al., 2016; Miller & Rasmussen, 2010; Nickerson et al., 2015). Some researchers have noted that resettlement stressors appear to moderate and/or mediate the association between trauma exposure and PTSD symptom levels (LeMaster et al., 2018; Riley et al., 2017).

Research into the influence of resource access and deprivation on mental health offers explicit implications for intervention. Optimized treatments for trauma-exposed individuals with symptoms of PTSD and mood disorders may include ameliorating chronic stressors rather than focusing primarily on the cognitive processing of past traumatic experiences (Budosan et al., 2016; LeMaster et al., 2018; Miller & Rasmussen, 2010, 2017). However, a major weakness in this literature is that most studies examining the association between resettlement stressors and mental health employ cross-sectional designs. This impedes knowledge about the long-term impact of resettlement stressors on mental health (James et al., 2019). Although some longitudinal research has demonstrated a long-term association between stressors and mental health among resettled refugees (James et al., 2019; Wu et al., 2021), it is unclear if these associations reflect reciprocal, bidirectional relations or a stress exposure model wherein stressors trigger the development of mental health symptoms that may not abate with the remission of those stressors (Keles et al., 2017). Keles and colleagues (2017) found evidence for a reciprocal association among unaccompanied minor refugees in Norway such that general hassles and depressive symptoms were reciprocally predictive of one another. Determining whether similar reciprocal, bidirectional relationships in other high-income countries (HICs) and among adults exist would inform what sort of interventions would best ameliorate refugees’ high rates of psychological distress.

An additional complication is the questionable validity of HIC-created psychopathology measures in cross-cultural populations. There is wide variability in study designs that assess for the prevalence of mental health problems or changes in distress levels in conflict-affected individuals. Although constructs such as depression, anxiety, and PTSD exist cross-culturally, their factor structures are not invariant (Rasmussen et al., 2015). This means that Western measures used to assess for these disorders may not optimally capture symptom topography in cross-cultural samples (Moore et al., 2020). Optimizing the measurement of distress symptoms is another important component of determining which interventions are most successful at improving these symptoms.

Despite research showing the influence of resource access on mental health outcomes, the longitudinal impact of resource deprivation remains unstudied. Considering the clinical implications of this research for refugees and other displaced populations, this oversight is a large weakness. For this study, we hoped to better understand the complex effects of difficulty accessing resources on various mental health outcomes through testing a reciprocal model that utilized three points of data collection over a 1-year period: Time (T) 1 (baseline), T2 (6 months), and T3 (12 months). Additionally, we hoped to further understand the utility of recently created culturally specific measures of depression and anxiety by including both culturally specific and Western measures of mental health symptoms in the reciprocal model (Choe et al., 2023).

We made several hypotheses. First, we posited that difficulty accessing resources at T1 and T2 would predict three mental health constructs (i.e., PTSD symptoms, depressive symptoms, and anxiety symptoms) and culturally specific distress at T2 and T3, respectively. Next, we hypothesized that all three mental health constructs at T1 and T2 would predict difficulty accessing resources at T2 and T3, respectively. We also expected that symptoms of depression, anxiety, and PTSD would show reciprocal associations with culturally specific emotional distress across time. Finally, we expected age to moderate the effects between mental health constructs longitudinally such that older age would increase the strength of the associations among mental health symptoms over time.

METHOD

Participants and procedure

A randomized controlled trial testing an intervention that brought together university students enrolled in a two-semester course and recently resettled refugees to improve community and systems responsiveness to refugees was conducted. Over 12 months, data were collected from 290 refugees who were randomly assigned to the intervention or a stress management waitlist control group. Participants were from three regions: Afghanistan (n = 103, 35.5%), Iraq and Syria (n = 95, 32.8%), and the Great Lakes region of Africa (n = 92, 31.7%). The term Great Lakes Africans refers to several Eastern and Central African countries that surround Lake Kivu, Lake Tanganyika, and Lake Victoria, including Burundi, Democratic Republic of Congo, Republic of the Congo, and Rwanda. Refugees from this region come from multiple ethnic groups but, like Iraqis and Syrians, have similar cultural backgrounds, perspectives on well-being, and experiences of forced displacement and resettlement. Similarly, despite having distinct cultural and ethnic backgrounds, Iraqis and Syrians also share cultural, linguistic, and religious similarities and were analyzed together on the suggestion of team members from those regions due to the much smaller number of Syrian participants.

Participation was open to all refugees 18 years of age and older from Afghanistan, the Great Lakes Region of Africa, and Iraq and Syria who had arrived in the United States within the past 3 years and were living in Albuquerque, New Mexico, where the study took place. Recruitment was coordinated through bilingual and bicultural research team members from the three regions. These team members contacted all households on a complete list of refugees resettled in the given time frame, which was compiled in coordination with the two resettlement organizations in the city as well as community networks. A household was only considered to have refused participation if no adults from the household agreed to participate. Of 161 households contacted, 143 (88.8%) agreed to participate in the study; 17 households were not interested in participating, and one household was ineligible because the one adult in the household was unavailable. Participants were excluded if they reported severe cognitive functioning difficulties, imminent suicide risk, severe mental illness, or participation in existing trauma-focused treatment. Only two participants were excluded due to these criteria and were referred for other treatment. Just over half (52.4%) of the participants were women. In total, there were 290 recently resettled refugees representing 143 households enrolled in the study from October 2013 to November 2016. Further participant demographic information can be found in Table 1.

TABLE 1.

