Abstract
Objectives. We studied mortality differentials between specific groups of foreign-born immigrants in Sweden and whether socioeconomic position (SEP) could account for such differences.
Methods. We conducted a follow-up study of 1 997 666 men and 1 964 965 women ages 30 to 65 years based on data from national Swedish total population registers. We examined mortality risks in the 12 largest immigrant groups in Sweden between 1998 and 2006 using Cox regression. We also investigated deaths from all causes, circulatory disease, neoplasms, and external causes.
Results. We found higher all-cause mortality among many immigrant categories, although some groups had lower mortality. When studying cause-specific mortality, we found the largest differentials in deaths from circulatory disease, whereas disparities in mortality from neoplasms were smaller. SEP, especially income and occupational class, accounted for most of the mortality differentials by country of birth.
Conclusions. Our findings stressed that different aspects of SEP were not interchangeable in relation to immigrant health. Although policies aimed at improving immigrants’ socioeconomic conditions might be beneficial for health and longevity, our findings indicated that such policies might have varying effects depending on the specific country of origin and cause of death.
Like many other Western countries, Sweden has experienced an increase in migration and an associated change in its ethnic mix. At the end of 2006, 12.9% of the population had been born outside Sweden.1 Sweden has a migrant population of approximately the same order as the United States, with a growing number of native-born children of migrants.2 Because the number of migrants residing in Sweden constitutes a considerable share of the population, and because migrants have been singled out as a social category with troublesome living and social conditions, especially in the aftermath of the severe recession during the first half of the 1990s,3 the study of migrants’ health and socioeconomic position (SEP) adjacent to this period of time deserves further attention.
Findings have shown higher all-cause mortality,4–6 cardiovascular disease (CVD) mortality,4,7–10 risk of poor self-rated health,11–13 higher rates of long-term illness,14 higher risks of stroke,15 elevated suicide rates,16 and higher cancer rates17,18 among groups of foreign-born immigrants in Sweden. More specifically, studies have found an increased risk of mortality from all causes among Finnish-born men and an increased risk of CVD among women from Finland and Eastern Europe,4 whereas another study found higher mortality among Nordic immigrants.5 It has also been found that Finnish immigrants have the highest and Middle Easterners the lowest odds for suicide death.19 However, higher all-cause and cause-specific mortalities have not been found in all immigrant categories.18,20 Lower overall death rates have been found among immigrants born in Turkey, Southern Europe, Latin America, Asia, and Africa,21 whereas cancer risks by country of birth depended on the type of cancer.18 However, most of the aforementioned studies examined very broad categories of immigrants that might “mask” mortality differentials by specific country of birth. They also primarily studied the contribution of only 1 single aspect of SEP, such as education4 occupational status,5,20 or income,21 although it has been shown that that the SEP–mortality relationship varied by the dimension of the SEP examined.22
There is voluminous empirical support for health inequalities by SEP, even in an egalitarian society such as Sweden.12,22–24 People with higher SEP (education, occupational class, and income) live longer and are healthier than those in disadvantaged positions in society. Findings indicate that many immigrant groups experience a disadvantaged SEP in Sweden.25 Compared with Swedes, most immigrant groups have lower income26 and a strikingly high prevalence of relative poverty,27 and immigrants born in Southern Europe and the Middle East are overrepresented among the lower occupational classes.19 Nevertheless, studies suggest that immigrants, such as those from Eastern and Southern Europe, as well as other Western countries, have a relatively high educational level compared with the Swedish-born population.4 However, immigrants may still experience a double burden in society, that is, they experience disadvantages from both their ethnic minority status and a low SEP.
Most studies on immigration and health regard SEP as a confounder that should be routinely adjusted for to accurately scrutinize the “pure” effect of immigrant status on health. Some claim that low SEP is instead a consequence of the immigration process, that is, that immigrant or ethnic minority status influences people’s SEP, at least with regard to income and social class.28 The high levels of occupational mismatch among immigrants, that is, a high percentage of immigrants are overeducated with respect to the educational requirements for their job, support the notion that low SEP may be a consequence of migration.29,30 Following this notion, SEP should be considered a mediator or mechanism linking immigrant status and health. This implies that a more integrative view on health disparities between immigrant groups and natives is required, for example, one that considers the contribution of different indicators of SEP to health differentials between immigrant groups. Accordingly, findings suggest that indicators of SEP are not interchangeable in relation to health.20
The contribution of SEP to mortality differentials between immigrants may also differ by cause of death.31 Some causes of deaths are more heavily stratified by adult SEP exposure.23 Conditions like diabetes and circulatory disease (CD) can be highly responsive to exposures that are influenced by SEP (e.g., quality of nutrition, housing, or living conditions) and preventive health care (health screening, disease detection, and long-term management). Acute external causes, such as accidents or injuries, may also directly respond to social and environmental conditions, whereas SEP may be less influential on diseases with a strong genetic component, such as certain types of neoplasms and inherited chronic diseases (e.g., cystic fibrosis).
Although several Swedish studies have examined immigration and health, a limitation of most previous studies is the fact that a broad and inconsistent categorization of immigrants was applied, which reduced the comparability between studies. The fact that the contribution of SEP to mortality differentials between immigrants may vary by cause of death and SEP indicator highlights the importance of mutually studying the relationships among the country of birth, SEP, and cause-specific mortality in the same study. Furthermore, the constant influx of immigrants to Sweden and changes in social and economic conditions in society stress the need for updated information on the health situation of new cohorts of immigrants.
Our study aim was to disentangle the relationships among immigrant status, different aspects of SEP, and mortality. We examined whether mortality differed between natives and the 12 largest groups of foreign-born immigrants residing in Sweden, and what the role of SEP had with regard to these differentials. We scrutinized relationships with all-cause and cause-specific mortality (CVD, neoplasms, and external causes) and examined whether education, income, and occupational class (both separately and jointly) accounted for mortality differentials in immigrant men and women.
METHODS
A great advantage offered by the use of the personal identity number in the Swedish setting is the possibility to study these differentials by examining registers covering the total population. This entailed a very large sample size, enabling longitudinal follow-up with information about mortality and other included variables. The large sample size gave us an opportunity to examine specific groups categorized by country of birth. The register-based data used in the present study came from a multiple-linked data of national routine registers, The Swedish Work and Mortality Data (HSIA). The data covered the total population of Sweden born before 1986, who were still alive in 1990 or 1980 (including individuals immigrating after 1990), with a possible follow-up between 1981 and 2009. Linkages were possible via the 10-digit personal identity number, which was replaced by a serial number to ensure anonymity. The baseline sample in this study consisted of individuals residing in Sweden in 1997, who did not emigrate during the follow-up period (1998–2006), and who were between 30 and 65 years old at baseline. We examined 1 997 666 men and 1 964 965 women. Deaths from all causes, CD (International Classification of Disease, 10th Revision32 codes I00-I99), neoplasms (C00-D48), and external causes (V01-Y98) were examined using Cox regression.
For each of the SEP indicators, we calculated the explained fraction (XF%) as the percentage of the excess risk, expressed as hazard ratios (HRs) that were explained by the respective indicator. The XF% estimated the proportion of excess risk explained by the mediating factors, and was used in previous studies.33,34 XF% was calculated from the HRs among groups of foreign-born individuals with native-born Swedes as the reference group, before adjustment (HR) and after adjustment for SEP (HR*):
 |
In cases where the HR did not suggest an excess risk, we decided not to calculate the XF%. XF% for immigrant groups with lower HRs than natives after adjustment for SEP indicators was set to 100% and referred to the fact that there was no excess risk left to be explained after adjustment for SEP.
Variables and Socioeconomic Position
Country of birth was categorized according to the country of birth of the 12 largest groups of foreign-born individuals in Sweden in 2010 (Finland, Iraq, the former Yugoslavia, Poland, Iran, Bosnia, Germany, Denmark, Norway, Turkey, Somalia, and Thailand), with the Swedish-born population as reference category.35 Because the registers distinguished migrants from Bosnia from other regions of the former Yugoslavia, we decided to follow this classification in our study.
