Sleep quality among internally displaced Georgian adolescents and population-based controls

Abstract

Sleep problems in children and adolescents are a significant public health concern and may be linked to a variety of psychoemotional difficulties. This study aimed to evaluate sleep quality and associated factors in conflict-affected Georgian adolescents after 9 months of forced displacement. Thirty-three internally displaced adolescents (mean age 11.4 years) and 33 adolescents (mean age 10.8 years) from the general population completed the Epworth Sleepiness Scale and the Children’s Depression Inventory (CDI). Parents completed the Children’s Sleep–Wake Scale and provided information on their socioeconomic status (SES) and the adolescents’ sleep behavior, academic performance, and peer social relationships. The groups differed significantly in sleep quality, peer relationships, SES, and CDI scores. In the internally displaced group, the only significant predictor of sleep quality was SES, which increased the predictive capacity of the model (demographic and psychosocial variables) by 20% in the hierarchical analyses. The most significant predictor in the non-internally displaced group was CDI. This research indicates that displacement may affect sleep quality and psychosocial functioning. The importance of family SES as a contributing factor to displaced adolescents’ poor sleep quality is highlighted. An integrated approach designed to improve the psychosocial environment of internally displaced adolescents is needed for their protection.

Keywords

Adolescent child behavior child protection sleep

Introduction

The current state of knowledge has confirmed that adolescence is a time of extensive maturational brain remodeling. Sleep is one aspect of behavior that changes greatly across adolescence and represents one of the components of the brain maturation process (Carskadon et al., 2004; Darchia and Cervena, 2014; Feinberg and Campbell, 2010). Most evidence supports the notion that insufficient sleep in adolescents affect their learning ability, memory processing, and involvement in social activities, as well as increases their vulnerability to mental illness (Carskadon et al., 2004; Gregory et al., 2008). Furthermore, increasing evidence indicates that sleep problems during childhood and early adolescence are linked to subsequent behavioral problems and emotional dysregulation in late adolescence and adulthood (Gregory et al., 2008).

Several studies have highlighted the importance of exposure to traumatic events in the study of sleep. Stressful conditions result in sleep disturbances that may appear separately or in conjunction with posttraumatic stress disorder, anxiety, and depression (Lavie, 2001). The immediate and long-term effects of traumatic events on sleep have been reviewed in several studies. It has been reported that people subjected to traumatic events, such as escaping from a war zone, exposure to violence, loss of family members, and forced displacements, regularly suffer from sleep disturbances (Basishvili et al., 2012; Germain et al., 2008; Lavie, 2001).

Worldwide, the number of refugees and internally displaced (ID) persons has increased considerably in the past decades and still continues to rise. Scientific evidence indicates that children and adolescents are the most vulnerable population with respect to trauma-related health effects (Caffo et al., 2005; Lustig et al., 2004). Sleep problems and nightmares have been reported in children exposed to war and other types of violence (Montgomery and Foldspang, 2001). A study of war experience and distress symptoms in Bosnian children also revealed a strong correlation between traumatic events and sleep disturbances, sadness, and anxiety (Goldstein et al., 1997).

Despite the growing interest in this field and the increasing number of ID children and adolescents, the impact of stress on sleep and psychobehavioral development in this population has not been systematically documented. Therefore, the evaluation of psychological and sleep problems among ID adolescents has become an important issue. Georgia is one of the countries that has experienced the problem of ID persons. In 1992 in Abkhazia and in 2008 in Shida Kartli (South Ossetia), thousands of people were forced to escape to other parts of Georgia because of war. As a result, thousands of people were affected, including children and adolescents. The prevalence of insomnia in an adult Georgian ID population from Abkhazia has been recently explored (Basishvili et al., 2012). Data on the sleep quality in a young population subjected to forced displacement, however, are scarce. To detect and understand how adolescents respond to traumatic events, it is important to examine sleep and psychological problems in adolescents exposed to war-related forced internal displacement. The purpose of the present study was (1) to compare sleep quality between healthy adolescents in ID families from Shida Kartli and population-based controls after 9 months of forced displacement, and (2) to evaluate predictors of sleep quality in ID and non-ID adolescents. We hypothesize that sleep quality of ID adolescents will be lower than that of the non-ID group, whereas predictors might be the same factors with different strength of association.

