Ethnicity & Health

This article was downloaded by:[Karolinska Institute] On: 15 August 2007 Access Details: [subscription number 779857346] Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Ethnicity & Health Publication details, including instructions for authors and subscription information: http://www.informaworld.com/smpp/title~content=t713421971

Self-Reported Anxiety, Sleeping Problems and Pain Among Turkish-Born Immigrants in Sweden Online Publication Date: 01 September 2007 To cite this Article: Steiner, Kristin Hjörleifsdottir, Johansson, Sven-Erik, Sundquist, Jan and Wändell, Per E. (2007) 'Self-Reported Anxiety, Sleeping Problems and Pain Among Turkish-Born Immigrants in Sweden', Ethnicity & Health, 12:4, 363 - 379 To link to this article: DOI: 10.1080/13557850701300673 URL: http://dx.doi.org/10.1080/13557850701300673

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf This article maybe used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material. © Taylor and Francis 2007

Downloaded By: [Karolinska Institute] At: 15:03 15 August 2007

Ethnicity and Health Vol. 12, No. 4, September 2007, pp. 363 379

Self-Reported Anxiety, Sleeping Problems and Pain Among Turkish-Born Immigrants in Sweden Kristin Hjo¨rleifsdottir Steiner, Sven-Erik Johansson, Jan Sundquist & Per E. Wa¨ndell

Objectives. To study whether symptoms of self-reported anxiety, sleeping problems and severe pain are more common among Turkish-born immigrants in Sweden than among Swedes, and whether age and socio-economic status can explain this hypothesised difference. Design. Two random samples were studied*the Swedish National Board of Health and Welfare Immigrant Survey, and the Swedish Annual Level-of-Living Survey, both from 1996. A total of 526 Turkish-born immigrants in Sweden were compared with 2,854 Swedish controls, all aged between 27 and 60 years. Data were analysed by sex, in an age-adjusted model; and a full model also included age, education, marital status, employment and country of origin (logistic regression). Results. In the full model, odds ratios were 2.12 (1.433.15) for anxiety, 2.60 (1.823.72) for sleeping problems, and 2.14 (1.503.05) for severe pain among Turkish-born men, and 2.44 (1.693.53) for anxiety, 3.01 (2.094.33) for sleeping problems, and 2.59 (1.803.71) for severe pain among Turkish-born women, using the Swedish controls as references. Conclusions. Being a Turkish-born immigrant in Sweden significantly increases the risks for self-reported anxiety, sleeping problems and severe pain, even after adjusting for age and socio-economic status (education, marital status and employment). Keywords: Turkish-Born; Immigrants; Sweden; Self-Reported; Anxiety; Sleeping Problems; Pain Correspondence to: Kristin Hjo¨rleifsdottir Steiner, Center for Family and Community Medicine, Karolinska Institutet, Alfred Nobels alle´ 12, 141 83 Huddinge, Sweden. Tel.: 46705406529; Email: kristin_ [email protected] ISSN 1355-7858 (print)/ISSN 1465-3419 (online) # 2007 Taylor & Francis DOI: 10.1080/13557850701300673

364 K. H. Steiner et al.


Introduction Although international migration has increased during the recent decades, and the interest in migration effects on health has resulted in many studies, there is still limited knowledge on estimated anxiety, sleeping problems and pain in immigrants, based on random sampling. In the Swedish population, the frequency of foreign-born people is now 12.0% (Official Statistics of Sweden 2005). Immigrants in Sweden are a very heterogeneous group. While some immigrants originate from Nordic and other Western countries, with a similar background to Swedish-born people, a large proportion of immigrants originate from all continents, with various socio-economic and cultural backgrounds, and different reasons for migration. One of the largest immigrant groups in Sweden is the Turkish-born group, who came to Sweden in the 1960s and 1970s, mainly as labour migrants, and in the 1980s because of family ties or as refugees. Today, there are 34,000 Turkish-born immigrants in Sweden (Official Statistics of Sweden 2005), many with a low attained educational level, a well-known risk factor for poor mental health. Immigrants adapting and integrating to a new country, culturally different from the country of birth, could be faced with difficulties to be acculturated, which in turn could add risk factors to poor mental health. The increased risk of poor mental health among immigrants has been confirmed by several previous studies (Sundquist 1993; Rosmond et al. 1996; Ritsner & Ponizovsky 1999; Soares & Grossi 1999; Bayard-Burfield et al. 2001). Associations between psychological stress/distress and pain have also been shown (Jorgensen et al. 2000; Okifuji & Turk 2002). Furthermore, pain behaviour (Lofvander & Furhoff 2002), as well as somatisation (Diefenbacher & Heim 1994; Kirmayer & Young 1998; Li & Browne 2000; Ritsner et al. 2000) is common among immigrants. Psychological distress could be one underlying cause (Soares & Grossi 1999; Baarnhielm & Ekblad 2000; Sabbioni & Eugster 2001; Lofvander & Furhoff 2002). Psychological distress among immigrants in Sweden is also related to sleeping problems. For example, in one study, the non-European group experienced frequent insomnia (Rosmond et al. 2000), and in another study, immigrants reported more frequent sleeping problems compared with native Swedes (Rosmond et al. 1996), a predictor of long-term disability (Eriksen et al. 2001). We have chosen to study selfreported anxiety, sleeping problems and pain, as these can be expressions of emotional stress. Although there are many studies supporting the migrationmorbidity hypothesis, for example, studies from other countries showing Turkish immigrant morbidity (Gunay & Haag 1990; Ulusahin et al. 1994), there are also exceptions, the ‘healthy migrant’ hypothesis, studies that do not support the migrationmorbidity hypothesis (Klimidis et al. 1994; Pernice & Brook 1994). Many previous immigration studies in Sweden have been performed on unspecified populations of immigrants. Although immigrant groups are heterogeneous, and significant ethnic group differences in self-reported health have been


