1. Introduction
By 2019, 6.6 million people had fled Syria due to the crisis that has lasted for nine years and that is still continuing [
1]. Refugees are often affected by loss of status, income, home and social network as well as interpersonal losses [
2,
3]. In a population-based sample of Syrian refugees resettled in Sweden, the loss or disappearance of family members or loved ones was reported by 64%, forced separation from family or close friends were reported by 68% [
3]. There are representative numbers showing that family separation is common among refugees in Germany [
4]. A link between separation from marital partners and lower quality of life is reported in a study of Syrian refugees in Germany [
5]. In addition, some affected persons describe the separation from the family as “traumatizing” [
6]. However, to our knowledge, there are no reliable figures on how many refugees in Germany have missing relatives. The International Organization for Migration’s Missing Migrants Project recorded more than 30,000 border deaths globally, whilst the numbers of disappeared or missing migrants and refugees reach into the thousands. This however only includes fatalities at international borders and explicitly excludes a number of instances (such as deaths and disappearances) that occurred in the country of origin, after deportation or within refugee camps. Thus, the numbers cannot serve as overall information on how many migrants and refugees have disappeared [
7].
Ambiguous loss (AL) is a sparsely researched topic in the field of refugees. It describes the phenomenon of loss without confirmation and thus potentially without closure [
8,
9,
10]. The definition of AL is often broad, from incidents ranging from diseases like Alzheimer’s or addiction (physical presence with psychological absence) to disappearances during war and terrorism, incarceration, vanishing at sea or migration (physical absence with psychological presence [
10,
11]). Additionally, separation from family can be understood as AL when those persons are at permanent risk, e.g., when family members remain in Syria or are separated during the flight [
12,
13]. Utržan and Northwood [
14] discussed the asylum process itself as a form of AL, since family members are physically absent and also psychologically not available due to uncertainty, fear and often guilt at not being able to support family members in severe situations. In our paper, we define ambiguous loss as the physical disappearance of a significant other during war, flight or resettlement without knowing whether the person is still alive. We focus on AL experienced by Syrian refugees based in Germany, since refugees from war zones often have missing family members or friends without having confirmation of their death [
15,
16].
As a consequence of AL, the lack of confirmation of death is often accompanied by the absence of rituals that may be important for the grieving process (e.g., funeral) [
17]. Economic, social or legal problems can correlate with the missing of family members who have provided for family income [
18]. Avoidant behavior may be encouraged in families and individuals due to fear of discovering the truth about the whereabouts of their loved ones [
19]. On an emotional level, AL can be accompanied by guilt and helplessness [
20]. Furthermore, the lack of clarity about the whereabouts of family members may complicate emotional or practical decisions that need to be taken for the future. This goes along with the phenomenon of not knowing who is in or out of a certain system (e.g., core family), which is described as boundary ambiguity [
9]. Boundary ambiguity can be a consequence of AL, is stated to confuse family roles, determine the family’s level of stress and was associated with depression and anxiety earlier [
21,
22,
23,
24].
Research on the link between AL and mental health is still limited. AL was associated with prolonged grief, anxiety, depression, post-traumatic stress disorder (PTSD), substance abuse and stress-related illnesses and hence was considered a risk factor for mental health [
8,
9,
10,
12,
20,
25]. Boss [
10] distinguished the experience of AL from PTSD in the sense that the trauma of uncertainty about loss persists. A systematic review by Lenferink, Keijser, Wessel, Vries and Boelen [
26] that investigated disappearances due to war or state terrorism reported widely varying prevalence rates in a small number of studies assessing PTSD (1–67%), depression (3–88%), anxiety (1–65%), prolonged grief (7–23%) and somatic complaints (43%) in persons experiencing AL. Moreover, the authors did not find a statistically significant difference regarding psychopathology between relatives of disappeared persons and relatives of homicide victims. In contrast, Isuru et al. [
12] found higher prevalence rates of depression and prolonged grief in persons who did not receive the mortal remains of the disappeared person in comparison to those who did after a tsunami in Sri Lanka. A study with relatives of persons who disappeared or were killed during the war in Bosnia–Herzegovina showed higher depression and traumatic grief scores in relatives of disappeared persons [
19]. Furthermore, forced separation, loss or disappearance of family members or loved ones and the lack of current information about family members were predictors for mental health in refugees in Europe [
3,
27]. Other studies highlight that the disappearance of a significant other does not go hand in hand with psychopathology and that there are different ways of coping. For example, a study with relatives of disappeared persons in the Netherlands showed that acceptance of the disappearance, mental disengagement, emotional social support and venting emotions with others were helpful coping strategies [
28].
