Abstract

Background Refugee immigrants have poorer health than other immigrant groups but little is known about their mortality. A comparison of mortality among refugees and non-refugee immigrants is liable to exaggerate the former if the latter includes labour migrants, whose mortality risk may be lower than that of the general population. To avoid bias, labour migrants are not included in this study. The aim was to investigate mortality risks among refugees compared with non-labour non-refugee immigrants in Sweden.

Methods Population-based cohort design, starting 1 January 1998 and ending with death or censoring 31 December 2006. Persons included in the study were those aged 18–64 years, had received a residence permit in Sweden 1992–98 and were defined by the Swedish Board of Migration as either a refugee or a non-labour non-refugee immigrant. The outcomes were all-cause and cause-specific mortalities and the main exposure was being a refugee. Cox-regression models estimated hazard ratios (HRs) of mortality.

Results The study population totalled 86 395 persons, 49.3% women, 24.2 % refugees. Adjusted for age and origin, refugee men had an over-risk of cardiovascular mortality (HR = 1.58, 95% CI = 1.08–2.33). With socio-economic factors added to the model, refugee men still had an over-risk mortality in cardiovascular disease (HR = 1.53, 95% CI = 1.04–2.24) and external causes (HR = 1.59, 95% CI = 1.01–2.50).

Conclusion Refugee men in Sweden have a higher mortality risk in cardiovascular and external causes compared with male non-labour non-refugee immigrants. This study suggests that the refugee experience resembles other stressors in terms of the association with cardiovascular mortality.

Introduction

In 2010, there were 214 million migrants worldwide, including 15 million refugees.1 The United Nations Refugee Convention defines a refugee as a person who: ‘owing to a well-founded fear of being persecuted for reasons of race, religion, nationality, membership of a particular social group or political opinion, is outside the country of his nationality, and is unable to, or owing to such fear, is unwilling to avail himself of the protection of that country’.

Health, including mental health, among refugees and other immigrants is largely associated with the same factors as among natives. In addition, there are immigrant- and refugee-specific determinants: pre- and post-migration factors. Pre-migration factors include income level in the immigrant’s country of origin,2 reason for migration3 and, for refugees, fear of persecution, experience or threat of torture and injuries and cumulative exposure to trauma such as torture.4,5 Post-migration factors include structural factors such as working and living conditions2 and, for refugees, the asylum process.6 Post-migration factors also include personal factors such as time in host country, social networks, language and acculturation status.2 Refugees to high-income countries have poorer health, especially mental health, than other immigrant groups, including non-refugee immigrants from the same countries.3

Refugees have an over-risk of depression and post-traumatic stress disorder (PTSD).46 Depression and PTSD are both independent risk factors for coronary heart disease.7,8 Refugee-specific pre- and post-migration determinants of health, such as fear of persecution and the asylum process, are stressors. Both acute stress and chronic stressors are associated with cardiovascular diseases.9 A study of refugees to the USA concluded that there is a strong public health concern for cardiovascular disease among refugees.10 Repeated studies have found that suicidal behaviour is common among refugees.11,12

Residence permits for refugees are based on the need of asylum, whereas labour migrants are accepted if there is a need of labour. As participation in the labour force requires good health, the health of labour migrants on arrival is better on average than that of the general population.2 Therefore, if labour migrants are included in the group of non-refugee immigrants, a difference in mortality between this group and refugees might be due to labour migrants’ good health and low mortality instead of to refugee-related mortality among the refugees.

An 18-year follow-up study in Canada by DesMeules and colleagues found lower mortality rates among migrants, both refugees and non-refugees, compared with natives, but higher among refugees than among other immigrants. However, as the study compared refugees with labour immigrants, the mortality difference might have been due to the immigrant selection process.13 The preferred reference group when studying refugee mortality is non-refugee non-labour immigrants from the same countries as the refugees. To our knowledge, this is the first study in which mortality among refugees is compared with mortality among non-labour non-refugee immigrants (referred to hereafter as non-refugees) in a high-income country.

Ethnic differences in mortality patterns have been studied extensively13–17 but little is known about refugees’ mortality patterns.

We hypothesize that being a refugee is a risk factor for cardiovascular mortality and mortality from external causes, for both men and women. Furthermore, we hypothesize that the increased risk is not explained by age differences, ethnic differences or differences in socio-economic factors.

The aim of this study is to investigate mortality risks among refugees compared with non-refugee immigrants adjusted for age differences, ethnic differences and differences in socio-economic factors.