Ethnicity & Health 365

shown (Lindstrom et al. 2001), studies on specific groups of immigrants are rare. In this study, we have exclusively studied the Turkish-born immigrant group in Sweden. While Turkish-born immigrants in Sweden are an ethnically diverse group, we have chosen to study them as one group, mainly because the size of the group would otherwise be to small, but also because this group reflects the Turkish-born population in Sweden. This is also a study in a larger project concerning ill-health among Turkish-born immigrants in Sweden, where stress-related diseases, such as diabetes and cardio-vascular events, seems to be more common among Turkish women in Sweden compared to Swedes (Wandell et al. 2003; Gadd et al. 2003). According to Statistics Canada, there are three fundamental ways of measuring ethnicity*origin or ancestry, race and identity. According to the English Dictionary, ethnicity is defined as an ethnic quality or affiliation resulting from racial or cultural ties, and an ethnic group is a group of people of the same race or nationality who share a distinctive culture. In this study we categorise ethnicity as origin (shared nationality, culture and traditions). We have not found any studies on Turkish immigrants in Sweden focusing on anxiety, sleeping problems and severe pain, symptoms common in limiting chronic illness, and one of the main causes of disability in adults of working age in Sweden. The first aim is to study whether there is a difference in self-reported anxiety, sleeping problems and pain, between Turkish-born immigrants and Swedes. The second aim is to study whether this hypothesised difference remains after adjusting for socio-economic status, i.e. age, education, marital status and employment. Material and Methods This study is based on data from the first Swedish National Survey of Immigrants (i.e. migrants born in Iran, Chile, Turkey, and Poland), conducted as a joint project by the National Board of Health and Welfare, the Swedish Immigration Board, the National Institute of Public Health, and the Swedish Government. The questionnaire used in the first Swedish national survey of immigrants was the same as the Swedish Survey of Living Conditions, with questions on housing conditions, health, employment, leisure, economy, education, safety and security, contacts with relatives and friends, and questions about children’s health (Statistics Sweden 1996), with a supplement of immigrant-specific questions (National Board of Health and Welfare 2000). The participants were interviewed face-to-face in their homes by trained interviewers from September 1996 to January 1997 by Statistics Sweden, the Swedish Government-owned statistics bureau. The interview was in Swedish, and the interviewer had a questionnaire translated into Turkish, Spanish, Farsi or Polish, only used as a support when interpretation was needed. The immigrant groups were compared with a sample of native Swedes (n2975) of the same age and gender, interviewed during 1996 as part of the annual Swedish



Survey of Living Conditions. Data collected resulted in four Swedish reports (National Board of Health and Welfare 2000). Data concerning self-reported anxiety, sleeping problems and severe pain among Turkish-born immigrants and Swedes are further analysed in this study. Description of the Turkish Immigrant Group A simple random national sample of 809 immigrants born in Turkey, aged 2760 years, arriving in Sweden between 1980 and 1989, was drawn from the Swedish population register. The response rate was about 66%, resulting in 526 interviews of 285 Turkish-born men and 241 Turkish-born women. A large proportion (59%) of the Turkish-born immigrant group studied belonged to minorities, 33% Kurds, 15% Assyrians and 11% other minorities. Approximately 58% of the studied group came from smaller cities or rural areas. Of the 526 interviews performed, 410 interviews were in Swedish, 16 with professional interpreters, 56 with interpretation by a household member, 27 with interpretation by a child, seven with interpretation by a neighbour or friend, and nine with information about interpretation was missing (National Board of Health and Welfare 2000). Analysis of Non-Respondents The non-response rate was 34% for Turkish-born immigrants. About half of all nonrespondents could not be located, one possible reason being repatriation without informing the Swedish authorities (i.e. the population registry) of their departure (Qvist 1999). The other half did not participate for reasons unknown. The age distribution among respondents and non-respondents was about the same. Nonrespondents were also more prevalent in large cities, such as Stockholm and Gothenburg, and had fewer persons with registered income than respondents. Owing to such non-response, it is probable that the prevalence of anxiety, sleeping problems and severe pain in the study population has been underestimated to some extent, since poor health is more common among people without registered income. Relative risks for the outcome are probably less influenced by non-response than by absolute measures, such as prevalence (National Board of Health and Welfare 2000). Outcome Variables Anxiety (anxiety, alarm, dread, fear, worry) was defined as present or not during the interview. In the questionnaire, the question in Swedish was: Har du mo¨jligen a¨ngslan, oro eller a˚ngest? (Do you possibly have alarm/anxiety, anxiety/worry, anxiety/dread/fear?), and in Turkish the words for anxiety were endise and huzursuzluk. The original response alternatives*1: yes, severe, 2: yes, mild, and 3: no*were categorised in this study into two levels, yes or no.