Education in the register-based data were divided into the categories “postgraduate studies,” “postgymnasium, 3 years or more,” “postgymnasium, less than 3 years,” “upper secondary school, 2–3 years,” “upper secondary school, maximum 2 years,” “compulsory school, 9 years,” “compulsory school, less than 9 years,” and “not classified” (i.e., missing information from the educational system). Occupational class included the categories “higher non-manual,” “intermediate non-manual,” “lower non-manual,” “skilled manual worker,” “unskilled manual worker,” “farmer,” “self-employed,” and “not classified.”36 The last category included unemployed individuals, people without an occupation (e.g., students and early retired), and those who otherwise lacked occupational information. The variable was based on the Swedish socioeconomic classification (labeled SEI), which first separates employees from self-employed (including farmers), then manual from nonmanual occupations, and then within these groups divides further based on the education level normally required. Disposable income (i.e., income from all sources and taxes deducted) was based on data from various income registers and was analyzed as a continuous variable based on information on the individual part of the household’s annual disposable income. Our income measure reflected the mean income of each individual between 1998 and 2006. The aforementioned categorization of education and occupational class is used in Tables 1 through 4. For the sake of brevity, we used broader occupational classes and educational groups as data available as a supplement to the online version of this article at http://www.ajph.org.
TABLE 1—
Hazard Ratios of Mortality From All Causes by Country of Birth Among Immigrant Men and Women in Sweden Aged 30–65 Years in 1997: Follow-Up 1998–2006
| Model 2b |
Model 3c |
Model 4d |
Model 5e |
|||||||
| Country | Model 1,a HR (95% CI) | HR (95% CI) | XF%f | HR (95% CI) | XF%f | HR (95% CI) | XF%f | HR (95% CI) | XF%f | Deaths, No. |
| Men | ||||||||||
| Sweden (Ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 96 658 | ||||
| Finland | 1.67 (1.63, 1.72) | 1.47 (1.43, 1.51) | 30 | 1.40 (1.35, 1.42) | 40 | 1.40 (1.37, 1.44) | 40 | 1.30 (1.26, 1.33) | 55 | 6549 |
| Iraq | 0.79 (0.70, 0.91) | 0.89 (0.78, 1.02) | … | 0.63 (0.57, 0.73) | … | 0.59 (0.52, 0.67) | … | 0.49 (0.43, 0.57) | … | 214 |
| Yugoslavia | 1.25 (1.19, 1.32) | 1.27 (1.21, 1.34) | … | 1.05 (0.99, 1.01) | 80 | 1.01 (0.96, 1.06) | 96 | 0.90 (0.85, 0.95) | 100 | 1430 |
| Poland | 1.08 (0.97, 1.20) | 1.18 (1.06, 1.30) | … | 0.91 (0.82, 1.00) | 100 | 0.95 (0.85, 1.05) | 100 | 0.86 (0.78, 0.95) | 100 | 365 |
| Iran | 0.66 (0.59, 0.73) | 0.72 (0.64, 0.81) | … | 0.49 (0.44, 0.55) | … | 0.49 (0.44, 0.55) | … | 0.41 (0.37, 0.46) | … | 307 |
| Bosnia | 1.23 (1.14, 1.33) | 1.35 (1.24, 1.46) | … | 1.03 (1.95, 1.12) | 87 | 0.87 (0.80, 0.94) | 100 | 0.73 (0.67, 0.79) | 100 | 615 |
| Germany | 0.98 (0.92, 1.05) | 1.05 (0.98, 1.13) | … | 0.96 (0.90, 1.03) | … | 0.97 (0.90, 1.04) | … | 0.94 (0.88, 1.01) | … | 787 |
| Denmark | 1.17 (1.09, 1.24) | 1.08 (1.01, 1.16) | 53 | 1.01 (0.95, 1.08) | 94 | 1.01 (0.95, 1.08) | 94 | 0.96 (0.90, 1.02) | 100 | 900 |
| Norway | 1.19 (1.10, 1.30) | 1.13 (1.05, 1.23) | 32 | 1.08 (1.00, 1.17) | 58 | 1.04 (0.96, 1.12) | 79 | 1.00 (0.92, 1.08) | 100 | 623 |
| Turkey | 0.93 (0.83, 1.04) | 0.96 (0.86, 1.07) | … | 0.74 (0.67, 0.83) | … | 0.81 (0.72, 0.90) | … | 0.76 (0.55, 0.69) | … | 319 |
| Somalia | 1.63 (1.27, 2.10) | 1.75 (1.36, 2.26) | 19 | 1.23 (0.97, 1.62) | 63 | 1.16 (0.90, 1.49) | 75 | 0.93 (0.72, 1.20) | 100 | 59 |
| Thailand | 1.42 (0.81, 2.50) | 1.25 (0.71, 2.19) | 40 | 0.90 (0.51, 1.59) | 100 | 1.03 (0.59, 1.82) | 93 | 0.80 (0.46, 1.42) | 100 | 12 |
| Women | ||||||||||
| Sweden (Ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 62 028 | ||||
| Finland | 1.19 (1.15, 1.22) | 1.09 (1.06, 1.13) | 53 | 1.08 (1.06, 1.12) | 58 | 1.06 (1.02, 1.09) | 68 | 1.05 (1.01, 1.09) | 74 | 4160 |
| Iraq | 0.94 (0.78, 1.14) | 0.70 (0.58, 0.85) | … | 0.78 (0.65, 0.95) | … | 0.64 (0.53, 0.77) | … | 0.51 (0.42, 0.61) | … | 107 |
| Yugoslavia | 1.10 (1.03, 1.19) | 0.93 (0.86, 1.00) | 100 | 0.93 (0.87, 1.00) | 100 | 0.81 (0.75, 0.88) | 100 | 0.74 (0.69, 0.80) | 100 | 692 |
| Poland | 1.08 (1.00, 1.18) | 1.09 (1.00, 1.18) | … | 0.93 (0.85, 1.09) | 100 | 0.88 (0.80, 0.96) | 100 | 0.85 (0.78, 0.93) | 100 | 511 |
| Iran | 0.57 (0.48, 0.67) | 0.49 (0.41, 0.58) | … | 0.45 (0.38, 0.54) | … | 0.38 (0.32, 0.45) | … | 0.34 (0.29, 0.40) | … | 138 |
| Bosnia | 1.16 (1.06, 1.28) | 0.85 (0.77, 0.95) | 100 | 0.91 (0.83, 1.01) | 100 | 0.75 (0.69, 0.83) | 100 | 0.59 (0.53, 0.65) | 100 | 428 |
| Germany | 0.85 (0.79, 0.93) | 0.87 (0.80, 0.95) | … | 0.80 (0.73, 0.87) | … | 0.79 (0.73, 0.86) | … | 0.78 (0.72, 0.85) | … | 556 |
| Denmark | 1.33 (1.23, 1.45) | 1.27 (1.17, 1.38) | 18 | 1.19 (1.10, 1.29) | 42 | 1.20 (1.10, 1.30) | 39 | 1.14 (1.05, 1.24) | 68 | 567 |
| Norway | 1.21 (1.11, 1.31) | 1.11 (1.02, 1.21) | 48 | 1.09 (1.00, 1.19) | 57 | 1.06 (0.97, 1.15) | 71 | 1.03 (0.95, 1.12) | 58 | 561 |
| Turkey | 0.88 (0.76, 1.03) | 0.67 (0.57, 0.78) | … | 0.68 (0.59, 0.80) | … | 0.64 (0.55, 0.75) | … | 0.49 (0.42, 0.57) | … | 167 |
| Somalia | 1.19 (0.84, 1.70) | 0.71 (0.50, 1.02) | 100 | 1.08 (0.76, 1.53) | 58 | 0.78 (0.55, 1.11) | 100 | 0.60 (0.42, 0.86) | 100 | 31 |
| Thailand | 1.15 (0.88, 1.49) | 0.83 (0.64, 1.09) | 100 | 0.82 (0.63, 1.07) | 100 | 0.83 (0.64, 1.08) | 100 | 0.65 (0.50, 0.84) | 100 | 55 |
Note. CI = confidence interval; HR = hazard ratio; XF% = explained fraction.