Methods

Study population

As a signatory to the above-mentioned UN Refugee Convention, as well as by law, Sweden is obliged to give asylum to refugees. Asylum applications are accepted if there is sufficient evidence for the claim; the burden of proof rests on the applicant.18 The Swedish Board of Migration and, in complicated cases, the Swedish Migration courts assess these claims. Moreover, each year the Swedish Parliament provides funding to enable the Swedish Migration Board to transfer refugees from their native country, or a refugee camp elsewhere, to Sweden. These persons have permission to stay in Sweden from the day of arrival. The quota is made up of persons selected by either the Swedish Board of Migration or the United Nations High Commissioner for Refugees (UNHCR).

The Swedish Board of Migration offers asylum seekers a place to stay, restricted working permits and, if needed, a small subsistence allowance. An asylum seeker is entitled by law to subsidised emergency medical care; other care is at own expense. When granted a residence permit, refugees and non-refugee immigrants are offered the same health care and social benefits as other immigrants and native Swedes.18

The non-refugee immigrant group was made up of persons accepted by Sweden as family of refugees (but not family of labour migrants) or for humanitarian reasons. Humanitarian reasons imply that the immigrant has been granted a residence permit due to circumstances in the current life situation, such as that children of the immigrant would suffer badly on being returned or that the immigrant’s health status would deteriorate severely.

In order to identify refugees and non-refugee immigrants, we used the Swedish Board of Migration’s un-aggregated classifications of reasons for a permanent residence permit.

Selection process for the study population

For a schematic view of the selection process, see Figure 1. Included in the study were persons born abroad, aged 18–64 years at the time of the study, who had received a residence permit in Sweden during 1992–98. The exclusion criteria were: have left Sweden prior to death or the end of the study (censoring) in 2006, according to the Swedish Population Registration System, or assumed to have left the country without informing the Swedish tax authorities, according to criteria described by Weitoft.19

Figure 1

The selection process

Figure 1

The selection process

Design

The study used a retrospective cohort design to compare mortality risks among refugees and non-refugee immigrants with time defined as years from the start of the study on 1 January 1998 to death or censoring on 31 December 2006.

Data sources

Register-based studies use data in official registers collected for generic purposes. Sweden has a high standard of official registers adapted for research purposes.20 The Population Registration System assigns a personal identity number to Swedish citizens and people living in Sweden with a permanent residence permit. This number was used as a means of identification. After ethical approval and permission, it is possible to link registers for research purposes and the data are anonymized after linkage. This study was performed with data from Statistics Sweden, the National Board of Health and Welfare and the Swedish Board of Migration.

Deaths are included in the registry irrespective of whether they occur in Sweden or abroad.21 Deaths abroad are registered either with the help of Swedish missions using Swedish death certificates or with domestic death certificates. International regulations for death certificates oblige physicians to report deaths according to WHO standards.21 Unregistered deaths abroad can lead to invalidly low mortality rates compared with natives.19 The age of retirement in Sweden is 65 years. Hence, after 64 years of age, the chance of spending long periods abroad with an increased risk of dying due to age creates a possibility of unregistered deaths. Statistics Sweden studied potential unregistered deaths and found a problem among immigrants over 85 years of age.20 We addressed the risk of unregistered deaths by including persons aged 18–64 years only. Differences in unregistered deaths abroad by origin were addressed by adjusting for origin.

Variables

Outcome

The outcome was mortality. Causes of death were coded according to ICD 9 and ICD 10, and included: all-cause mortality, neoplasms (ICD 9: 140–239) (ICD 10: C00−D48), cardiovascular disease (ICD 9: 390–459) (ICD 10: I00−I99), external causes (ICD 9: E800–E999) (ICD 10: S00−T98, V01−Y98) and all other causes. These causes were chosen as being the most common causes of death in the age group.

Time

From 1998 until death or censoring in 2006, measured in years.

Main exposure

Reason for migration; refugee or non-refugee immigrant, as explained above.

Explanatory variables

Age at the beginning of the study in 1998, age group was coded into five dummy age groups: 18–24, 25–34, 35–44, 45–54 and 55–64 years.

Country of origin: Iraq, Iran and former Yugoslavia were countries from which large refugee groups arrived in Sweden during 1992–2006. Other countries of origin were categorized as Others.

Arrival was coded into three dummy variables: 1992–93, 1994–95 and 1996–98.

Marital status at immigration was either living with a partner or living alone.

Statistics Sweden converted education completed outside Sweden into equivalent levels of schooling in Sweden. Education at immigration was coded into three dummy groups: <11 years of schooling, ≥11 years of schooling and unknown length of schooling.