Sleeping problems (or trouble sleeping) were defined as present or not during the last two weeks before the interview. The question in Swedish was: Har du haft besva¨r med so¨mnen? (Have you had trouble sleeping?) and in Turkish: Uyku ile bir probleminiz oldu mu? The original response alternatives*1: yes and 2: no*were used in this study. Severe pain (severe pain or ache in neck, shoulder, lower back and/or extremities) was defined as present or not during the interview. The questions in Swedish were: Har du mo¨jligen: (1) va¨rk i skuldror, nacke eller axlar? (2) ryggsma¨rtor, ryggva¨rk, ho¨ftsma¨rtor eller ishias? (3) va¨rk eller sma¨rtor i ha¨nder, armba˚gar, ben eller kna¨n? (Do you possibly have: (1) ache in shoulders or neck? (2) backpain, backache, hip pain or sciatica? (3) Ache or pain in hands, elbows, legs or knees?) The Turkish translation was: Asagida belirtilen rahatsizliklar sizde de var mi? (1) ku¨rek kemiklerinde, ense ve ouzlarda agri? (2) sirt sacilari, sirt agrilari, kalca agrilari veya siyatik? (3) eller, dirsekler, bacaklar ve dizlerde agri ya da sanci? The original response alternatives*1: yes, severe, 2: yes, mild, 3: no*were categorised in this study into two levels, severe or not. Explanatory Variables Age was categorised into the following groups: 2739, 4049, and 5060 years of age. Educational status was divided into three categories indicating the extent of school attendance, i.e. years at school: (1) compulsory school level, B10 years, (2) at least two years of high school, between 10 and 12 years, and (3) three years of high school or university studies, 12 years. Marital status was dichotomised as single or cohabiting/married. Employment status was dichotomised as employed or not employed. Not employed included unemployed persons, students, household workers, persons with chronic sickness, and persons with early retirement pension. Employed were those with gainful occupation. Psychoactive drugs included regular use of soporific, antidepressant and/or sedative medicine. Analgesic drugs included regular use of analgesic medicine with or without prescription. Knowledge in Swedish was divided into three categories, poor, intermediate and good. Five questions were asked regarding respondents’ knowledge in Swedish: the ability to: understand news reports and debates on radio and television, discuss at meetings, communicate with authorities over the telephone (e.g. Department of Health, unemployment centre and social security), read books and complete a written application for employment. Experience of discrimination comprised three categories, low, middle and high. Questions asked were if the respondent, comparing him/herself with Swedes, experienced better, the same or worse treatment: when applying for work, when



working, when applying for housing, in the neighbourhood, when applying for a loan in the bank, when seeing a doctor or visiting a hospital, by the unemployment centre, by the social security, by the police, and when visiting restaurants. Validation of the First Swedish National Survey of Immigrants Based on the International Classification of Diseases (ICD)-ninth revision, medical experts developed a list with common symptoms of illness. Questions about anxiety, sleeping problems and pain were added to the questionnaire in 1987 and 1988, and since then, have been used consistently. Five follow-up studies have been carried out in order to estimate reliability. The reliability of the dependent and independent variables was analysed in 1989 in re-interviews (test-retest method), about four weeks after the main interview of a random sample of 410 respondents (response rate 88.4%) for the following variable included in the present study: educational status, giving kappa coefficients between 0.7 and 0.9, indicating a high level of reliability (Warneryd 1990). Statistics Sweden has also carried out validation studies in which self-declared information on certain types of illnesses are checked against corresponding information from medical records. One study of rheumatic diseases showed a 68% concurrence between self-reported lower-back ailments and medical diagnose. Selfreported musculo-skeletal symptoms were generally underestimated, where one possible explanation could be effective treatment (Official Statistics of Sweden, Medical Validation of The Swedish Survey of Living Conditions 1992). Statistical Analysis Data was analysed using logistic regression (Kleinbaum & Klein 2002). The results are shown as odds ratios (OR) with a 95% confidence interval (CI). Of the two models taken into consideration, the first was adjusted for age (age-adjusted model), and the second also for the other explanatory variables (main effect model). The fit of the models was judged by the Hosmer-Lemeshow goodness-of-fit test. If the p-value of the test was 0.05, the fit was considered satisfactory. A test of interaction was performed, and we did not find any interaction of interest. Results The distribution of socio-demographic variables, anxiety, sleeping problems, severe pain and the use of psychoactive and analgesic drugs is demonstrated in Table 1. Turkish-born immigrants showed lower educational level and lower employment rate, especially among women, than among the Swedish controls. The crude prevalence of anxiety, sleeping problems, severe pain and the use of psychoactive and analgesic drugs was at least twice as high among the Turkish-born immigrants compared with Swedes. The higher prevalence among Turkish-born immigrants


Ethnicity & Health 369 Table 1 Distribution of Self-Reported Socio-Demographic Data* (in%), Aged 27 60 Years, by Sex and Country of Origin Variable

Men Sweden

n

1,425

Women Turkey 285

Sweden 1,429

Turkey 241

Age, years 27 39 40 49 50 60

39.7 30.1 30.3

59.7 33.7 6.7

40.0 30.0 30.0

59.3 31.1 9.5

Education, years 0 9 10 12 12

20.8 51.4 27.9

50.5 27.4 22.1

17.3 51.2 31.6

70.5 21.2 8.3

Marital status Single Cohabiting/married

24.7 75.3

11.9 88.1

21.7 78.3

23.2 76.8

Employment Yes No Anxiety Sleeping problems Severe pain Psychoactive drugs, regular use Analgesic drugs, regular use

88.6 11.4 10.3 12.3 13.8 2.7 3.3

67.4 32.6 19.3 27.0 31.2 5.6 8.1

84.3 15.7 17.8 18.3 19.6 4.9 7.1

36.1 65.9 40.7 45.6 47.7 11.6 20.3

Knowledge in Swedish Poor Intermediate Good

20.4 29.8 49.8

44.4 26.1 29.5

Discrimination experience Low Middle High

33.0 33.0 34.0

44.8 28.2 27.0

*The first Swedish National Survey of Immigrants (1996). Face-to-face interviews were performed from September 1996 to January 1997.