Age.
Age and education.
Age and income.
Age and occupational class position.
Age and all socioeconomic position variables.
XF% is not calculated (…) for immigrant groups with lower HR than native-born because there is no excess risk to “explain” by socioeconomic position. XF% for immigrant groups with lower HRs of mortality after adjustment for socioeconomic position indicators is set to 100% and refers to the fact there is no excess risk left to be explained after adjustment for socioeconomic position.
TABLE 2—
Hazard Ratios of Mortality From Circulatory Diseases by Country of Birth Among Immigrant Men and Women in Sweden Aged 30–65 Years in 1997: Follow-Up 1998–2006
| Model 2b |
Model 3c |
Model 4d |
Model 5e |
|||||||
| Country | Model 1,a HR (95% CI) | HR (95% CI) | XF%f | HR (95% CI) | XF%f | HR (95% CI) | XF%f | HR (95% CI) | XF%f | Deaths, No. |
| Men | ||||||||||
| Sweden (Ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 28 196 | ||||
| Finland | 1.81 (1.73, 1.89) | 1.57 (1.50, 1.64) | 30 | 1.47 (1.40, 1.53) | 42 | 1.49 (1.43, 1.56) | 40 | 1.37 (1.31, 1.43) | 54 | 2103 |
| Iraq | 1.02 (0.80, 1.29) | 1.18 (0.93, 1.50) | … | 0.78 (0.62, 1.00) | 100 | 0.76 (0.60, 0.96) | 100 | 0.63 (0.50, 0.80) | 100 | 67 |
| Yugoslavia | 1.24 (1.13, 1.37) | 1.27 (1.15, 1.40) | … | 1.01 (0.91, 1.11) | 96 | 0.99 (0.90, 1.10) | 100 | 0.88 (0.80, 0.97) | 100 | 411 |
| Poland | 1.20 (1.00, 1.45) | 1.35 (1.12, 1.62) | … | 0.98 (0.81, 1.18) | 100 | 1.05 (0.87, 1.26) | 75 | 0.95 (0.79, 1.14) | 100 | 112 |
| Iran | 0.66 (0.53, 0.83) | 0.75 (0.60, 0.94) | … | 0.47 (0.38, 0.59) | … | 0.49 (0.39, 0.61) | … | 0.40 (0.32, 0.51) | … | 77 |
| Bosnia | 1.29 (1.11, 1.49) | 1.43 (1.23, 1.67) | … | 1.02 (0.88, 1.18) | 93 | 0.91 (0.78, 1.05) | 100 | 0.75 (0.65, 0.88) | 100 | 180 |
| Germany | 0.88 (0.77, 1.00) | 0.97 (0.85, 1.11) | … | 0.86 (0.76, 0.99) | … | 0.88 (0.77, 1.10) | … | 0.87 (0.76, 0.99) | … | 218 |
| Denmark | 1.17 (1.04, 1.32) | 1.10 (0.97, 1.23) | 41 | 1.00 (0.88, 1.12) | 100 | 1.02 (0.90, 1.15) | 88 | 0.95 (0.85, 1.07) | 100 | 278 |
| Norway | 1.08 (0.93, 1.25) | 1.04 (0.90, 1.21) | 50 | 0.98 (0.84, 1.14) | 100 | 0.94 (0.81, 1.10) | 100 | 0.92 (0.79, 1.07) | 100 | 171 |
| Turkey | 0.96 (0.78, 1.18) | 1.01 (0.82, 1.24) | … | 0.74 (0.60, 0.91) | … | 0.83 (0.67, 1.02) | … | 0.61 (0.50, 0.75) | … | 90 |
| Somalia | 1.29 (0.71, 2.33) | 1.41 (0.78, 2.55) | … | 0.95 (0.52, 1.78) | 100 | 0.92 (0.51, 1.65) | 100 | 0.73 (0.41, 1.33) | 100 | 11 |
| Thailand | 1.93 (0.72, 5.14) | 1.66 (0.62, 4.42) | 29 | 1.12 (0.42, 2.98) | 87 | 1.33 (0.50, 3.54) | 65 | 1.02 (0.38, 2.70) | 98 | 4 |
| Women | ||||||||||
| Sweden (Ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 10 978 | ||||
| Finland | 1.46 (1.36, 1.56) | 1.30 (1.21, 1.39) | 35 | 1.30 (1.21, 1.39) | 35 | 1.22 (1.14, 1.31) | 52 | 1.21 (1.13, 1.29) | 54 | 930 |
| Iraq | 1.58 (1.09, 2.28) | 1.02 (0.69, 1.50) | 97 | 1.24 (0.85, 1.79) | 59 | 0.91 (0.63, 1.32) | 100 | 0.71 (0.48, 1.03) | 100 | 28 |
| Yugoslavia | 1.51 (1.29, 1.77) | 1.17 (1.00, 1.37) | 67 | 1.21 (1.04, 1.41) | 59 | 0.95 (0.81, 1.11) | 100 | 0.86 (0.73, 1.01) | 100 | 161 |
| Poland | 1.06 (0.85, 1.32) | 1.08 (0.87, 1.35) | … | 0.84 (0.68, 1.05) | 100 | 0.77 (0.62, 0.96) | 100 | 0.77 (0.62, 0.96) | 100 | 80 |
| Iran | 0.48 (0.30, 0.76) | 0.38 (0.24, 0.60) | … | 0.35 (0.22, 0.56) | … | 0.27 (0.17, 0.43) | … | 0.24 (0.15, 0.38) | … | 18 |
| Bosnia | 1.43 (1.17, 1.76) | 0.92 (0.73, 1.16) | 100 | 1.04 (0.85, 1.28) | 91 | 0.79 (0.64, 0.97) | 100 | 0.60 (0.48, 0.75) | 100 | 92 |
| Germany | 0.81 (0.67, 0.98) | 0.85 (0.70, 1.03) | … | 0.74 (0.61, 0.90) | … | 0.74 (0.61, 0.89) | … | 0.75 (0.62, 0.90) | … | 105 |
| Denmark | 1.27 (1.04, 1.54) | 1.19 (0.98, 1.45) | 30 | 1.09 (0.89, 1.32) | 67 | 1.09 (0.89, 1.32) | 67 | 1.02 (0.84, 1.24) | 93 | 100 |
| Norway | 1.40 (1.17, 1.68) | 1.26 (1.05, 1.50) | 35 | 1.24 (1.04, 1.48) | 40 | 1.16 (0.97, 1.39) | 60 | 1.13 (0.94, 1.35) | 68 | 120 |
| Turkey | 1.25 (0.92, 1.72) | 0.84 (0.61, 1.16) | … | 0.89 (0.65, 1.22) | 100 | 0.78 (0.57, 1.06) | 100 | 0.56 (0.41, 0.78) | 100 | 39 |
| Somalia | 1.16 (0.48, 2.78) | 0.60 (0.25, 1.46) | … | 1.01 (0.42, 2.42) | 94 | 0.66 (0.27, 1.57) | 100 | 0.50 (0.20, 1.19) | 100 | 5 |
| Thailand | 2.19 (1.27, 3.78) | 1.36 (0.79, 2.35) | 70 | 1.37 (0.79, 2.35) | 69 | 1.35 (0.78, 2.33) | 71 | 0.98 (0.57, 1.69) | 100 | 13 |
Note. CI = confidence interval; HR = hazard ratio; XF% = explained fraction.
Age.
Age and education.
Age and income.
Age and occupational class position.
Age and all socioeconomic position variables.
