Coronary heart disease risks in first- and second-generation immigrants in Sweden: a follow-up study
Dr Kristina Sundquist MD, PhD, Karolinska Institutet, Center for Family and Community Medicine, Alfred Nobels allé 12, SE-141 83 Huddinge, Sweden.
(fax: +46 8 524 887 06; e-mail: firstname.lastname@example.org).
Objectives. To analyse whether there is an association between country of birth in first-generation immigrants and first hospitalization for or death from coronary heart disease (CHD) and to analyse whether this association remains in second-generation immigrants.
Design. In this follow-up study, the MigMed database at the Karolinska Institute, Stockholm, was used to identify all hospital diagnoses of and deaths from incident CHD in first- and second-generation immigrants in Sweden between 1 January 1987 and 31 December 2001. Incidence ratios standardized by age, geographical region and socio-economic status were estimated by sex in first- and second-generation immigrants; the reference group was Swedish-born people whose parents were both born in Sweden.
Subjects. The total Swedish population aged 25–69 years.
Results. First-generation immigrants from Finland, central European countries, other eastern European countries and Turkey had higher rates of CHD than men or women in the reference group. First-generation immigrant women born in southern Europe, other western European countries and Baltic countries had lower CHD risks than the reference group. Sons of both male and female first-generation immigrants showed CHD risks similar to or slightly higher than those of their parents. Amongst second-generation women, only subjects with Finnish fathers or mothers had higher risks of developing CHD than the reference.
Conclusions. Increased risks of CHD found in some first-generation immigrant groups often persist in second-generation immigrant men. Healthcare professionals and policy makers should take this into account when designing and undertaking measures to prevent CHD.
Sweden, like many countries, has experienced dramatic demographic changes because of increasing global migration, and has become a multicultural society in the new millennium. Today, approximately 20% of all people living in Sweden are first- or second-generation immigrants .
International studies have found that risk factors for coronary heart disease (CHD) and mortality due to CHD increase following immigration, for example, in South Asian immigrants settled in Glasgow and in Pacific Atoll male immigrants settled in New Zealand [2, 3]. In a study of Japanese men in Japan, Hawaii and California , CHD rates were significantly higher in Japanese-American men aged 45–69 years than in men in the corresponding age group in Japan.
A previous study has established that female immigrants from Finland and Eastern Europe have about twice the risk of developing CHD as their Swedish-born counterparts . To our knowledge, however, no earlier studies have attempted to ascertain whether or not increased risk of developing CHD in first-generation immigrant parents in Sweden is associated with increased risk of developing CHD in the next generation, i.e. in second-generation immigrants.
Careful study of rates of first hospitalization for or death from CHD in first- and second-generation immigrants may help identify novel risk factors for CHD in these groups and contribute to the formulation of new preventive strategies. The MigMed database provides a unique opportunity to study rates of CHD amongst first- and second-generation immigrants. The database incorporates information on the entire national population over a period of 15 years, including all first- and second-generation immigrants registered in Sweden. It includes CHD incidence data, for the entire population. Additionally, the database incorporates population-wide documentation regarding concomitant factors like geographical region and socio-economic status (SES). Socio-economic status is a clear and widely accepted individual potential confounding variable in studies of CHD [6–10].
The aim was to analyse whether there is an association between country of birth in first-generation immigrants and first hospitalization for CHD or death from CHD and, further, to analyse whether this association remains in second-generation immigrants after accounting for age, geographical region and SES.
Materials and methods
MigMed research database
Data used in this study were retrieved from the MigMed database, located at the Center for Family and Community Medicine at the Karolinska Institute in Stockholm. MigMed is a single, comprehensive database that contains individual-level information on all adult people in Sweden, including age, sex, occupation, geographical region of residence, hospital diagnoses and dates of hospital admissions in Sweden (1986–2001), country of birth, parents’ countries of birth, date of emigration and date and cause of death. This unique database was constructed using several national Swedish data registers, including but not limited to the total population register, the multigeneration register , the death register , the Swedish hospital discharge register (1986–2001)  and the immigration register.
Information retrieved from the various registers in the MigMed database was linked at the individual level via the national 10-digit civic registration number assigned to each person in Sweden for his or her lifetime. Prior to inclusion in the MigMed database, civic registration numbers were replaced by serial numbers to ensure the anonymity of all individuals. In addition to using the serial numbers to track all records in the database at the individual level, these numbers were used to check that individuals with hospital diagnoses of CHD only appeared once in the data set, for their first hospital diagnosis of CHD in the study period.
