Sense of coherence and depression in the framework of immigration: Turkish patients in Germany and in Turkey

Abstract

The present study explores sense of coherence (SOC) levels in two clinical samples (outpatients with neurotic disorders) with the same Turkish cultural background in comparison to the German reference values as well as the association between SOC and depression and the protective role of SOC. A total of 96 Turkish patients in Germany (36.67 ± 9.52 years) as well as 60 local Turkish patients (38.57 ± 10.15 years) have been examined for SOC measured with the Sense of Coherence scale (SOC-29) and depression with the Beck Depression Inventory (BDI). Both samples scored significantly lower for SOC compared to the normal Turkish and German population and to German subjects with psychiatric symptoms (p < 0.001) but did not differ significantly from each other. Negative significant correlations were found between SOC and the degree of depressiveness in both groups (immigrants: r = −0.59, p < 0.001; Turks: r = −0.51, p < 0.001). Multiple regression analyses including SOC, age, gender, education, marital and employment status have demonstrated SOC to be the strongest predictor for depressiveness. SOC can be regarded as a protective factor for depression in patients with Turkish migration background in Germany and in local Turkish patients. However, further studies are needed to clarify if the concept SOC can be used adequately in collectivistic cultures as, for example, the Turkish one.

Introduction

Over the last decades the focus of scientific interest has changed from the search of potential pathogenic factors and conditions to the identification of aspects with positive impact on health (Antonovsky, 1979). In this context, sense of coherence (SOC) has gained widespread attention in public health research (Eriksson & Lindström, 2006; Suominen et al., 2001; Surtees et al., 2006). As the central construct of Antonovsky's (1987) theory of salutogenesis, SOC represents ‘a person's view of life and capacity to respond to stressful situations’ (Lindström & Eriksson, 2005, p. 441), a tendency towards seeing the world as comprehensible, manageable and meaningful. There is empirical evidence of a strong relationship between perceived health, especially mental health, and SOC in such a way that individuals with a strong SOC report a better health status in general, at least those with an initial high SOC; this protective role of high SOC could be substantiated regardless of sex, age, nationality, ethnicity and study design (Eriksson & Lindström, 2006).

A wide variety of cross-sectional studies in different countries with clinical and non-clinical populations exploring the relationship between the salutogenic construct of SOC and depressive symptoms in adults (Carstens & Spangenberg, 1997; Cederblad et al., 2003; Konttinen et al., 2008; Larsson & Kallenberg, 1996; Matsuura et al., 2003; Stankunas, 2009) as well as adolescents (Blom et al., 2010) have consistently demonstrated a significant negative association between these variables. The stronger the level of SOC the less are the perceived symptoms of depression. Two longitudinal studies (Schnyder et al., 2000) on the stability of SOC and its relationship with depression and anxiety in severely injured accident victims and in patients with rheumatoid arthritis have also indicated significant negative correlations of a moderate to high degree between SOC and levels of depression and anxiety in both samples at three measurement points. A prospective cohort study among Japanese workers found that SOC may be able to predict onset of depression (Sairenchi et al., 2011).

An inverse correlation between SOC and depressive symptoms was also observed in studies with immigrants. A prospective investigation in Sweden on a sample of mass-evacuated adults from Kosovo has detected a strong negative correlation between SOC and depressiveness at three follow-up times; however, SOC could neither predict the diagnosis nor the symptoms of depression (Roth & Ekblad, 2006). A study with 2,234 south-east Asian refugees (Ying et al., 1997) has provided empirical support for the mediating effect of SOC on the association between resistance deficits and resources on the one hand and psychological distress on the other. SOC was found to be the most powerful predictor of depression, anxiety and psychosocial dysfunction. Bayard-Burfield et al. (2001) have shown low SOC to be a risk factor of self-reported longstanding psychiatric illness and for intake of psychotropic drugs in five ethnic minorities (Iranian, Chilean, Turkish, Kurdish and Polish) in Sweden.