Employment rates differ greatly between newly arrived immigrants and immigrants who have been in Sweden for a long time.22 However, the differences between those who have been in Sweden for 5 years and those who have been in Sweden for longer are minor.23 Economic activity was measured after 5 years in Sweden or, if a person died after <5 years in Sweden, as employment status at the year of death. Economic activity was coded in three dummy variables: employed (including parental leave and full-time students), sick leave/long-term sick leave (including disability pension) and unemployed.

At immigration, area of residence in Sweden was coded into large city (ref), comprising Stockholm, Gothenburg and Malmö, and smaller cities, towns or rural, comprising all other areas.

Statistical analyses

All analyses were done separately for men and women. Incidence rates were calculated as deaths per 100 000 person-years. Demographic variables were compared between refugees and non-refugee immigrants using chi-square and t-tests. Cox-regression models were used to estimate hazard risk ratios for all-cause mortality and cause-specific mortality. These models were adjusted by age, age and origin, and age, origin and additional explanatory variables, respectively. The explanatory variables included in the final model were selected in a stepwise process. Tests of statistical power, as well as graphical and statistical tests of fulfilment of the Cox-regression assumption of proportional hazards, were performed as suggested by Hosmer.24 Results are presented as hazard ratios (HR) with 95% confidence interval (CI). The analyses were conducted with the SAS software package 9.2.

Ethical approval

This study was approved by the Stockholm Regional Ethical Review Board (2008/732–31 and 2010/1983–32).

Results

A total of 93 843 persons met the inclusion criteria, of whom 23.7% were refugees. Some were excluded because they had left Sweden before death or censoring (non-refugee immigrants 8.5%, refugees 6.2%). Chi-square tests indicated that the groups were equally large except for men from Iraq and Other countries, among whom the difference between excluded refugees and excluded non-refugee immigrants was explained by year of arrival in Sweden; more years since arrival in Sweden gave a greater probability of having left Sweden.

The study population totalled 86 395, of whom 49.3% were women and 24.2% refugees. Population characteristics are outlined in Table 1. Mean age in 1998 was: male non-refugees 34.19 (SD 9.67) years, male refugees 34.89 (SD 9.27), female non-refugees 33.24 (9.27) and female refugees 35.40 (10.22). The largest groups were from Iraq and former Yugoslavia. The largest group of non-refugees was from former Yugoslavia, as was the largest group of female refugees, whereas the largest group of refugee men came from Iraq. t-test and chi-square tests indicated that the means and proportions of refugees and non-refugee immigrants were different for all explanatory variables. Hence, if any of the explanatory factors was related to the outcome, there was a need to adjust for this in the Cox-regression models.

Table 1

Percentage and number of deaths (cases) by year of arrival, country of origin, education, marital status, economic activity and area of residence in Sweden for refugee and non-refugee immigrants by sex

 Men
 
Women
 
 Non-refugee Refugee Non-refugee Refugee 
Arrival % (Cases) % (Cases) % (Cases) % (Cases) 
1992–93 43.76 (170) 37.42 (51) 41.18 (113) 33.82 (23) 
1994–95 38.43 (177) 34.75 (79) 37.94 (110) 42.21 (45) 
1996–98 17.80 (62) 27.83 (39) 20.88 (48) 23.97 (17) 
Origin     
    Iran 2.63 (9) 10.73 (8) 4.71 (7) 11.04 (2) 
    Iraq 7.67 (22) 46.24 (51) 10.91 (15) 31.28 (20) 
    Former Yugoslavia 65.28 (283) 27.42 (92) 56.36 (181) 44.49 (57) 
    Other 24.41 (95) 15.61 (18) 28.02 (68) 13.19 (6) 
Education     
    ≥11 years 29.88 (270) 27.86 (98) 23.31 (129) 22.09 (28) 
    <11 years 8.48 (77) 8.86 (34) 12.21 (71) 12.76 (29) 
    Missing 61.64 (62) 63.28 (37) 64.47 (71) 65.16 (28) 
Marital status     
    With partner 58.43 (306) 51.26 (121) 72.87 (204) 29.93 (73) 
    Alone 41.57 (103) 48.74 (48) 27.13 (67) 70.07 (12) 
Economic activity     
    Employed 67.36 (269) 65.46 (108) 70.15 (195) 71.06 (72) 
    Sick leave 3.22 (31) 4.93 (16) 3.61 (28) 4.75 (3) 
    Unemployed 29.42 (109) 29.61 (45) 26.15 (48) 24.19(10) 
Area of residence     
    Smaller 66.56 (259) 60.07 (104) 63.93 (139) 63.91 (51) 
    Large 33.44 (150) 39.93 (65) 36.07 (132) 36.09 (34) 
    n-Total (cases) 30 951 (409) 12 802 (169) 34 550 (271) 8092 (85) 
 Men
 