showed a proportional increase between illness and medical use. Turkish-born women showed a lower knowledge in Swedish compared with Turkish-born men. Discrimination experience differed between the Turkish-born men and women (p0.020), with a higher rate of men experiencing high discrimination. In Table 2, the prevalences of outcomes variables are demonstrated. Sleeping problems and severe pain increased with age. Living alone and being unemployed was associated with more symptoms, especially among women. Among Turkish-born women there was an increasing prevalence of severe pain (p0.01) with decreasing knowledge in Swedish, and increasing prevalence of anxiety (p 0.01) with increasing



Table 2(a) Prevalence (%) of Self-Reported Data* (Anxiety, Sleeping Problems, Severe Pain) Among Men, Aged 27 60 Years, by Country of Origin Variable

Men, Sweden

Men, Turkey

Anxiety

Sleeping problems

Severe pain

Anxiety

Sleeping problems

Severe pain

Age, years 27 39 40 49 50 60

10.1 11.4 9.5

12.6 11.9 12.3

10.8 14.5 17.5

13.5 28.1 26.3

23.5 29.2 47.4

20.6 33.3 52.6


8.1 10.4 11.8

10.5 14.1 10.3

18.2 15.0 8.3

16.6 25.6 19.1

22.9 30.8 31.8

31.9 23.1 20.6


16.2 8.4

15.1 11.4

14.5 13.6

26.5 18.3

29.4 26.7

23.5 27.5

Employment Yes No

8.8 22.2

11.0 22.2

12.4 25.3

15.6 26.9

20.3 40.9

25.5 30.1


27.6 20.0 15.5

29.3 27.1 26.1

25.9 36.5 21.8


18.1 18.1 21.7

20.2 29.8 30.9

25.5 29.8 25.8


experience of discrimination. Among the Turkish-born men these associations were not significant. The age-adjusted models are shown in Table 3. Low education was associated with a higher risk of severe pain among men and women, and among women also with anxiety and sleeping problems. Living alone was associated with anxiety among men and women, and among women also with sleeping problems and severe pain. Unemployment was associated with higher rates of anxiety sleeping problems and severe pain among both men and women. Turkish-born men and women showed higher odds ratios of anxiety sleeping problems and severe pain, than Swedish controls. Among Turkish-born men odds ratios varied between 2 and 3, and for Turkish-born women between 3 and 5. Results from the full models are shown in Table 4. Among men, lower educational level was associated with increased risks of severe pain, living alone for anxiety, being unemployed for anxiety, sleeping disorders and severe pain. After adjustment for age,


Ethnicity & Health 371 Table 2(b) Prevalence (%) of Self-Reported Data* (Anxiety, Sleeping Problems, Severe Pain) Among Women, Aged 27 60 Years, by Country of Origin Variable

Women, Sweden

Women, Turkey

Anxiety

Sleeping problems

Severe pain

Anxiety

Sleeping problems

Severe pain

Age, years 27 39 40 49 50 60

16.3 16.6 21.0

16.4 14.0 24.9

14.2 18.7 27.7

35.0 48.0 52.7

39.9 52.0 60.9

42.6 46.7 82.6


22.7 17.4 15.7

22.3 16.8 18.4

30.8 19.6 13.5

40.0 47.1 30.0

48.8 35.3 45.0

52.9 41.2 20.0


23.9 16.1

23.2 16.9

23.2 18.6

62.5 34.1

53.6 43.2

55.4 45.4

Employment Yes No

15.4 30.4

16.2 29.5

16.9 34.4

36.8 42.9

39.1 49.4

46.0 48.7


42.1 41.3 38.0

49.5 47.6 38.0

55.1 52.4 32.4


31.5 45.6 50.8

44.4 44.1 49.2

45.4 48.5 50.8


education, marital status and employment, Turkish-born men showed odds ratio of around 2 for anxiety, sleeping problems and severe pain, using Swedish men as reference. Among women, older age was associated with increased risks of sleeping problems and severe pain, lower education for severe pain, and living alone as well as being unemployed for anxiety, sleeping problems and severe pain. Using Swedish women as control, Turkish-born women showed an odds ratio between 2 and 3 for anxiety, sleeping problems and severe pain after adjustment for age, education, marital status and employment. In Table 5, the age-adjusted model, poor knowledge in Swedish showed an association with anxiety among the Turkish-born men, and severe pain among the Turkish-born women, while high discrimination experience showed an association with sleeping problems among the Turkish-born men and anxiety among the Turkish-born women.