Immigration can be regarded as a stressful life event (Grinberg & Grinberg, 1990) to which a pathogenic effect is often ascribed in psychiatric-epidemiologic research (Häfner, 1980). It is associated with various stressors and losses – for example the loss of the accustomed surroundings, social contacts, native language, and quite often the profession and social status. The situation-related factors prior to and during immigration as well as in the accommodating community may vary extremely and influence the mental state of those involved (Bhugra, 2004). Although in a large number of studies the migration status has been shown to be a risk factor for health problems (Bayard-Burfield et al., 2001; Bermejo et al., 2010; Merbach et al., 2008), there are many immigrants who stay healthy despite immigration-associated encumbrances. Due to the fact that immigration has become an integral phenomenon of extending scale in the globalized world it is worth exploring which factors may play a protective role for mental health of immigrants and, in this context, the protective function of SOC. In Germany, individuals with a migration background make up almost one fifth of the population as reported by the Federal Statistical Office (Statistisches Bundesamt, 2010), forming a highly heterogeneous group – among them the Turkish minority as the biggest one.

It is important to point out that most of the studies on SOC and health have been correlational, so it remains open to debate whether a low SOC has some effect on the health status or whether the health affects SOC (Geyer, 1997). Another criticism concerns the discriminant validity of SOC questionnaires. The similarity of the items used in SOC questionnaires and those in depression scales as well as the strong association between SOC and depression suggest that these instruments may measure the same phenomenon, but with inverse signs (Geyer, 1997). However, by means of confirmatory factor analyses Konttinen et al. (2008) have provided support for the assumption that it is possible to differentiate between depressive symptoms, anxiety variables and SOC. Thus, SOC and mental health can be regarded as two independent but correlated constructs (Eriksson & Lindström, 2006). Furthermore, the transcultural applicability of the SOC concept has also been questioned. Antonovsky (1997) claimed SOC to be a universally human way of coping with stress regardless of culture. According to a systematic review (Eriksson & Lindström, 2005) the SOC scale has been used in at least 33 languages in 32 countries on persons from both western individualistic cultures and collectivistic countries such as South Africa, Japan, China and Thailand. Based on their research synthesis including 458 publications and 13 doctoral theses the authors conclude that SOC seems to be a cross-culturally applicable instrument.

The aim of the study was to examine SOC levels in two clinical samples with the same (Turkish) cultural background in comparison to the German reference values and therefore to be able to work out special effects of immigration and culture on SOC. Another purpose of this study was to investigate the association between SOC and depression and the function of SOC as a protective factor in Turkish outpatients with neurotic disorders in Germany and Turkey.

Methods

Research design and strategy

In the present cross-sectional study a two group design was used consisting of Turkish immigrants and local Turkish patients. Quantitative measurement (Rasch et al., 2010) of SOC and depression was applied.

The study was presented to and approved by the Ethics Committee of the Medical Faculty of the University of Duisburg-Essen and was conducted in the outpatient unit of the Department for Psychosomatic Medicine and Psychotherapy. In Germany departments for psychosomatic medicine traditionally focus on the treatment of neurotic disorders, the ICD-10 categories F4 and F5, partially F3. Local Turkish patients were recruited in a corresponding setting, in the outpatient unit for neurotic disorders in Istanbul. Patients of the Turkish-language outpatient clinic were given appointments for an initial interview and a written explanation of the study in which they were also asked to participate; the local Turkish patients (every fourth patient) were recruited by a Turkish-speaking physician (H.A.) after their initial interview. Both patient groups were presented with a Turkish questionnaire consisting of demographic (and immigration-specific) items, SOC-29 and BDI on the day of their initial investigation and after they had agreed to participation in the study. Filling out the questionnaires took approximately 30 to 45 minutes depending on the patient; where indicated, questions pertaining to the individual items were answered by the Turkish-speaking investigator (S.A. or H.A.).