The MigMed database includes data on more than 3.2 million families with 6.9 million Swedish-born second-generation individuals. The multigeneration register, which is included in the MigMed database, includes the identities of the biological parents of index persons who have been registered (folkbokförd) in Sweden any time since 1961 and were born in 1932 or later. Thus, data of all second-generation immigrants’ parents are included in the MigMed database. The latest version of the multigeneration register has been incorporated in the MigMed database; it includes supplementary data from church records on index persons domiciled in Sweden between 1947 and 1961, including information about biological parents, children, siblings and adoptions.
The follow-up period started on 1 January 1987 and proceeded until hospitalization for CHD, death from CHD, death from another cause, emigration or the end of the study period on 31 December 2001. Information on hospitalizations prior to 1 January 1986 was unavailable.
The 9th and 10th revisions of the WHO's International Classification of Diseases (ICD-9 and ICD-10) were used to define the outcome variable, first hospitalization for CHD or death outside the hospital from CHD (hereafter referred to as incident CHD). ICD-9 codes were used in Swedish hospitals before 1997, and ICD-10 codes from 1997 onwards. ICD-9 codes used in this study included 410 (acute myocardial infarction), 411 (other acute and subacute forms of CHD), and 413 (angina pectoris). ICD-10 codes used in this study included I20 (angina pectoris), I21 (acute myocardial infarction), I23 (complications due to acute myocardial infarction) and I24 (other acute forms of CHD).
Individual variables included sex, age at the start of the study, geographical region and SES of the men and women. Age at the start of the study ranged between 25 and 69 years in both the first- and the second-generation in order to match the age range in the second-generation population. Geographical region was divided into (i) big cities (cities with a population of more than 200 000, i.e. Stockholm, Gothenburg and Malmö), (ii) Southern Sweden and (iii) Northern Sweden. Sweden is divided into 25 counties. The border between northern and southern Sweden has traditionally been drawn at the Dalälven River. Therefore, all counties north of that river were defined as part of northern Sweden. Geographical region was included as an individual variable to adjust for possible differences between geographical regions in Sweden regarding hospital admissions for CHD.
Statistics Sweden, the Swedish governmental statistics bureau, uses a socio-economic classification system based on occupation, which they developed in 1974 and revised in 1982 and 1984 . This system was used to classify SES in the current study. Occupations of persons in the economically active sector of the population were divided into different categories, which were constructed by combining occupations in which individuals had the same type of working situations. Prestige of individual jobs was not taken into account when constructing the categories. Socio-economic status of men and women was thus divided into five categories: (i) manual workers, (ii) lower level employees, (iii) middle level employees and professionals, (iv) the self-employed and farmers and (v) all others. Homemakers and students without an occupation were categorized based on their father's or mother's occupation; if that was not possible, they were included in the ‘all others’ category. In addition, people without paid employment are also included in the ‘all others’ category.
It is of great importance to account for socio-economic factors in studies on CHD and immigrants, because even though ethnic, cultural and social heterogeneity amongst immigrants is large, in Sweden as in most countries, poor socio-economic circumstances are more common amongst immigrants than amongst the non-immigrant population .
The predictor variable was first- or second-generation immigrant status. First-generation immigrants were defined as all people born abroad in one of the countries and regions of parental birth included in the statistical analysis. Second-generation immigrants were defined as all people with at least one first-generation immigrant parent.
The MigMed database includes people from 64 countries and regions of birth (or, in the case of second-generation immigrants, of parental birth). Immigration from a number of these countries and regions began relatively recently, so members of the second generation have not yet reached the age when people start to be hospitalized for CHD in greater numbers. For this reason (few cases of CHD in the second generation), these countries and regions of birth were excluded from the study. As a result, only 10 of 64 countries and regions were included in our analysis. However, the large majority of all immigrants in Sweden come from these countries: Denmark, Finland, Norway, southern Europe (Andorra, Cyprus, Greece, Israel, Monaco, Portugal, San Marino, France, Spain, Italy and the Vatican City State), other western European countries (Austria, Belgium, Germany, Liechtenstein, Luxembourg, the Netherlands, Switzerland and the UK and Ireland), Baltic countries (Estonia, Latvia and Lithuania), central European countries (Poland, the Czech Republic, Slovakia and Hungary), other eastern European countries (Bosnia-Herzegovina, Croatia, Yugoslavia, Rumania, Bulgaria, and Albania), Russia and Turkey.