Participant demographic characteristics

Variable n % M SD
Age (years) 34.6 11.5
Nationality
 Iraqi/ Syrian 95 32.6
 Afghan 103 35.4
 Great Lakes African 92 31.7
Gender
 Male 138 48.5
 Female 152 51.5
Marital status
 Single 96 32.9
 Married 169 58.4
 Widowed 19 6.5
 Divorced 6 2.1
Number of people in household 5.0 2.1
Number of children in household 2.6 2.4
Years of education 10.0 5.5
Open in a new tab

Quantitative data were collected in each participant’s native language (i.e., Arabic, Dari, French, Kiswahili, and Pashto) via personal interviews in their homes with a bilingual, bicultural interviewer who was a native speaker of the select language. These languages included Arabic for Iraqi and Syrian participants, in their respective dialects; French and Kiswahili for Great Lakes African participants; and Dari and Pashto for Afghan participants. Using the Translation, Review, Adjudication, Pretesting, and Documentation (TRAPD) process, measures were translated and back-translated from English into each of the utilized languages (Survey Research Center, 2016). All data were collected through interviews because of the high illiteracy rate in the sample. All measures used in the study have a history of successful implementation with refugees and culturally diverse groups. See Goodkind et al., (2020) for further information on the overall study design and further results from and registration for this study. Approval for this study was granted by the Institutional Review Board at the University of New Mexico Human Research Protections Office. We complied with all ethical standards in the treatment of individuals participating in the research, and all research participants provided informed consent.

Measures

Access to resources

Difficulty accessing resources (Sullivan & Bybee, 1999) was measured using 10 questions that evaluated participants’ perceptions regarding the level of difficulty in accessing resources in their community (e.g., “How difficult has it been to access housing?,” “How difficult has it been to access childcare resources?”). Responses were rated on a 4-point Likert-type scale ranging from 1 (not at all difficult) to 4 (very difficult). We used the mean T1 score for all 10 items (range: 1–4 across waves). In the present sample, internal consistency was moderately strong across waves, Cronbach’s αs = .80–0.85.

Depressive and anxiety symptoms

Symptoms of depression and anxiety were measured using the Hopkins Symptom Checklist–25 (HSC-25; Derogatis, 1974). The HSC-25 is a self-report measure of anxiety and depressive symptoms that has been used and validated in populations throughout the world (Hollifield et al., 2002; Wind et al., 2017), and the measure has demonstrated validity in detecting depression and anxiety symptoms in multiple populations, including refugees (Hollifield et al., 2002). Items are rated on a Likert-type scale ranging from 1 (not at all) to 4 (extremely), with higher scores indicating more severe symptoms. The HSC-25 produces three scores: a total score, calculated as the mean of all 25 items; a Depression subscale, calculated as the mean of 15 items related to depressive symptoms; and an Anxiety subscale score, calculated as the mean of 10 items related to anxiety symptoms. Prior research has shown that the total score is highly correlated with severe depression and anxiety of unspecified diagnosis and that the Depression subscale score is correlated with major depression as defined in the Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychological Association [APA], 1994). Due to cultural inappropriateness and based on feedback from bicultural team members, we removed one item from the depression scale (i.e., “loss of sexual interest or pleasure”) and calculated the total score as the mean of the 24 remaining items. Participant scores ranged from 1 to 3.75. In the present sample, internal consistency was strong across waves, Cronbach’s αs = .96–.97.

PTSD symptoms

PTSD symptom severity was measured using the 17-item, self-report PTSD Symptom Checklist–Civilian Version (PCL-C; Weathers et al., 1993). Responses are scored on a scale of 1 (not at all) to 5 (extremely), with higher scores indicating higher PTSD symptom levels. A total score is calculated (possible range: 17–85). The PCL-C has been used or validated in all three geocultural groups represented in this study (Fodor et al., 2015; Ibrahim et al., 2018). An advantage of the PCL-C is that it enquires about symptoms in relation to generic “stressful experiences,” which allows for the inclusion of multiple PTEs that refugees may have experienced and does not require symptoms to be anchored to a specific index traumatic event. In the present sample, participant scores ranged from 17 to 81, and internal consistency was strong across waves, Cronbach’s αs = .95–.97.

Culturally specific distress related to resettlement

Resettlement-related culturally specific distress was measured using the Iraqi, Great Lakes African, and Afghan Newcomer Symptom Scales, which were created for this sample through qualitative methods and were used to assess postmigration mental health symptoms using culturally specific terminology for refugees resettled in the United States (Choe et al., 2023). Scale development was based on the Afghan Symptom Checklist described by Miller and colleagues (2006) in response to research that symptoms may present differently across cultures, especially among non-Western migrants in Western countries. Participants were asked to rate items on a 5-point Likert-type scale ranging from 0 (never) to 4 (almost always). Total scores were averaged (possible range: 0–4). It is important to note that these measures are meant to assess both culturally and contextually specific distress among resettled refugees. For example, an item on the Afghan scale asks, “In the past month: how often have you felt fishar bala [emotional pressure and agitation] or fishar payinasab [low energy or motivation]?; an item on the African Great Lakes scale asks, “How often have you experienced agahinda [stress]?”; and an item on the Iraqi scale asks, “How often have you felt like difficulties learning English were negatively affecting your well-being?” In the present sample, internal reliability was very good for the 13-item Great Lakes African Newcomer Symptom Scale, Cronbach’s α = .91; 18-item Afghan Newcomer Symptom Scale, Cronbach’s α = .92; and 25-item Iraqi Newcomer Symptom Scale, Cronbach’s α = .91.

Sociodemographic covariates

Sociodemographic covariates were included and retained based on statistical significance and model performance. These controls included gender, age, national origin, time in the United States, household size, and monthly household income. Gender was measured as a dichotomous variable (1 = female, 0 = male). Age was measured continuously ranging from 18 to 71 years. National origin was dummy-coded into three separate items to assess origin: Iraqi/Syrian (yes = 1, no = 0), Afghan (yes = 1, no = 0), or Great Lakes African (yes = 1, no = 0). Time in the United States, in weeks, and monthly income were measured as continuous variables.