Sample

Between March and April 2005, local Turkish patients of the outpatient unit for neurotic disturbances of the Psychiatric Hospital Bakirköy-Istanbul, and between May and December 2005 patients of the Turkish language outpatient unit of the Department of Psychosomatic Medicine and Psychotherapy in Essen were recruited consecutively. The native language outpatient unit at the Department of Psychosomatic Medicine and Psychotherapy in Essen has provided a bilingual outpatient unit for Turkish immigrants since 1995. This service is being made available to those patients who either do not have sufficient knowledge of the German language to convey their complaints or who wish to speak to someone acquainted with their cultural background. Specific characteristics of this sample compared to German patients of the same outpatient clinic were first published in 2000 (Erim, 2001; Erim-Frodermann et al., 2000). Most of the Turkish probands live permanently in and near Essen. The Psychiatric Hospital in Bakirköy-Istanbul with 3000 beds attends to a large neighbourhood of this city.

Inclusion criteria for the study were age of consent (18 years) as well as agreement to participate in the study; exclusion criteria were the diagnosis psychosis, organic brain disorders, addiction. Of the original 158 questionnaires, two had to be omitted from evaluation because of the diagnosis delusional disorder. The sample, therefore, consisted of 156 probands.

Instruments

Antonovsky's orientation to life questionnaire/SOC scale (SOC-29). The SOC-29 is a self-report questionnaire consisting of 29 items to record the extent of coherence; according to Antonovsky (1987) this represents a basic human orientation consisting of three components. (1) Comprehensibility refers to the extent to which stimuli or events are perceived as being structured, orderly or predictable. (2) Manageability refers to the extent to which a person perceives adequate personal and social resources to manage internal and external demands. (3) Meaningfulness refers to the extent to which a person perceives life to be a source of satisfaction. Representative norm values for the German population are available for the SOC-29 (Hannöver et al., 2004), the norm values being M = 5.34 (Cronbach's alpha = 0.91). A translation of the SOC-29 into Turkish and the validation for Turkish probands were carried out by Uslucan (2010), the norm values for and Turkish version being M = 4.7 (Cronbach's alpha = 0.89).

Beck Depression Inventory (BDI). The Beck Depression Inventory created by Beck (1961) is one of the most widely used self-report inventories for measuring the severity of depression. It consists of 21 items that represent the most important symptoms of depressiveness. Each item has four possible responses ranging from 0 to 3, indicating the severity of the symptom. Scores between 0 und 63 can be achieved. A total score of 18 points and above indicates a clinically relevant depression. In our study a Turkish version of the BDI (Hisli, 1989) was employed (Cronbach's alpha = 0.80).

Statistical analysis

Data calculation was carried out by means of SPSS version 18 (2009). Descriptive statistics, t-tests for a sample and independent samples, non-parametric tests (chi-square test, Fisher's exact test, Mann–Whitney U test), correlations according to Pearson and Spearman as well as linear regression analyses were carried out. A level of significance of p < 0.05 was predetermined.

Results

Demographic and immigration-specific data

Table I contains the demographic and immigration-specific data. The total sample (n = 156) consisted of 96 Turkish immigrant and 60 local Turkish patients, of these 76.9% were women. The average age at the time of the study was 37.40 years (SD = 9.78), the youngest patient was 19, the oldest 71 years. With regard to the gender distribution (χ² (1) = 0.109, p = 0.741) and age (t (154) = −1.183, p = 0.239), the two groups did not differ significantly.

Table I. Demographic and immigration-specific characteristics.