Previous studies found mixed results regarding predictors for mental distress in persons with AL. Individual studies found female sex and higher age in persons experiencing AL to predict anxiety and depression but not prolonged grief or PTSD [
12,
29,
30,
31]. The relation to the missing person is one of the better-evidenced predictors of mental distress in relatives of disappeared persons. Thereby, spouses and parents seemed to be most affected by AL [
12,
25,
26,
29,
32]. In addition, mental distress seemed to be related to a higher exposure to traumatic events (TE) [
19,
33]. On an individual level, emotion-focused coping strategies were associated with higher levels of anxiety, depression and stress in persons with AL compared to problem-focused strategies [
27]. Regarding family strategies after AL, cooperative problem solving and being in control of dealing with adverse life events were associated with fewer avoidant PTSD symptoms and lower prolonged grief [
29]. Ambivalence regarding whether the disappeared person died and a higher extent of hope that the person was alive showed higher prolonged grief [
12,
31]. Concurrently, Wayland, Maple, McKay, and Glassock [
34] postulated hope as useful for dealing with AL. Lenferink et al. [
26] formulated the hypothesis that violent disappearances might be associated with higher levels of grief compared to non-violent disappearances.
The field of AL is under-researched, studies have shown contradictory findings and there is a call for more insights into correlates of psychopathology in relatives of disappeared persons [
10,
26,
31]. To our knowledge, though refugees are often exposed to AL, there are hardly any studies addressing AL in refugee populations. There is a need to identify important factors that influence mental health and to derive clinical implications to enhance well-being of those affected. Hence, the aim of this study was to identify predictors for prolonged grief, anxiety, depression, PTSD and somatization in Syrian refugees with post-traumatic stress symptoms (PTSS) in Germany who experienced AL.
4. Discussion
The aim of this study was to identify predictors for prolonged grief, anxiety, depression, PTSD and somatization in Syrian refugees with post-traumatic stress symptoms in Germany experiencing AL. Our results indicate that boundary ambiguity and missing family members are predictors for prolonged grief. The overall model for somatization was statistically significant while no predictor independently reached statistical significance. Boundary ambiguity showed a statistically significant positive association with depression, while the overall model showed no statistically significant associations.
In attempting to interpret the results of the analyses, it is first necessary to point out some specific characteristics of the sample. On one hand, sample size was small and hence may have led to an underestimation of the impact of individual variables or the overall models. It is possible that small effects (e.g., on anxiety, depression, PTSD or somatization) could not show up. On the other hand, due to the need in the Sanadak trial to include only individuals with at least mild PTSS, we cannot make any statements about Syrian refugees without PTSS. The exclusion of persons with severe depression and PTSD may have led to an underestimation of the predictors’ impact on the outcomes. The very specific inclusion- and exclusion-criteria of the study make the predictive value of our findings high for the very specific group of Syrian refugees with PTSS experiencing AL, however the findings may not be generalizable to other refugee populations. Additionally, it should be noted that while boundary ambiguity has been the subject of multidisciplinary research for over 40 years, the definition of the concept has been very broad and varied. In contrast, the selection of valid measurement instruments is very limited. The most commonly used measurement instrument is the BAS, which has been adapted many times. In the literature there is a call for a standardized version with good psychometric properties [
22]. In addition to these difficulties in capturing the concept of boundary ambiguity, no version is adapted to the realities of refugees’ lives after experiencing war and flight, so we had to adapt the instrument. A validation of the instrument would have been desirable, however, the study lacked adequate measures due to the fact that we used baseline data of a RCT that primarily aimed at evaluating the efficacy of a self-help app. Therefore, results must be interpreted with caution.
To our knowledge, our study is the first to report boundary ambiguity as a relevant predictor for prolonged grief in Syrian refugees with PTSS in Germany. Heeke et al. [
31] reported the degree of hope for survival of the missing person to be associated with prolonged grief but not with depression earlier. Isuru et al. [
12] found not being sure whether a person was dead or alive and the belief that the missing person was still alive to predict higher prolonged grief. Since the extent of hope and the belief whether a person is still alive are both components of the concept of boundary ambiguity [
21], our results support and connect these findings. However, Isuru et al. [
12] also found an association between the belief of the status of the missing person and depression. Since the overall depression model in our study was not statistically significant, we could not replicate this result.
Regarding the relationship to the missing person, missing persons being a close family member (i.e., partner, child, parent, sibling) as compared to other relatives (i.e., grandparent, parent-in-law, others) predicted higher prolonged grief in this study while the missing person being a friend did not. This is an important finding since most previous research was on family members only or did not distinguish between family and friends [
12,
25,
26,
28]. Furthermore, the finding points in the same direction as a study by Georgiadou et al. [
5], which showed that family separation is associated with lower quality of life among Syrian refugees. Interestingly, the closeness felt to the missing person did not predict higher prolonged grief in our sample. One possible hypothesis could be that these findings illustrate that organizational and legal needs and problems, which are greater when a member of the nuclear family is missing, are more likely to contribute to psychological distress than emotional closeness to the missing person.