Data analysis

The sample size provided .80 power to detect small-to-medium differences in postintervention intercept (d = 0.25) at an alpha level of .05 (two-tailed), using three-level multilevel modeling with randomization at Level 3. Optimal Design software was used for power estimation (Spybrook et al., 2011).

A cross-lagged, autoregressive path model was fit using maximum likelihood estimations with robust standard errors (MLR). These models examined difficulty in accessing resources and indicators of psychological distress over three waves of data collected over 12 months. Cross-lagged models can assess for (a) concurrent associations between variables at one time point; (b) autoregressive correlations reflecting variable stability over multiple time points; and (c) cross-lagged associations, or the extent to which a variable at one time point predicts a variable at a subsequent time point, adjusted for the prior two effects. Three psychological distress variables were included (i.e., depression and anxiety symptoms, PTSD symptom severity, and culturally specific distress). A parameter reflecting difficulty accessing resources was also included. Skew and kurtosis were assessed and appeared within relatively normal ranges (i.e., plus or minus 2 standard deviations). No variables were adjusted for skew.

In each model, autoregressive effects were estimated. Specifically, each variable at one time point was allowed to predict the same variable at the subsequent assessment point (e.g., depression and anxiety symptoms at T1 were allowed to predict depression and anxiety symptoms at T2). Controlling for autoregressive effects helps provide information about the stability of each construct over time, reduces bias in parameter estimates, and provides insight into the temporal sequence of associations between study variables (Selig & Little, 2012). In the model, psychological distress variables were allowed to predict psychological distress variables at the following assessment point. Difficulty accessing resources was also allowed to predict each of the psychological distress variables at the following assessment. Reciprocal effects were estimated by allowing each prior variable to predict variables at the subsequent wave. Estimating these paths allowed for the assessment of cross-lagged cascade effects over time. Residual variances of exogenous variables within the same wave were allowed to correlate.

Interaction terms were tested to assess whether participant age and time in the United States moderated the effects between psychological distress variables and difficulty accessing resources at T1 on the subsequent assessment point. Each of the interaction terms was independently tested in separate models to reduce multicollinearity (Babyak, 2004) rather than fitting a model that included all interaction terms simultaneously. Age, in years, and time in the United States were both treated as continuous variables. Significant interactions were plotted using Preacher et al.’s (2006) interaction tool. For brevity, additional moderation analyses were conducted in relation to national origin and monthly income; however, no significant interaction effects were identified. National origin was retained as a covariate due to its significant contribution to final analytic models and significant associations with variables at T3.

The cross-lagged and autoregressive path model using MLR estimation was fit in STATA (Version 14.2). Satorra–Bentler (Satorra, 1992; Satorra & Bentler, 1994) scaled chi-squared test adjustment was used to account for data nonnormality. The Satorra–Bentler scaled chi-squared test provides robust standard errors, p values, and confidence intervals (CIs). Models were considered to have an acceptable fit if they satisfied at least two of the three following criteria: a chi-square/degree of freedom (df) value less than 3.0, a comparative fit index (CFI) value of .95 or higher, and a root mean square error of approximation (RMSEA) value of .08 or less. Additional estimates of model fit were also assessed, including the goodness of fit index (GFI) and Tucker–Lewis index (TLI), for which values should be .95 or higher (values of .90 and below are considered inadequate) and can exceed 1.00 (Huang, 2017).

RESULTS

Means, standard deviations, and bivariate correlations among continuous study variables are presented in Table 1. All main analytic variables were correlated.

Reciprocal effects between difficulty accessing resources and psychological distress

The cross-lagged and autoregressive path model to examine reciprocal relations across the assessment points (i.e., 12 months) fit the data well, χ2(48, N = 290) = 91.94, p < .001, χ2/df = 1.68, GFI = .97, TLI = .98, CFI = .99; RMSEA = .05, 95% CI [.04, .07] (see Figure 1). Several controls were included in the final model and related to other primary study variables (not depicted in the figure for clarity). At T2, age was positively associated with PTSD symptom severity at T2, B = 0.18, β = .16, p = .015, and female gender identity was negatively associated with PTSD symptom severity, B = −0.54, β = −.25, p = .023. Iraqi/Syrian national identity was negatively correlated with culturally specific distress at T3, B = −0.75, β = −.14, p < .001, and difficulty accessing resources at T3, B = −0.11, β = −.05, p < .001.

FIGURE 1. Reciprocal relations between accessing resources and measures of psychological distress among Afghan, Iraqi/Syrian, and Great Lakes African refugees in the United States.

FIGURE 1

Open in a new tab

Note: Unstandardized and standardized (in parentheses) beta coefficients are provided. All analyses controlled for Iraqi/Syrian nationality and age at T1. Residual variances among endogenous variables within each wave of data collection were allowed to correlate. Mode-to-data fit: χ2(48, N = 290) = 91.94, p < .001, χ2/df = 1.68, goodness of fit index = .97, Tucker–Lewis index = .98, comparative fit index = .99; root mean square error of approximation = .05, 95% confidence interval [.04, .07]

An examination of the main analytic variables revealed several significant effects. Difficulty accessing resources at T1 was a significant positive predictor of depressive and anxiety symptoms, PTSD symptom severity, and culturally specific distress at T2. Difficulty accessing resources at T2 was a significant positive predictor of culturally specific distress at T3. Autoregressive paths were positive and significant across time points. T1 depressive and anxiety symptoms were positively associated with T2 PTSD symptom severity. T2 depressive and anxiety symptoms displayed no significant associations with T3 variables outside of autoregressive effects. PTSD symptom severity at T2 displayed no significant effects on T3 variables outside of autoregressive effects. Culturally specific distress had a positive impact on depressive and anxiety symptoms and PTSD symptom severity at T3; autoregressive effects were present at T3. In total, the model explained 70% of the variance in difficulty accessing resources, 55% of the variance in depressive and anxiety symptom severity, 56% of the variance in PTSD symptom severity, and 65% of the variance in culturally specific distress at T2. In addition, the model explained 75% of the variance in difficulty accessing resources, 56% of the variance in depressive and anxiety symptom severity, 53% of the variance in PTSD symptom severity, and 63% of the variance in culturally specific distress at T3.

