Results

Table 1 illustrates that the proportion with poor self reported health was 15.1% among men and 17.2% among women. The two largest country of origin subgroups apart from those born in Sweden were those born in other Western countries and all other countries. No clear differences in country of origin between men and women were seen. People in older age brackets had answered the questionnaire to a higher extent than younger people. The proportion of participants with more than 12 years of education was 25.6% among men and 25.5% among women. Somewhat less than half of both the male and female population had nine years of education or less. Approximately a third of both men and women were exposed to low social participation, low social anchorage, low emotional support and low instrumental support. The proportion of people unable to pay bills every month was 8.5% among men and 7.8% among women, while 67.2% of all men and 64.6% of all women never had such problems.

Table 2 shows that the odds ratio of having poor self reported health was significantly higher among men born in other Western countries, Yugoslavia, Arabic speaking countries and the category all other countries compared with men born in Sweden. Men born in Poland had the odds ratio that was closest to the Swedish born reference group, odds ratio 1.3 (0.6, 2.8). Women born in Yugoslavia, Poland, Arab countries and other countries all had significantly higher odds ratios regarding poor self reported health than women born in Sweden. In contrast, women born in other Western countries had approximately the same odds ratio as women born in Sweden, 1.4 (0.9, 2.2).

Table 3 illustrates that the odds ratio of having poor self reported health increased in higher age groups, and increased with lower educational level. The people with low social participation had increased odds ratios of having poor self reported health, 3.1 (2.6, 3.8). The odds ratios of having bad self reported health were also increased for both men and women who scored low on the other three psychosocial variables (social anchorage, emotional support and instrumental support), although not to the same extent as for social participation. The people who could not pay their bills every month had an increased odds ratio of having poor self reported health, 4.0 (3.0, 5.2).

Table 4 illustrates that adjustment for age in fact increased the odds ratios of having poor self reported health in most ethnic groups, compared with people born in Sweden. This was particularly clear for men born in Arabic speaking countries, when odds ratio changes from 3.0 (1.6, 5.5) to 3.8 (2.0, 7.1), and women born in Arabic speaking countries, odds ratio changes from 1.4 (0.6, 3.0) to 1.6 (0.7, 3.6)). Men born in other Western countries, Yugoslavia, the Arab countries and the all other countries category had significantly increased odds ratios of having poor self reported health compared with men born in Sweden. Women in all groups except Western countries and Arab countries had significantly increased odds ratios of having bad self reported health compared with women born in Sweden. The inclusion of education in the multivariate analyses only slightly changed the odds ratios for both men and women in the different ethnic groups. Finally, the introduction of social participation, social anchorage, emotional support, instrumental support and economic stress into the multivariate model reduced the odds ratios of having bad self reported health for all ethnic groups compared with the Swedish born reference group, but the reduction was particularly pronounced for both men and women born in Yugoslavia, Arab countries and all other countries, and for women born in Poland. For men born in Arab countries the odds ratio was reduced from 3.8 (2.0, 7.2) to 1.1 (0.5, 2.3), for men born in all other countries from 2.8 (1.8, 4.3) to 1.6 (1.0, 2.6), and for men born in Yugoslavia from 2.7 (1.5, 4.8) to 1.4 (0.7, 2.8). Similar reductions of the odds ratios were seen for women born in these countries. The odds ratio for women born in Poland was also reduced from 3.6 (2.1, 6.0) to 2.3 (1.3, 4.0).

Discussion

The main finding of this study was the augmented odds ratios of having poor self reported health among men born in other Western countries, Yugoslavia, Arab countries and in the category all other countries as well as among women born in Yugoslavia, Poland and the category all other countries compared with those in the Swedish born reference group. However, adverse psychosocial and economic living condition factors seemed to have a strong influence on poor health in these groups because a large reduction of the odds ratios was observed for both men and women born in Yugoslavia, Arab countries and all other countries, and for women born in Poland after the introduction of these factors into the multivariate model.

Self reported health is an important independent predictor of mortality.4 28 Both self reported health and mortality are closely related to age, but their mutual relation may not be the same in all age groups as among the old. However, even for the younger and middle aged, the respondent's perceived global health is a consistent and significant predictor of mortality.4 It has been proposed that cultural differences between ethnic groups may imply that they perceive their combined physical and psychological health differently.29 However, there was a strong association between self reported health in different ethnic groups in the USA and the total mortality of these ethnic groups.1 This association consequently seems to be universal rather than culturally determined.

The finding that psychosocial and economic conditions had the strongest influence on the increased risks of having poor self reported health among those who had immigrated to Sweden from countries that were geographically most distant and culturally most dissimilar to Swedish society—that is, the groups born in Arabic speaking countries, Yugoslavia and all other countries—agreed with another Swedish study.30 It was found that economic difficulties in exile, a low sense of coherence, poor acculturation (men only), and poor sense of control, were stronger risk factors for self reported psychological distress in this group than exposure to violence before migration.30 Moreover, a British study found that poor social support in exile was a stronger predictor for depression than trauma factors.31

