Post-Traumatic Stress Disorder in Kosovo Veterans

Objective: The objective of this study was to explore the prevalence of post-traumatic stress disorder (PTSD) at veterans 8 years after war, to find out relation of PTSD with other demographic and health related variables and discover the impact of depression and trauma on PTSD on 687 veterans from six municipalities in Kosovo. Method: Participants were 687 war veterans selected from six regions of Kosovo during 2008. The Harvard Trauma Questionnaire (HTQ-40), was administered to measure PTSD and Hopkins Symptom Checklist (HSCL-25) for depression and anxiety. Pearson chi-square, analysis of variance (ANOVA), and multiple regressions were used to analyze the data. Results: Results indicated that 11.2 % of veterans even 8 years after the war ended were suffering from PTSD. Six percent of veterans with PTSD did not seek medical help. They reported to have had emotional problems and physical problems, but they did not seek medical help. The findings suggest that self-medication may be one way of veterans dealing with PTSD symptoms. Veterans with PTSD symptoms were more concerned with “family issues” than those without PTSD symptoms. Conclusions: The study found that 8 years after the war the veterans of the war in Kosovo suffer PTSD symptoms and that a good number of them do not seek help for this problem. The establishment of adequate services by the state would transform these veterans’ dealing with PTSD not into a moral challenge but into a fundamental right to equal and high-quality services.


Introduction
War, as an experience that directly or indirectly affects the life of the members of the armed forces, is considered a powerful source of stress and particularly of the post-traumatic stress disorder (PTSD). In the current global political circumstances, the risk of experiencing trauma by involvement in combat and hostile environments is extremely high (Hoge et al., 2004). According to the Institute of Medicine of the National Academies, 13% to 20% out of 2.6 million U.S. military servants have displayed potential post-traumatic stress symptoms since 2001 (IOM, 2012). Different studies have reported serious mental health consequences of genocides and wars worldwide, affecting not only the civilian population but veterans as well (Cardozo et al., 2004;Dohrenwend et al., 2007;Dursa, Reinhard, Barth, & Schneiderman, 2014;Pham, Weinstein, & Longman, 2004).
The Balkans was also affected by wars during the 1990s, which caused mental health problems to the population in general and to vulnerable groups, such as war veterans. Indications of high level of post-traumatic stress and mental problems in individuals of all age groups, who were directly or indirectly involved in high emotional intensity events during wars and genocides, were reported by Cardozo, Vergara, Agani, and Gotway (2000), Cardozo et al. (2004), and Pham et al. (2004). Nearly 16 years after the 1998-1999 war, which had left a large part of the population with mood disorders (47.6%) and anxiety (41.8%; Priebe et al., 2010), a number of studies were conducted in Kosovo on post-traumatic stress in war victims. Those studies were focused on hospitalized persons (Ahern et al., 2004), relatives of the victims (Morina, Rudari, Bleichhardt, & Prigerson, 2010), victims of torture (Wang et al., 2012), refugees (Ai, Peterson, & Ubelhor, 2002), and adolescents (Gordon, Staples, Blyta, Bytyqi, & Wilson, 2008).
In spite of all the studies on the recognition of mental health problems, particularly PTSD, not much has been understood about its correlation with variables associated with the functioning of veterans. A good proportion of veterans do not seek help, or are provided inadequate help (Hoge, Riviere, Wilk, Herrell, & Weathers, 2014)  reasons, which are important to understand to create a healthy environment not only for their own functioning but also for the functioning of their families.
The goals of this study were (a) to identify the prevalence of PTSD in war veterans 8 years after the end of their service and (b) to study the correlation between PTSD and a number of variables associated with seeking medical help, their economic and social status, and the influence of a number of variables on PTSD.

Participants
The participants in this study were selected from the list of war veterans provided by the Association of War Veterans. The war veterans were involved in the study during 2008 through a random selection from the list, involving six regions, which geographically cover the whole territory of Kosovo. Of the 1,000 veterans selected from the list, 687 completed the questionnaire (response rate 72.3 %); 97.7% (n = 671) of them were male, while 2.3% (n = 16) were female. Mean age of participants was 42 years (SD = 9). Mean age of males was 42 years (SD = 9.1) and females was 36.2 years (SD = 5.5). For number of members living together in the family, responses ranged from 1 to 37 people, with an average of 7.8 (SD = 3.8). The response range for education was 4 to 23 years of school, with an average of 11.8 (SD = 2.3). Only 16.3% of participants have more than 250 euro per month, and the average number of children was 3.4 (SD = 1.9). Most of them (65.1%) were living in nuclear family and 34.9 % in extended family. Before the interview, the participants were read the letter of information and the informed consent. Each participant was informed of the voluntary nature of the study, their right to withdraw at any time, and the confidentiality of their responses.