We show two models, one age-adjusted and a full model that adjusted for SES and geographical region. Age-standardized incidence rates were calculated for the whole follow-up period, divided into five 3-year periods (1987–1989, 1990–1992, 1993–1995, 1996–1998 and 1999–2001). Relative weights used to calculate the incidence rates were based on the 1970 European standard population. Population constitutes risk time.
Standardized incidence ratios (SIRs) with 95% confidence intervals were calculated as the ratio of observed to expected number of cases . The expected number of cases was based on the observed number of cases in the reference category. The expected number of cases was calculated by age (in 5-year groups), sex, period (in 3-year groups), SES and geographical region. Risks of incident CHD in first-generation immigrants were figured separately for men and women. Person-years were calculated beginning when people were included in the study (in 1986 or later) until first hospitalization for CHD, death, emigration or the end of the study on 31 December 2001. Risks of incident CHD in second-generation immigrants were also calculated separately for men and women, first by father's country of birth and then by mother's country of birth.
The reference groups were the same in all analyses, and included all people aged 25–69 years (categorized in 5-year groups) during the study period who were born in Sweden to parents born in Sweden and of the same sex as the people in the corresponding case group. Confidence intervals (95%) were determined assuming a Poisson distribution.
This study was approved by the Ethics Committee at Karolinska Institute, Stockholm.
In total, 221 260 men (7.2% of the Swedish male population) and 242 523 women (7.9% of the Swedish female population) were born in the 10 countries and regions under study. The mean age at diagnosis of CHD was 69.6 years for fathers, 74.5 years for mothers, 54.4 years for sons and 55.4 years for daughters (data not shown). During the study period (1987–2001), there were 11 768 cases of CHD amongst the foreign-born men in the study and 5056 cases of CHD amongst the foreign-born women. Male manual workers and men and women living in northern Sweden had high incidence rates of CHD. Men born in Finland, other eastern European countries and Turkey had the highest incidence rates of CHD (Table 1).
Table 1. Population, number of cases and incidence rates of coronary heart disease in 25-to 69-year-old first- and second-generation immigrants in Sweden
|Manual workers||1 344 179||241.3||1 214 305||88.4||1 556 250||241.6||1 402 517||100.9|
|Lower level employees|| 728 012||184.3|| 897 775||57.2|| 910 519||192.2|| 990 651|| 63.1|
|Middle level employees and professionals|| 87 953||160.9|| 73 086||42.7|| 88 237||172.5|| 83 108|| 45.0|
|Self-employed and farmers|| 530 302||211.5|| 478 632||69.3|| 514 918||212.2|| 459 391|| 86.6|
|Others|| 377 645||215.1|| 419 860||80.4|| 498 450||187.1|| 463 321|| 73.3|
|Big cities||1 063 345||202.1||1 051 259||65.9|| 588 227||217.0|| 561 911|| 75.6|
|Southern Sweden||1 413 915||217.4||1 432 438||71.4||2 321 275||206.0||2 211 609|| 73.9|
|Northern Sweden|| 590 831||242.3|| 599 961||84.5|| 658 872||245.3|| 625 468|| 96.3|
| || || || || ||By father's birth country||By mother's birth country||By father's birth country||By mother's birth country|
|Sweden||2 754 792||209.4||2 771 515||69.1||3 071 216||211.7||3 088 409||214.4||2 926 982|| 77.6||2 945 465|| 78.1|
|Denmark|| 20 087||234.8|| 18 027||80.2|| 26 082||238.7|| 23 317||288.6|| 24 745|| 42.2|| 21 878|| 74.0|
|Finland|| 84 437||312.8|| 111 722||115.9|| 106 049||265.0|| 137 730||252.8|| 100 654||130.4|| 130 444||100.5|
|Norway|| 17 805||239.9|| 24 997||80.0|| 21 988||192.7|| 29 999||252.8|| 20 823|| 90.1|| 28 412|| 92.4|
|Southern Europea|| 6848||165.1|| 4217||54.0|| 8370||90.1|| 5398||157.8|| 8099|| 8.4|| 5212|| 14.3|
|Other Western Europeanb|| 26 700||208.5|| 24 424||69.7|| 32 170||161.6|| 29 283||153.7|| 30 511|| 67.7|| 27 686|| 64.0|
|Baltic countriesc|| 7430||181.2|| 6348||53.1|| 7951||159.1|| 7102||161.9|| 7807|| 62.9|| 6893|| 50.6|