Women
 
 Non-refugee Refugee Non-refugee Refugee 
Arrival % (Cases) % (Cases) % (Cases) % (Cases) 
1992–93 43.76 (170) 37.42 (51) 41.18 (113) 33.82 (23) 
1994–95 38.43 (177) 34.75 (79) 37.94 (110) 42.21 (45) 
1996–98 17.80 (62) 27.83 (39) 20.88 (48) 23.97 (17) 
Origin     
    Iran 2.63 (9) 10.73 (8) 4.71 (7) 11.04 (2) 
    Iraq 7.67 (22) 46.24 (51) 10.91 (15) 31.28 (20) 
    Former Yugoslavia 65.28 (283) 27.42 (92) 56.36 (181) 44.49 (57) 
    Other 24.41 (95) 15.61 (18) 28.02 (68) 13.19 (6) 
Education     
    ≥11 years 29.88 (270) 27.86 (98) 23.31 (129) 22.09 (28) 
    <11 years 8.48 (77) 8.86 (34) 12.21 (71) 12.76 (29) 
    Missing 61.64 (62) 63.28 (37) 64.47 (71) 65.16 (28) 
Marital status     
    With partner 58.43 (306) 51.26 (121) 72.87 (204) 29.93 (73) 
    Alone 41.57 (103) 48.74 (48) 27.13 (67) 70.07 (12) 
Economic activity     
    Employed 67.36 (269) 65.46 (108) 70.15 (195) 71.06 (72) 
    Sick leave 3.22 (31) 4.93 (16) 3.61 (28) 4.75 (3) 
    Unemployed 29.42 (109) 29.61 (45) 26.15 (48) 24.19(10) 
Area of residence     
    Smaller 66.56 (259) 60.07 (104) 63.93 (139) 63.91 (51) 
    Large 33.44 (150) 39.93 (65) 36.07 (132) 36.09 (34) 
    n-Total (cases) 30 951 (409) 12 802 (169) 34 550 (271) 8092 (85) 

Unadjusted incidence rates differed between male and female refugees and non-refugee immigrants (Table 2).

Table 2

Number of deaths (cases) and unadjusted incidence rates; deaths per 100 000 person-years (incidence) for all-cause, neoplasm, cardiovascular, external and other causes of mortality for refugee and non-refugee immigrants by sex

 Men
 
Women
 
 Non-refugee Refugee Non-refugee Refugee 
Causes Cases (incidence) Cases (incidence) Cases (incidence) Cases (incidence) 
All causes 409 (16.76) 169 (16.62) 271 (10.00) 85 (13.25) 
Neoplasm 183 (7.50) 55 (5.41) 135 (4.98) 45 (7.01) 
Cardiovascular 84 (3.44) 51 (5.02) 43 (1.59) 20 (3.12) 
External causes 67 (2.75) 40 (3.93) 37 (1.37) 8 (1.25) 
Other 75 (3.07) 23 (2.26) 56 (2.07) 12 (1.87) 
 Men
 
Women
 
 Non-refugee Refugee Non-refugee Refugee 
Causes Cases (incidence) Cases (incidence) Cases (incidence) Cases (incidence) 
All causes 409 (16.76) 169 (16.62) 271 (10.00) 85 (13.25) 
Neoplasm 183 (7.50) 55 (5.41) 135 (4.98) 45 (7.01) 
Cardiovascular 84 (3.44) 51 (5.02) 43 (1.59) 20 (3.12) 
External causes 67 (2.75) 40 (3.93) 37 (1.37) 8 (1.25) 
Other 75 (3.07) 23 (2.26) 56 (2.07) 12 (1.87) 

The Cox-regression assumption was fulfilled. The statistical power was good for all-cause mortality and weaker for some specific causes of death, especially cardiovascular diseases and external causes for women.

In the first Cox-regression model, adjusted for age, mortality risk did not differ between refugees and non-refugee immigrants, either for all-cause mortality or for any of the specific causes apart from a lower neoplasm mortality risk for refugees (Table 3). Adding country of origin to the model resulted in differences in the hazard ratios between male refugees and non-refugee immigrants for cardiovascular mortality. Adding country of origin also increased the hazard ratios of cardiovascular mortality among women and of external causes among men. The results appeared to be negatively confounded by origin.