XF% is not calculated (…) for immigrant groups with lower HRs than native born because there is no excess risk to “explain” by socioeconomic position. XF% for immigrant groups with lower HRs of mortality after adjustment for socioeconomic position indicators is set to 100% and refers to the fact there is no excess risk left to be explained after adjustment for socioeconomic position.
TABLE 3—
Hazard Ratio of Mortality From Neoplasms by Country of Birth Among Immigrant Men and Women in Sweden Aged 30–65 Years in 1997: Follow-Up 1998–2006
| Mode 2b |
Model 3c |
Model 4d |
Model 5e |
|||||||
| Country | Model 1,a HR (95% CI) | HR (95% CI) | XF%f | HR (95% CI) | XF%f | HR (95% CI) | XF%f | HR (95% CI) | XF%f | Deaths, No. |
| Men | ||||||||||
| Sweden (Ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 28 386 | ||||
| Finland | 1.27 (1.20, 1.33) | 1.19 (1.13, 1.25) | 30 | 1.16 (1.10, 1.22) | 41 | 1.16 (1.10, 1.22) | 41 | 1.11 (1.06, 1.17) | 59 | 1480 |
| Iraq | 0.92 (0.72, 1.18) | 0.99 (0.77, 1.27) | … | 0.81 (0.63, 1.04) | … | 0.77 (0.60, 0.99) | … | 0.72 (0.56, 0.92) | … | 62 |
| Yugoslavia | 1.30 (1.19, 1.43) | 1.30 (1.18, 1.43) | 0 | 1.17 (1.07, 1.29) | 43 | 1.14 (1.04, 1.25) | 53 | 1.08 (0.98, 1.20) | 73 | 434 |
| Poland | 0.98 (0.80, 1.20) | 1.03 (0.84, 1.26) | … | 0.89 (0.72, 1.09) | … | 0.90 (0.74, 1.11) | … | 0.88 (0.72, 1.08) | … | 92 |
| Iran | 0.60 (0.47, 0.76) | 0.63 (0.50, 0.80) | … | 0.51 (0.41, 0.65) | … | 0.50 (0.40, 0.64) | … | 0.47 (0.37, 0.59) | … | 71 |
| Bosnia | 1.78 (1.58, 2.02) | 1.89 (1.66, 2.15) | … | 1.57 (1.39, 1.78) | 27 | 1.43 (1.26, 1.62) | 45 | 1.34 (1.18, 1.52) | 56 | 252 |
| Germany | 1.04 (0.93, 1.18) | 1.09 (0.96, 1.23) | … | 1.04 (0.92, 1.17) | 0 | 1.03 (0.91, 1.17) | 25 | 1.04 (0.92, 1.17) | 0 | 259 |
| Denmark | 1.31 (1.17, 1.46) | 1.27 (1.13, 1.42) | 13 | 1.22 (1.09, 1.36) | 29 | 1.21 (1.08, 1.35) | 32 | 1.19 (1.06, 1.33) | 39 | 311 |
| Norway | 1.12 (0.97, 1.30) | 1.12 (0.96, 1.29) | 0 | 1.01 (0.94, 1.26) | 92 | 1.06 (0.92, 1.23) | 50 | 1.06 (0.92, 1.23) | 50 | 179 |
| Turkey | 0.96 (0.78, 1.18) | 0.96 (0.78, 1.17) | 0 | 0.84 (0.68, 1.03) | … | 0.86 (0.70, 1.05) | … | 0.76 (0.61, 0.93) | … | 91 |
| Somalia | 1.71 (1.03, 2.80) | 1.79 (1.08, 2.97) | … | 1.47 (0.89, 2.44) | 34 | 1.39 (0.84, 2.30) | 45 | 1.27 (0.77, 2.11) | 62 | 15 |
| Thailand | 0.94 (0.24, 3.77) | 0.91 (0.23, 3.63) | … | 0.76 (0.19, 3.04) | … | 0.80 (0.20, 3.20) | … | 0.73 (0.18, 2.90) | … | 2 |
| Women | ||||||||||
| Sweden (Ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 28 258 | ||||
| Finland | 0.99 (0.94, 1.04) | 0.95 (0.90, 1.00) | … | 0.93 (0.89, 0.98) | … | 0.94 (0.89, 0.99) | … | 0.94 (0.89, 0.99) | … | 1577 |
| Iraq | 0.92 (0.70, 1.22) | 0.87 (0.65, 1.16) | … | 0.79 (0.59, 1.04) | … | 0.80 (0.60, 1.06) | … | 0.73 (0.55, 0.97) | … | 48 |
| Yugoslavia | 0.96 (0.86, 1.09) | 0.90 (0.80, 1.02) | … | 0.85 (0.75, 0.95) | … | 0.85 (0.76, 0.96) | … | 0.82 (0.73, 0.92) | … | 276 |
| Poland | 1.03 (0.91, 1.19) | 1.05 (0.92, 1.20) | … | 0.93 (0.81, 1.06) | 100 | 0.95 (0.83, 1.09) | 100 | 0.93 (0.81, 1.06) | 100 | 224 |
| Iran | 0.55 (0.43, 0.71) | 0.53 (0.41, 0.69) | … | 0.46 (0.36, 0.60) | … | 0.47 (0.37, 0.61) | … | 0.44 (0.34, 0.57) | … | 62 |
| Bosnia | 1.09 (0.94, 1.26) | 1.03 (0.88, 1.20) | 67 | 0.90 (0.78, 1.04) | 100 | 0.93 (0.80, 1.07) | 100 | 0.84 (0.72, 0.98) | 100 | 184 |
| Germany | 0.85 (0.75, 0.97) | 0.86 (0.76, 0.97) | … | 0.82 (0.72, 0.93) | … | 0.83 (0.73, 0.94) | … | 0.82 (0.72, 0.92) | … | 251 |
| Denmark | 1.37 (1.21, 1.54) | 1.37 (1.18, 1.51) | 0 | 1.27 (1.12, 1.43) | 27 | 1.31 (1.16, 1.48) | 16 | 1.27 (1.12, 1.43) | 27 | 265 |
| Norway | 1.06 (0.93, 1.21) | 1.03 (0.90, 1.17) | 50 | 1.00 (0.87, 1.14) | 100 | 1.01 (0.88, 1.15) | 83 | 0.99 (0.87, 1.13) | 100 | 225 |
| Turkey | 0.82 (0.65, 1.03) | 0.75 (0.59, 0.95) | … | 0.67 (0.53, 0.85) | … | 0.73 (0.58, 0.92) | … | 0.63 (0.50, 0.80) | … | 71 |
| Somalia | 0.67 (0.34, 1.34) | 0.59 (0.29, 1.18) | … | 0.61 (0.30, 1.22) | … | 0.57 (0.28, 1.14) | … | 0.55 (0.28, 1.13) | … | 8 |
| Thailand | 1.03 (0.69, 1.55) | 0.91 (0.60, 1.37) | 100 | 0.82 (0.55, 1.24) | 100 | 0.90 (0.60, 1.36) | 100 | 0.79 (0.52, 1.19) | 100 | 23 |
Note. CI = confidence interval; HR = hazard ratio; XF% = explained fraction.
Age.
Age and education.
Age and income.
Age and occupational class position.
Age and all socioeconomic position variables.
XF% is not calculated (…) for immigrant groups with lower HRs than native born because there is no excess risk to “explain” by socioeconomic position. XF% for immigrant groups with lower HRs of mortality after adjustment for socioeconomic position indicators is set to 100% and refers to the fact there is no excess risk left to be explained after adjustment for socioeconomic position.
