Posttraumatic stress disorder correlates among Iraqi internally displaced persons in Duhok, Iraqi Kurdistan

In 2014, the terrorist militant group the Islamic State of Iraq and Syria (ISIS) took over one-third of Iraq. This study measured the rate of posttraumatic stress disorder (PTSD) among Iraqi internally displaced persons (IDPs) and examined associated demographic and traumatic risk factors and comorbid psychiatric symptoms. A cross-sectional survey was carried out in April-June 2015 at the Khanke camp, northern Iraq. Trauma exposure and PTSD were measured by the Harvard Trauma Questionnaire (Iraqi version), and psychiatric comorbidity was measured by the General Health Questionnaire (GHQ-28). Of 822 adult IDPs, 33.8% screened positive for PTSD. Associated factors included exposure to a high number of traumatic events, unmet basic needs and having witnessed the destruction of residential or religious areas. Being a widow was the only linked demographic factor (OR = 14.56, 95% CI: 2.93–72.27). The mean scores of anxiety/insomnia and somatic symptoms were above the average cutoff means (M = 3.74, SD = 1.98, R = 0–7 and M = 3.69, SD = 2.14, R = 0–7, respectively) among the IDPs with PTSD. High traumatic exposure, specically unmet basic needs and having witnessed destruction, was an important predictor of PTSD among IDPs. Psychiatric comorbidity was high among those with PTSD. Anxiety, insomnia and somatic symptoms were strongly associated with PTSD. These ndings are important for mental health planning for IDPs in camps.

Although many studies have investigated the psychosocial consequences of mass traumas and displacement in different population samples, we aimed to investigate the mental health impact of mass displacement in a population with a long history of multiple genocides added to decades of exposure to internal and external con icts and political instability. Our study highlights the necessity of mental health assessment, care, and follow-up among this speci c population group because although refugee camps provide safe accommodations, displaced people need physical and mental health care. We examined the prevalence of PTSD and investigated the possible correlated factors of PTSD among IDPs resettled in refugee camps in Duhok, Iraqi Kurdistan. Furthermore, we investigated the association between PTSD and symptoms such as depression, anxiety, somatic symptoms and social dysfunction.

Study design and participants
This is a cross-sectional study performed in April-June 2015 as part of a larger community cross-sectional survey among Iraqi IDPs and Syrian refugees resettled in refugee camps in Duhok Province. This research was supported by the Directorate of Health and the Italian nongovernmental organization Association for Solidarity among People (AISPO).
The participants were adults of both genders living in the Khanke camp, which is located 20 km west of the city of Duhok, northern Iraq. In 2015, the number of IDPs in Duhok Province was 483,068, constituting 92,024 families [13]. They were living in the camps and in un nished structures and school buildings. At the time of the study, the camp population comprised 16,460 individuals living in 3,120 tents, most of whom were Yezidi IDPs from Sinjar [14].
The inclusion criterion was that the individual should have a history of displacement following the 2014 ISIS attacks in Iraq. Individuals suffering from mental disorders that affect insight, disturb normal communication or deteriorate cognitive abilities, such as intellectual disabilities, dementia or psychotic disorders, were excluded.
The tent numbers were entered into an Excel sheet, and 822 tents were randomly selected. From each selected tent, one eligible adult tting the sample inclusion criterion was asked to participate voluntarily. The participant was randomly selected by putting eligible family members' names on small pieces of paper in a bag and choosing one blindly.
The scienti c committee of the College of Medicine/University of Duhok and the Research Ethics Committee of the Duhok Directorate of Health approved the study. The participants provided written informed consents before we conducted the interviews. The interviewer assured the participants of the con dentiality of the gathered data. Of the 822 IDPs selected to participate in the current study, only 8 did not consent to participate in the study; in those cases, other family members were randomly selected and asked to participate in the study. A psychiatrist familiar with training in such instruments trained six counselors in the use of the study tools full time for 5 days. The project supervisor, his assistant, and the 6 interviewers were hired locally. The face-to-face interviews lasted for two months, beginning on April 15, 2015.