		Total	Turkish immigrant patients	Local Turkish patients
		N = 156	N = 96	N = 60
Age	M	37.40	36.67	38.57
	SD	9.78	9.52	10.15
	Range	19–71	19–71	19–63
Gender	Women	120 (76.9%)	73 (76.0%)	47 (78.3%)
	Men	36 (23.1%)	23 (24.0%)	13 (21.7%)
Marital status	Married	110 (70.5%)	70 (72.9%)	40 (66.7%)
	Single/divorced/widowed	46 (29.5%)	26 (27.1%)	20 (33.3%)
Education
	No education	8 (5.1%)	8 (8.3%)	0
	Primary school (grades 1–5) no certificate	55 (35.3%)	27 (28.1%)	28 (46.7%)
	Middle school (grades 6–8) with certificate	27 (17.3%)	20 (20.8%)	7 (11.7%)
	Secondary/vocational school	63 (40.4%)	41 (42.7%)	25 (41.7%)
Occupation	Household	78 (50.0%)	46 (47.9%)	32 (53.3%)
	Blue-collar worker	52 (33.3%)	37 (38.5%)	15 (25.0%)
	Self-employed	13 (8.3%)	8 (8.3%)	5 (8.3%)
	White-collar worker/ public servant	12 (7.7%)	4 (4.2%)	8 (13.3%)
	Student	1 (0.6%)	1 (1.0%)	0
Employment status	Employed	50 (32.1%)	30 (31.3%)	20 (33.3%)
	Unemployed (unemployed/on sick leave/jobless/pensioner)	106 (67.9%)	66 (68.8%)	40 (66.7%)
Length of residence in Germany	M (in years)	–	14.15	–
	SD	–	10.81	–
	Range		1–35

The majority of the patients were married (70.5%); a large proportion (42.3%) had qualified for higher education; only 32.1% of the patients were employed. With regard to marital status, no significant difference could be found between the two groups investigated (χ² (1) = 0.694, p = 0.405). However, the groups differed significantly with regard to education (Fisher's exact test: p = 0.012): the local Turkish group showed a higher degree of formal education than was the case for the Turkish immigrants. No significant difference could also be found with regard to their vocation (Fisher's exact test: p = 0.123) and employment (χ² (1) = 0.074, p = 0.786).

The average duration of residence of Turkish patients in Germany was 14.15 years (SD = 10.81) at the time of data collection (Table I ).

Sense of coherence

The mean value for the SOC was 3.74 (SD = 0.78) for the Turkish immigrants and 3.87 (SD = 0.60) for the local Turkish patients (Figure I). Both values were comparable (t (144.781) = −1.165, p = 0.246), but they differed highly significantly from the German norm population (M = 5.34; immigrants: t (89) = −19.460, p < 0.001; local Turkish patients: t (59) = −18.900, p < 0.001) as well as from the norm value of the general Turkish population (M = 4.7; immigrants: t (89) = −11.674, p < 0.001; local Turkish patients: t (59) = −10.660, p < 0.001). Compared to German probands, who fulfilled test-psychological psychiatric diagnoses according to the DSM-IV (M = 5.04), both patient groups demonstrated a significantly lower SOC (immigrants: t (89) = −15.811, p < 0.001; local Turkish patients: t (59) = −15.037, p < 0.001). No significant gender differences could be found (immigrants: Z = −0.878, p = 0.380, local Turkish group: Z =−1.248, p = 0.212). Significant differences could neither be found between female immigrants and local Turkish women (t (113.278) =−0.861, p = 0.391) nor between male immigrants and male Turkish locals (Z = −0.727, p = 0.467).

Figure 1. Mean values of sense of coherence: comparison of patients of Turkish origin in Germany, local Turkish patients, norm values for healthy Turkish and German persons and a German sample with mental symptoms.

Correlation between sense of coherence, depressive symptoms, socio-demographic and immigration-specific variables

A highly significant negative correlation between SOC scores and depressive symptoms (total BDI scores) could be demonstrated for both the Turkish immigrants (r = −0.592, p < 0.001) and the Turkish locals (r = −0.512, p < 0.001) (Table II). The correlation between the SOC total score of the immigrants and their educational background, age and length of stay in Germany was not significant; nor did the socio-demographic variables age and education correlate significantly with SOC for the local Turkish patients.

Table II. Correlation coefficients between sense of coherence, depression (BDI), socio-demographic and immigration-specific variables.