Methods

Subjects

ID adolescents were selected from the population that escaped from Shida Kartli, Georgia, in August 2008. All adolescents lived with their families in a new rural community near Tbilisi (the capital). Each family possessed their own house (specially built for the displaced population after the war conflict), and their overall living conditions were also limited but relatively better compared to the other ID population settlements (in former hotels, kinder gardens, etc.) in Georgia. The control group consisted of non-ID adolescents from the general population recruited from five different public schools in Tbilisi.

The inclusion criteria for both groups were good general health, absence of current psychiatric illness, and absence of all types of psychotropic medication.

The first phase of the recruitment process involved advertising in schools (non-ID population) and providing direct information on the study to the displaced population. The second step was the randomized selection of participants who fulfilled the inclusion criteria (in both groups). Finally, 33 ID adolescents (11.4 ± 1.47 years old) and an equal number of adolescents from the general population (non-ID group) of a similar age (10.8 ± 1.29 years) and gender were selected.

Procedures

The study was conducted May-June 2009 (nine months after the displacement). All participants and their parents received detailed information regarding the purpose of the study, methodology, and signed an informed consent form. The adolescents were paid for their participation. Subjects were interviewed at home. In addition to the interview, adolescents and their parents were asked to complete validated self-report questionnaires. The study was approved by the local ethics committee and was conducted in accordance with the Helsinki Declaration.

Measures

Adolescents were interviewed with regard to demographics (age, gender, and parents’ marital status) and were asked to complete the two measures. The Epworth Sleepiness Scale (ESS) for adolescents, a slightly adapted version of the ESS for adults (Johns, 1991), measures daytime sleepiness in children and adolescents. The adapted version focuses on tendency to sleepiness in nine different situations that makes the questionnaire a 27-point scale rather than the original 24-point scale (for details see Campbell et al., 2007). The scale was used to compare average sleepiness ratings between groups. The Children’s Depression Inventory (CDI) is a 27-item inventory designed to assess cognitive, affective, and behavioral signs of depression in children and adolescents (Kovacs, 1992).

Parent interviews included questions regarding the adolescents’ general health status, typical bedtime and wake time during school days and weekends, and sleep latency using a one-month reference period. Weekend oversleep was identified by calculating the difference between the weekend and school night nocturnal sleep duration. Information on academic performance (4-point scale ranging from 1 (bad) to 4 (excellent)), based on the average school records, and the level of peer social relationships (3-point scale ranging from 1 (good) to 3 (bad)), based on the acceptance/friendship with classmates (having multiple friends, reliance on friendship), were also collected.

The most difficult task was the evaluation of the ID families’ socioeconomic status (SES) after nine months of displacement. Most of the ID subjects were highly dissatisfied with their new living conditions and social opportunities in light of their past backgrounds. In addition, education level was not used as an SES indicator. Education qualifications, obtained in Georgia during 1990s, do not accurately reflect actual socioeconomic opportunities (Ministry of Education and Science of Georgia, 2007). Furthermore, Georgia is among the leaders with the unemployed workforce with secondary or higher education levels (Glonti, 2011). The evaluation of SES was based on a single question about the household’s current economic condition (with a particular focus on living standards and income level) and social position (with a particular focus on social well-being/opportunities) using a 3-point scale ranging from 1 (low) to 3 (high). The approach to assess SES based on the part of the main SES indices, especially in the displaced/refugee and immigrant population, have been used in previous research (Gregory et al., 2008; Karunakara et al., 2004; Maisuradze et al., 2010; Stevens et al., 2003).

In addition, parents were asked to complete a 26-item version of the children’s Sleep–Wake Scale (CSWS), which is based on the original 40-item version of the CSWS (LeBourgeois and Harsh, 2001; LeBourgeois et al., 2001). The CSWS measures sleep quality in 2- to 12-year-old children across five behavioral dimensions: going to bed, falling asleep, maintaining sleep, reinitiating sleep, and returning to wakefulness. Using a one-month reference interval, parents reported the frequency of their child’s sleep behaviors using a six-point response set (always, frequently-if not always, often, sometimes, not often, and never). The CSWS provides five subscale scores and an overall sleep quality score. Scores on each of the five CSWS subscales and the CSWS total sleep quality scale (average of subscales) range from 1 (poor sleep quality) to 6 (good sleep quality). Psychometric assessments show that CSWS subscale and total scale scores have good reliability (α = .72 to α = 0.93) for research instruments (LeBourgeois and Harsh, 2001; LeBourgeois et al., 2001, 2013).