Studies of migrants are prone to mistaken conclusions, as both culturally determined factors inherited in the country of origin as well as current social and economic living conditions in the new country may affect the behaviour and self perceived health of the immigrants.32 The higher risks of a poor self reported health status may theoretically have been determined by culturally inherited differences in perceiving health, or by events that occurred in the native countries resulting from oppression, violence, torture or economic exploitation. The increased risks may alternatively have been determined by current psychosocial and material conditions in the new country. Our results suggest that socioeconomic, psychosocial and economic conditions in the new country were important determinants of self rated health of ethnic minority immigrant groups in Malmö in southern Sweden in the 1990s. The importance of the psychosocial and economic determinants increased with the geographical and cultural distance between the country of origin and Sweden. However, men born in other countries, particularly refugees from non-European countries, still had an increased risk of poor self reported health after adjustment for socioeconomic, psychosocial and economic conditions in the new country. Moreover, the finding of a strong risk of poor health status among Polish women partly agreed with a Swedish study that found an increased risk of suicide among Polish women, higher than in their countries of birth.33

The people who migrate from their country of birth to another country or even another part of the world, are generally more healthy than those who do not migrate from their native country. However, this “healthy migrant effect” tends to wear off with time.34 One important reason could be that psychosocial and economic conditions in the new country affect the health status in a detrimental and unhealthy way. The results of this study further strengthen the notion that such structural improvements as the integration of immigrants in Swedish society and their increased participation on the Swedish labour market are absolutely crucial for achieving health equity between Swedish born people and immigrants.

LIMITATIONS AND STRENGTHS

This study has several limitations. To begin with, its cross sectional design makes it difficult to draw inferences about causal pathways between migration, psychosocial and economic conditions and self reported health. Next, there is the possibility of bias from self reported data. Although self reported health status is a useful indicator of the health conditions of a population, it is a subjective and imprecise measure of health, which reflects a person's general perception of health more than well defined health outcomes. Doubt about the significance of self rated health status have been suggested by an Australian study that revealed unexpected gender differences in health ratings and survival.35 In addition, the reference point for assessment of self rated health is not absolute and varies with demography and social context. However, the stability over time of self reported health status was demonstrated, as well as the fact that this kind of subjective health assessment was a valid indicator of health in middle aged populations and could be used in cohort studies.10 Moreover, the test-retest reliability of self rated health status has been good.36

The classification of immigrants is general in this paper. Some of the respondents of foreign origin may be refugees, others may be people who have moved as a step in their occupational career. There may thus be different underlying social and economic circumstances that cause migration from one country to another. However, from the end of the 1940s until the beginning of the 1970s Sweden almost exclusively had a labour market immigration from countries within the category Other Western countries and the Yugoslavia category. In the 1970s there was a transition from labour market immigration to refugee immigration from the countries within the Arabic speaking countries and Other countries categories. The Yugoslavia category has also been a source of refugee immigration from 1992 and onwards, but only a few were domiciled in the Swedish population registry in 1994.37 This pattern is also reflected by the fact that only 14% of the Arabic speaking men and 4.5% of the Arabic speaking women in the 1994 public health survey had moved to Sweden before 1985, while 89% in the Other Western countries category had moved to Sweden before 1985.

The questionnaire was sent to the respondents without any translations or any translation help. Sweden provides courses in Swedish for immigrants, “Svenska för invandrare”, and some of the introductory courses are even mandatory. As the persons who were chosen to be respondents were randomly chosen from the population register of Malmö, they must also have been living in Sweden for a year.

The Arabic speaking world is not culturally homogeneous. However, compared with the Western world and compared with Sweden, it is rather homogeneous both when it comes to language and religion. This categorisation is also as close we can get by using this empirical material.

Country of origin does not confer any particular legal status within Sweden. Even foreign citizens (11% of the Malmö population) have the right to vote in municipal elections, although not in elections to the National parliament. It is thus not plausible that legal status have any impact on the health status of the migrants of this study.

No item concerning previous/current disease was included in the 1994 public health survey questionnaire. This possible confounder may account for some of the ethnic differences in self reported health. This may infer a reversed causality in the direction from disease and thus poor self reported health to social and economic conditions. However, this reversed causal pathway is probably not quantitatively important, and most importantly not in the younger age brackets were, for example, the Arabic speaking group is concentrated.

The results of this study could also be biased by selection bias and confounding. The participation rate is 71%, which is fairly high. Furthermore, the proportion of people born in countries other than Sweden was 24% in Malmö in 1994 according to the population register covering the entire population. The 26.7% proportion of people born in other countries than Sweden rather seems to be a slight overrepresentation of the foreign born part of the Malmö population in the age brackets included in this study. The distribution of the country of origin categories is also similar to that observed in the general Malmö population.25 Non-participation is thus not likely to have produced a serious selection bias in this study. Age, sex, and education might be confounders of the associations between country of origin and self reported health. Adjusting for these possible confounders, however, only marginally affected the estimates.

There are other positive aspects of this study. The 1994 public health survey in Malmö on which we base our findings is highly representative both concerning the whole group born in other countries than Sweden, and concerning the proportions of the different ethnic subgroups, compared with official register data concerning the composition of the whole population of the city of Malmö in 1994. The comparably high proportion of people of, for example, Arab extraction combined with the good representation of this group within the study makes it possible to draw inferences and conclusions concerning the health of an ethnic minority group that has rarely been studied.

The results of this study suggest that psychosocial and economic conditions should be considered as determinants of poor health in certain ethnic minority groups, and thus a much broader social perspective should be taken into account when facing these patients.

To conclude, this study presents support for the notion that psychosocial and economic conditions are highly important as causal determinants of self rated health among important ethnic minority groups in Swedish society. Ethnic differences in self reported health are thus not primarily a consequence of cultural differences between different immigrant groups.