TABLE 4—
Hazard Ratios of Mortality From External Causes by Country of Birth Among Immigrant Men and Women in Sweden Aged 30–65 Years in 1997: Follow-Up 1998–2006
| Model 2b |
Model 3c |
Model 4d |
Model 5e |
|||||||
| Country | Model 1,a HR (95% CI) | HR (95% CI) | XF%f | HR (95% CI) | XF%f | HR (95% CI) | XF%f | HR (95% CI) | XF%f | Deaths, No. |
| Men | ||||||||||
| Sweden | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 10 223 | ||||
| Finland | 2.03 (1.88, 2.18) | 1.70 (1.58, 1.83) | 32 | 1.53 (1.42, 1.65) | 49 | 1.61 (1.50, 1.74) | 41 | 1.45 (1.34, 1.56) | 56 | 756 |
| Iraq | 0.61 (0.43, 0.86) | 0.81 (0.57, 1.14) | … | 0.47 (0.36, 0.63) | … | 0.47 (0.33, 0.66) | … | 0.37 (0.26, 0.53) | … | 32 |
| Yugoslavia | 0.90 (0.75, 1.09) | 1.00 (0.83, 1.21) | … | 0.70 (0.59, 0.85) | … | 0.73 (0.61, 0.88) | … | 0.62 (0.52, 0.75) | … | 115 |
| Poland | 1.00 (0.74, 1.36) | 1.11 (0.81, 1.51) | … | 0.73 (0.57, 0.99) | … | 0.84 (0.62, 1.14) | … | 0.70 (0.51, 0.95) | … | 41 |
| Iran | 0.63 (0.48, 0.82) | 0.75 (0.57, 0.97) | … | 0.42 (0.32, 0.54) | … | 0.46 (0.35, 0.60) | … | 0.37 (0.30, 0.49) | … | 54 |
| Bosnia | 0.36 (0.24, 0.53) | 0.52 (0.35, 0.78) | … | 0.31 (0.21, 0.47) | … | 0.27 (0.18, 0.40) | … | 0.24 (0.16, 0.36) | … | 24 |
| Germany | 1.02 (0.80, 1.31) | 1.09 (0.85, 1.40) | … | 0.92 (0.72, 1.19) | 100 | 0.97 (0.75, 1.24) | 100 | 0.91 (0.71, 1.17) | 100 | 62 |
| Denmark | 0.71 (0.54, 0.95) | 0.66 (0.49, 0.87) | … | 0.58 (0.44, 0.77) | … | 0.59 (0.45, 0.79) | … | 0.51 (0.42, 0.74) | … | 47 |
| Norway | 1.49 (1.18, 1.88) | 1.38 (1.09, 1.74) | 22 | 1.19 (0.94, 1.50) | 61 | 1.22 (0.96, 1.54) | 55 | 1.12 (0.89, 1.41) | 76 | 72 |
| Turkey | 0.47 (0.32, 0.71) | 0.56 (0.37, 0.84) | … | 0.34 (0.23, 0.50) | … | 0.45 (0.30, 0.67) | … | 0.31 (0.20, 0.46) | … | 24 |
| Somalia | 1.19 (0.64, 2.21) | 1.55 (0.83, 2.88) | … | 0.80 (0.43, 1.47) | 100 | 0.85 (0.46, 1.59) | 100 | 0.64 (0.35, 1.20) | 100 | 10 |
| Thailand | 1.85 (0.60, 5.75) | 1.74 (0.56, 5.38) | 13 | 1.00 (0.32, 3.10) | 100 | 1.33 (0.43, 4.14) | 61 | 0.93 (0.30, 2.90) | 100 | 3 |
| Women | ||||||||||
| Sweden | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 3673 | ||||
| Finland | 1.70 (1.52, 1.91) | 1.53 (1.37, 1.72) | 24 | 1.49 (1.32, 1.67) | 30 | 1.46 (1.30, 1.63) | 34 | 1.45 (1.01, 1.62) | 36 | 318 |
| Iraq | 0.48 (0.20, 1.16) | 0.51 (0.21, 1.23) | … | 0.42 (0.17, 1.00) | … | 0.34 (0.14, 0.82) | … | 0.30 (0.26, 0.74) | … | 5 |
| Yugoslavia | 0.90 (0.65, 1.24) | 0.87 (0.63, 1.20) | … | 0.75 (0.54, 1.03) | … | 0.66 (0.48, 0.91) | … | 0.65 (0.47, 0.89) | … | 38 |
| Poland | 1.68 (1.30, 2.18) | 1.60 (1.23, 2.08) | 12 | 1.30 (1.00, 1.69) | 56 | 1.23 (0.95, 1.60) | 66 | 1.13 (0.87, 1.47) | 81 | 58 |
| Iran | 0.51 (0.28, 0.92) | 0.51 (0.28, 0.93) | … | 0.40 (0.22, 0.72) | … | 0.33 (0.18, 0.59) | … | 0.31 (0.17, 0.56) | … | 11 |
| Bosnia | 0.60 (0.36, 0.99) | 0.66 (0.40, 1.11) | … | 0.51 (0.31, 0.85) | … | 0.41 (0.25, 0.68) | … | 0.38 (0.23, 0.63) | … | 15 |
| Germany | 0.92 (0.62, 1.38) | 0.91 (0.61, 1.36) | … | 0.80 (0.54, 1.20) | … | 0.81 (0.54, 1.21) | … | 0.74 (0.50, 1.11) | … | 24 |
| Denmark | 1.36 (0.95, 1.96) | 1.28 (0.89, 1.84) | 22 | 1.16 (0.81, 1.67) | 56 | 1.17 (0.81, 1.67) | 53 | 1.10 (0.77, 1.57) | 72 | 30 |
| Norway | 1.45 (1.05, 2.02) | 1.32 (0.95, 1.83) | 29 | 1.23 (0.88, 1.72) | 49 | 1.19 (0.86, 1.66) | 58 | 1.15 (0.83, 1.60) | 67 | 36 |
| Turkey | 0.60 (0.31, 1.15) | 0.58 (0.30, 1.13) | … | 0.48 (0.25, 0.92) | … | 0.47 (0.25, 0.91) | … | 0.41 (0.21, 0.79) | … | 9 |
| Somalia | 0.86 (0.21, 3.42) | 0.73 (0.18, 2.94) | … | 0.81 (0.20, 3.25) | … | 0.54 (0.13, 2.15) | … | 0.60 (0.15, 2.41) | … | 2 |
| Thailand | 1.01 (0.46, 2.23) | 0.81 (0.36, 1.82) | 100 | 0.66 (0.30, 1.48) | 100 | 0.68 (0.30, 1.51) | 100 | 0.57 (0.26, 1.29) | 100 | 6 |
Note. CI = confidence interval; HR = hazard ratio; XF% = explained fraction.
Age.
Age and education.
Age and income.
Age and occupational class position.
Age and all socioeconomic position variables.
XF% is not calculated (…) for immigrant groups with lower HRs than native born because there is no excess risk to “explain” by socioeconomic position. XF% for immigrant groups with lower HRs of mortality after adjustment for socioeconomic position indicators is set to 100% and refers to the fact there is no excess risk left to be explained after adjustment for socioeconomic position.
Modeling Strategy
The modeling strategy in the analyses examined whether associations between country of birth and mortality from all causes (Table 1), CD (Table 2), neoplasms (Table 3), and external causes (Table 4), respectively, existed and to what extent they remained if we adjusted for education (model 2), income (model 3), and occupational class (model 4) and made a mutual adjustment for all 3 indicators of SEP (model 5). HRs for risk of mortality during the follow-up (1998–2006) were estimated using Cox proportional models and displayed in Tables 1 through 4. By using variables on in- and out-migration available in the registers, we excluded individuals who had returned to their country of origin from the analyses. Information on education and social class position was derived from registers from 1997, and therefore, was treated as fixed covariates. The income variable was based on the averaged disposable income between 1998 and 2006 (or the year up to death). Age was treated as a time-varying variable in the analyses. We presented nonsignificant results for some of the immigrant groups in Tables 1 through 4 because results were based on total population data.