The moderating role of age

Next, interactions among difficulty accessing resources, mental health symptom variables, and age at T1 were examined in relation to T2 variables; the path models fit to examine moderation effects are not depicted for brevity. The interaction between depressive and anxiety symptoms and age at T1 had a significant positive effect on PTSD symptom severity, B = 0.230, β =.177, p < .001. An examination of simple slopes yielded significant and positive gradients (m) for younger age in years, m = 5.91, t(289) = 4.98, p < .001, and older age in years, m = 18.10, t(289) = 3.98, p < .001, such that the positive correlation between PTSD symptom severity and depressive and anxiety symptoms was stronger for older participants. Similarly, the interaction between depressive and anxiety symptoms and age at T1 had a significant and positive effect on culturally specific distress, B = 0.179, β =.107, p < .001. An examination of simple slopes yielded significant positive gradients for younger age, m = 4.99, t(289)= 3.18, p = .002, and older age, m = 14.48, t(289) = 3.19, p < .001, such that the positive correlation between depressive and anxiety symptoms and culturally specific distress was stronger for older participants. See Table 2 for the slopes of the interaction effects.

TABLE 2.

Age, in years, as a moderator of the associations between psychological distress variables at Time (T) 1 and psychological distress at T2

T1 variable T2 variable Lower age (−1 SD) Higher age (+1 SD)
m t(289) p m t(289) p
PTSD symptom severity Depressive and anxiety symptoms 5.91 4.98 .000 18.10 3.98 .000
Depressive and anxiety symptoms Culturally specific distress 4.99 3.18 .002 14.48 3.19 .000
PTSD symptom severity Culturally specific distress 3.58 3.03 .003 13.65 3.01 .003
Culturally specific distress PTSD symptom severity 7.22 4.78 .003 12.88 4.37 .000
Open in a new tab

The interaction between PTSD symptom severity and age at T1 had a significant positive predictive effect on culturally specific distress at T2, B = 0.190, β =.115, p < .001. The examination of simple slopes yielded significant positive gradients for younger age, m = 3.58, t(289) = 3.03, p = .003, and older age, m = 13.65, t(289) = 3.01, p = .003, such that both relatively younger and older age demonstrated a positive effect of PTSD symptom severity on culturally specific distress, but this effect was stronger for older participants. The interaction between culturally specific distress and age at T1 was significantly associated with PTSD symptom severity, B = 0.247, β =.173, p < .001. An examination of simple slopes yielded significant positive gradients for younger age, m = 7.22, t(289)= 4.78, p = .003, and older age, m = 12.88, t(289) = 4.37, p < .001, such that younger age and older age demonstrated a positive effect of PTSD symptom severity on culturally specific distress, but this association was stronger for older participants. See Table 2 for the slopes of the interaction effects.

DISCUSSION

As expected, refugees’ difficulty accessing resources at baseline contributed to increased depressive and anxiety symptoms and PTSD symptoms 6 months later. However, difficulty accessing resources at this 6-month time point did not predict depressive and anxiety symptoms or PTSD symptom severity at the 12-month assessment. These findings suggest that the link between difficulty accessing resources and mental health may not persist longitudinally for recently resettled refugees. The results of this study have two important implications for responses to and assessment of refugee mental health. First, although access to resources may play a causal role in the development of mental health disorders among resettled refugees, the link between those constructs may wane over time. This finding is somewhat consistent with the limited longitudinal research on the associations between mental health and stressors. Whereas Keles and colleagues (2017) found a more consistent reciprocal relationship between depressive symptoms and “general hassles” in a sample of refugee children in Norway, they found relatively stable depressive symptoms over time even as these general hassles decreased. However, it is worth noting that although we found no reciprocal relationship between resource access and psychological symptoms, resource access at T1 and T2 did predict culturally specific distress at T2 and T3, respectively. Although it was not a reciprocal relationship, as we expected, this does pose the possibility that for some mental health symptoms, resource access may predict symptoms longitudinally. This finding is also likely related to the contextually specific component of the culturally specific distress measures, which were created for recently resettled refugees and, thus, included items related to resource access. It is also possible that the culturally specific measures of distress we created have superior construct validity for measuring symptoms of PTSD and psychological distress in their respective cultures and are better suited to show the link between mental health and social stressors; alternatively, the culturally specific measures of distress may lack discriminant validity with some postmigration stressors. Future research should examine these constructs at more time points to better answer these questions.

The potential short-term predictive role of access to resources has important implications for treatment recommendations. The findings suggest that although quickly addressing resource needs may reduce the development of emotional disorders, meeting resource needs after mental health symptoms have already developed is still essential but may not successfully treat the current mental health needs of recently resettled refugees. Thus, alleviating resource stress may act like putting a stop sign in a dangerous intersection: There will likely be fewer accidents, but the stop sign will not improve the health of individuals who got into an accident there before the sign was added. Once patterns of depression and anxiety are in motion, alleviating the initial cause of those symptoms may not alleviate the symptoms themselves. Alternatively, it could be that the level of material support needed to affect mental health was far higher than participants were receiving. Multimodal interventions in which specialized mental health treatments are implemented alongside resource support to offset negative cascading may be more optimal than a sequential model for resettled refugees who are already experiencing mental health concerns.