Measures
The demographic questionnaire requested the age, gender, religion, marital status, education level, work status, number of siblings, past psychiatric history, and past family history of psychiatric disorders.
The Harvard Trauma Questionnaire (HTQ) is a widely used simple and reliable checklist that measures traumatic experiences, torture exposure, and symptoms [15]. It is bene cial in the assessment of the severity and types of premigration or displacement traumas suffered by survivors of mass violence and common trauma-related psychiatric disorders, such as depression, PTSD and anxiety [16]. Parts I and IV of the Iraqi version of the HTQ were used [17]. Part I comprises 43 items assessing the traumatic events experienced and witnessed [18]. The rst 16 items of part IV assess PTSD symptoms [19]. For each item, the individual selects from a 4-point severity scale: "not at all", "a little", "quite a bit", and "extremely". In the present study, the HTQ symptom scale received a Cronbach's alpha of 0.799, which indicates good internal consistency. If the individual's HTQ symptom score is ≥2.5, it indicates a likelihood of clinical PTSD.
The General Health Questionnaire (GHQ-28) was developed by Goldberg in 1978 as a screening instrument for those at risk of developing psychiatric disorders, especially emotional distress, in medical settings [20]. Although it is designed to screen but not to diagnose psychological well-being, it can detect possible cases of psychiatric morbidity, and the total score can be used as an index of severity [21]. It comprises four factor subscales: somatic symptoms (items 1-7); anxiety/insomnia (items [8][9][10][11][12][13][14]; social dysfunction (items [15][16][17][18][19][20][21], and severe depression (items 22-28) [22]. The respondent is asked to report alterations in his/her mood, feelings and behaviors in the previous 4-week period. The individual responds on a 4-point scale: "less than usual", "no more than usual", "rather more than usual", and "much more than usual" [23]. A binary method of scoring was used in which the rst 2 response selections were scored as 0 and the last 2 as 1 [21]. A score above 4 indicates the presence of distress [21].

Statistical Analysis
The gathered data were analyzed using SPSS (software statistical computer package version 22). The preliminary descriptive analysis used frequency tables, including means (M) and standard deviations (SDs), for quantitative data, and percentages were used for qualitative data. The categorical data were tested by chi square. Principal component analysis (PCA) with oblique rotation (δ = 0) was applied to variables of the HTQ part I to designate component groups of traumatic events. The oblique rotation was carried out because we anticipated that the events were correlated with one another. The items with a primary loading greater than 0.30 on the same component were combined. Multiple linear regression analysis was adopted for the types of experienced traumatic event components as predictors of PTSD. Logistic regression analysis was used to probe the contribution of demographic factors to participants' PTSD symptom levels. Signi cance was assumed at P values <0.05, and high signi cance at P <0.001. The independent sample t test was applied for differences in means. Table 1 demonstrates the sociodemographic characteristics and mental disturbances of the participants at the time of the study. The mean age was 33.79 (SD = 12.74) years, and the range was 77 (18-95) years, but most of the participants were young (18-40 years). There were slightly more males than females (56.4% vs 43.6%). The largest group was from the Yezidi religion (n = 812, 98.8%), and most of them were married (77.6%). A total of 364 (44.3%) of the participants were illiterate, 33.1% had completed primary school, and a few had completed high school or held a higher academic degree. The majority were unemployed (n = 694, 86.2%). Most of the participants were from large families, and the mean number of siblings was 7.22 (SD = 3.26). Past family psychiatric history Positive 53 6.5 PTSD, PTSD-rst 16 questions of HTQ part IV Only 3.9% had a positive past psychiatric history, and 6.5% had a family history of psychiatric disorders. The PTSD scores measured on the HTQ are also shown in Table 1. PTSD had a mean score of 2.26 (SD = 0.51, R = 1-4). Intrusion and avoidance symptoms had the highest means (M = 2.63, SD = 0.59, R = 1-4 and M = 2.5, SD = 0.63, R = 1-4, respectively) compared to numbing and hyperarousal symptoms. Table 2 displays typing and subtyping of common traumatic events, item loading, and frequencies and percentages of experiencing or witnessing among the participants. Prior to the PCA step, the frequencies and percentages of exposure to traumatic events were studied. Traumatic events that were extremely rare (experienced by less than 5%) were deleted because they did not provide enough information to maintain a meaningful grouping of items. From a total of 48 items, 25 were deleted. Examples of uncommon traumatic events were sexual violence; brainwashing; forced labor; witnessing the torture, murder, arrest or execution of others; witnessing chemical attacks; being con ned to home; being forced to pay for bullets used to kill family members; receiving the body of a family member and being prohibited from mourning and burial rites; and having someone inform against a participant. Another item (suffering ill health without access to medical care) was also canceled because its primary loading was less than 0.30 on the same factor. As a result, 22 variables remained that were suitable for PCA.