The CSWS, ESS, and CDI questionnaires were translated into Georgian by bilingual translators using a standard protocol via translation followed by back-translation. A review of the translated questionnaires and pretesting were performed to ensure that the meaning of the questions remained consistent with the original version.

Statistical analysis

Comparisons of demographics, sleep, and psychological measures were conducted with one-way analysis of variance (ANOVA). Categorical data were analyzed by the χ ² test. Paired t-tests were used to evaluate differences in sleep duration between weekend and school days. The strength of the relationship between sleep quality and other variables was analyzed using Pearson or Spearman correlation analysis as appropriate in the combined groups of adolescents and for each group separately. Hierarchical multiple regression (95% confidence interval) was applied to assess the relationship between sleep quality (CSWS total score) and potential predictors. The significance level for comparisons was set at .05. Results are reported as the mean ± standard error of mean unless otherwise noted. All statistical analyses were performed using PASW Statistics 18 (Chicago, Illinois, USA).

Results

Table 1 compares the demographic and psychosocial characteristics of ID and non-ID adolescents. As shown, groups were significantly different with respect to the level of peer relationships, SES, and the CDI score. Although the average scores for CDI remained within the nonclinical range in both groups, the score was higher in the ID group (p < .001). The SES of ID families was significantly lower than that of non-ID families (p < .001). The level of peer social relationships was significantly higher in the non-ID group, as reported by the parents.

Table 1.

Demographic and psychosocial characteristics of ID and non-ID adolescents.

Variables	ID group (n = 33)	non-ID group (n = 33)	Statistic, p value
Age	11.4 ± .26	10.8 ± .23	F(1, 64) = 3.73, p = .058
Gender—female, % (n)	45.5 (15)	48.5 (16)	χ ² = .061, p = .805
Academic performance^a	2.48 ± .12	2.76 ± .17	F(1, 64) = 1.696, p = .197
Peer social relationships^a	1.45 ± .13	1.15 ± .06	F(1, 64) = 4.324, p = .042*
Family marital status, % (n)
Married	93.9 (31)	78.8 (26)	χ ² = 3.216, p = .073
Divorced	6.1 (2)	21.2 (7)	χ ² = 3.216, p = .073
SES^a	1.52 ± .88	2.15 ± .77	F(1, 64) = 29.523, p = .000***
CDI	9.64 ± .66	5.30 ± .67	F(1, 64) = 20.992, p = .000***

Note: SES: socioeconomic status; CDI: Child Depression Inventory; ID: internally displaced.

^aThose ordinal variables were treated as continuous variables ranging from 1 (good) to 3 (bad) for peer relationships, 1(bad) to 4 (excellent) for academic performance, and 1 (low) to 3 (high) for SES.

*p < .05; ***p < .001.

The mean scores for sleep variables are presented in Table 2. Statistical analysis revealed significant group differences in all the CSWS dimensions except the returning to wakefulness subscale, with non-ID adolescents presenting higher scores. Overall sleep quality score was also higher in the non-ID group (p < .01), as hypothesized. Regarding the CSWS subscales, the most significant difference was detected in the falling asleep subscale (p < .01). Accordingly, the sleep onset latency was significantly higher in the ID adolescents (p < .01). The nocturnal sleep duration on school nights as well as on weekend nights was nearly identical in ID and non-ID subjects, with relatively higher values found in the ID group. Therefore, the difference in the weekend oversleep between groups did not reach statistical significance. However, within-group differences between sleep time on week nights and weekend nights were significant for each group (p < .001, paired t-test). Average sleepiness ratings did not differ between groups.

Table 2.

Mean scores for CSWS subscales and other sleep variables in ID and non-ID adolescents.