Table 3 Logistic Regression by Age-Adjusted Model, with Odds Ratio (95% Confidence Interval) for Self-Reported Data* (Anxiety, Sleeping Problems, Severe Pain), Aged 27 60 Years, by Sex Variable

Education, years 0 9 10 12 12 Marital status Single Cohabit Employment No Yes Country of origin Turkey Sweden

Men

Women

Anxiety

Sleeping problems

Severe pain

Anxiety

Sleeping problems

Severe pain

OR CI

OR CI

OR CI

OR CI

OR CI

OR CI

0.83 1.14 2.59 2.19 2.04 4.06 (0.55 1.25) (0.78 1.67) (1.77 3.78) (1.59 3.03) (1.50 2.77) (2.93 5.64) 0.94 1.23 1.71 1.24 0.92 1.70 (0.66 1.33) (0.88 1.71) (1.19 2.45) (0.92 1.67) (0.68 1.23) (1.24 2.33) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1.81 1.16 0.96 1.86 1.48 1.37 (1.31 2.50) (0.85 1.59) (0.70 1.32) (1.43 2.42) (1.13 1.93) (1.05 1.78) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 2.97 2.94 2.27 2.77 2.89 3.12 (2.12 4.16) (2.15 4.02) (1.66 3.12) (2.14 3.59) (2.24 3.73) (2.42 4.03) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 2.10 2.72 2.70 3.45 4.37 4.69 (1.48 2.99) (1.99 3.74) (1.97 3.71) (2.56 4.88) (3.24 5.89) (3.48 6.34) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref)


The associations were diminished when adjusting for education, marital status and employment, as seen in Table 6, however, anxiety remained in association with poor knowledge of Swedish for the Turkish-born men, and in association with middle and high discrimination experience among the Turkish-born women. Discussion The main findings of this study are that Turkish-born immigrants have higher selfreported risks of anxiety, sleeping problems and severe pain than Swedish controls. Although Turkish-born immigrants reported lower attained education and a lower employment rate than Swedish controls, after adjusting for age and socio-economic status, the increased risks for anxiety sleeping problems and severe pain were only partially reduced. These higher self-reported risks can, therefore, only partially be explained by age and socio-economic status, i.e. education, marital status and employment, more so among the Turkish-born women.


Ethnicity & Health 373 Table 4 Multivariate Logistic Regression, Including All Variables in Full Model, with Odds Ratio (95% Confidence Interval) Regarding Self-Reported Data* (Anxiety, Sleeping Problems and Severe Pain), Aged 27 60 Years, by Sex Variable

Men Anxiety

Age, years 27 39 40 49 50 60 Education, years 0 9 10 12 12 Marital status Single Cohabitation Employment No Yes Country of origin Turkey Sweden

Sleep problem

Women Severe pain

Anxiety

Sleep problem

Severe pain

0.65 0.92 0.64 0.81 0.98 0.74 (0.46 0.92) (0.67 1.27) (0.47 0.89) (0.60 1.09) (0.72 1.32) (0.55 1.00) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 0.79 1.07 1.13 1.19 1.76 1.68 (0.53 1.20) (0.74 1.57) (0.80 1.59) (0.86 1.65) (1.27 2.42) (1.23 2.29) 0.66 0.84 2.13 1.23 1.05 2.32 (0.43 1.07) (0.56 1.26) (1.44 3.15) (0.85 1.78) (0.73 1.49) (1.62 3.35) 0.95 1.26 1.76 1.17 0.84 1.60 (0.66 1.35) (0.90 1.76) (1.22 2.53) (0.86 1.59) (0.63 1.14) (1.16 2.27) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1.84 1.21 1.03 1.78 1.39 1.29 (1.32 2.56) (0.88 1.69) (0.74 1.43) (1.35 2.34) (1.05 1.84) (0.97 1.71) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 2.39 2.34 1.81 1.87 1.90 1.88 (1.67 3.42) (1.68 3.25) (1.30 2.53) (1.40 2.51) (1.42 2.53) (1.41 2.50) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 2.12 2.60 2.14 2.44 3.01 2.59 (1.43 3.15) (1.82 3.72) (1.50 3.05) (1.69 3.53) (2.09 4.33) (1.80 3.71) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref)

Model fit Hoswer-Lemeshow p -Value 0.72

0.72

0.65

0.55

0.95

0.07


This study has some important limitations. Regarding the first Swedish National Survey of Immigrants, the response rate was only 66% among the Turkish-born immigrants, which could underestimate the prevalence of poor health. Half of the non-responders were not found, and it is possible that these people left Sweden without reporting to the authorities. Furthermore, data were self-reported with a possibility of bias. However, Bayard-Burfield et al. (1998), for example, revealed that self-reported, chronic psychiatric illness was a strong risk factor for all cause mortality and violent death. In addition, requirements for ‘good Swedish’ rating were very high. Respondent were required to score ‘very good’ or ‘good’ on all five items which could underestimate their knowledge in Swedish.



Table 5 Logistic Regression by Age-Adjusted Model, with Odds Ratio (95% Confidence Interval) for Self-Reported Data* (Anxiety, Sleeping Problems and Severe Pain) Among Turkish-Born Men and Women, Aged 27 60 Years, by Sex Variable

Men, Turkey Anxiety

Sleeping problem

Women, Turkey Severe pain

Anxiety

Sleeping problem

Severe pain

Knowledge in Swedish Poor 2.05 0.95 0.92 0.93 1.29 2.06 (1.12 3.77) (0.46 1.97) (0.29 2.88) (0.39 2.24) (0.77 2.14) (1.16 3.65) Intermediate 1.37 1.02 2.03 1.19 1.54 2.33 (0.79 2.39) (0.56 1.88) (0.84 4.92) (0.47 3.09) (0.89 2.64) (1.27 4.29) Good 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) Discrimination experience Low 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) Middle 1.06 1.81 1.35 2.00 1.06 1.25 (0.46 2.45) (0.92 3.57) (0.65 2.80) (1.18 3.38) (0.61 1.82) (0.78 2.01) High 1.48 2.10 1.24 2.49 1.31 1.37 (0.65 3.35) (1.06 4.15) (0.59 2.62) (1.46 2.25) (0.75 2.27) (0.85 2.22) *The first Swedish National Survey of Immigrants (1996). Face-to-face interviews were performed from September 1996 to January 1997.