Results
PCA yielded 6 trauma components, which altogether produced a cumulative variance of 50.1%. Every traumatic event subtype had a loading greater than 0.30 on one component and did not have a loading greater than 0.30 on the next component. The traumatic events were sorted as follows: 1. Trauma to or persecution of self, 2. Trauma to or abduction of family member or friend, 3. Forced immigration, 4. Lack of basic necessities, 5. Witnessed destruction, and 6. Coercion. who did not. All these differences were statistically signi cant (X 2 = 8.66, P < 0.05).   OR, Odd ratio, CI, Con dence interval, *P-value <0.05 Table 6 demonstrates the comparison of the means and SDs of the GHQ-28 and its components among Iraqi IDPs with or without the diagnosis of PTSD. The scores of the total GHQ-28 and all its components were signi cantly higher in those diagnosed with PTSD than in those without the diagnosis (P <0.001 in all domains). The total GHQ-28 score was near the average cutoff mean (M = 13.32, SD = 4.71, R = 0-28) among the IDPs with a diagnosis of PTSD. The mean scores of anxiety/insomnia and somatic symptoms were above the average cutoff means (M = 3.74, SD = 1.98, R = 0-7 and M = 3.69, SD = 2.14, R = 0-7, respectively) among the IDPs with PTSD. IDPs, internally displaced persons, PTSD, post-traumatic stress disorder mean score ≥ 2.5, t, independent sample t test, df, degree of freedom, P, P value.

