Abstract
The clinical relevance of negative changes in cognitions about oneself, others, and the world is reflected in the diagnostic criteria for posttraumatic stress disorder (PTSD) in the DSM-5 and complex posttraumatic stress disorder in the ICD-11. Although such changes in cognition have been posited to be especially relevant for traumatised refugees, few studies have examined this in refugee populations. The present study used a cross-sectional design to compare negative cognitions among 213 adult treatment-seeking refugees with those in previously published samples from the general population, veterans with combat-related PTSD, and whiplash victims. Measures included the World Assumptions Scale (WAS) and the Events and DSM-IV PTSD subscales of the Harvard Trauma Questionnaire (HTQ). Path models examined the relation of the WAS subscales to five demographic and trauma-related variables. Results showed that world assumptions were especially negative with regard to Benevolence of World, Benevolence of People, and Luck subscales, on which refugees scored lower than all reference samples. Differences between the refugee sample and the reference samples were smallest with regard to self-worth and self-controllability. World assumptions were associated with gender and PTSD symptom severity but not with age, length of residence in the Netherlands, and number of traumatic event types. The DSM-5 criterion of negative changes in belief about oneself, others, and the world appears more applicable to refugees than the more narrowly formulated ICD-11 criterion of diminished and defeated sense of self. Prevention and treatment efforts with refugees may need to be especially aimed at preventing a further decline of trust as well as restoration of trust in others and the world.
Introduction
The occurrence of changes in assumptions about oneself, others, and the world as a result of traumatic experiences has long been clinically acknowledged (e.g., Brewin & Holmes, 2003). Basic assumptions about oneself as valuable, others as trustworthy, and the world as a safe place may be violated by personal experiences of threatened death, serious injury, or sexual violence, or by witnessing such atrocities happen to others. Posttraumatic adaptation requires an assimilation of traumatic experiences into existing assumptions or a change of basic assumptions (Janoff-Bulman, 1989). Posttraumatic negative assumptions are related to posttraumatic functional impairment and have been found to predict chronicity and severity of posttraumatic stress disorder (PTSD; American Psychiatric Association [APA], 2013; Friedman, 2013).
The contribution of negative assumptions to posttraumatic stress reactions has been diagnostically formalised only recently. Research into the association between negative assumptions and posttraumatic stress has resulted in the inclusion of negative assumptions in the criteria for PTSD in the DSM-5 (American Psychiatric Association, 2013) and the proposed revision of the World Health Organization (WHO) International Classification of Diseases (ICD-11). The PTSD diagnostic criteria in the DSM-5 include the symptom of “persistent and exaggerated negative beliefs or expectations about oneself, others, or the world” (D2; APA, 2013, p. 272), as part of the new symptom cluster “Negative Alterations in Cognitions and Mood.” The proposal for ICD-11 contains the new diagnosis of complex PTSD, which includes “beliefs about oneself as diminished, defeated or worthless” (Maercker et al., 2013, p. 200) or “sustained and pervasive … negative beliefs about self” (First, Reed, Hyman, & Saxena, 2015, p. 87).
The clinical utility of these criteria in different trauma-exposed populations is a matter of interest. In particular, the applicability of the PTSD label to traumatised refugees has long been the subject of debate. The main approach to mental distress in refugees (termed trauma-focused psychiatric epidemiology by Miller, Kulkarni, & Kushner, 2006) focuses on the prevalence of PTSD in refugee populations as a result of experiencing potentially traumatic events (PTEs). Following this approach, refugees have been shown to be at increased risk of developing PTSD in comparison to general, nonrefugee populations (Fazel, Wheeler, & Danesh, 2005) and to labour migrants (Lindert, von Ehrenstein, Priebe, Mielck, & Brähler, 2009). This increased risk is chiefly explained by exposure to multiple PTEs, in particular torture (Steel et al., 2009). Demographic factors associated with greater risk of PTSD among refugees include older age, female gender, more education, higher socioeconomic status, and rural residence in their countries of origin (Porter & Haslam, 2005).