Another limitation is the issue of conceptual equivalence, the instrument measuring the same theoretical construct in each culture. It is important that translations must be in lay terminology to be understandable to responders (Westermeyer & Janka 1997). Unfortunately, there is no description of how issues of conceptual equivalence were handled, only that questionnaires were translated by professional translators. Considering that almost 80% of the interviews were performed in Swedish, the question also is how the Turkish-born immigrants comprehended the Swedish words used in the questionnaire. Another issue is the validation and reliability of studied outcome variables. It is difficult to test the reliability of conditions ‘experienced at the moment’ with reinterviews, since conditions may change over time. Official Statistics of Sweden has carried out several validation studies. We have not found any validation studies regarding anxiety and sleeping problems, only that self-reported musculo-skeletal symptoms seem to be under-reported. One must bear in mind that this is a study comparing self-reported, self-experienced, non-objective, symptoms among two groups. The increased risks of self-reported anxiety, sleeping problems and severe pain are supported by the proportional increase in the use of psychoactive and analgesic drugs. We argue that the risk of bias is low, that this study represents the clinical reality, and that the measurements, widely used internationally and constructed to give an objective response, are comparable between the studied groups.


Ethnicity & Health 375 Table 6 Multivariate Logistic Regression in a Model Adjusted for Age, Education, Marital Status and Employment, with Odds Ratio (95% Confidence Interval) for Self-Reported Data* (Anxiety, Sleeping Problems and Severe Pain) Among Turkish-Born Men and Women, Aged 27 60 Years, by Sex Variable

Men, Turkey Anxiety

Sleeping problem

Women, Turkey Severe pain

Anxiety

Sleeping problem

Severe pain

Knowledge in Swedish Poor 2.72 1.04 0.70 1.01 1.13 1.71 (1.01 7.38) (0.43 2.54) (0.26 1.89) (0.35 2.91) (0.51 2.52) (0.66 4.43) Intermediate 1.69 1.14 1.68 1.38 1.46 2.00 (0.71 4.03) (0.55 2.35) (0.79 3.57) (0.48 3.97) (0.66 3.27) (0.77 5.16) Good 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) Discrimination experience Low 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) 1 (Ref) Middle 1.00 1.75 1.34 2.19 1.13 1.46 (0.41 2.44) (0.78 3.91) (0.65 2.75) (1.06 4.54) (0.50 2.55) (0.75 2.82) High 1.33 1.92 1.37 2.58 1.33 1.52 (0.55 3.22) (0.86 4.32) (0.65 2.88) (1.23 5.40) (0.58 3.02) (0.78 2.46) *The first Swedish National Survey of Immigrants (1996). Face-to-face interviews were performed from September 1996 to January 1997.

There are no data on Turkish-born people in Turkey regarding our outcome variables, which also would have been an interesting comparison when this could show possible effects of migration itself. Even if this study has several limitations, they are balanced by the strengths. The Annual Swedish Survey of Living Conditions has been consistent over the years, and the first (1996) survey of Swedish immigrants is one of the largest and most comprehensive surveys outside the USA. The long tradition and experience of this well validated survey with high reliability is a considerable strength. Many studies have found associations between low socio-economic status, poor mental health and pain. For example, significant ethnic group differences in selfreported health were greatly reduced by psychosocial and economic factors (Lindstrom et al. 2001). Furthermore, a study among British Asians showed that long periods of unemployment were associated with lower psychological well-being (Shams & Jackson 1994). Another study showed that immigrant patients live under more strained psychosocial conditions and experience a deeper impact of pain than Swedish-born (Soares & Grossi 1999). Finally, one study showed a strong association between ethnicity and poorly reported health, which seemed to be mediated by socioeconomic status, poor acculturation and discrimination (Wiking et al. 2004). In this study, socio-economic status, poor acculturation and discrimination can only partly explain the higher risks among the Turkish-born immigrants indicating that there must be other factors not studied here that are of importance.



Ethnicity has been shown to be an independent, powerful, social dimension compared with social class in relation to self-rated illness, where there was a strong association between being a Latin American refugee and ill-health (Sundquist 1995). Ethnicity, defined as foreign-born people, was in another study, strongly associated with limiting long-term illness when controlling for social, material, and lifestyle factors (Sundquist & Johansson 1997). Furthermore, ethnicity was the strongest independent risk mediator for poor mental health among Latin American refugees when matched with Swedish controls (Sundquist 1993). In addition, being born outside Sweden was significantly associated with longer consultation time and an increased prevalence of musculo-skeletal disease compared with people born in Sweden (Sundquist et al. 1994). The gender difference has been shown earlier. For example, Ritsner et al. (2001) found evidence that the higher psychological distress among female immigrants stems from women’s greater exposure to specific psychosocial stressors. Another study showed that the prevalence of probable posttraumatic stress syndrome varied between ethnic groups in Sweden, with a prevalence of 53 among the Turks and 29 among the Swedes, where the probable post-traumatic stress syndrome outcome was associated with multiplicity of relatives’ traumas, multiplicity of own traumas, and belonging to an ethnic minority, but not with gender (Al-Saffar 2003). Our results are in strong association with all these studies. We chose to study the three variables, anxiety, sleeping problems and pain, since these variables can express emotional stress, and there were no previous studies performed regarding the Turkish-born immigrant group. Migration to a socially and culturally different society is likely to cause dramatic life changes, where difficulty in acculturation, i.e. learning of the ideas, conventions and behaviour that characterise a social group, could affect mental and physical health. The immigrant situation implies possible problems, i.e. with adaptation to the new home country, owing to differences in culture, religion and language, or due to low education, i.e. the illiteracy is high in some groups. The social position could be affected, due to unemployment, loss of professional status and social prestige. Many Turkish-born immigrants in Sweden come from rural areas in Turkey, and according to the United Nations Human Development Report of 2001, a majority of Turkish rural women work in agriculture. In Sweden, Turkish-born immigrants predominantly live in suburbs with many immigrants, preferably with the same background, preserving their culture, which could diminish the acculturation process. In our study, we found that a large proportion of the Turkish-born immigrants had poor knowledge in Swedish, especially the Turkish-born women. Compared with immigrant groups in Sweden from Chile, Iran and Poland, the Turkish-born immigrant group, especially the Turkish-born women, showed poorer knowledge in Swedish (National Board of Health and Welfare 2000). The Swedish Sos-report (2000, p. 3) showed that health among Swedes and four immigrant groups (immigrants born in Iran, Chile, Turkey and Poland) is connected with similar