These findings make critical contributions to the longstanding debate over trauma-focused and psychosocial treatments, primarily because most studies that look at the impact of resource stress on mental health have done so using cross-sectional data (Scharpf et al., 2021). This may be especially pertinent for older refugees, who demonstrated stronger reciprocal associations among mental health symptoms but not significantly different associations between resource access and mental health. The stronger association between older age and increased mental health symptoms is consistent with prior research (Alemi et al., 2015; Bogic et al., 2015; Pumariega et al., 2005), and although the exact mechanisms of this are unclear, potential causes of the association between older age and worse mental health may be increased health concerns, stronger memories of safety and life premigration, and increased difficulty adapting to new environments.

The lack of observed reciprocal relationships should not be taken to mean there is no need to improve access to resources for refugees. First, there is some prior support for reciprocal associations between mental health and resettlement stressors among youth (Keles et al., 2017), so further study is needed. However, if our results are replicated and the remission of stressors does not necessarily lead to improved mental health after postresettlement onset of distress symptoms, improving access to resources immediately upon resettlement should be emphasized to stave off the development of depression and anxiety, as delaying immediate resource access may result in the development of chronic, hard-to-treat mental health disorders. Rather than discount the importance of resource access with regard to mental health, these findings suggest that preventing psychological distress through resource access is time-sensitive and urgent (Miller & Rasmussen, 2010). Additionally, access to resources is related to a multitude of health outcomes, including life expectancy, cardiac health, and health behaviors (e.g., smoking, substance use; Baptiste et al., 2020). Further, our findings suggest that interventions that include the transference of skills related to resource access may result in later ease of access or increased comfort for accessing critical health care. The fact that psychological distress can persist independent of difficulty accessing resources does not undermine the importance of resources to these other outcomes. However, the results do suggest that specialized mental health interventions are not of secondary importance.

A secondary key finding relates to the use of the culturally specific distress measures used in this study, which were created to measure distress in the context of resettlement. The findings showed that T1 PTSD symptom severity predicted T2 culturally specific distress, which then predicted T3 PTSD symptom severity. Contrary to our expectations, depressive and anxiety symptoms did not show these reciprocal effects. The results suggest that these measures of culturally specific distress could offer culturally valid alternatives to approximate PTSD symptoms in diverse populations. Although previous research has explored the validity, reliability, and factor structure of these measures (Choe et al., 2023), this study supports the potential convergent validity of culturally specific distress and PTSD symptom measures, though not necessarily with respect to anxiety and depression symptoms.

Given the limited studies that have examined resource access and mental health longitudinally, the results should be interpreted cautiously. At least two studies that employed longitudinal data sets found long-term associations between difficulty accessing resources and mental health, though the samples were among unaccompanied minors in Europe, whereas our study exclusively examined adults (Keles et al., 2017; Vervliet et al., 2014). Thus, it is possible that the findings do not generalize to all resettled refugees. Additionally, although we did not observe a reciprocal relationship between difficulty accessing resources and culturally specific distress, difficulty accessing resources at T2 did predict culturally specific distress—but not either of the other two mental health variables—at T3. This could mean that culturally specific distress is a more accurate measure of mental health in these populations and that difficulty accessing resources continues to play a role in mental health over time following resettlement or that these new measures emphasized current stressors more heavily, leading to some discriminant invalidity. Finally, data from this study were collected prior to the 2016 U.S. presidential election and the onset of the COVID-19 pandemic. Following the election, policies and public opinion toward refugees shifted substantially in the United States, creating increasingly stressful and negative environments (Norris & Inglehart, 2019). During the pandemic, newcomers were also disproportionately affected by resource inequities, such as a higher likelihood of job loss and reduced access to health care (Mengesha et al., 2022). These changing contexts highlight the importance of research on mental health and resource access among refugees, and although they could have an impact on the associations demonstrated in the findings, it is likely that significant reciprocal effects would persist.

Researchers working with refugee populations have criticized the reliance on Western assessments of mental health and pointed out the weaknesses of these measures, but few have rigorously created culturally valid symptom assessments (Moore et al., 2020). Importantly, the reciprocal findings reported here support the validity of these new culturally and contextually specific measures of distress being used as measurements for latent posttraumatic stress symptoms in resettled Afghan, Great Lakes African, and Iraqi and Syrian refugees. Researchers working with these populations should consider using measures created for the symptoms these groups are experiencing. It would also be useful to examine the experiences of Iraqi and Syrian participants, as well as the experiences of Burundian, Congolese, and Rwandan participants, separately from one another. We were not able to do so due to sample size constraints.

In addition, future researchers should analyze these reciprocal relationships in lower- and middle-income countries (LMICs) to see if the results are replicated in those contexts. As discussed, resource access is lower and the risk of mental health symptoms higher in LMIC contexts (Giacco et al., 2018; Patanè et al., 2022); thus, understanding whether the dynamics of the observed associations are similar or different in varied environments would be useful to contextualize these findings. Given how coping often occurs at the collective level (i.e., family and/or community) level rather than the individual level (Nickerson et al., 2010; Sireau et al., 2019), the individual focus of these analyses is also a weakness of this study. Future studies should seek to account for the role of family, household, and community networks in outcomes.

Researchers who examine reciprocal dynamics may want to consider including random-effect cross-lagged models (Berry & Willoughby, 2017), as these analyses would help decompose scores within the models into fluctuating within-component dynamics and stable between components (Mulder & Hamaker, 2021). Last, our longitudinal design included a 6-month lag between each assessment point. Differing patterns may emerge across shorter or longer periods of time. Moreover, these processes may need to be studied on a longer-term basis. Newcomers’ sense of acuity around accessing resources and mental health may change over time as they are better able to understand new systems (e.g., health care, housing). Qualitative data from the overall study suggests that as newcomers’ experiences in the United States extend over time, their understanding of the meaning and impact of cultural differences and the effects these have on their lives deepens. In addition, global news media and contact with friends and family in a refugee’s home country and abroad provide constant reminders of ongoing violence and suffering.