The assumptions of trauma-focused psychiatric epidemiology among refugees are challenged by two approaches, one focusing on culture and the other on stressors (Summerfield, 1999; Warfa & Bhui, 2007). With respect to culture, Hinton and Lewis-Fernández (2011) conclude that while PTSD has substantial cross-cultural validity it also shows substantial variability in the prevalence of specific symptoms (especially avoidance and somatic symptoms; see also Rohlof, Knipscheer, & Kleber, 2014) as well as in the interpretation of PTEs and posttraumatic symptoms. Some authors have argued that the cross-cultural focus on PTSD should be replaced by a focus on local idioms of distress (Miller et al., 2006).
The posttraumatic stress construct also has been nuanced by an emphasis on the influence of current stress on refugee mental health (e.g., Murray, Davidson, & Schweitzer, 2010). Postmigration stressors, including living in institutional accommodations and having restricted economic opportunities (Porter & Haslam, 2005), as well as ongoing political terror in the country of origin (Steel et al., 2009), have been shown to contribute significantly to the development of PTSD in refugees. Consequently, refugees generally have poorer mental health status than compatriots who stayed in the country of origin (Porter & Haslam, 2005).
The applicability to refugees of the DSM-5 and ICD-11 symptoms of negative changes in belief has been the subject of similar debate. On the one hand, the new DSM-5 cluster of negative alterations in cognitions and mood has been posited to be especially applicable to traumatised refugees because it facilitates a distinction between refugees with and without PTSD (Schnyder et al., 2015). Moreover, the ICD-11 complex PTSD symptom of diminished and defeated sense of self also may apply to refugees in particular because complex PTSD is believed to develop following exposure to prolonged or repetitive traumatic events from which escape is difficult or impossible, such as torture or genocide campaigns (see Maercker et al., 2013). On the other hand, authors such as Hinton and Lewis-Fernández (2011) question the phrasing of the DSM-5 D2 criterion and suggest adding examples that may be more applicable to non-Western traumatised populations, such as “I’ve lost my soul forever” and “my whole nervous system is permanently ruined” (p. 795).
To date, structural assessment of negative assumptions in refugees has been limited. One of the most widely used instruments for assessing changed assumptions is the World Assumptions Scale (WAS; Janoff-Bulman, 1989). Although the WAS has been used with a wide variety of traumatised populations, it has been used only rarely in studies on refugee mental health. In a study with the WAS, Mooren and Kleber (2001) found diminished belief in the benevolence of world and people in Bosnian refugees. However, not all WAS items were administered in this study. Paunovic and Öst (2001) reported the outcomes of a randomised trial, which showed the positive effect of cognitive-behavioural treatment on world assumptions in refugees. However, they only reported overall endorsement and not the different subscales.
The current study aimed to increase knowledge about world assumptions among treatment-seeking refugees. Such knowledge may aid in understanding the extent to which new symptoms of negative cognitions recognised in DSM-5 and ICD-11 may be a potential focus for psychotherapeutic and psychosocial interventions among refugees. Given the association between negative assumptions and posttraumatic functional impairment, alleviating negative assumptions about self, others, and the world may contribute to improved posttraumatic functioning of refugees on individual, family (Weine et al., 2004), and societal levels (Kivling-Bodén & Sundbom, 2002).
Our study aimed to answer the following questions: First, how do world assumptions of treatment-seeking refugees compare to those of appropriate reference populations? Second, is there a relationship between world assumptions and demographic characteristics (sex, age, and length of residency in the Netherlands) as well as trauma-related characteristics (the number of traumatic event types and PTSD symptom severity) among treatment-seeking refugees?
The study hypotheses were that: (1) given the high exposure to potentially traumatic events among treatment-seeking refugees, the level of endorsement of positive world assumptions would be lower in refugees than in reference populations; (2) variance in world assumptions could be accounted for partly by demographic factors; and (3) variance in world assumptions could be accounted for partly by number of traumatic event types and PTSD symptom severity.