|Central European countriesd|| 19 162||270.6|| 21 328||90.2|| 19 743||286.4|| 22 390||309.4|| 18 641|| 96.7|| 21 112|| 92.9|
|Other Eastern European countriese|| 24 539||304.2|| 19 693||126.7|| 51 699||201.8|| 46 804||274.5|| 48 332|| 80.9|| 43 589|| 77.6|
|Russia|| 3044||243.9|| 3113||88.8|| 3932||198.1|| 4577||166.8|| 3828|| 74.4|| 4387|| 88.7|
|Turkey|| 11 208||360.3|| 8654||141.0|| 19 503||205.4|| 17 022||188.6|| 18 494|| 53.8|| 16 062|| 0.0|
Table 2 shows SIRs of CHD in first-generation immigrant men and women compared with persons of the same sex who were born in Sweden and whose both parents were born in Sweden, after accounting for age, SES and geographical region. Men and women born in Finland, central European countries, other eastern European countries and Turkey had increased risks of incident CHD. Male immigrants from Norway also had slightly increased SIRs. In contrast, female immigrants from southern Europe, other western European countries and Baltic countries had decreased risks of CHD; their SIRs were 0.69 (95% CI 0.48–0.96), 0.91 (95% CI 0.83–0.99) and 0.68 (95% CI 0.56–0.83), respectively.
Table 2. Standardized incidence ratios of coronary heart disease in first-generation male and female immigrants, aged 25–69 years
|Socio-economic status (adjusted for age)|
| Manual workers||56 404||1.50||1.49–1.51||18 487||2.07||2.04–2.10|
| Lower level employees||24 458||1.15||1.14–1.16||12 060||1.35||1.32–1.37|
| Middle level employees and professionals||1709||1 (reference)||461||1 (reference)|
| Self-employed and farmers||15 431||1.32||1.30–1.34||5270||1.62||1.58–1.67|
| Others||12 812||1.29||1.27–1.31||3859||1.77||1.71–1.82|
|Region (adjusted for age)|
| Big cities||36 660||0.93||0.92–0.94||12 670||0.92||0.91–0.94|
| Southern Sweden||50 159||1 (reference)||18 119||1 (reference)|
| Northern Sweden||23 995||1.11||1.10–1.13||9348||1.18||1.16–1.21|
|Birth country (adjusted for age, SES and region)|
| Sweden|| ||1 (reference)|| ||1 (reference)|
| Southern Europea||188||0.86||0.74–1.00||34||0.69||0.48–0.96|
| Other western European countriesb||1213||0.99||0.93–1.05||498||0.91||0.83–0.99|
| Baltic countriesc||349||0.90||0.81–1.00||100||0.68||0.56–0.83|
| Central European countriesd||1083||1.29||1.21–1.37||305||1.18||1.05–1.32|
| Other eastern European countriese||1529||1.55||1.47–1.63||399||1.77||1.60–1.95|
A greater risk of developing CHD was found in second-generation immigrant men than in men in the reference group (Table 3). Overall, 1603 first hospitalizations for CHD were observed in men with foreign-born fathers, and 2208 in men with foreign-born mothers (data not shown).
Table 3. Standardized incidence ratios of coronary heart disease in second-generation male and female immigrants aged 25–69 years by father's and mother's country of birth
|Other western European countriesb||178||0.93||0.80–1.07||205||0.91||0.79–1.05||60||1.04||0.79–1.33||69||1.02||0.80–1.30|
|Central European countriesd||135||1.45||1.21–1.71||124||1.32||1.10–1.57||31||1.02||0.69–1.45||34||1.08||0.75–1.51|
|Other eastern European countriese||64||1.85||1.42–2.36||69||2.04||1.58–2.58||10||0.95||0.45–1.75||11||1.08||0.54–1.95|
|Turkey||21||2.62||1.62–4.02||20||2.86||1.75–4.43||1||0.38||0.00–2.16||0|| || |
In general, in second-generation immigrant men, the risks of developing CHD were similar to those in their first-generation parents, and only small differences were found when risk was calculated based on fathers’ countries of birth as opposed to when it was calculated based on mothers’ countries of birth. In addition, second-generation men born to mothers from southern Europe had an increased risk of developing CHD. Six groups of second-generation men had CHD risk levels higher than those of their first-generation parents: men with fathers or mothers born in Denmark, Norway, southern Europe, central European countries, other eastern European countries and Turkey. The increased risk in first-generation immigrants from Finland declined in second-generation immigrants from this country but remained significantly higher than in the Swedish reference group.