In contrast to previous research [
12,
29,
30,
31], we did not find female sex and older age to predict anxiety and depression. Analogous to these previous results, it was also no predicting factor for PTSD or prolonged grief in our study. In line with Heeke et al. [
31], the educational level was unrelated to prolonged grief. It is noteworthy that we did not find sociodemographic factors as predictors for mental distress in our sample at all. Since the effect of sociodemographic factors on mental health (e.g., prolonged grief) is well studied [
46], the question arises as to the discrepancy with our research. One explanation might once again be the small sample size so that minor differences could not show in the regression analysis. The small to mediate correlations between sociodemographic variables and mental health outcomes in the bivariate analysis support this hypothesis. Further, it is possible that men were more frequently involved in military operations in Syria and thus potentially traumatic events than women, whereby the higher burden on women caused by gender-related factors would no longer be visible due to a war-based higher burden on men. However, the reports on wartime burdens specific to women speak against this hypothesis. By analogy, it can be hypothesized that younger persons were more likely to have been actively involved in the war, thus leveling out increasing age as a risk factor. Another explanation could also be the specific help-seeking sample, as only individuals who were interested in an online mental health support intervention responded. One assumption would be that rather “young at heart” persons are interested in such an innovative form of intervention. Regarding socioeconomic status, it could be hypothesized that the high overall burden in the help-seeking sample levels out the protective function of a higher income.
4.1. Strengths and Limitations
This study takes an important step by examining the still largely unexplored field of AL in the under-researched group of refugees with PTSS. The highly specific target group of Syrian refugees in Germany with PTSS experiencing AL allows the derivation of concrete implications for supporting a vulnerable subgroup of the largest group of refugees in Germany since 2013 [
46]. The study may also contribute to a better understanding of the role that boundary ambiguity plays in prolonged grief, which may provide valuable information for clinical practice. Additionally, we adapted the original version of the boundary ambiguity measure by Boss et al. [
21] to better serve our target group.
As discussed above, the most relevant limitation of our study is the small sample size, which may have led to an underestimation of the impact of included variables and did not allow including further possibly important factors in the analyses. Already the existing number of predictors could be quite high for the regression models with our small sample size. In addition, the exclusion of severely depressed persons and the inclusion criterion of having experienced at least one traumatic event may have led to bias. Further sources of bias could be the cross-sectional nature of our data and the use of self-report questionnaires. The sample consists of help-seeking individuals and is not representative for the general population of refugees based in Germany; therefore, generalization of the results is not appropriate. Although our measure of boundary ambiguity was adapted to the target population, it still lacks proper evaluation of test criteria, was not validated for the target group and the adaption process could use expert-based refinement. Additionally we would like to point out that the cultural, political, and social context of the authors has most likely influenced the research question as well as the interpretation of the results.
4.2. Implications
Since having lost a family member and boundary ambiguity are risk factors for prolonged grief in this study, we suggest supporting the search for missing relatives at political and organizational levels, for example through tracing services and reunification programs (e.g., by the German Red Cross). Additionally, psychosocial support groups including psychoeducation on boundary ambiguity and mental health could help to establish and expand coping mechanisms for dealing with ambiguous loss and strengthen social support structures. A systemic approach for the whole family system dealing with AL could be helpful, since prolonged grief is also defined as a type of AL [
11] and can hence have adverse effects on the mental health of other family members.
4.3. Further Research
In our help-seeking sample of Syrian refugees with PTSS in Germany, 35.6% (
n = 47) of a total of
N = 133 persons experienced AL. When interpreting this data, it should be kept in mind, that the sample is not representative and may be affected by a help-seeking bias. Therefore, future studies with representative samples are necessary to shed more light on the prevalence of AL among Syrian refugees. Moreover, further research should be carried out with a greater sample size to also enable detection of smaller effects of loss-related variables on anxiety, depression, PTSD, and somatization. A comparison between refugees with and without symptoms of post-traumatic stress regarding their boundary ambiguity after AL would be interesting to better understand the association between boundary ambiguity and psychopathology and to make results more generalizable. Further research should also include more potential predictors for psychopathology such as the time since, or the type of disappearance, but also potential resources such as social support or religiosity. Given that the whereabouts of many family members of Syrian refugees is unclear, family separation and ambiguous loss are probably closely connected in this population. Family separation was described as important mental health stressor before [
5,
6], hence, an investigation of the relationship between family separation and boundary ambiguity could contribute to a better understanding of AL in refugee populations. A group comparison between refugees who have experienced confirmed loss and ambiguous loss would be a further step towards understanding the impact of AL on mental health. Additionally, validation of the adapted version of the boundary ambiguity measure would be desirable.