Variables	ID group (n = 33)	Non-ID group (n = 33)	Statistic, p value
CSWS 1—going to bed subscale	3.58 ± .19	4.12 ± .19	F(1, 64) = 4.073, p = .048*
CSWS 2—falling asleep subscale	4.18 ± .15	4.70 ± .12	F(1, 64) = 7.737, p = .007**
CSWS 3—maintaining asleep subscale	4.63 ± .13	5.04 ± .11	F(1, 64) = 5.550, p = .022*
CSWS 4—reinitiating asleep subscale	4.82 ± .15	5.30 ± .11	F(1, 64) = 6.523, p = .013*
CSWS 5—returning to wakefulness subscale	4.07 ± .15	4.35 ± .14	F(1, 64) = 1.839, p = .180
CSWS total sleep quality score	4.26 ± .10	4.70 ± .08	F(1, 64) = 10.237, p = .002**
ESS score	4.24 ± .45	4.27 ± .54	F(1, 64) = .002, p = .966
Sleep latency (minutes)	27.88 ± 3.46	17.27 ± 1.32	F(1, 64) = 8.161, p = .006**
Nocturnal sleep duration on school days (hours)	9.38 ± .19	9.56 ± .22	F(1, 64) = .386, p = .537
Nocturnal sleep duration on weekend days (hours)	10.42 ± .19	10.79 ± .14	F(1, 64) = 2.425, p = .124
Weekend oversleep (hours)	1.04 ± .15	1.23 ± .14	F(1, 64) = .817, p = .369

Note: ESS: Epworth Sleepiness Scale; CSWS: children’s Sleep–Wake Scale; ID: internally displaced.

*p < .05; **p < .01.

Table 3 illustrates the association between sleep quality and demographic and psychosocial variables. Analyses of all adolescents (both samples combined) demonstrated that SES and the CDI score had the most significant bivariate correlation with the sleep quality index in adolescents. However, separate analyses in each cohort showed that SES was no longer correlated with sleep quality in non-ID adolescents. In addition, analyses of each sleep quality subscale showed that the strength of the relationship between SES and sleep quality differed across samples. SES was the only variable significantly correlated with all dimensions in the ID group and only with falling asleep (r = .39, p < .05) in the non-ID adolescents.

Table 3.

Correlation coefficients between sleep quality and demographics and psychosocial characteristics of adolescents.

Variables	CSWS total score all subjects (n = 66)		CSWS total score ID group (n = 33)		CSWS total score non-ID group (n = 33)
Variables	r	p	r	p	r	p
Group^a	.37	.02*
Age^b	−.13	.30	.20	.91	−.18	.33
Gender^a	−.13	.31	.58	.75	−.39	.024*
Academic performance^a	.37	.002**	.49	.004**	.25	.17
Peer social relationships^a	−.28	.024*	−.36	.039*	−.19	.45
Family marital status^a	.12	.32	.13	.94	.01	.95
CDI^b	−.57	.000***	−.39	.025*	−.47	.006**
SES^a	.62	.000***	.73	.000***	.31	.08

Note: SES: socioeconomic status; CDI: Child Depression Inventory; ID: internally displaced.

^aSpearman correlation.

^bPearson correlation.

*p < .05; **p < .01; ***p < .001.

To identify predictors of sleep quality and find out whether SES could account for the group differences in sleep quality, hierarchical multiple regression analyses were performed. Because of the relatively small sample size, the variables were limited to those who showed a significant correlation with sleep quality in either group of adolescents. We found that demographic and psychosocial variables (gender, CDI, social relationship, and academic performance) accounted for 39% of the variance in sleep quality (R ² = .393, F(4, 28) = 4.538, p < .01) in the ID group and 40% of the variance in the non-ID group (R ² = .404, F(4, 28) = 4.753, p < .01). The SES score was added in step 2. The total variance in the CSWS score explained by the model was significantly increased. However, in the non-ID group, SES explained only an additional 4.5% of the variance in the total CSWS score (R ² = .449, F(5, 27) = 4.405, p < .01), whereas it explained an additional 20% in the ID group (R ² = .593, F(5, 27) = 7.854, p < .001). When all variables were entered in step 2, the only significant predictor of sleep quality in the ID group was SES (β = .576, p = .001); In the non-ID group, the significant predictors were CDI (β = −.571, p = .02) and gender (β = −.329, p = .035).

Furthermore, hierarchical regression analyses of all subjects indicated that the displacement status accounted for only 0.6% of the variance in predicting sleep quality after controlling for all other variables (R ² change = .006, F(5, 60) = 10.550, p < .001).

Discussion

War-related health problems and their consequences are particularly important when studying children and adolescents (Caffo et al., 2005; Lavie, 2001). A review of the mental health literature confirms that child and adolescent refugees are the group most vulnerable to suffering from significant conflict-related exposures (Lustig et al., 2004). In the same way, displacement is an important topic that impacts many regions around the world. Displacement leads to a remarkable decline in living conditions, changes the individual’s habitual activities in both social and professional contexts, and can negatively influence health in numerous ways, especially in the developing countries.