Instruments
For the purpose of this study, participants filled out a demographic questionnaire prepared from the main investigator and the Harvard Trauma Questionnaire-40 (HTQ) for measuring PTSD. HTQ is a self-report questionnaire, which consists of three sections. In the first section are 48 questions related to the traumatic event(s) that veterans may have experienced or witness during the war. This section of the questionnaire was evaluated with "yes" and "no." The second section has open-ended question asking for subjectivity description of most traumatic experiences during the war. The third section of the questionnaire has 40 questions and refers to three main groups of symptoms associated with PTSD: intrusion symptoms associated with the traumatic event(s), persistent avoidance of stimuli associated with the traumatic event(s), and psychological hyper arousal. The participants were asked to rate each item on a 4-point Likert scale (1 = not at all, 2 = a little, 3 = quite a bit, and 4 = extremely). The first 16 items such as "recurrent thoughts or memories of the most hurtful or terrifying events" or "feeling as though the event is happening again" attempt to assess the accepted symptoms of PTSD diagnosis and following 24 items assess the impact of the traumatic experiences. The respondents with total PTSD score >2.5 were categories with "positive PTSD" while those <2.5 in HTQ were the "negative PTSD" group. In this study, the HTQ was found to have an internal consistency of .97(Cronbach's alpha).
The Hopkins Symptoms Checklist (HSCL-25) is a widely used screening instrument designed to measure symptoms of anxiety and depression. It consists of 25 items. The first part is composed of 10 items that measure anxiety, and the second part is composed of 15 items that measure depression. The scale for each question includes four categories of response (not at all = 1, a little = 2, quite a bit = 3 and extremely = 4. Two scores are calculated separately, and the Total score includes both subscales. The Cronbach's alpha for HSCL-25 for the Total score in this study was α = .956, for anxiety subscale, α = .905, and for depression subscale, α = .933.

Data Analysis
Internal consistencies of the HTQ and HSCL-25 were examined using Cronbach's alphas. Pearson chi-square test was used to explore the relation between PTSD and other variables in the study, whereas to find out the differences between groups in several variables, the t test for independent groups was used. Analysis of variance (ANOVA) was conducted to asses whether there were differences in the PTSD among groups and multiple linear regressions to asses whether depression and anxiety predict PTSD. The statistical package used in the present study was SPSS Version 19.0.

PTSD Scores and Its Correlation With Variables Asking Medical Assistance, Self-Medication, Their Concerns
The prevalence rate of PTSD in war veterans in Kosovo using HTQ was found in 11.2% of the sample. For PTSD scale, responses ranged from minimum of 40 to maximum of 152, with a mean of 64.9 (SD = 22.8).
To assess the relationship between PTSD and gender and some variables related to health seeking issues, we used the Pearson chi-square analyzes. The results for gender and PTSD groups were not significant, χ 2 (1, N = 687) = .02, p = .698. We found a significant relation between PTSD groups and how much they have been "working in the last week, χ 2 (3, N = 687) = 31.7, p = .0001. There were more than expected participants in the "not working" group and the "positive PTSD" and fewer participants grouped into "full time" and "positive PTSD." Also, a significant relation was found between "are you using medication" and PTSD groups, χ 2 (1, N = 687) = 31.5, p = .00. There were more participants into categories "yes" and "positive PTSD" than expected, and less participants in categories "no" and "positive PTSD" than expected. A significant relation was found between groups that "asked for medical assistance after the war" and PTSD groups, χ 2 (1, N = 687) = 31.5, p = .00. We found that participants in categories "no" and "positive PTSD" had less than expected participants. Chi-square shows a significant relation between the reason of seeking help and PTSD groups, χ 2 (3, N = 687) = 21.4, p = .001. Results show more than expected participants into groups "emotional problems," "other," and "more than one problem" in "positive PTSD" compared with "negative PTSD." Table 1 presents chi-square results for these variables.
Based on the results of the independent t test for PTSD according to whether they were using medication without subscription, we found significant differences t(667) = 11.1, p < .001. Participants that responded "yes" had higher mean scores in PTSD (M = 75.6, SD = 23.3) than those who responded "no" (M = 57.2, SD = 19.2). Significant differences on PTSD mean scores were found in group that sought medical assistance after war (M = 72, SD = 23.6), compared with those group that did not seek medical assistance (M = 54.8, SD = 17.2), t(667) = 10.3, p < .0001.
ANOVA was conducted to asses whether there were differences in the PTSD by "what is the most concern for you now" as a fixed factor. Participants' percentage responses are shown in Figure 1 Post hoc analyses using the Bonferroni procedure were conducted to assess the differences between group concerns. Significant differences were found between "family concerns" with all other groups. Significant differences were found between "political issues" with "more than one problem" (p < .001) but not with other groups such as "financial issues" (p = .496), "other problems" (p = .337), and "medical issues" (p = .170).
ANOVA (4 × 2) was used to explore effects of reasons for seeking medical assistance and using medication without prescription on PTSD. The effect of reasons for seeking help was significant for PTSD, F(3, 380.) = 8.7, p < .001. In post hoc analyses, the significant effect was found between "emotional problems" and "physical problems" (p < .001) and "other problems" group (p < .001). The highest scores were in emotional problems (M = 88.7, SD = 28.1) and those with more than one problem (M = 80.6, SD = 25.9). No interaction effect was found between these two variables with PTSD, F(3, 380) = .13, p < .938.
A Pearson correlation matrix was used among PTSD, age, incomes, and education. It was shown in the scatterplot matrix ( Figure 2) that PTSD was significantly correlated with these three variables. A weak negative correlation was found between PTSD and incomes, r(687) = −.183, p <.001. Participants with more scores in PTSD have fewer incomes. When years of education increase, the scores in PTSD decrease weakly, r(687) = −.104, p < .007. A weak positive correlation was found between PTSD and the age of participants, r(687) = .121, p < .002. Scores in PTSD scale increase weakly with the age of participants.
A multiple linear regression was conducted to asses whether depression and anxiety predict PTSD. The results were significant, F(2, 686) = 615.5, p < .0001, R 2 = .64, suggesting that depression and anxiety accounted for 64% of the variance in PTSD. Both these factors were accounted further. Depression was a significant predictor for PTSD, with the value of B = 1.3, p < .001, suggesting that for every one unit increase in depression, PTSD increases with 1.3. Anxiety also was a significant predictor (B = 1.1, p < .001) for PTSD.