RESULTS
The data available as a supplement to the online version of this article at http://www.ajph.org shows the distribution of SEP indicators by country of birth. Generally, these data suggested that most immigrant groups were disadvantaged in terms of occupational class and income. Immigrant men were overrepresented in the class of blue-collar workers compared with Swedish-born men, whereas a lower percentage of immigrant men were white-collar workers; this pattern was fairly similar among women. All groups of foreign-born men and women had lower average income compared with Swedish-born individuals. These data showed a different pattern pertaining to educational level. Several groups of foreign-born men had a higher percentage of education at the post-gymnasium level compared with Swedish-born individuals. By contrast, relatively few groups of immigrant women had a higher educational level compared with those born in Sweden. The profound differences in the distribution of different dimensions of SEP across immigrant categories suggested that measures of SEP should not be considered interchangeable.
Table 1 shows the HR of mortality by country of birth among men. The age-adjusted model (model 1) suggested excess mortality risks among immigrants born in Finland (HR = 1.67), the former Yugoslavia (HR = 1.25), Bosnia (HR = 1.23), Somalia (HR = 1.63), Thailand (HR = 1.42), Denmark (HR = 1.17), and Norway (HR = 1.19), whereas lower mortality risks were found among immigrants from Iraq (HR = 0.79) and Iran (HR = 0.66). Many of the excess risks remained after adjustment for education (model 2) and disappeared after further adjustment by income (model 3) and occupational class (model 4). The XF%, for instance, suggested that most or all of the excess mortality among immigrants from the former Yugoslavia (XF% = 80), Poland (XF% = 100), Bosnia (XF% = 87), Denmark (XF% = 94), Norway (XF% = 58), Somalia (XF% = 63), and Thailand (XF% = 100) was explained by income alone. Model 4 also suggested much lower mortality among immigrants from Iraq, Iran, Bosnia, and Turkey after adjustment for occupational class. Finally, model 5 showed no excess mortality risk after adjustment for all 3 indicators of SEP, with the exception of immigrants from Finland. By contrast, adjusting for all 3 indicators of SEP led to an even greater mortality advantage for immigrants from Iraq, Iran, Bosnia, and Turkey.
The HRs among women showed a similar pattern, but were generally lower. Women born in Finland (HR = 1.19), the former Yugoslavia (HR = 1.10), Bosnia (HR = 1.16), Denmark (HR = 1.33), Norway (HR = 1.21), Somalia (HR = 1.19), and Thailand (HR = 1.15) had higher age-adjusted all-cause mortality risks, whereas women from Iran had a much lower mortality risk. According to the XF%, the independent contribution of the SEP indicators to the mortality risk of those who were foreign-born seemed about equally important. After mutual adjustment for all indicators of SEP (model 5), no excess mortality risks were found among most immigrant groups. Lower mortality risks were found among many groups of immigrant women, especially those born in Iran.
Table 2 suggested higher age-adjusted mortality from CD among men born in Yugoslavia, Poland, Bosnia, Somalia, Thailand, and Finland, although men from Iran had a lower CD mortality risk. The XF% suggested that occupational class (model 4) and income (model 3) were the primary contributors to such mortality differentials. Higher mortality risks were only found among immigrants born in Finland after mutual adjustment for SEP in model 5.
Women born in Finland, Iraq, Yugoslavia, Bosnia, Norway, Thailand, Denmark, Turkey, and Somalia had higher mortality from CD compared with Swedish-born individuals. The XF% suggested that education, income, and occupational class independently accounted for a relatively large share of the mortality differentials. Model 5 showed that only immigrant women from Finland had higher mortality risks from CD after joint adjustment for the indicators of SEP, whereas foreign-born women from several countries had lower mortality.
Table 3 shows higher mortality risks from neoplasms among foreign-born immigrant men from Bosnia, Somalia, Yugoslavia, Denmark, and Finland. The XF% suggested that SEP accounted for much less of the mortality differentials for neoplasms. Although occupational class and income accounted for some of the excess risks (models 4 and 3), higher mortality risks remained among most of the immigrant groups (model 5).
There were modest differentials between immigrant groups and Swedish-born women in mortality from neoplasms. Only women from Denmark showed higher mortality from neoplasms. Some immigrant groups, such as women from Iran and Germany, showed lower mortality after mutual adjustment for SEP.
Table 4 shows that higher mortality from external causes was found among men born in Norway, Somalia, Thailand, and Finland, whereas lower mortality risks were found among men from Iran, Bosnia, Denmark, and Turkey. The XF% suggested that occupational class and income contributed to some of the mortality differentials from external causes, whereas the contribution by education was smaller. Most of the mortality differentials disappeared after mutual adjustment for all 3 indicators of SEP (model 5).
Higher mortality from external causes was found among immigrant women from Poland, Denmark, Norway, and Finland. The XF% suggested that occupational class and income accounted for some of these differentials, whereas the contribution by education was modest. Mutual adjustment for SEP factors erased most of the excess mortality risks in external causes. Women born in Yugoslavia, Iran, Iraq, Bosnia, Turkey, Somalia, and Thailand had lower mortality after mutual adjustment for SEP indicators.
DISCUSSION
There is still limited information on mortality in specific groups of migrants and whether different aspects of SEP account for mortality differentials. In this study, we showed how key indicators of SEP differed between the largest groups of foreign-born immigrants in Sweden and further scrutinized whether SEP accounted for mortality differentials by country of birth.
In our age-adjusted analyses, we found higher mortality from all causes and CD, respectively, among many of the studied immigrant categories, which conformed to previous Swedish findings.4–18 The relatively modest differentials in mortality from neoplasms between immigrant groups also corresponded to previous empirical evidence.18 Previous research primarily found higher mortality among Nordic and Eastern European immigrants and lower mortality among immigrants born in Turkey, Southern Europe, Latin America, Asia, and Africa.4,5,18,20,21 Our results corroborated previous findings on Nordic and European immigrants, although the results by specific country of birth and cause of death also provided additional information. Although we found lower all-cause mortality among some non-European immigrant groups (i.e., Iraq, Iran, and Turkey), higher mortality was found in others (i.e., Somalia and Thailand). This indicated that studying broad categories of immigrants might “mask” profound health variations by country of birth or that mortality among some non-European immigrants might have increased over time. The fact that lower mortality from neoplasms and external causes were found among several immigrant groups, whereas all-cause and CD mortality was higher, highlights the importance of studying cause-specific mortality by country of birth.
The contribution of SEP to the observed mortality differentials was significant. When SEP was mutually adjusted for, excess risks of mortality from all causes and CD disappeared in most immigrant groups. However, the contribution of SEP to mortality differentials from neoplasms was much more modest. This might indicate that mortality from neoplasms largely originated from factors that were more weakly related to SEP, such as genetic vulnerability and early life conditions in the country of origin. In contrast, mortality from CD was more responsive to factors that were influenced by SEP,31 including material conditions, preventive health care, and lifestyle factors. Furthermore, we even found lower mortality risks among many groups of immigrants after adjustment for SEP, which indicated that improving immigrants’ SEP to the level of natives could actually lead to health advantages. However, we found relatively modest disparities in mortality from external causes by country of birth. Accordingly, previous studies found lower suicide rates primarily among non-European immigrants.19
We also found that the contribution of education to mortality differentials by immigrant groups was relatively modest compared with income and occupational class. This finding might be seen to contradict some evidence from the United States that suggested that education tended to be a stronger predictor of health than income.37 It could be that the education gradient was flattened among immigrants because of occupational mismatch— education is simply not as good a predictor of overall SEP among immigrants who face occupational barriers because of credentialing problems and discrimination.38 Education played a particularly modest role in explaining mortality differentials by immigrant groups for women. This might be cause of even greater occupational barriers for highly educated immigrant women. It was suggested that immigrant women faced additional constraints in the labor market because of country-, religious-, and cultural-specific values and gender stereotypes.39,40
Our findings highlighted that indicators of SEP should not be used interchangeably. The fact that income and occupational class were more influential in our analysis highlighted the importance of socioeconomic conditions in the country of destination. Accordingly, education was, to a greater extent, achieved in the country of origin, whereas income and occupational class were acquired in the country of destination. Overall, our findings suggested that SEP might be one of the most important factors that explained mortality differentials by country of birth. The fact that SEP was also a determinant of several other risk factors for ill health, such as adverse health behaviors, poor working conditions, stress, residential segregation, etc., could further contribute to the strong influence of SEP.3
Although our findings indicated that occupational class position and income accounted for most of the mortality differentials by immigrant category, it is important to emphasize that policies should not be exclusively aimed at improving these conditions. Education is often a prerequisite for a job with high status, which in turn provides a high income and is therefore important to take into account from a policy perspective.