Despite this worldwide problem, sleep in conflict-affected populations subjected to forced displacement has received little attention. The present study provides supporting evidence of the negative impact of forced displacement on adolescents’ sleep and psychosocial functioning. Although the overall sleep quality level (the global CSWS score) was adequate to good in both groups, the total sleep quality and all the sleep quality subscale scores, except returning to wakefulness, were significantly lower in the ID group compared with the non-ID group. Considering that we evaluated healthy ID children who had not been personally affected by war-related violence, the difference found between the ID and non-ID subjects only nine months after displacement warrants attention.

We found that sleep quality was significantly correlated with many variables, including the depression score, SES, academic performance, and social relationships. However, when ID and non-ID subjects were separately analyzed, SES was a significant correlate in the ID group only. Regression analyses also showed that the only significant predictor of sleep quality in ID adolescents was SES that accounted for the difference in sleep quality between the groups. These data provide some evidence that a family’s SES during the resettlement phase influences children’s adjustment. It has been reported that the family environment is particularly important for the development of childhood sleep disturbances in response to war-related traumatic events (Montgomery and Foldspang, 2001). The study of Croatian refugees showed that parental well-being plays an important role in stress reactions among resettled refugees. There was an inverse relationship between the stress levels in Croatian children and their mother’s coping abilities (Ajduković and Ajduković, 2014). Evidence indicates that parental well-being affects child functioning (Lustig et al., 2004) and that SES is linked to a wide array of health, cognitive, and socioemotional outcomes in children (Bradley and Corwyn, 2002). The fact that SES was the strongest predictor of sleep quality in ID adolescents which increased the model’s (demographic and psychosocial variables) predictive capacity in the hierarchical analyses by 20% further indicates that living and social conditions are particularly important for displaced individuals and may negatively affect children’s sleep quality and associated health consequences. Moreover, the finding that the displacement status accounted for only 0.6% of the variance in predicting sleep quality after controlling for all other variables suggests that SES could be a reliable indicator of displacement-related consequences that impact sleep problems in ID children.

Several studies have reported that conflict exposure and resettlement lead to mental health problems in children and adolescents. Refugee children exhibit emotional symptoms, such as anxiety, depression, and aggression (Abdalla & Elklit, 2001; Lustig et al., 2004). Children’s responses to major stress are similar to those of adults (Caffo et al., 2005). Our data also support the idea that displacement during childhood affects psychosocial functioning even in the absence of clinically significant psychiatric disturbance. Despite the fact that only healthy adolescents were included in our study, the CDI score was significantly higher in ID adolescents. The level of peer relationships was also significantly lower in the ID group. However, academic performance in ID group was as good as in the non-ID group. We presume that the recovery power of children plays a significant role in helping them cope with difficulties associated with displacement. The results of our study suggest that this ability is more likely to be directed toward successful academic functioning in this age group than toward psychosocial functioning or sleep behavior.

Advantages and limitations of the study

The major limitations of our study are the small number of subjects and the parent-reported data on sleep (which are often used in this young age-group) that may not accurately reflect children’s sleeping pattern. Another limitation is the lack of assessment of previous sleep problems, individual psychological factors (e.g. temperament and resilience), and a family history of sleep problems in both populations. Finally, findings about the primary importance of SES for sleep quality in ID adolescents raise the possibility that SES appraisal by ID population with a single question was less precise and that more detailed/thoughtful assessment is needed for further studies. However, the careful interview of the included subjects limits the influence of these factors on the final results.

Despite these limitations, to the best of our knowledge, our study presents the first data on sleep quality in an ID adolescent population in Georgia. Additional studies with more subjects are necessary to better understand the nature of sleep disturbances and how family status and environmental factors influence sleep and well-being in displaced individuals.

The results of this study suggest that, even in healthy ID adolescents, lower sleep quality and worse psychosocial functioning occur. Because such problems are likely to have further consequences for health and behavior, interventions that target various levels of resettlement should be evaluated and developed. An integrated approach designed to improve the psychosocial environment for ID adolescents is necessary for their protection.

Footnotes

Acknowledgments

We thank the subjects and their parents who participated in the study for their cooperation.

Conflict of interest

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 work was supported by the Georgian National Science Foundation [grant number GNSF/ST07/6-237] and Swiss National Science Foundation, program SCOPES [grant number IZ74Z0_137415].