Discussion
Our study provides evidence of mental distress after exposure to traumatic events endured by Iraqi IDPs resettled in IDP camps following the violent attacks by ISIS in 2014. It showed a 33.8% prevalence rate of PTSD and a high rate of psychiatric symptoms, especially anxiety, insomnia, and somatic symptoms. A higher number of traumatic events was associated with a higher rate of PTSD. Predictive traumatic events for PTSD were unmet basic needs and having witnessed destruction. Demographic factors did not predict PTSD except that being widowed increased the rate of PTSD by 14 times.
The high prevalence rate of PTSD (33.8%) found in our study is consistent with the rates of PTSD revealed in other studies performed on Iraqi refugees and displaced people after the events of 2014. A systematic review of the literature describing the prevalence rates of PTSD among resettled Iraqi refugees in Western countries nds a range from 8 to 37.2% [24]. Asylum seekers, refugees, and IDPs displaced to Iraqi Kurdistan showed a PTSD prevalence rate of 48.7% [25]. Tekin et al. [26] showed a rate of 42.9% among displaced Iraqi Yazidis. The rate was 36.4% among Yazidi children and adolescents who had immigrated to Turkey from Iraq [27]. Among other refugees inhibiting Iraqi Kurdistan camps, such as Kurdish Syrian refugees, the estimated levels of PTSD symptoms ranged between 35 and 38% [28]. This nding indicates the impact of mass con ict and displacement and the severity of the traumatic events that these respondents had experienced, including a combination of war and political, religious and ethnic violence. Several factors, such as different methodological approaches and demographic characteristics, may contribute to the differences in PTSD rates reported in other studies.
Among the 4 categories of speci c PTSD symptoms, the intrusion symptoms suffered by the Iraqi IDPs participating in our study showed the highest means. Symptoms of avoidance were also relevant, with high means compared to numbing and hyperarousal symptoms. Iraqi refugees in Germany presented similarly, with nightmares as the most prominent symptom of PTSD [29].
Traumatic events among refugees and IDPs are not presented individually, but they are usually accumulated and interconnected [27]. There is a positive correlation between the number of experienced traumatic events and vulnerability to developing mental disorders [6]. Our study presented the effect of the number of traumatic events on the rate of PTSD. Approximately 66.7% of those who had experienced more than 20 traumatic events developed symptoms of PTSD. This indicates the cumulative effect of exposure to multiple traumatic events before and during the displacement time. The cumulative trauma index accounts for the increased prevalence rate of PTSD among Iraqi refugees [19]. This nding supports the dose-response relationship between traumatic exposure and PTSD among refugees in postcon ict periods [30]. Among Syrian refugees in Turkey, experiencing 2 or more traumatic events was a signi cant predictor of PTSD [31].
Different subtypes of experienced traumatic events had different effects as predictors of PTSD. There were signi cant positive correlations between traumatic events and PTSD symptoms among Kurdish Syrian refugees in Iraqi Kurdistan camps [28]. In our study, the most signi cant predictors for the development of PTSD were unmet basic needs and having witnessed destruction. Unmet basic needs included becoming homeless with no access to health care and lack of food and water. This nding indicates that being deprived of basic elements of survival can have a predictive value in the development of PTSD among IDPs because of its direct impact on individuals. Having witnessed destruction included witnessing the destruction of religious shrines; the shelling, burning, or razing of residential areas; and rotting corpses. Witnessing these violent events, which are associated with losing the homeland and religious persecution, exacerbates war impacts. The most painful or terrifying traumatic event recounted by the survivors of "Anfal", a military operation against the Kurds of northern Iraq, was witnessing murder [32]. Additionally, trauma subtypes provided even more entropy than the cumulative trauma effect in the prediction of PTSD [19]. The study showed that unmet basic needs was a stronger risk factor for depression than for PTSD.
When we searched for sociodemographic predictors of PTSD among Iraqi IDPs, marital status was the only component that acted as a predictive factor. The severity of experienced traumatic events and the short period between traumatic exposures and the assessment (less than 1 year) made the IDPs perceive the traumas more collectively and more similarly than respondents in other studies. A study conducted on Yazidis found that no sociodemographic predictors for PTSD among the surviving Yezidi women and girls except the number of family members directly affected by ISIS [4]. Being a widow increased the risk of developing PTSD by 14 times. Not having a partner is associated with poor mental health owing to the lack of the social support provided by a partner and the sense of increased responsibility for raising the children and increased worry about the future. Additionally, widows may have witnessed the killing of their husbands by ISIS solders.
This nding indicates that poor social support is a strong predictor of PTSD among Iraqi IDPs. Those who were widowed prior to the 2014 events may have been affected by prior loneliness and its mental consequences, aggravating their vulnerability to PTSD [33]. Following traumatic events, the incidence of PTSD is signi cantly higher in widows than in married women with living spouses. Among the Rwandan population, 87% of widows suffered from panic disorders and had a higher incidence of psychopathology, including PTSD [34].
Our study clari ed the occurrence of mental health disturbances in addition to PTSD among Iraqi IDPs. The comorbid psychiatric problems detected among those suffering from PTSD in comparison to the non-PTSD-affected group were somatic symptoms, anxiety/insomnia, social dysfunction and severe depression. The most frequently observed psychological problem with higher means was anxiety/insomnia.
The reason could be the severe stress experienced during the attacks and during displacement, or it could be PTSD and depression. Anxiety and insomnia can be part of or can aggravate the hyperarousal component of PTSD psychopathology [35]. Although the relationship between sleep and PTSD appears to be more complex than can be explained by the current PTSD paradigm [36], the research suggests that experiencing traumatic events can lead to sleep problems, and posttraumatic stress can interfere with sleep. Another explanation is that hypervigilance, an important symptom of PTSD, can lead to regional arousal states and in turn produce insomnia [37].
In addition to anxiety insomnia, somatic symptoms were more prominent among IDPs. The relationship between traumatic experiences, PTSD, and somatic symptoms has been well reported [38]. Biologically, PTSD is associated with instability in the limbic system and interferes with the hypothalamic-pituitary-adrenal and sympatho-adrenal medullary axes, which interfere with neuroendocrine functions, resulting in pseudoneurological somatic symptoms [37]. Psychologically, there are two somatic complaint factors related to PTSD psychopathology: the rst is related to the physical perceptions of bodily dysfunctions and is called "weakness," and the second is related to sympathetic hyperactivation and is called "arousal [38]." A high correlation between the somatic and syndrome scales and the PTSD checklist was found among Cambodian refugees [39].
Our study showed that the IDPs with HTQ scores above the PTSD cutoff score have higher depression means than those without PTSD.
This high rate of depression accompanying PTSD may be due to the effect of loss, which is usually accompanied by trauma, especially in con ict-affected populations, and often involves the deaths of family members, relatives, or friends [40]. On the other hand, exposure to traumatic events and PTSD can be a risk factor for depression. Furthermore, the poor nancial situation of the IDPs can explain the high rates of depression. Among IDPs in Kaduna, northwestern Nigeria, comorbid PTSD was a predictive factor of depression [41].
This study is not devoid of limitations. Because the sample was taken from only one IDP camp, omitting IDPs inhabiting informal settlements and within cities, it is di cult to determine whether it is representative. Those with a history of psychiatric disorders prior to displacement were excluded from participating in the study. The strength of this study is the careful assessment of PTSD and other mental problems by six well-trained Kurdish psychological counselors who did not need the help of interpreters and did not use self-reported questionnaires. Additionally, they were able to select probable cases and refer them to suitable mental health services. Furthermore, this is a rare study that highlights the cumulative and qualitative effects of trauma exposure in IDPs who have recently escaped from severe terroristic attacks to displacement camps.