We found lower mortality from all causes among immigrants from Iran and lower mortality from all causes and external causes among people born in Iraq, which corroborated some earlier Swedish studies.18 The lower mortality could reflect healthier behaviors among these groups. Smoking and alcohol consumption were strong risk factors for mortality from CVD and neoplasms, whereas alcohol consumption could additionally influence mortality from external causes. Previous Swedish findings suggested that smoking was particularly low among women from the Middle East, whereas alcohol consumption was low among both men and women from these countries.41,42 Accordingly, some scholars suggested that the relationship between health risk behaviors might be reversed in countries that have yet to undergo the Western “lifestyle transition,” and they may therefore “import” healthy behaviors to the country of destination.43 These findings could also suggest that immigrants from Iran and Iraq who migrated and remained in Sweden were the healthiest and strongest members of their population of origin.44 It was possible that a paradox similar to the US “Hispanic paradox in health” could apply to these groups. Future studies should more closely investigate explanations for lower mortality among these groups in Sweden.
Despite the obvious strengths of this study, such as the use of total population data, longitudinal follow-up, reliable information on mortality, and other included variables, some limitations should be noted. A shortcoming of our data was that essential information could not be collected via registers, which might cause omitted variable bias. Register data, for example, did not include information on health behaviors, social support, working conditions, etc., that might influence immigrants’ health and contribute to the significance of SEP in explaining the observed associations. Another problem of population registers was the risk of “overcoverage,” that is, people registered in the data leave Sweden, most often returning to their country of origin, without this being registered.21,45 This was shown to be a problem especially among non-European immigrants, meaning that we might have underestimated their death risks. To minimize this problem, we performed additional analyses (not shown) that excluded individuals with no registered income for 2 years or more. These tests did not change the overall findings. We also performed simulations in which we increased the number of deaths among the immigrant groups with lower mortality rates until parity with Swedes was achieved. This allowed us to determine how likely it was that our findings of reduced mortality risk resulted from overcoverage. These analyses showed that a rather large number of deaths had to go unreported until mortality was in parity with Swedes. For instance, the number of deaths would have to be about 50% higher among Iranian men and almost doubled among Iranian women.
Another source of bias was that those excluded in this study, that is, those who emigrated during the follow-up period or immigrants never officially registered upon arrival to Sweden, had different mortality rates than their compatriots. Although our study included all deaths during the follow-up period that were not subject to sampling error, it might have still been affected by random variation, especially for some immigrants with very small number of deaths, such as those born in Thailand, Iraq, or Somalia. Finally, we could not rule out the possibility of reversed causality. Poorer health might have led to downward social mobility and poorer SEP among immigrants, which, in turn, could contribute to higher mortality. Consequently, it was not immigrants’ lower SEP, but rather their poorer health status, that primarily explained their mortality rates.
To summarize, the findings of this study stressed the fact that different aspects of SEP were not interchangeable in relation to immigrant health. Although policies aiming to improve migrants’ SEP might be beneficial for their health and longevity, our findings indicated that such policies might have varying effects depending on the specific country of birth and cause of death.
Acknowledgments
We would like to thank the Swedish Research Council (grants 2011-1649 and 2008-1677) for funding.
We would also like to thank Hélio Manhica for help with the statistical analyses.
Human Participant Protection
This study was approved by the Central Ethical Review Board of Stockholm (decision no. 2012/1260-31).
References
- 1.Sweden Statistics. Statistical Yearbook of Sweden 2008. Stockholm, Sweden: Statistics Sweden; 2008. [Google Scholar]
- 2.Organisation for Economic Co-operation and Development. International Migration. Paris: Organisation for Economic Co-operation and Development; 2010. [Google Scholar]
- 3.Palme J, Bergmark Å, Bäckman O . Welfare in Sweden: The Balance Sheet for the 1990s. Stockholm, Sweden: Fritzes; 2002. [Google Scholar]
- 4.Sundquist J, Johansson SE. The influence of country of birth on mortality from all causes and cardiovascular disease in Sweden 1979-1993. Int J Epidemiol. 1997;26(2):279–287. doi: 10.1093/ije/26.2.279. [DOI] [PubMed] [Google Scholar]
- 5.Albin B, Hjelm K, Ekberg J, Elmståhl S. Mortality among 723 948 foreign and native born Swedes 1970-1999. Eur J Public Health. 2005;15(5):511–517. doi: 10.1093/eurpub/cki026. [DOI] [PubMed] [Google Scholar]
- 6.Rostila M. Birds of a feather flock together and fall ill? Migrant homophily and health in Sweden. Sociol Health Illn. 2010;32(3):382–399. doi: 10.1111/j.1467-9566.2009.01196.x. [DOI] [PubMed] [Google Scholar]
- 7.Gadd M, Johansson SE, Sundquist J, Wändell P. The trend of cardiovascular disease in immigrants in Sweden. Eur J Epidemiol. 2005;20(9):755–760. doi: 10.1007/s10654-005-1047-2. [DOI] [PubMed] [Google Scholar]
- 8.Gadd M, Johansson SE, Sundquist J, Wändell 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;6:102. doi: 10.1186/1471-2458-6-102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Hedlund E, Lange A, Hammar N. Acute myocardial infarction incidence in immigrants to Sweden. Country of birth, time since immigration, and time trends over 20 years. Eur J Epidemiol. 2007;22(8):493–503. doi: 10.1007/s10654-007-9147-9. [DOI] [PubMed] [Google Scholar]
- 10.Hedlund E, Pehrsson K, Lange A, Hammar N. Country of birth and survival after a first myocardial infarction in Stockholm, Sweden. Eur J Epidemiol. 2008;23(5):341–347. doi: 10.1007/s10654-008-9240-8. [DOI] [PubMed] [Google Scholar]
- 11.Lindström M, Sundquist J, Östergren P-O. Ethnic differences in self reported health in Malmö in southern Sweden. J Epidemiol Community Health. 2001;55(2):97–103. doi: 10.1136/jech.55.2.97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Fritzell J, Lennartsson C, Lundberg O. Health inequalities in Sweden: long and short-term perspectives. In: Fritzell J, Lundberg O, editors. Health Inequalities and Welfare Resources. Bristol, UK: Policy Press; 2007. pp. 19–42. [Google Scholar]
- 13.Rostila M. Social capital and health in the Swedish welfare state. In: Fritzell J, Lundberg O, editors. Health Inequalities and Welfare Resources. Bristol, UK: Policy Press; 2007. pp. 157–178. [Google Scholar]