References

Abdalla

Elklit

(2001) A nationwide screening survey of refugee children from Kosovo. Torture 11: 45–49.

Ajduković

(2014) Psychological well-being of refugee children. Child Abuse & Neglect 17(6): 843–854.

Basishvili

Eliozishvili

Maisuradze

. (2012) Insomnia in a displaced population is related to War-associated remembered stress. Stress and Health 28(3): 186–192.

Bradley

Corwyn

(2002) Socioeconomic status and child development. Annual Review of Psychology 53: 371–399.

Caffo

Forresi

Strik Lievers

(2005) Impact, psychological sequelae and management of trauma affecting children and adolescents. Current Opinion in Psychiatry 18(4): 422–428.

Campbell

Higgins

Trinidad

. (2007) The increase in longitudinally measured sleepiness across adolescence is related to the maturational decline in low-frequency EEG power. Sleep 30(12): 1677–1687.

Carskadon

Acebo

Jenni

(2004) Regulation of adolescent sleep: implications for behavior. Annals of the New York Academy of Sciences 1021: 276–291.

Darchia

Cervena

(2014) The journey through the world of adolescent sleep. Reviews in the Neurosciences 25(4): 585–604.

Feinberg

Campbell

(2010) Sleep EEG changes during adolescence: an index of a fundamental brain reorganization. Brain and Cognition 72(1): 56–65.

10.

Germain

Buysse

Nofzinger

(2008) Sleep-specific mechanisms underlying posttraumatic stress disorder: Integrative review and neurobiological hypotheses. Sleep Medicine Reviews 12(3): 185–195.

11.

Glonti

(2011) Employment and Unemployment Trends in Georgia. Democracy @ Freedom Watch, 15 November, 2011.

12.

Goldstein

Wampler

Wise

(1997) War experiences and distress symptoms of Bosnian children. Pediatrics 100(5): 873–878.

13.

Gregory

Van der Ende

Willis

. (2008) Parent-reported sleep problems during development and self-reported anxiety/depression, attention problems, and aggressive behavior later in life. Archives of Pediatrics and Adolescent Medicine 162(4): 330–335.

14.

Johns

(1991) A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep 14(6): 540–545.

15.

Karunakara

Neuner

Schauer

. (2004) Traumatic events and symptoms of post-traumatic stress disorder amongst Sudanese nationals, refugees and Ugandans in the West Nile. African Health Sciences 4(2): 83–93.

16.

Kovacs

(1992) Children’s Depression Inventory (CDI): Technical Manual. North Tonawanda: Multi-Health Systems, Inc.

17.

Lavie

(2001) Sleep disturbances in the wake of traumatic events. New England Journal of Medicine 345(25): 1825–1832.

18.

LeBourgeois

Harsh

(2001) A new research measure for children’s sleep. Sleep 24: A213–A214.

19.

LeBourgeois

Hancock

Harsh

(2001) Validation of the children’s sleep-wake scale (CSWS). Sleep 24: A218–A219.

20.

LeBourgeois

Wright

Jr LeBourgeois

. (2013) Dissonance between parent-selected bedtimes and young children’s circadian physiology influences nighttime settling difficulties. Mind, Brain, and Education 7(4): 234–242.

21.

Lustig

Kia-Keating

Knight

. (2004) Review of child and adolescent refugee mental health. Journal of the American Academy of Child and Adolescent Psychiatry 43(1): 24–36.

22.

Maisuradze

Lortkipanidze

Eliozishvili

. (2010) Insomnia and Posttraumatic Stress Disorder Development in Individuals Escaped from Shida Kartli, Georgia. Georgian Medical News 3(180): 64–69.

23.

Ministry of Education and Science of Georgia (2007) Short Country Survey Findings—Georgia. Survey for the BFUG Employability Working Group. Available at: http://www.ond.vlaanderen.be/hogeronderwijs/bologna/actionlines/documents/Employability_Georgia.pdf (accessed 28 April 2015).

24.

Montgomery

Foldspang

(2001) Traumatic experience and sleep disturbance in refugee children from the Middle East. European Journal of Public Health 11(1): 18–22.

25.

Stevens

Pels

Bengi-Arslan

. (2003) Parent, teacher and self-reported problem behavior in The Netherlands: comparing Moroccan immigrant with Dutch and with Turkish immigrant children and adolescents. Social Psychiatry and Psychiatric Epidemiology 38: 576–585.