Few cases of CHD were observed in second-generation immigrant women (Table 3). Only daughters of Finnish fathers or mothers had a significantly increased risk of developing CHD, 1.46 (95% CI 1.27–1.68) and 1.29 (95% CI 1.15–1.44), respectively.
The risk of developing CHD amongst second-generation immigrants with two foreign-born parents was also analysed (data not shown). The results show the same pattern as was found when the father and mother were treated separately (Table 3). For example, when both parents were born in Finland, the risk was 1.41 for men (95% CI 1.28–1.54) and 1.48 for women (95% CI 1.27–1.73). Only people with both parents from southern Europe showed higher risk (3.64; 95% CI 1.73–6.71), and even in this case, the confidence intervals of the results overlapped (please see Table 3 for comparison).
The main and worrying finding of this study is that the increased risk of CHD amongst certain immigrant groups is also present in the next male generation. First-generation immigrant men and women born in Finland, central European countries, other eastern European countries and Turkey had increased risks of developing CHD that persisted in second-generation men regardless of the sex of the immigrant parent.
The results of our study partly agree with those of another Swedish study that focused on rates of CHD in first-generation immigrants, although on different foreign-born groups. In that study, the age-adjusted risk of developing CHD was higher in nine of 12 male foreign-born groups than in Swedes and in seven of 12 female groups. Even after adjustment for educational level and employment status, the risks still increased . The results of the current study are also consistent with previous results from Sweden with regard to the increased risk of CHD in first-generation Finnish immigrants . In addition, the current study showed that the increased risk of developing CHD present in first-generation Finnish immigrants to Sweden persisted in second-generation Finnish immigrants (both men and women). Although the increased risk of developing CHD declined somewhat in second-generation Finnish immigrants the persisting risk could be explained by a combination of environmental and genetic factors. One striking finding of the study; namely, that Finnish women evinced raised levels of CHD risk in both the first and second generations, whereas CHD risk did not persist in the other second-generation female immigrant groups, may in part be due to environmental factors and in part to relatively small numbers in the other second-generation female immigrant groups. However, CHD risk was lower than that of the reference groups in three first-generation female immigrant groups, including women from southern Europe, other western European countries and Baltic countries.
This study, as well as previous studies, paint a relatively complex, sometimes positive and sometimes negative role of immigrant status in CHD rates in immigrant populations compared with the majority population [2–4, 16–18]. The explanation for the higher or lower CHD risk found in certain immigrant groups in the current study may be a combination of genetic and environmental effects. Environmental effects may include risk factors for CHD, low SES, stress caused by lack of acculturation [19, 20] and discrimination.
Some studies show that immigrant groups have better cardiovascular health profiles than the majority population [21, 22], whereas others indicate that immigrant groups have poorer cardiovascular health profiles than the majority population regarding individual risk factors for CHD such as physical inactivity , smoking [24, 25] and obesity [25, 26]. Many risk factors for CHD are also associated with low SES .
It is reasonable to assume that socio-economic disadvantage plays a role in the development of CHD in both first- and second-generation immigrants. In Australia, both first- and second-generation immigrants were at a disadvantage in the labour market  and in Belgium, second-generation Moroccan immigrants had high rates of unemployment . In our study, the SIRs were still significantly high after adjustment for SES, which indicates that socio-economic factors alone cannot explain the increased risk of incident CHD in certain immigrant groups. Other factors must play a role. For example, there is an ongoing discussion about whether lifestyle patterns in the country of origin are transmitted from the first generation to the second generation of immigrants . In addition, international literature shows that with increasing time since immigration, immigrants typically lose health advantages stemming from protective social and cultural factors from their country of origin [22, 31]. In the USA, first-generation Mexican-American men and women had healthier cardiovascular profiles than those born in the US . Second-generation US-born English-speaking Mexican-Americans had healthier cardiovascular profiles than second-generation US-born Spanish-speaking Mexican-Americans. This was likely due to higher levels of acculturation amongst the English-speaking Mexican-Americans, which may result in positive benefits, including higher levels of education, as well as increased access to preventive health services. Second-generation US-born Spanish-speaking Mexican-Americans had the least healthy profiles, possibly because they had lost the protective influences of their native culture but had not yet gained the protective influences associated with the dominant English-speaking culture.