- 14.Sundquist J, Johansson S-E. Long-term illness among indigenous and foreign-born people born in Sweden. Soc Sci Med. 1997;44(2):189–198. doi: 10.1016/s0277-9536(96)00145-1. [DOI] [PubMed] [Google Scholar]
- 15.Khan FA, Zia E, Janzon L, Engstrom G. Incidence of stroke and stroke subtypes in Malmö, Sweden, 1990-2000. Marked differences between groups defined by birth country. Stroke. 2004;35(9):2054–2058. doi: 10.1161/01.STR.0000135761.18954.0b. [DOI] [PubMed] [Google Scholar]
- 16.Johansson LM, Sundquist J, Johansson SE, Bergman B, Qvist J, Träskman-Bendz L. Suicide among foreign-born minorities and native Swedes: an epidemiological follow-up study of a defined population. Soc Sci Med. 1997;44(2):181–187. doi: 10.1016/s0277-9536(96)00142-6. [DOI] [PubMed] [Google Scholar]
- 17.Westerling R, Rosén M. ‘Avoidable’ mortality among migrants in Sweden. Eur J Public Health. 2002;12(4):279–286. doi: 10.1093/eurpub/12.4.279. [DOI] [PubMed] [Google Scholar]
- 18.Hemminki K, Li X. Cancer risks in first-generation immigrants to Sweden. Int J Cancer. 2002;99(2):218–228. doi: 10.1002/ijc.10322. [DOI] [PubMed] [Google Scholar]
- 19.Hjern A, Allebeck P. Suicide in first- and second-generation immigrants in Sweden. A comparative study. Soc Psychiatry Psychiatr Epidemiol. 2002;37(9):423–429. doi: 10.1007/s00127-002-0564-5. [DOI] [PubMed] [Google Scholar]
- 20.Sundquist K, Li X. Coronary heart disease risks in first- and second-generation immigrants in Sweden: a follow-up study. J Intern Med. 2006;259(4):418–427. doi: 10.1111/j.1365-2796.2006.01630.x. [DOI] [PubMed] [Google Scholar]
- 21.Weitoft GR, Gullberg A, Hjern A, Rosén M. Mortality statistics in migrant research: a method for adjusting underestimation of mortality. Int J Epidemiol. 1999;28(4):756–763. doi: 10.1093/ije/28.4.756. [DOI] [PubMed] [Google Scholar]
- 22.Geyer S, Hemström Ö, Vågerö D. Education, income, and occupational class cannot be used interchangeably in social epidemiology. Empirical evidence against a common practice. J Epidemiol Community Health. 2006;60(9):804–810. doi: 10.1136/jech.2005.041319. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Erikson R, Torssander J. Social class and cause of death. Eur J Public Health. 2008;18(5):473–478. doi: 10.1093/eurpub/ckn053. [DOI] [PubMed] [Google Scholar]
- 24.Fritzell J. Socioekonomiska skillnader i hälsa [Socioeconomic inequalities in health] In: Rostila M, Toivanen S, editors. Den orättvisa hälsan. Om socioekonomiska skillnader i hälsa och livslängd [Health Injustice. On Socioeconomic Differences in Health and Life Expectancy] Stockholm, Sweden: Liber; 2012. pp. 46–61. [Google Scholar]
- 25. SOU 2001:53 Välfärd och arbete i arbetslöshetens årtionde [Welfare and Work in a Decade of Unemployment]. Stockholm, Sweden: Fritzes, 2001.
- 26.Tinghög P, Hemmingsson T, Lundberg I. To what extent may the association between immigrant status and mental illness be explained by socioeconomic factors? Soc Psychiatry Psychiatr Epidemiol. 2007;42(12):990–996. doi: 10.1007/s00127-007-0253-5. [DOI] [PubMed] [Google Scholar]
- 27.Fritzell J, Bäckman O, Ritakallio V-M. Income inequality and poverty: do the Nordic countries still constitute a family of their own? In: Kvist J, Fritzell J, Hvinden B, Kangas O, editors. Changing Social Equality. The Nordic Model in the 21st Century. Bristol, UK: Policy Press; 2012. pp. 165–185. [Google Scholar]
- 28.Kawachi I, Daniels N, Robinson DE. Health disparities by race and class: why both matter. Health Aff. 2005;24(2):343–352. doi: 10.1377/hlthaff.24.2.343. [DOI] [PubMed] [Google Scholar]
- 29.Peter R, Gassler H, Geyer S. Socioeconomic status, status inconsistency and risk of ischaemic heart disease: a prospective study among members of a statutory health insurance company. J Epidemiol Community Health. 2007;61(7):605–611. doi: 10.1136/jech.2006.047340. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.JUSEK. Invandrade akademiker - hur tufft är det [Immigrant Academics–How Tough Is It]? Stockholm, Sweden: Jusek; 2010. [Google Scholar]
- 31.Howard G, Anderson RT, Russell G, Howard VJ, Burke GL. Race, socioeconomic status, and cause-specific mortality. Ann Epidemiol. 2000;10(4):214–223. doi: 10.1016/s1047-2797(00)00038-7. [DOI] [PubMed] [Google Scholar]
- 32.International Classification of Diseases, 10th Revision. Geneva, Switzerland: World Health Organization; 1992. [Google Scholar]
- 33.Lynch JW, Kaplan GA, Cohen RD, Tuomilehto J, Salonen JT. Do cardiovascular risk factors explain the relation between socioeconomic status, risk of all-cause mortality, cardiovascular mortality, and acute myocardial infarction? Am J Epidemiol. 1996;144(10):934–942. doi: 10.1093/oxfordjournals.aje.a008863. [DOI] [PubMed] [Google Scholar]
- 34.Yngwe MA, Diderichsen F, Whitehead M, Holland P, Burström B. The role of income differences in explaining social inequalities in self rated health in Sweden and Britain. J Epidemiol Community Health. 2001;55(8):556–561. doi: 10.1136/jech.55.8.556. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Statistics Sweden. Available at: http://www.scb.se. Accessed June 4, 2012.
- 36.Statistics Sweden. Occupations in Population and Housing Census 1985 (FOB85) according to Nordic Standard Occupational Classification and Swedish Socio-economic Classification. Stockholm, Sweden: Statistics Sweden; 1989. [Google Scholar]
- 37.Cutler DM, Lleras-Muney A, Vogl T. NBER Working Paper No. 14333. Cambridge, MA: The National Bureau of Economic Research; 2008. Socioeconomic status and health: dimensions and mechanisms. [Google Scholar]
- 38.Rostila M, Toivanen S. Den orättvisa hälsan: Den orättvisa hälsan. Om socioekonomiska skillnader i hälsa och livslängd. [Unjust Health. On Socioeconomic Differences in Health and Longevity.] Stockholm, Sweden: Liber; 2012. [Google Scholar]
- 39.Antecol H. An examination of cross-country differences in the gender gap in labour force participation rates. Labour Econ. 2000;7(4):409–426. [Google Scholar]
- 40.Nielsen HS, Rosholm M, Smith N. Centre for Labour Market and Social Research Working Paper 00-6. Aarhus, Denmark: University of Aarhus; 2000. Hit Twice? Danish evidence on the double-negative effect on the wages of immigrant women. [Google Scholar]
- 41.Gadd M, Sundquist J, Johansson S-E, Wändell P. Do immigrants have an increased prevalence of unhealthy behaviours and risk factors for coronary heart disease? Eur J Cardiovasc Prev Rehabil. 2005;12(6):535–541. doi: 10.1097/01.hjr.0000174829.25388.ed. [DOI] [PubMed] [Google Scholar]
- 42.Hjern A, Allebeck P. Alcohol-related disorders in first- and second-generation immigrants in Sweden: a national cohort study. Addiction. 2004;99(2):229–236. doi: 10.1046/j.1360-0443.2003.00562.x. [DOI] [PubMed] [Google Scholar]
- 43.Razum O, Twaedella D. Time travel with Oliver Twist–towards an explanation for paradoxically low mortality among recent immigrants. Trop Med Int Health. 2002;7(1):4–10. doi: 10.1046/j.1365-3156.2002.00833.x. [DOI] [PubMed] [Google Scholar]
- 44.Franzini L, Ribble JC, Keddie AM. Understanding the Hispanic paradox. Ethn Dis. 2001;11(3):496–518. [PubMed] [Google Scholar]
- 45.Qvist J. Täckningsproblem i Registret over totalbefolkningen RTB–Skattning av övertäckning med en indirekt metod [Problems of coverage in the Register of Total Population (RTB)—estimation of overcoverage by an indirect method]. R & D Report. Stockholm, Sweden: Statistics Sweden; 1999. [Google Scholar]