Variables	Age	Education	BDI score	Length of residence in Germany
SOC (both groups)	−0.016	0.130	−0.573*	–
SOC (local Turkish patients)	0.091	0.041	−0.512*	–
SOC (Turkishimmigrants)	−0.090	0.193	−0.592*	−0.012

*p ≤ 0.001.

Regression analysis

To examine the influence of SOC as well as socio-demographic and immigration-related variables on the BDI total score, step-wise multiple regression analyses were calculated (Table III). Variables were entered into the model when p < 0.05 and were removed from the model when p > 0.1. A model (model 1) was calculated for both groups; here socio-demographic variables such as gender, age, education, marital and employment status (both dichotomized) as well as group affiliation and SOC as predictors and the total BDI score as criterion variable were included. SOC demonstrated the strongest significant influence (β =−0.547; p < 0.001). A further significant predictor was employment status (β =−0.155; p = 0.027). The explanation of variance of BDI scores was 34.2%. In model 2 (local Turkish patients), SOC proved to be the only significant predictor (β = −0.512; p < 0.001); the explanation of variance was 24.9%. The third regression model (Turkish immigrants) has demonstrated a considerable increase of the explanation of variance (51.5%) with significant predictors being SOC (β = −0.609; p < 0.001) and employment status (β = −0.318; p = 0.003).

Table III. Influences on the total BDI score.

Independent variables	Model 1	Model 2	Model 3
	(Both groups)	(Local Turkish patients)	(Turkish immigrants)
Group	−0.117	–	–
Gender	−0.084	−0.064	0.005
Age	0.074	−0.132	0.157
Education	0.065	0.105	−0.028
Employed (yes/no)	−0.155*	−0.004	−0.318**
Married (yes/no)	−0.027	−0.115	0.053
Length of stay in Germany	–	–	0.161
Sense of coherence	−0.547***	−0.512***	−0.609***
Adjusted R-square	0.342	0.249	0.515
F	37.662***	20.585***	27.561***

*p ≤ 0.05; * *p ≤ 0.01; *** p ≤ 0.001.

The results of the regression analyses demonstrate SOC to be the best predictor for depressiveness according to the BDI – with an increase in SOC, the total BDI score decreases.

Discussion

The main purpose of the present study was to investigate the level of SOC in two clinical populations with the same Turkish cultural background (patients of Turkish origin in Germany and local Turkish patients) in comparison to a reference group of local German patients. Another aim was to explore the association between SOC and depressive symptoms as well as the protective function of SOC. The main results can be summarized under the following three points: (1) The immigrants and the indigenous Turkish patients scored significantly lower for SOC compared to the normal Turkish and German population and to German patients with mental symptoms but did not differ significantly from each other. (2) A strong significant negative association was observed between SOC and depression in both samples. (3) SOC was found to be the strongest predictor for depressiveness in both groups of patients followed by having an employment for immigrants.

The low SOC both in local Turkish patients as well as Turkish immigrants points to a high mental vulnerability in the investigated samples. We postulated SOC to be lower in the sample of the Turkish immigrants than in the local healthy Turkish population, but it was revealed to be also lower than in the German controls that demonstrated psychiatric symptoms. The local Turkish patients also demonstrated a significantly lower SOC compared to the German control sample with mental problems; however, it did not differ from the SOC of the Turkish immigrants. We may assume a culture-specific effect, i.e. the collectivistic orientation of Turkish culture. Turkish patients may attribute skills measured in the SOC questionnaire to their families and not to themselves as individuals and therefore score low for these items. Compared to German patients, Turkish patients, parallelized according to diagnosis, age and gender, demonstrate more pessimistic orientations concerning the course of as well as the external attributions to illness (Franz et al., 2007). One reason for the low SOC could be the pessimistic expectations in Turkish patients (Franz et al., 2007).