Table 3

HRs and 95% CIs (lower and upper limits) for three models (adjusted firstly for age group, secondly for age group and origin and thirdly for age group, origin and socio-economic factors) for all-cause, neoplasm, cardiovascular, external causes and other mortality for refugees compared with non-refugee immigrants by sex

Causes of death  Men
 
Women
 
  Age group Age and origin Age, origin and socio-economic factorsa Age group Age and origin Age, origin and socio-economic factorsa 
All causes Non- refugee  
 Refugee 0.96 (0.80–1.15) 1.15 (0.95–1.40) 1.10 (0.91–1.35) 1.00 (0.78–1.28) 1.09 (0.84–1.40) 1.05 (0.82–1.35) 
Neoplasms Non-refugee   
 Refugee 0.70 (0.52–0.95) 0.91 (0.66–1.26) 0.86 (0.62–1.20) 1.05 (0.74–1.47) 1.08 (0.76–1.53) 1.05 (0.74–1.50) 
Cardiovascular Non- refugee 
 Refugee 1.40 (0.99–1.99) 1.58 (1.08–2.33) 1.53 (1.04–2.24) 1.33 (0.78–2.28) 1.55 (0.89–2.68) 1.49 (0.86–2.59) 
External causes Non-refugee 
 Refugee 1.40 (0.95–2.08) 1.52 (0.97–2.38) 1.59 (1.01–2.50) 0.90 (0.42–1.94) 0.96 (0.43–2.13) 0.91 (0.41– 2.04) 
Other causes Non-refugee   
 Refugee 0.71 (0.44–1.14) 0.88 (0.53–1.45) 0.83 (0.50–1.37) 0.67 (0.36–1.26) 0.79 (0.42–1.50) 0.75 (0.39–1.42) 
Causes of death  Men
 
Women
 
  Age group Age and origin Age, origin and socio-economic factorsa Age group Age and origin Age, origin and socio-economic factorsa 
All causes Non- refugee  
 Refugee 0.96 (0.80–1.15) 1.15 (0.95–1.40) 1.10 (0.91–1.35) 1.00 (0.78–1.28) 1.09 (0.84–1.40) 1.05 (0.82–1.35) 
Neoplasms Non-refugee   
 Refugee 0.70 (0.52–0.95) 0.91 (0.66–1.26) 0.86 (0.62–1.20) 1.05 (0.74–1.47) 1.08 (0.76–1.53) 1.05 (0.74–1.50) 
Cardiovascular Non- refugee 
 Refugee 1.40 (0.99–1.99) 1.58 (1.08–2.33) 1.53 (1.04–2.24) 1.33 (0.78–2.28) 1.55 (0.89–2.68) 1.49 (0.86–2.59) 
External causes Non-refugee 
 Refugee 1.40 (0.95–2.08) 1.52 (0.97–2.38) 1.59 (1.01–2.50) 0.90 (0.42–1.94) 0.96 (0.43–2.13) 0.91 (0.41– 2.04) 
Other causes Non-refugee   
 Refugee 0.71 (0.44–1.14) 0.88 (0.53–1.45) 0.83 (0.50–1.37) 0.67 (0.36–1.26) 0.79 (0.42–1.50) 0.75 (0.39–1.42) 

aFor all-cause, neoplasms, cardiovascular and other mortality, the socio-economic factors associated with the outcome were economic activity, education and marital status. For external causes, the only socio-economic factor associated with the outcome was economic activity.

To explore whether or not all-cause and cause-specific mortality were associated with being a refugee and whether any of the explanatory variables was associated or interacted with this, we tested different Cox-regression models. The possible explanatory variables were age group, country of origin, year of arrival, marital status, education, economic activity and area of residence in Sweden. After testing various models, we concluded that all outcomes except external causes were associated with age group, country of origin, economic activity, education and marital status. Apart from the reason for migration, the external causes outcome was only associated with origin and economic activity. No interaction effects were found.

In the final model, being a refugee or not was associated with increased mortality in cardiovascular diseases for both sexes and in external causes for men (Table 3). The size of the difference in the risk of cardiovascular mortality between refugees and non-refugee immigrants was similar among women and men, but the statistical power was lower among women. Suicide accounted for about 40% of the external causes of death among men and about half of those among women. The other external causes were injuries from accidents, assaults and events of undetermined intent.

Discussion

The main finding of this study is that male refugees in Sweden have a higher risk of mortality from cardiovascular and external causes compared with male non-refugee immigrants. Socio-economic factors did not have a major impact on the over-risks. Female refugees also had a higher cardiovascular mortality, with an effect of similar size to that for men. However, possibly due to fewer deaths among refugee women, the statistical power was low.