The authors of a 1976 study on CHD in Japanese-Americans found that the most traditional group of Japanese-Americans had a CHD prevalence rate as low as that observed in Japan. The group that was most acculturated to Western culture, on the other hand, had a three- to five-fold excess in CHD prevalence . The authors of the 1975 NI-HO-SAN study, who found that CHD mortality was higher in Japanese-American men than in men in the same age group in Japan, hypothesized that the higher rates might be due at least in part to the adoption of American dietary habits by Japanese immigrants in the United States .
In the current study, we lacked information about language spoken, food habits or other indicators of acculturation status. In addition, it is important to note that caution must be exercised when comparing the results of studies on immigrants in one country with the results of studies on immigrants in another country. Groups that make up the majority of the immigrant population of one country may be nearly or entirely absent in others, and conditions in both countries of emigration and countries of immigration vary widely. In Sweden, persons from Finland, the other Nordic countries and other European countries have dominated the immigrant population for many years.
The present study has some limitations. First, residual socio-economic confounding likely exists. In a previous study of SES in Blacks and Whites, residual confounding was present . Secondly, though the national database includes data on the entire Swedish population, it only incorporates information about hospital admissions for CHD. However, in Sweden most people with CHD events are hospitalized. In addition, deaths from CHD outside hospitals are also included in the study. Thirdly, we had no data on risk factors for CHD, like smoking, diet and physical activity. Furthermore, we had no data from countries of origin about background CHD risk factors or morbidity. However, we did adjust for SES, which is associated with several risk factors for CHD, such as smoking . Fourthly, we only had information on hospital diagnoses of CHD for 1987–2001, which means that the first hospitalization for CHD in the study period might not have been the individual's first hospitalization for CHD. However, we excluded ICD codes 412 (old myocardial infarction), 414 (other forms of chronic CHD), I22 (reinfarction, within 4 weeks) and I25 (chronic CHD), which means that we could exclude at least some people with chronic CHD. Finally, we did not have individual data on date of arrival in Sweden, so we do not know how long first-generation immigrants had lived in Sweden.
This study also has a number of strengths. For example, our study population included a well-defined open cohort of first- and second-generation immigrants. Because of the civic registration number assigned to each individual in Sweden, it was possible to track the records of every person for the whole follow-up period. Data about occupational status were almost 100% (99.2%) complete (1985 and 1990 censuses), which enabled us to adjust our models for SES. The validity of the CHD diagnosis from the hospital registers has been shown to be high in an evaluation by the National Board of Health and Welfare . In 2001, the main diagnosis was missing in 0.9% . Finally, very few pieces of data are missing from the multigenerational part of the MigMed database; it includes information about parents, children, siblings and adoptions for index persons born in 1932 or later and domiciled in Sweden any time between 1947 and 2001. For example, for Swedish-born index persons (including second-generation immigrants), data on mothers are missing for only 3% of persons, and on fathers for 5% .
Though for people in some immigrant groups, the risk of developing CHD was similar to or even lower than the risk for people in the Swedish reference groups, other first- and second-generation immigrant groups exhibited statistically significantly higher rates of first hospitalization for or death from incident CHD. The detrimental effects caused by CHD are many and serious, and healthcare resources are limited. It makes sense to use the information from this study to help target these limited resources to where they are most needed, such as the prevention of CHD amongst specific first- and second-generation immigrant groups known to be at higher risk of developing the disease.
Conflict of interest statement
There are no conflicts of interest.
This work was supported by grants from the National Institutes of Health (grant no. R01-H271084-1), the Swedish Research Council (grant no. K2005-27X-15428-01A to K.S.) and the Swedish Council for Working Life and Social Research (grant no. 2001-2373). The authors wish to thank Scientific Editor Kimberly Kane at the Center for Family and Community Medicine for useful comments on the text, and Sanna Sundquist at the University of California, San Diego, for technical assistance.