Conclusions
The results of this study provide evidence of the high prevalence rate of PTSD among Iraqi IDPs living in displacement camps in northern Iraq. The number of experienced traumas was positively correlated with the symptom severity of PTSD. Unmet basic needs and having witnessed destruction were the most impactful traumatic events. Anxiety, insomnia, and somatic symptoms had high levels of comorbidity with PTSD. The results of our study provide a better understanding of the mental health of Iraqi IDPs and a cross-cultural understanding of the effects of mass con icts and displacement. Furthermore, the results of this study have possible applications for governmental and nongovernmental organizations and those who supply psychosocial support for Iraqi IDPs.

Consent for publication
Not applicable Availability of data and materials All data generated or analysed during this study are included in this published article and its supplementary information les.

Competing interests
No con ict of interest is declared by the authors and the funders.

Funding
The AISPO nancially supported the interviewers' participation in data collection.
Authors' contributions PHT, NIT, and HMS were all contributors to the design, planning and implementation of the current study. PHT conceptualized and designed the study, review, revised the manuscript, carried out the statistical analyses and preparation of submission les. NIT participated in the design of the study, coordinated and supervised data collection, interpreted the data. HMS coordinated data collection, reviewed statistical analysis, and edited the draft manuscript. All authors contribute to the writing of the manuscript and all read and approved the nal manuscript.