socioeconomic factors, such as age, sex, marital status, education, employment, economic situation, security and social support. Among the immigrants there was also an association between acculturation status (knowledge in Swedish) and health. However, socioeconomic factors and acculturation status could not explain the difference between the four immigrant groups (National Board of Health and Welfare 2000). This could indicate that ethnical/ cultural differences are of importance. Poor knowledge in Swedish and high experience of discrimination could add to psychological distress factors, and diminish the acculturation process in Sweden. Difficulties in acculturation could be one possible explanation for the higher risks of anxiety, sleeping problems and pain among Turkish-born immigrants in Sweden, but there seems to be other factors not studied here. Further studies regarding cultural beliefs/traditions, measuring cultural differences in valuing health, illness and living conditions, could bring important knowledge. Conclusion Being a Turkish-born immigrant in Sweden significantly increases the risks for selfreported anxiety, sleeping problems and severe pain even after adjusting for age and socio-economic status (education, marital status and employment). Immigration to Sweden with difficulty in adaptation, poor knowledge in Swedish and experience of discrimination, could probably add to these higher prevalences, but does not fully explain the increased risks. Further studies are, therefore, needed. Nevertheless, these findings indicate a greater demand on the health care system, and should be taken into consideration among politicians, health analysts, social workers, planners and providers when developing intervention programmes for reducing inequalities in health between native-born Swedes and Turkish-born immigrants.

References Al-Saffar, S. (2003) ‘The aetiology of posttraumatic stress disorder in four ethnic groups in outpatient psychiatry’, Social Psychiatry and Psychiatric Epidemiology, vol. 38, pp. 456 462. Baarnhielm, S. & Ekblad, S. (2000) ‘Turkish-born migrant women encountering health care in Stockholm: A qualitative study of somatization and illness meaning’, Culture, Medicine and Psychiatry, vol. 24, pp. 431 452. Bayard-Burfield, L., Sundquist, J. & Johansson, S. E. (1998) ‘Self-reported long-standing psychiatric illness as a predictor of premature all-cause mortality and violent death: A 14-year follow-up study of native Swedes and foreign-born migrants’, Social Psychiatry and Psychiatric Epidemiology, vol. 33, pp. 491 496. Bayard-Burfield, L., Sundquist, J. & Johansson, S. E. (2001) ‘Ethnicity, self reported psychiatric illness, and intake of psychotropic drugs in five ethnic groups in Sweden’, Journal of Epidemiology and Community Health , vol. 55, pp. 657 664. Diefenbacher, A. & Heim, G. (1994) ‘Somatic symptoms in Turkish-born and German depressed patients’, Psychosomatic Medicine , vol. 56, pp. 551 556.