Method
Setting, procedure, and sample
Data were collected in Foundation Centrum ’45, a Dutch centre specialising in diagnostics and treatment of the psychosocial consequences of war, persecution, profession-related traumatic events, and other complex traumatic events. Data collection took place between July 2002 and October 2011. The WAS was administered as part of a routine diagnostic self-report assessment to all refugees who applied for treatment in that period, who had temporary or permanent refugee status or Dutch nationality, and whose language proficiency was sufficient to complete a diagnostic assessment in Dutch (at the centre, refugees who have not yet obtained a refugee status and who speak little Dutch are given a diagnostic assessment that does not include the WAS).
Patients were asked for informed consent. Upon consultation, the Medical Ethics Committee of the University of Leiden stated that no review of the ethical merits of the study was needed because assessments were conducted primarily for diagnostic purposes within the institution and only secondarily for data analysis. In our analyses, we included those assessments that had been completed within 12 months of referral. WAS questionnaires that contained more than three missing items were excluded from analysis. The dataset at our disposal contained a small number of demographic variables (gender, age, country of origin, and length of residency in the Netherlands).
The final sample 1 consisted of 213 patients: 153 men (71.8%) and 60 women (28.2%). Ninety-three percent of assessments had been completed within 6 months of referral. Patients were from 30 countries of origin, most frequently former Yugoslavia (n = 67, 31.4%), Iran (n = 37, 17.4%), Iraq (n = 33, 15.5%), Afghanistan (n = 17, 8.0%), and Turkey (n = 11, 5.2%). Age at assessment ranged between 24 and 72 years (M = 42.9, SD = 8.7). Length of residency in the Netherlands ranged between 4 and 38 years (M = 13.4, SD = 5.0; n = 166). Types of traumatic event most frequently experienced by the participants were being close to death (n = 142/192, 74.0%), threat to be physically tortured (n = 130/192, 67.7%), and forced separation from family members (n = 126/191, 66.0%). Participants reported a mean of 10.2 different types of self-experienced traumatic events (SD = 5.5; n = 153). The mean PTSD symptom severity score on the Harvard Trauma Questionnaire (HTQ; Mollica et al., 1992) was 3.1 (SD = 0.6; n = 192), with 82.3% of the participants (n = 158/192) scoring above the cut-off score of 2.5 (see Instruments section).
Instruments
World Assumptions Scale (WAS)
The WAS is a self-report questionnaire consisting of 32 items that are rated on a 6-point Likert scale ranging from 1 (strongly disagree) to 6 (strongly agree; see Janoff-Bulman, 1989, for full scale). The items form eight subscales consisting of four items each: Benevolence of (impersonal) World, Benevolence of People, Randomness, Justice, Controllability, Self-Worth, Self-Controllability, and Luck. Although some studies have grouped those eight subscales into three secondary dimensions, a confirmatory factor analysis using a large sample of trauma-exposed participants considered an eight-factor model to be better than a three-factor model (Elklit, Shevlin, Solomon, & Dekel, 2007). Therefore, in the current study we used the eight subscales. A higher endorsement of subscale items reflects a higher belief in positive world assumptions, with the exception of items in the Randomness subscale, which are negatively worded and are interpreted the other way around.
Based on research with a trauma-exposed sample, Elklit et al. (2007) suggest that the WAS has adequate psychometric properties for use in clinical and research settings. Based on research with a general sample, however, Kaler et al. (2008) question the temporal stability and construct validity of the WAS. To our knowledge, the only study on the cross-cultural validity of the WAS was conducted with a Turkish sample (Yilmaz, 2008). This study showed a six-factor structure, an acceptable internal consistency, but a poor test–retest reliability.