To our knowledge, this is the first longitudinal study in which mortality is compared between refugees and non-refugee immigrants in high-income countries where the latter do not include labour migrants. Our findings confirm that an increased mortality risk exists among refugees but only as regards cardiovascular and external causes for men. Contrary to the study by DesMeules and colleagues,13 these results cannot be explained by the migrant selection process.

Studies have reported high rates of suicidal behaviour among refugees.11,12 Our study found a higher risk of death from external causes among male refugees but as the number of deaths was too small for a separate analysis, it is not clear whether this was due to an over-risk of suicide or any of the other external causes of deaths.

Limitations

The number of deaths in the different groups of origin was too small to allow separate analyses. A large part of the study population originated from former Yugoslavia. Sweden had an influx of labour market immigrants from Yugoslavia before the war in former Yugoslavia created a need for people to seek asylum in Sweden. Hence, refugees from former Yugoslavia might have better networks than other refugees in Sweden. Male and female refugees differ in their patterns of migration and asylum seeking; hence origin differs between the largest group of male refugees (from Iraq) and the largest group of female refugees (from former Yugoslavia).

The non-refugee group included persons who had been granted residence permits for family reunification with a refugee and for humanitarian reasons. Refugee status is determined by the Swedish Board of Migration, the Swedish Migration courts and the UNHCR. Each member of a family is supposed to be assessed individually in terms of asylum needs.18 However, when one member of a family arrives in Sweden before the other/s and is granted asylum, the rest of the family can obtain residence permits for family reunion. There might be persons who have been granted a residence permit for family reunification who would have been granted asylum on their own if they had applied. There might also be immigrants accepted for humanitarian reasons who are misclassified refugees.11 Refugees misclassified as family of refugees or refugees admitted for humanitarian reasons would dilute the association between reason for migration and mortality risk.

Less than 10% of the study population was excluded for leaving Sweden before death or censoring. As the exclusion rate was relatively small, there is little reason to believe that these exclusions affected the results.

Experiencing wars and disasters is not necessarily associated as such with severe psychological reactions.11,23 Rather, it seems to be personal experience of persecution, such as torture, that is associated with mental ill health.5 A meta-study found substantial differences in mental ill health outcome between refugees and non-refugees (refugees had the higher risk) even when the non-refugees had been exposed to some violence and had previous experience of war.4 This study’s finding of a higher mortality risk among refugee men points in the same direction as findings of differences in mental health between refugees and non-refugee immigrants.

Studies have shown ethnic differences in mortality in the general immigrant population in Sweden,15 including cardiovascular mortality.15 This was confirmed by the present study’s finding of origin-based differences. Origin actually confounded the age-adjusted results negatively. Both being a refugee and ethnicity seem to play a major role for mortality risks. One potential interpretation could be that in spite of different reasons for seeking refuge in Sweden, the pre- and post-migration factors associated with being a refugee are similarly harmful for refugees regardless of origin. The study highlights that among refugees, origin-based differences in mortality may conceal over-risks of mortality in cardiovascular and external causes.

What could explain the over-risk in cardiovascular mortality and external causes for men? Refugees have a higher prevalence of depression and PTSD4,5 than the general population. Depression and PTSD are independent risk factors for cardiovascular disease.7,8 Hence, the higher prevalence of mental ill health could be one explanation of the over-risk of cardiovascular mortality. Another explanation could be that exposure to refugee-specific pre- and post-migration factors is generally harmful in a manner similar to other stressors known to be associated with cardiovascular mortality,9 including war combat injury.25

Cardiovascular diseases are clearly related to lifestyle factors such as smoking habits, exercise and diet. Some lifestyle factors can be a part of strategies for coping with pre- and post-migration factors and are likely to play a moderating or mediating role in the association between exposure to refugee-specific pre- and post-migration factors and cardiovascular mortality.

In order to prevent cardiovascular diseases among refugee men, it is important that health-care facilities become aware of refugees as a risk group not just for poor mental health but also for cardiovascular mortality. A method for enhancing health literacy among refugees has been tested in the south of Sweden with promising results.26

Sweden is obliged to give asylum to refugees according to the UN Refugee Convention. Although many countries have ratified this Convention, refugees are treated differently depending on the host country. As there are few studies on refugee mortality, it is difficult to tell whether the results of this study are applicable to refugees in other high-income countries.

Mortality from external causes included not only cases of suicide but also other causes such as injuries from accidents and assault. The higher risk of mortality from external causes among refugees compared with non-refugees was evident only among refugee men and needs to be further investigated.