Eriksen, W., Natvig, B. & Bruusgaard, D. (2001) ‘Sleep problems: A predictor of long-term work disability? A four-year prospective study’, Scandinavian Journal of Public Health , vol. 29, pp. 23 31. Gadd, M., Johansson, S. E., Sundquist, J. & Wa¨ndell, P. (2003) ‘Morbidity in cardiovascular diseases in immigrants in Sweden’, Journal of Internal Medicine , vol. 254, pp. 236 243. Gunay, E. & Haag, A. (1990) [‘Illness in emigration-a study of Turkish-born patients in general practice from the psychosomatic viewpoint]’, Psychotherapie, Psychosomatik medizinische Psychologie , vol. 40, pp. 417 422. Jorgensen, C. K., Fink, P. & Olesen, F. (2000) ‘Psychological distress and somatisation as prognostic factors in patients with musculoskeletal illness in general practice’, British Journal of General Practice , vol. 50, pp. 537 541. Kirmayer, L. J. & Young, A. (1998) ‘Culture and somatization: Clinical, epidemiological, and ethnographic perspectives’, Psychosomatic Medicine , vol. 60, pp. 420 430. Kleinbaum, D. G. & Klein, M. (2002) Logistic Regression , Springer, New York. Klimidis, S., Stuart, G., Minas, I. H. & Ata, A. W. (1994) ‘Immigrant status and gender effects on psychopathology and self-concept in adolescents: A test of the migration-morbidity hypothesis’, Comprehensive Psychiatry, vol. 35, pp. 393 404. Li, H. Z. & Browne, A. J. (2000) ‘Defining mental illness and accessing mental health services: Perspectives of Asian Canadians’, Canadian Journal of Community Mental Health , vol. 19, pp. 143 159. Lindstrom, M., Sundquist, J. & Ostergren, P. O. (2001) ‘Ethnic differences in self reported health in Malmo in southern Sweden’, Journal of Epidemiology and Community Health , vol. 55, pp. 97 103. Lofvander, M. B. & Furhoff, A. K. (2002) ‘Pain behaviour in young immigrants having chronic pain: An exploratory study in primary care’, European Journal of Pain , vol. 6, pp. 123 132. National Board of Health and Welfare (2000) Different Conditions-Different Health. A Study Among Immigrants from Chile, Iran, Poland and Turkey (in Swedish) , National Board of Health and Welfare, Stockholm. Official Statistics of Sweden (1992) Medical Validation of the Swedish Survey of Living Conditions , Statistics Sweden, Stockholm. Official Statistics of Sweden (1996) The Swedish Survey of Living Conditions. Design and Method . ¨ rebro. Official Statistics of Sweden (2005) Population Statistics 2005 , Statistics Sweden, O Okifuji, A. & Turk, D. C. (2002) ‘Stress and psychophysiological dysregulation in patients with fibromyalgia syndrome’, Applied Psychophysiology and Biofeedback , vol. 27, pp. 129 141. Pernice, R. & Brook, J. (1994) ‘Relationship of migrant status (refugee or immigrant) to mental health’, The International Journal of Social Psychiatry, vol. 40, pp. 177 188. Qvist, J. (1999) Problems of Coverage in the Register of Total Population (RTB). Estimation of Overcoverage by an Indirect Method . Ritsner, M. & Ponizovsky, A. (1999) ‘Psychological distress through immigration: The two-phase temporal pattern?’, The International Journal of Social Psychiatry, vol. 45, pp. 125 139. Ritsner, M., Ponizovsky, A., Kurs, R. & Modai, I. (2000) ‘Somatization in an immigrant population in Israel: A community survey of prevalence, risk factors, and help-seeking behavior’, American Journal of Psychiatry, vol. 157, pp. 385 392. Ritsner, M., Ponizovsky, A., Nechamkin, Y. & Modai, I. (2001) ‘Gender differences in psychosocial risk factors for psychological distress among immigrants’, Comprehensive Psychiatry, vol. 42, pp. 151 160. Rosmond, R., Lapidus, L. & Bjorntorp, P. (1996) ‘A comparative review of psychosocial and occupational environment in native Swedes and immigrants’, Scandinavian Journal of Social Medicine , vol. 24, pp. 237 242. Rosmond, R., Nilsson, A. & Bjorntorp, P. (2000) ‘Psychiatric ill health and distribution of body fat mass among female immigrants in Sweden’, Public Health , vol. 114, pp. 45 51.


Ethnicity & Health 379 Sabbioni, M. E. & Eugster, S. (2001) ‘Interactions of a history of migration with the course of pain disorder’, Journal of Psychosomatic Research , vol. 50, pp. 267 269. Shams, M. & Jackson, P. R. (1994) ‘The impact of unemployment on the psychological well-being of British Asians’, Psychological Medicine , vol. 24, pp. 347 355. Soares, J. J. & Grossi, G. (1999) ‘Experience of musculoskeletal pain. Comparison of immigrant and Swedish patient’, Scandinavian Journal of Caring Sciences , vol. 13, pp. 254 266. Sundquist, J. (1993) ‘Ethnicity as a risk factor for mental illness. A population-based study of 338 Latin American refugees and 996 age-, sex- and education-matched Swedish controls’, Acta Psychiatrica Scandinavica , vol. 87, pp. 208 212. Sundquist, J. (1995) ‘Ethnicity, social class and health. A population-based study on the influence of social factors on self-reported illness in 223 Latin American refugees, 333 Finnish and 126 south European labour migrants and 841 Swedish controls’, Social Science & Medicine , vol. 40, pp. 777 787. Sundquist, J. & Johansson, S. E. (1997) ‘Long-term illness among indigenous and foreign-born people in Sweden’, Social Science & Medicine , vol. 44, pp. 189 198. Sundquist, J., Rose´n, U. & Johansson, S.-E. (1994) ‘Ethnicity as a risk indicator for musculoskeletal disease, long consultation times and care utilization’, Scandinavian Journal of Social Welfare , vol. 3, pp. 212 217. Ulusahin, A., Basoglu, M. & Paykel, E. S. (1994) ‘A cross-cultural comparative study of depressive symptoms in British and Turkish-born clinical samples’, Social Psychiatry and Psychiatric Epidemiology, vol. 29, pp. 31 39. Wandell, P. E., Hjo¨rleifsdottir Steiner, K. & Johansson, S. E. (2003) ‘Diabetes mellitus in Turkishborn immigrants in Sweden’, Diabetes & Metabolism , vol. 29, pp. 435 439. Warneryd, B. (1990) Levnadsfo¨rha˚llanden. A˚terintervjustudie i underso¨kningen av levnadsfo¨rha˚llanden (ULF) 1989 . (In Swedish) (Living Conditions. Reinterview in ULF 1989) , Statistics Sweden, Stockholm. Westermeyer, J. & Janka, A. (1997) ‘Language, culture and psychopathology: Conceptual and methodological issues’, Transcultural Psychiatry, vol. 34, pp. 291 311. Wiking, E., Johansson, S. E. & Sundquist, J. (2004) ‘Ethnicity, acculturation, and self reported health. A population based study among immigrants from Poland, Turkey, and Iran in Sweden’, Journal of Epidemiology and Community Health , vol. 58, pp. 574 582.