Our institute translated the WAS from English into Dutch as part of a larger project of translating a number of diagnostic instruments for use in the standard intake procedure (Kleijn, Hovens, & Rodenburg, 2001). Validation of the Dutch language version of the WAS is being undertaken. In the current study, reliability of the subscales was reasonable with Cronbach’s alphas ranging between .65 (Self-worth) and .77 (Benevolence of the World).
Harvard Trauma Questionnaire (HTQ)
The HTQ (obtained from the Harvard Program in Refugee Trauma) is a self-report questionnaire that was designed especially to assess traumatic experiences and posttraumatic symptoms in refugees. In the first part of the HTQ, participants are asked which types of traumatic events they experienced, witnessed, or heard of, out of a total list of 20. For each participant a total score was computed by counting the number of self-experienced traumatic event types (range 0–20). The second and third parts of the HTQ assess PTSD according to the DSM-IV (16 items) and associated features (14 items). In this study we used only the DSM-IV part.
Symptoms are rated on a 4-point scale ranging from 1 (not at all) to 4 (extremely). We computed mean scores, allowing for a maximum number of three missing values. A mean score of 2.5 was used as a cut-off score for likelihood of a PTSD diagnosis. The HTQ has been shown to have good psychometric properties in studies with refugees (Hollifield et al., 2002). The Dutch language version, although carefully translated (Kleijn et al., 2001), has not been separately validated.
Statistical analyses
First, one-sample t-tests in SPSS (Version 20) were used to test for differences in WAS subscale mean scores between our sample of treatment-seeking refugees and three reference samples: (a) participants who experienced PTSD after exposure to a single traumatic event, (b) participants who experienced PTSD after exposure to multiple traumatic events, and (c) the general population. The first reference sample (Elklit et al., 2007) consisted of 722 Danish participants who had sustained whiplash, chiefly as a result of motor vehicle accidents, and who scored above the HTQ cut-off score for PTSD. The second reference sample (Dekel, Solomon, Elklit, & Ginzburg, 2004) consisted of 59 Israeli military veterans, all of whom were clinically diagnosed with PTSD. The third reference sample (Kaler et al., 2008) consisted of 742 undergraduate psychology students in the United States.
Second, a path model in which the subscales of the WAS were simultaneously regressed on gender, age, length of residency in the Netherlands, number of self-experienced traumatic event types, and PTSD symptom severity was fitted to the data in MPlus Version 7.3 (Muthén & Muthén, 1998–2012), making use of maximum likelihood estimation (see Figure 1).
Path model in which beliefs about world assumptions are regressed on gender, age, length of residency in the Netherlands, number of self-experienced traumatic event types, and PTSD symptom severity.
Regression paths as well as covariances (in Figure 1 depicted as multiheaded arrows) between the predictor variables and between the subscales of the WAS were modelled. Significance of the regression coefficients was tested.
Results
Mean scores and standard deviations on the World Assumptions Scale
In Figure 2, mean levels of the WAS are presented for our sample of treatment-seeking refugees, as well as the three reference samples.
Mean scores on the World Assumptions Scale for treatment-seeking refugees in comparison to three reference samples.
Compared to the sample exposed to single trauma, results were mixed, with refugees reporting significantly less belief in Benevolence of World, Benevolence of People, Meaningfulness of the World (i.e., the reversal of Randomness), and Luck, and significantly more positive assumptions about Justice, Self-Worth, and Controllability. In comparison to a sample exposed to combat, refugees reported significantly less belief in Benevolence of World, Benevolence of People, Self-Worth, Meaningfulness of the World (the reversal of Randomness), Luck, and Controllability. Compared to the general population sample, treatment-seeking refugees reported significantly less positive world assumptions in all domains. Assumptions were especially negative with regard to the Benevolence of World, Benevolence of People, and Luck subscales, on which refugees scored lower than all reference samples.
Differences among the refugee sample and the reference samples were smallest with regard to Self-Worth and Self-Controllability. It must be noted that with regard to sex and age, our sample (71.8% male, mean age = 42.9) differed from the reference samples of participants exposed to single trauma (21% male, mean age = 43.0), military veterans exposed to combat (100% male, mean age = 25.3), and the general reference sample (21% male, age between 18 and 21).