This study found a gender difference in that it was only male refugees who had a higher risk of mortality from cardiovascular and external causes. Besides low statistical power among women, other explanations are possible. One could be that women and men react in different ways to pre- and post-migration stressors. Another is that the results correspond to gender differences in the general population; with a higher risk of mental ill health for women but lower mortality among women compared with men.27 In a cross-sectional study on the same population as in this cohort, refugee women were found to have a higher likelihood of mental ill health than non-refugee immigrant women.28

Eight years of follow-up is a limited time and it would be interesting to follow the cohort for longer. Further research should focus on whether the difference in mental health between refugees and non-refugee immigrants is associated with higher mortality in cardiovascular diseases and external causes. Comparing mortality between refugees in different high-income countries would add to the knowledge of refugee mortality. Yet another focus in future studies would be on strategies to prevent mortality in refugees.

Refugees leave their countries to protect their life and freedom. This study shows that although life-threatening persecution is avoided by asylum in another country, for some the past violations continue to be a threat to life in the new host country. Still, it is not certain that the over-risk of cardiovascular mortality and mortality in external causes among refugee men is due only to the pre-migration factors. The risk might be induced by pre-migration stress, but be interlinked with the asylum process and post-migration factors. The asylum-seeking process seems to be particularly important, as some of the non-refugees are families of refugees who could have similar experiences to those of the refugees apart from the asylum process.

Conclusions

Our study confirms that there are differences in the risk of mortality between refugees and non-refugee immigrants and extends existing knowledge by showing that these differences are found in cardiovascular mortality and external causes for men. As bias was eliminated by not including labour migrants in the reference group, these findings could be explained by the refugee experience. This study suggests that the refugee experience resembles other stressors in terms of the association with cardiovascular mortality.

Funding

FAS, The Swedish Council for Working Life and Research (FAS-2007-1961).

Acknowledgement

Special thanks to Patrick Hort for language editing of the text.

Conflict of interest: None declared.

KEY MESSAGES

  • Refugee immigrants in high-income countries have poorer health than other immigrant groups but little is known about their mortality.

  • The main finding of this study is that refugee men in Sweden have a higher mortality risk in cardiovascular and external causes compared with male non-labour non-refugee immigrants.

  • This study suggests that the refugee experience resembles other stressors in terms of the association with cardiovascular mortality.

References

1
UN High commissioner for refugees (UNHCR)
 
http://www.unhcr.org/ (3 March 2012, date last accessed)
2
Malmusi
D
Borrell
C
Benach
J
Migration-related health inequalities: showing the complex interactions between gender, social class and place of origin
Soc Sci Med
 , 
2010
, vol. 
71
 (pg. 
1610
-
19
)
3
Jamil
H
Nassar-McMillanb
S
Lambert
R
Wangd
Y
Ager
J
Arnetz
B
Pre- and post-displacement stressors and time of migration as related to self-rated health among Iraqi immigrants and refugees in Southeast Michigan
Med Confl Surviv
 , 
2010
, vol. 
26
 (pg. 
207
-
22
)
4
Porter
M
Haslam
N
Predisplacement and postdisplacement factors associated with mental health of refugees and internally displaced persons: a meta-analysis
JAMA
 , 
2005
, vol. 
294
 (pg. 
602
-
12
)
5
Steel
Z
Chey
T
Silove
D
Marnane
C
Bryant
RA
van Ommeren
M
Association of torture and other potentially traumatic events with mental health outcomes among populations exposed to mass conflict and displacement: a systematic review and meta-analysis
JAMA
 , 
2009
, vol. 
302
 (pg. 
537
-
49
)
6
Mollica
RF
Sarajlic
N
Chernoff
M
Lavelle
J
Vukovic
IS
Massagli
MP
Longitudinal study of psychiatric symptoms, disability, mortality, and emigration among Bosnian refugees
JAMA
 , 
2001
, vol. 
286
 (pg. 
546
-
54
)
7
Glassman
AH
Depression and cardiovascular comorbidity
Dialogues Clin Neurosci
 , 
2007
, vol. 
9
 (pg. 
9
-
17
)
8
McFarlane
AC
The long-term costs of traumatic stress: intertwined physical and psychological consequences
World Psychiatry
 , 
2010
, vol. 
9
 (pg. 
3
-
10
)
9
Dimsdale
JE
Psychological stress and cardiovascular disease
J Am Coll Cardiol
 , 
2008
, vol. 
51
 (pg. 
1237
-
46
)
10
Kinzie
JD
Riley
C
McFarland
B
Hayes
M
Boehnlein
J
Leung
P
, et al.  . 
High prevalence rates of diabetes and hypertension among refugee psychiatric patients
J Nerv Ment Dis
 , 
2008
, vol. 
196
 (pg. 
108
-
12
)
11
Blight
JK
Ekblad
S
Lindencrona
F
Shahnavaz
S
Promoting mental health and preventing mental disorder among refugees in western countries
Int J Ment Health Promot
 , 
2009
, vol. 
11
 (pg. 
33
-
44
)
12
Ferrada-Noli
M
Asberg
M
Ormstad
K
Lundin
T
Sundbom
E
Suicidal behavior after severe trauma. Part 1: PTSD diagnoses, psychiatric comorbidity, and assessments of suicidal behavior
J Trauma Stress
 , 
1998
, vol. 
11
 (pg. 
103
-
12
)
13
DesMeules
M
Gold
J
McDermott
S
Cao
Z
Payne
J
Lafrance
B
, et al.  . 
Disparities in mortality patterns among Canadian immigrants and refugees, 1980–1998: results of a national cohort study
J Immigr Health
 , 
2005
, vol. 
7
 (pg. 
221
-
32
)
14
Gadd
M
Johansson
SE
Sundquist
J
Wandell
P
Are there differences in all-cause and coronary heart disease mortality between immigrants in Sweden and in their country of birth? A follow-up study of total populations
BMC Public Health
 , 
2006
, vol. 
6
 pg. 
102
 