One further reason for the low SOC in both Turkish groups could be their poor socio-economic situation: Both groups demonstrated a large percentage of low education level as well as unemployment. Moreover, unemployment was a significant predictor of depression in immigrants. A study on the SOC in asylum seekers from Turkey as well as other countries now living in Germany (n = 63) found a significantly low score in the SOC in refugees with a low education level and unemployment (Bay et al., 2009). Asylum seekers compared to the German norm sample demonstrated a highly significant lower score of SOC; furthermore, the kind of employment (high proportion of blue-collar workers in both samples) may also influence the extent of SOC. A survey in 4000 people in Sweden demonstrated a lower SOC score in blue-collar workers and farmers than in white-collar workers and the self-employed (Lundberg & Nyström, 1994). A positive relation between the level of income and the level of SOC has also been reported (Larsson & Kallenberg, 1996). The disadvantageous socio-economic factors in our sample could, for example, negatively influence successful manageability of life and thus lower the SOC.

Another main result of our study is the strong significant negative correlation between SOC and the degree of depression. This finding is in accordance with many previous cross-sectional (Carstens & Spangenberg, 1997; Cederblad et al., 2003; Larsson & Kallenberg, 1996; Stankunas, 2009) as well as longitudinal studies (Schnyder et al., 2000). Finally, multiple regression analyses have revealed SOC to be the strongest predictor of depressiveness in the present study. This close association between low levels of SOC with high scores of depression is consistent with previous studies (Matsuura et al., 2003; Sairenchi, 2011). However, in a longitudinal investigation with immigrants in Sweden, SOC was neither able to predict the diagnosis nor the symptoms of depression (Roth & Ekblad, 2006). Until now, little has been known about the way SOC influences variables of mental health. A possible mechanism underlying the association between SOC and depressive symptoms could be the stress-buffering effect of a strong SOC. A strong level of SOC might buffer stress in different ways. Amirkhan and Greaves (2003) have found evidence for perceptual and behavioural mechanisms which might explain the positive effect of a high SOC on a person's health. Individuals with a strong SOC seem to perceive stressful events as more benign and thus feel less stressed by them and also cope adequately with stressors using active and non-avoidant strategies. A high level of SOC has been shown to moderate the impact of recent stressful life events on self-reported health (Richardson & Ratner, 2005) and to be associated with fewer reports of stressful events and using more coping strategies (Kenne Sarenmalm et al., 2011).

Conclusions

Sense of coherence can be regarded as a protective factor for depression in patients with Turkish migration background in Germany and in local Turkish patients. Our study results give a strong hint at mental stabilizing properties of employment.

Limitations of the study

The small and selected (two health-care utilization groups) sample precludes generalizations to the population of immigrants in general as well as solely Turkish immigrants in particular. The correlational study design presents a further limitation because it does not allow drawing causal conclusions towards the influence of the measured variables that may be bi-directional. Prospective investigations are required including unselected, randomized samples of the immigrant populations from different cultures and their countries of origin to explore the effects of culture and immigration on the SOC of individuals. Furthermore the influence of the social adaptation in the host country as integration in working life, being a member of social security systems, feeling themselves socially accepted versus being confronted with social discrimination should be investigated. Differences between nationalities may be due to grades of the general social welfare perceived in a society.

Recommendations

Longitudinal (intercultural) studies should investigate whether a low SOC of and Turkish patients may be related to early life events, immigration-specific encumbrances or cultural differences. Further investigations should also be carried out to find out whether the individualistic or collectivistic shaping of the culture has a significant impact on SOC.

Take-home points

Patients from the Turkish (collectivistic) culture should be supplied with interventions which strengthen SOC, for example providing information, psycho-education, parental education for nurturing resilience in their children, etc. In individuals of Turkish origin a strong SOC is associated with lower levels of depression.

Future directions

Future research should focus on the compound structure of SOC including cultural differences (e.g. individualism–collectivism), immigration-specific encumbrances, etc. in larger and representative samples. For immigrants, culture-sensitive psychotherapy and psycho-education methods for nurturing resilience should be developed.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.