Unstandardised and standardised regression coefficients of the path model for the prediction of beliefs about world assumptions
Note. B = unstandardised regression coefficient; SE = standard error; β = standardised regression coefficient: 02 small, .13 medium, .26 large (Cohen, 1988).
*p < .05.
Variance in world assumptions was most strongly accounted for by PTSD symptom severity. Higher levels of PTSD symptom severity were related to lower beliefs in Benevolence of World, Benevolence of People, Justice, Self-Worth, and Luck. Variance in world assumptions also was accounted for by gender, with women reporting higher beliefs in Self-Worth than men, and men reporting higher beliefs in Self-Controllability and Controllability than women. Greater age was related only to a higher belief in Controllability. Length of residency and number of traumatic event types were not related to any world assumptions; beliefs in Randomness were not associated with any predictor.
Discussion
Endorsement of world assumptions
In this study, refugees generally showed less belief in positive world assumptions than other trauma-exposed samples and the general population. Assumptions were especially negative with regard to the Benevolence of World, Benevolence of People, and Luck subscales, on which refugees scored lower than all three reference samples. These results confirm our hypothesis that endorsement of positive world assumptions among refugees would be lower when compared to other populations.
Interestingly, refugees had less positive beliefs on most world assumption subscales than combat veterans, another population exposed to multiple traumatic events. Although this difference might be explained by exposure to a greater number of traumatic event types in refugees, it also may be related to exposure to events that do not meet the A-criterion for PTSD. Refugees are known to experience a broad range of stressful events, including postmigration experiences such as discrimination (Warfa & Bhui, 2007). Postmigration stressors may contribute to a vicious circle of mental health problems in which migration-related stressors such as low social status lead to a deterioration of mental health, which in turn may increase the chances of stigma and discrimination (Quinn, 2014). Although such experiences may not necessarily be classified as “traumatic,” they may add to the conviction that the world and other people are not as benevolent as previously assumed.
Contrary to our hypothesis, endorsement of Self-Worth and Self-Controllability in refugees was relatively high. There may be several reasons for this. First, the WAS may not adequately capture trauma-related assumptions about the self for refugees. The WAS contains items such as “I often think I am no good at all” and “I have a low opinion of myself,” which are an evaluation of character or mood state. Hinton and Lewis-Fernández (2011) give additional suggestions for assumptions about the self which they believe may be valid across some cultures, including “I’ve lost my soul” and “My nervous system is permanently ruined.” Qualitative research on world assumptions among refugees would be valuable to clarify the appropriate wording for such measures, because it seeks answers to questions about how and why world assumptions change (Green & Thorogood, 2013) and can detect unnamed and specific cultural and contextual factors associated with world assumptions.
Alternatively, high endorsement of Self-Worth and Self-Controllability by treatment-seeking refugees may reflect an actual high belief in self-worth and self-controllability. Jobson (2009) posits that in collectivistic cultures, the meaning of traumatic experiences is shaped by constructs such as relatedness to family and community, while in individualistic cultures, constructs such as autonomy and self determine the meaning of traumatic experiences. This implies that traumatised people in collectivistic cultures may be more vulnerable to changes in assumptions about relatedness, while having more robust assumptions about self. As the majority of our sample originated from cultures that may be considered predominantly collectivistic, their beliefs in self-worth and self-controllability may have been maintained in the face of multiple traumas.
The implication of this last explanation is that the ICD-11 complex PTSD symptom of diminished sense of self might be less applicable to refugee populations from collectivistic cultures. The prevalence of complex PTSD has been found to be relatively low in several cross-cultural studies as well as refugees (Ter Heide, Mooren, & Kleber, 2016). Further study of ICD-11 complex PTSD among refugees is needed. If the prevalence is found to be low (as found for example by Tay, Rees, Chen, Kareth, & Silove, 2015), it would be important to clarify which criteria refugees fail to meet.