Discussion
This study explored the prevalence of the PTSD in war veterans in Kosovo. The study has found a prevalence of 11.2% of veterans who had PTSD 8 years upon the end of the war. These data are consistent with other studies that found a prevalence of PTSD in veterans after the end of the war (15.2%, Kulka et al., 1990;10%, Dohrenwend et al., 2006;16%, Hoge, Terhakopian, Castro, Messer, & Engel, 2007;20.9%, O'Toole et al., 1996;12%, Erbes, Westermeyer, Engdahl, & Johnsen, 2007). In addition, this study explored the differences in relation to seeking medical help, self-medication, and their concerns over various issues related to life in general in veterans with positive symptoms versus those with negative symptoms. In the group of participants with positive PTSD symptoms, 6% of them did not seek medical help for their problems. These data are consistent with the study carried out by Vaughan, Schell, Tanielian, Jaycox, and Marshall (2014). But, at the same time, our results are lower than the results obtained by Hoge et al. (2004), which showed that only 23% to 40% of returning veterans who were positive for a mental disorder sought mental health care.
Many participants reported that they had emotional problems and more than one problem and that they did not seek medical help. One of the reasons that could explain this result may be the difficulties with accessing medical services in Kosovo, as there are no institutions specialized for working with veterans, and the relatively high level of stigma. Meanwhile, this leaves room for their self-medication. Our  results are consistent with other studies (Schell & Marshall, 2008), which found that barriers to care included, in rough order, concerns about confidentiality and discrimination, preferences for relying on family and friends, concerns about the effectiveness or side effects of treatments, and logistics.
The study found that low income and low level of education, although at moderate level, are related to the severity of PTSD. Income was also found to be a significant predictor of PTSD. The study has shown that the veterans in Kosovo are a neglected category of the society, because most of them live in less than optimal subsistence conditions, which increases even more their reluctance to seek help for their emotional problems. We assume that this status of theirs is consistent with the results indicating that their biggest concerns are family issues. Persons with PTSD often have difficulties in establishing relationships with family, which may affect both them and their families and even more specifically their children, because according to Dekel and Goldblatt (2008), clinical observations and empirical research have shown that the consequences of traumatic events are not limited to the persons immediately exposed to the event and that they often affect significant others in their environment, such as family members. Herzog, Everson, and Whitworth (2011) found that family members of combat exposed soldiers with high levels of PTSD are at risk for developing secondary traumatic stress. Many studies have reported family problems among veterans, including emotional problems and behavior problems with their spouse and children (Ahmadzadeh & Malekian, 2004;Caselli, 1995;Klarić, Kvesić, Mandić, Petrov, & Frančišković, 2013). According to Braga, Mello, and Fiks (2012), the model of transmitting trauma from one generation to another does not consist of only traumatic experiences but also resilience patterns that can be transmitted to and developed by the second generation.

Conclusion
The study has found that 8 years after the end of the war, veterans in Kosovo suffer from post-traumatic stress symptoms and that a good proportion of them do not seek help for this problem. The economic status appears to affect the retention of PTSD symptoms, by transforming them into a moral challenge for veterans, because of the lack of institutional support for this population with high risk of developing mental health problems. Health policies of a country should foresee specific services for veterans and war survivors, an in particular for the families of veterans, to reduce the stigma and ameliorate the impact of PTSD on their lives. Studies continue to raise a "hopeful possibility that PTSD may be reversible if patients can be helped to cope with stresses in their current life" (Friedman, 2004, p. 76). Therefore, the establishment of such services would transform the veterans' dealing with PTSD from a moral challenge into a fundamental right to equal and high-quality services.