Despite its limitations, the present study employed a rigorous design and the findings provide novel evidence of reciprocal relations between difficulty accessing resources and psychological distress. Future research should further tease out these associations to determine the most effective treatment approaches. Alleviating enduring and prevalent mental health symptoms among refugees should be a priority, and researchers should continue to explore the role of access to resources in mental health longitudinally to contribute to these important efforts.

Acknowledgments

This research was funded by a grant from the National Institute on Minority Health & Health Disparities (R01MD007712).

Footnotes

We have no conflicts of interest to declare.

OPEN PRACTICES STATEMENT

The data are not publicly available due to privacy or ethical restrictions. The data that support the findings of this study are available on request from the corresponding author.

REFERENCES

  1. Alemi Q, James S, Siddiq H, & Montgomery S (2015). Correlates and predictors of psychological distress among Afghan refugees in San Diego County. International Journal of Culture and Mental Health, 8(3), 274–288. 10.1080/17542863.2015.1006647 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Author. [Google Scholar]
  3. Aragona M, Pucci D, Mazzetti M, Maisano B, & Geraci S (2013). Traumatic events, post-migration living difficulties and post-traumatic symptoms in first-generation immigrants: A primary care study. Annali Dell’Istituto Superiore Di Sanità, 49(2), 169–175. 10.4415/ANN_13_02_08 [DOI] [PubMed] [Google Scholar]
  4. Babyak MA (2004). What you see may not be what you get: A brief, nontechnical introduction to overfitting in regression-type models. Psychosomatic Medicine, 66(3), 411–421. 10.1097/01.psy.0000127692.23278.a9 [DOI] [PubMed] [Google Scholar]
  5. Baptiste D, Commodore‐Mensah Y, Alexander KA, Jacques K, Wilson PR, Akomah J, Sharps P, & Cooper LA (2020). COVID‐19: Shedding light on racial and health inequities in the USA. Journal of Clinical Nursing, 29(15–16), 2734–2736. 10.1111/jocn.15351 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Berry D, & Willoughby MT (2017). On the practical interpretability of cross-lagged panel models: Rethinking a developmental workhorse. Child Development, 88(4), 1186–1206. 10.1111/cdev.12660 [DOI] [PubMed] [Google Scholar]
  7. Blackmore R, Boyle JA, Fazel M, Ranasinha S, Gray KM, Fitzgerald G, Misso M, & Gibson-Helm M (2020). The prevalence of mental illness in refugees and asylum seekers: A systematic review and meta-analysis. PLoS Medicine, 17(9), Article e1003337. 10.1371/journal.pmed.1003337 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Bogic M, Njoku A, & Priebe S (2015). Long-term mental health of war-refugees: A systematic literature review. BMC International Health and Human Rights, 15(1), Article 29. 10.1186/s12914-015-0064-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Budosan B, Aziz S, Benner MT, & Abras B (2016). Perceived needs and daily stressors in an urban refugee setting: Humanitarian Emergency Settings Perceived Needs Scale survey of Syrian refugees in Kilis, Turkey. Intervention Journal of Mental Health and Psychosocial Support in Conflict-Affected Areas, 14(3), 293–304. [Google Scholar]
  10. Choe R, Lardier D, Hess J, Blackwell MA, Amer S, Ndayisenga M, Deewa S, Isakson B, Goodkind J, (2023). Understanding Culturally and Contextually Specific Distress among Afghan, Iraqi, and Great Lakes African Refugees: A Mixed Methods Study to Develop and Test New Measures. Manuscript under review. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Derogatis LR, Lipman RS, Rickels K, Uhlenhuth EH, & Covi L (1974). The Hopkins Symptom Checklist (HSCL): A self-report symptom inventory. Behavioral Science, 19(1), 1–15. 10.1002/bs.3830190102 [DOI] [PubMed] [Google Scholar]
  12. Fasfous AF, Peralta-Ramírez I, & Pérez-García M (2013). Symptoms of PTSD among children living in war zones in same cultural context and different situations. Journal of Muslim Mental Health, 7(2), 47–61. 10.3998/jmmh.10381607.0007.203 [DOI] [Google Scholar]
  13. Fodor KE, Pozen J, Ntaganira J, Sezibera V, & Neugebauer R (2015). The factor structure of posttraumatic stress disorder symptoms among Rwandans exposed to the 1994 genocide: A confirmatory factor analytic study using the PCL-C. Journal of Anxiety Disorders, 32, 8–16. 10.1016/j.janxdis.2015.03.001 [DOI] [PubMed] [Google Scholar]
  14. Giacco D, Laxhman N, & Priebe S (2018). Prevalence of and risk factors for mental disorders in refugees. Seminars in Cell & Developmental Biology, 77, 144–152. 10.1016/j.semcdb.2017.11.030 [DOI] [PubMed] [Google Scholar]