Association between world assumptions and demographic variables
Confirming our hypothesis, we found an association between world assumptions and gender, with women reporting higher beliefs in Self-Worth than men, and men reporting higher Self-Controllability and Controllability than women. The relatively low Self-Worth score in men was especially interesting given the finding of Bleidorn et al. (2015) that, across cultures, men consistently report higher self-esteem than do women. Again, our finding may be related to postmigration stressors. Men commonly experience a greater loss of social status than women (Young & Chan, 2015) with migration, and this may contribute to a greater loss of self-esteem among male refugees. Almost no association was found between world assumptions and age. This contrasts with Bleidorn et al. (2015) who found that across cultures, self-worth tends to increase with age. Posttraumatic and postmigration stress may interfere with the universal tendency to gain self-confidence from late adolescence to middle adulthood.
No association was found between world assumptions and length of residency in the Netherlands, implying that, at least for this treatment-seeking sample, belief in positive world assumptions was not restored the longer an individual lived in relative safety. This finding confirms meta-analytic outcomes by Steel et al. (2009), who found no association between PTSD prevalence and refugees’ length of residence in countries of asylum. This suggests that to restore belief in self, others, and the world, active intervention is needed rather than simply hoping that “time will heal all.” Interventions may consist of psychotherapeutic treatment such as cognitive-behavioural therapy (see Diehle, Schmitt, Daams, Boer, & Lindauer, 2014, which includes Paunovic & Öst, 2001) or cognitive processing therapy (Schulz, Resick, Huber, & Griffin, 2006), as well as psychosocial interventions aimed at increased societal participation and acceptance (Kivling-Bodén & Sundbom, 2002).
Association between world assumptions and trauma-related variables
Contrary to expectation, the number of traumatic event types was not significantly related to endorsement of world assumptions. This might be explained by the generally high mean number of traumatic event types (10.2) in our sample, suggesting that in our sample a saturation point was reached after which world assumptions no longer deteriorate. Alternatively, other characteristics of trauma history, such as age of traumatisation or interpersonal versus noninterpersonal traumatic experiences, may have a greater impact on world assumptions. In this study, we did not check for such correlations because the high number of possible correlations (of 20 event types and eight subscales) was likely to result in false positives.
In contrast, higher PTSD symptom severity was associated with more negative world assumptions, including lower scores on Benevolence of World, Benevolence of People, Justice, Self-Worth, and Luck. This finding supports the inclusion of negative changes in belief in the DSM-5 and ICD-11 diagnoses of PTSD and complex PTSD, respectively. In other words, the correlation between high PTSD symptom severity and high level of negative beliefs among treatment-seeking refugees likely reflects the internal consistency of the PTSD construct.
Conclusion
To our knowledge, this is the first detailed report of world assumptions among refugees. Our findings show the relevance of world assumptions to diagnostic criteria and posttraumatic functioning in refugees. The WAS showed acceptable internal consistency in this large, culturally heterogeneous sample. Cultural heterogeneity, however, prevented us from drawing any firm conclusions on the role of culture in world assumptions. As the sample consisted only of refugees with a residency status, we were unable to determine the level of world assumptions among refugees without a residency status, who are likely to be more heavily burdened (Laban, Gernaat, Komproe, Schreuders, & De Jong, 2004). Finally, the limited number of demographic and trauma-related variables prevented us from differentiating between posttraumatic and postmigration influences on world assumptions. Determining which factors are most strongly associated with world assumptions might help shape efforts to prevent and treat negative beliefs in world assumptions among refugees. In conclusion, our study shows the importance of assessing, understanding, and restoring world assumptions among treatment-seeking refugees, whose trust in self, others, and the world has been disrupted by trauma and forced migration.
Footnotes
Acknowledgements
The authors thank Ania Fiksinski and Talitha Cornelisse for their help in preparing the dataset.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by Foundation Centrum ’45 | partner in Arq Psychotrauma Expert Group.