15
Albin
B
Hjelm
K
Ekberg
J
Elmstahl
S
Higher mortality and different pattern of causes of death among foreign-born compared to native Swedes 1970–1999
J Immigr Minor Health
 , 
2006
, vol. 
8
 (pg. 
101
-
13
)
16
Hajat
A
Blakely
T
Dayal
S
Jatrana
S
Do New Zealand’s immigrants have a mortality advantage? Evidence from the New Zealand Census-Mortality Study
Ethn Health
 , 
2010
, vol. 
15
 (pg. 
531
-
47
)
17
Razum
O
Zeeb
H
Akgun
HS
Yilmaz
S
Low overall mortality of Turkish residents in Germany persists and extends into a second generation: merely a healthy migrant effect?
Trop Med Int Health
 , 
1998
, vol. 
3
 (pg. 
297
-
303
)
18
The Swedish Board of Migration
 
http://www.migrationsverket.se/english (20 May 2011, date last accessed)
19
Weitoft
GR
Gullberg
A
Hjern
A
Rosen
M
Mortality statistics in immigrant research: method for adjusting underestimation of mortality
Int J Epidemiol
 , 
1999
, vol. 
28
 (pg. 
756
-
63
)
20
Statistics Sweden SCB
 
http://www.scb.se/ (20 Feb 2012, date last accessed)
21
The National Board of Health and Welfare
Dödsorsaksstatistik
  
http://www.sos.se/ (3 March 2012, date last accessed)
22
Nekby
L
How Long Does It Take to Integrate? Employment Convergence of Immigrants and Natives in Sweden
 , 
2002
Stockholm
FIEF
23
Silove
D
Ekblad
S
Mollica
R
The rights of the severely mentally ill in post-conflict societies
Lancet
 , 
2000
, vol. 
355
 (pg. 
1548
-
49
)
24
Hosmer
DW
Lemeshow
S
May
S
Applied Survival Analysis: Regression Modeling of Time-to-Event Data
 , 
2008
2nd
Hoboken, NJ
Wiley-Interscience
25
Kunnas
T
Solakivi
T
Renko
J
Kalela
A
Nikkari
ST
Late-life coronary heart disease mortality of Finnish war veterans in the TAMRISK study, a 28-year follow-up
BMC Public Health
 , 
2011
, vol. 
11
 pg. 
71
 
26
Eriksson Sjöö
T
Ekblad
S
Malmsten
J
Hälsoskola för nyanlända flyktingar
Migrationens utmaningar inom hälsa, omsorg och vård
 , 
2010
Malmö stad
FoU-rapport
(pg. 
33
-
52
)
27
Sen
G
Östlin
P
Gender Equity in Health: The Shifting Frontiers of Evidence and Action
 , 
2010
New York
Routledge
28
Hollander
AC
Bruce
D
Burstrom
B
Ekblad
S
Gender-related mental health differences between refugees and non-refugee immigrants: a cross-sectional register-based study
BMC Public Health
 , 
2011
, vol. 
11
 pg. 
180