  15. Gleeson C, Frost R, Sherwood L, Shevlin M, Hyland P, Halpin R, Murphy J, & Silove D (2020). Post-migration factors and mental health outcomes in asylum-seeking and refugee populations: A systematic review. European Journal of Psychotraumatology, 11(1), Article 1793567. 10.1080/20008198.2020.1793567 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Goodkind J, Ferrera J, Lardier D, Hess JM, & Greene RN (2021). A mixed-method study of the effects of post-migration economic stressors on the mental health of recently resettled refugees. Society and mental health, 11(3), 217–235. 10.1177/2156869320973446 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Goodkind JR, Bybee D, Hess JM, Amer S, Ndayisenga M, Greene RN, Choe R, Isakson B, Baca B, & Pannah M (2020). Randomized Controlled Trial of a Multilevel Intervention to Address Social Determinants of Refugee Mental Health. American Journal of Community Psychology, 65(3–4), 272–289. 10.1002/ajcp.12418 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Ibrahim H, Ertl V, Catani C, Ismail AA, & Neuner F (2018). The validity of Posttraumatic Stress Disorder Checklist for DSM-5 (PCL-5) as screening instrument with Kurdish and Arab displaced populations living in the Kurdistan region of Iraq. BMC Psychiatry, 18(1), 259–266. 10.1186/s12888-018-1839-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hollifield M, Warner TD, Lian N, Krakow B, Jenkins JH, Kesler J, Stevenson J, & Westermeyer J (2002). Measuring trauma and health status in refugees: A critical review. JAMA, 288(5), 611–621. 10.1001/jama.288.5.611 [DOI] [PubMed] [Google Scholar]
  20. Huang H, Zhou H, Wang J, Chang F, & Ma M (2017). A multivariate spatial model of crash frequency by transportation modes for urban intersections. Analytic Methods in Accident Research, 14, 10–21. 10.1016/j.amar.2017.01.001 [DOI] [Google Scholar]
  21. James P, Iyer A, & Webb TL (2019). The impact of post‐migration stressors on refugees’ depression and anxiety and health: A longitudinal analysis. European Journal of Social Psychology, 49(7), 1359–1367. 10.1002/ejsp.2589 [DOI] [Google Scholar]
  22. Johnson H, & Thompson A (2008). The development and maintenance of post-traumatic stress disorder (PTSD) in civilian adult survivors of war trauma and torture: A review. Clinical Psychology Review, 28(1), 36–47. 10.1016/j.cpr.2007.01.017 [DOI] [PubMed] [Google Scholar]
  23. Keles S, Idsøe T, Friborg O, Sirin S, & Oppedal B (2017). The longitudinal relation between daily hassles and depressive symptoms among unaccompanied refugees in Norway. Journal of Abnormal Child Psychology, 45(7), 1413–1427. 10.1007/s10802-016-0251-8 [DOI] [PubMed] [Google Scholar]
  24. Khamis V (2005). Post-traumatic stress disorder among school-age Palestinian children. Child Abuse & Neglect, 29(1), 81–95. 10.1016/j.chiabu.2004.06.013 [DOI] [PubMed] [Google Scholar]
  25. LeMaster JW, Broadbridge CL, Lumley MA, Arnetz JE, Arfken C, Fetters MD, Jamil H, Pole N, & Arnetz BB (2018). Acculturation and post-migration psychological symptoms among Iraqi refugees: A path analysis. American Journal of Orthopsychiatry, 88(1), 38–47. 10.1037/ort0000240 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Li SSY, Liddell BJ, & Nickerson A (2016). the relationship between post-migration stress and psychological disorders in refugees and asylum seekers. Current Psychiatry Reports, 18(9), 82–90. 10.1007/s11920-016-0723-0 [DOI] [PubMed] [Google Scholar]
  27. Mengesha Z, Alloun E, Weber D, Smith M, & Harris P (2022). “Lived the pandemic twice”: A scoping review of the unequal impact of the COVID-19 pandemic on asylum seekers and undocumented migrants. International Journal of Environmental Research and Public Health, 19(11), 1–16. 10.3390/ijerph19116624 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Miller KE, Omidian P, Quraishy AS, Quraishy N, Nasiry MN, Nasiry S, Karyar NM, & Yaqubi AA (2006). The Afghan Symptom Checklist: A culturally grounded approach to mental health assessment in a conflict zone. American Journal of Orthopsychiatry, 76(4), 423–433. 10.1037/0002-9432.76.4.423 [DOI] [PubMed] [Google Scholar]
  29. Miller KE, & Rasmussen A (2010). War exposure, daily stressors, and mental health in conflict and post-conflict settings: Bridging the divide between trauma-focused and psychosocial frameworks. Social Science & Medicine, 70(1), 7–16. 10.1016/j.socscimed.2009.09.029 [DOI] [PubMed] [Google Scholar]
  30. Miller KE, & Rasmussen A (2017). The mental health of civilians displaced by armed conflict: an ecological model of refugee distress. Epidemiology and Psychiatric Sciences, 26(2), 129–138. 10.1017/S2045796016000172 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Moore A, van Loenhout JAF, de Almeida MM, Smith P, & Guha-Sapir D (2020). Measuring mental health burden in humanitarian settings: A critical review of assessment tools. Global Health Action, 13(1), 54–62. 10.1080/16549716.2020.1783957 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Mulder JD, & Hamaker EL (2021). Three extensions of the random intercept cross-lagged panel model. Structural Equation Modeling: A Multidisciplinary Journal, 28(4), 638–648. 10.1080/10705511.2020.1784738 [DOI] [Google Scholar]
  33. Naal H, Nabulsi D, El Arnaout N, Abdouni L, Dimassi H, Harb R, & Saleh S (2021). Prevalence of depression symptoms and associated sociodemographic and clinical correlates among Syrian refugees in Lebanon. BMC Public Health, 21(1), 1–13. 10.1186/s12889-021-10266-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Nickerson A, Bryant RA, Schnyder U, Schick M, Mueller J, & Morina N (2015). Emotion dysregulation mediates the relationship between trauma exposure, post-migration living difficulties and psychological outcomes in traumatized refugees. Journal of Affective Disorders, 173, 185–192. 10.1016/j.jad.2014.10.043 [DOI] [PubMed] [Google Scholar]
  35. Nickerson A, Bryant RA, Steel Z, Silove D, & Brooks R (2010). The impact of fear for family on mental health in a resettled Iraqi refugee community. Journal of Psychiatric Research, 44(4), 229–235. 10.1016/j.jpsychires.2009.08.006 [DOI] [PubMed] [Google Scholar]
  36. Norris P, & Inglehart R (2019). Cultural backlash: Trump, Brexit, and authoritarian populism. Cambridge University Press. [Google Scholar]
  37. Onyut LP, Neuner F, Ertl V, Schauer E, Odenwald M, & Elbert T (2009). Trauma, poverty and mental health among Somali and Rwandese refugees living in an African refugee settlement – an epidemiological study. Conflict and Health, 3(1). 10.1186/1752-1505-3-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Patanè M, Ghane S, Karyotaki E, Cuijpers P, Schoonmade L, Tarsitani L, & Sijbrandij M (2022). Prevalence of mental disorders in refugees and asylum seekers: A systematic review and meta-analysis. Global Mental Health, 9, 250–263. 10.1017/gmh.2022.29 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Preacher KJ, Curran PJ, & Bauer DJ (2006). Computational tools for probing interactions in multiple linear regression, multilevel modeling, and latent curve analysis. Journal of Educational and Behavioral Statistics, 31(4), 437–448. 10.3102/10769986031004437 [DOI] [Google Scholar]
  40. Pumariega AJ, Rothe E, & Pumariega JB (2005). Mental health of immigrants and refugees. Community Mental Health Journal, 41(5), 581–597. 10.1007/s10597-005-6363-1 [DOI] [PubMed] [Google Scholar]
  41. Qouta S, & El Sarraj E (2004). Prevalence of PTSD among Palestinian children in Gaza Strip. Arab Psynet Journal, 2, 8–13. [Google Scholar]
  42. Rasmussen A, Verkuilen J, Ho E, & Fan Y (2015). Posttraumatic stress disorder among refugees: Measurement invariance of Harvard Trauma Questionnaire scores across global regions and response patterns. Psychological Assessment, 27(4), 1160–1170. 10.1037/pas0000115 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Riley A, Varner A, Ventevogel P, Taimur Hasan MM, & Welton-Mitchell C (2017). Daily stressors, trauma exposure, and mental health among stateless Rohingya refugees in Bangladesh. Transcultural Psychiatry, 54(3), 304–331. 10.1177/1363461517705571 [DOI] [PubMed] [Google Scholar]
  44. Selig JP, & Little TD (2012). Autoregressive and cross-lagged panel analysis for longitudinal data. In Laursen B, Little TD, & Card NA (Eds.), Handbook of developmental research methods (pp. 265–278). Guilford Press. [Google Scholar]
  45. Satorra A (1992). Asymptotic robust inferences in the analysis of mean and covariance structures. Sociological Methodology, 22, 249–278. 10.2307/270998 [DOI] [Google Scholar]
  46. Satorra A, & Bentler P (1994). Corrections to test statistics and standard errors in covariance structure analysis. In von Eye A, Clogg CC (Eds.), Latent variables analysis: Applications to developmental research. SAGE. [Google Scholar]
  47. Scharpf F, Kaltenbach E, Nickerson A, & Hecker T (2021). A systematic review of socio-ecological factors contributing to risk and protection of the mental health of refugee children and adolescents. Clinical Psychology Review, 83, 1–16. 10.1016/j.cpr.2020.101930 [DOI] [PubMed] [Google Scholar]
  48. Sierau S, Schneider E, Nesterko Y, & Glaesmer H (2019). Alone, but protected? Effects of social support on mental health of unaccompanied refugee minors. European Child & Adolescent Psychiatry, 28(6), 769–780. 10.1007/s00787-018-1246-5 [DOI] [PubMed] [Google Scholar]
  49. Spybrook J, Bloom H, Congdon R, Hill C, Martinez A, & Raudenbush SW (2011). Optimal design plus empirical evidence: Documentation for the “Optimal Design” software Version 3.0. http://wtgrantfoundation.org/FocusAreas#tools-for-group-randomized-trials
  50. Sullivan CM, & Bybee DI (1999). Reducing violence using community-based advocacy for women with abusive partners. Journal of Consulting and Clinical Psychology, 67(1), 43–53. 10.1037/0022-006X.67.1.43 [DOI] [PubMed] [Google Scholar]
  51. Survey Research Center. (2016). Guidelines for best practice in cross-cultural surveys. http://www.ccsg.isr.umich.edu
  52. Vervliet M, Lammertyn J, Broekaert E, & Derluyn I (2014). Longitudinal follow-up of the mental health of unaccompanied refugee minors. European Child & Adolescent Psychiatry, 23(5), 337–346. 10.1007/s00787-013-0463-1 [DOI] [PubMed] [Google Scholar]
  53. Weathers F, Litz B, Herman D, Huska J, & Keane T (1993, October). The PTSD Checklist (PCL): Reliability, validity, and diagnostic utility [Conference presentation]. Annual meeting of the International Society for Traumatic Stress Studies, San Antonio, TX, USA. [Google Scholar]
  54. Wind TR, van der Aa N, de la Rie S, & Knipscheer J (2017). The assessment of psychopathology among traumatized refugees: measurement invariance of the Harvard Trauma Questionnaire and the Hopkins Symptom Checklist–25 across five linguistic groups. European Journal of Psychotraumatology, 8(sup2), Article 1321357. 10.1080/20008198.2017.1321357 [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. World Health Organization. (2022). World Mental Health Report: Transforming mental health for all. https://www.who.int/publications/i/item/9789240049338
  56. Wu S, Renzaho AMN, Hall BJ, Shi L, Ling L, & Chen W (2021). Time-varying associations of pre-migration and post-migration stressors in refugees’ mental health during resettlement: A longitudinal study in Australia. The Lancet Psychiatry, 8(1), 36–47. 10.1016/S2215-0366(20)30422-3 [DOI] [PubMed] [Google Scholar]

RESOURCES