Migrant mortality differences in the 2000s in Belgium: interaction with gender and the role of socioeconomic position

Many studies have already assessed the health status of migrant communities. These studies often show a migrant mortality paradox, at least among first-generation (FG) migrants [1‐8]. Compared to the native population, FG migrants seem to have a mortality advantage which is quite paradoxical as they tend to belong to the lower socioeconomic strata of society [2, 5]. In literature, apart from being the result of possible data artefacts [1, 8], this paradox is explained in several ways. Firstly, to be able to cope with the challenges of the journey of migration, migrants are a selective group of healthy people, i.e. the healthy migrant effect [1, 5‐12]. Secondly, migrants can take advantage of the healthy lifestyle of the home country [1, 5‐8, 11] and of the well-developed and efficient healthcare system of the host country [1, 7, 12, 13]. In this way migrants combine the good things of the home and host country, leading to a double advantage. Migrants not only combine the advantages of both worlds, they also have been confronted with different exposures in both the home and host country [4, 9, 14, 15]. This could imply that FG migrants pass through the epidemiological transition at an accelerated pace, at least for migrants from non- or less industrialised countries [8, 15, 16]. The shift from unfavourable hygienic conditions in the home country to better hygienic conditions in the host country could lead to a decline in infectious-related mortality. At the same time migrants may adopt a western lifestyle which could increase chronic-disease mortality, yet only after a certain time lag; the so-called migration-as-rapid-health-transition hypothesis.

Studying cause-specific mortality patterns of the migrant population relative to that of the host country population may reveal important clues on disease aetiology and on the different hypotheses concerning migrant mortality [1, 9, 17, 18]. Belgium is a particularly apt setting to study this as it has a long history of migration [1, 2, 4], with nowadays a proportion of people with a migration background of about 20% [19]. As previous research has already probed this question [1, 4, 18], we will go a step further by adding gender and socioeconomic position (SEP) into the picture. Therefore, the objectives of this paper are to assess whether there is indeed a migrant mortality paradox in Belgium, to probe into the interaction with gender, and to assess the role of different SEP indicators in this paradox. More specifically, we will document migrant mortality patterns from 2001 to 2011 for all important migrant groups in Belgium for the five most important causes of death (COD). Since men and women have a different migration history [5, 20] we will focus on the interaction between migrant origin and gender in the analyses. Additionally, as life chances may be very different for migrants, as well as for migrant men and women, we will focus on the role of different SEP indicators in explaining these mortality differences.

Data were derived from the individual linkage of three data sources: the Belgian census, the Belgian National Register and death certificates. In a first stage, a link was made by Statistics Belgium between the Belgian census of October 1st, 2001 and Register data for the follow-up period of 2001–2011. The census contains demographic and socioeconomic information for the total de jure population living in Belgium. The linkage with the National Register enabled us to account for all emigration and mortality during the study period. For all Belgian residents that died during the study period, cause-specific mortality was added by individual linkage with death certificates. The study population consisted of all Belgian residents with an attained age during follow-up between 25 and 64 years. The lower age limit was set to include a sufficient number of persons and deaths. Furthermore, the lower and upper age limit were set to be able to assess the various SEP indicators. For example when including employment status, it is essential to exclude persons outside the economically age range.

In this paper, our aim is to study the interaction between migrant mortality and gender, as well as the contribution of SEP to this association. More specifically, we intended to compare the mortality patterns of the native Belgian population with those of the largest migrant groups in Belgium, i.e. migrants originating from the neighbouring countries (the Netherlands (i.e. Dutch migrants) and France), Spain, Italy, Eastern Europe (Poland, Hungary, Romania, Czechoslovakia, Bulgaria, Czech Republic and Slovakia), Turkey, Morocco and Sub-Saharan Africa (SSA) (Congo (Zaire), Burundi and Rwanda). A migrant was defined using a stepwise method, based on both the individual and his/her parents’ country of origin, hereby maximizing the population with migrant roots. If an individual could be linked to his/her parents at the census of 2001, and one of his/her parents had its roots (i.e. nationality at birth) outside Belgium, we used this country as the individual’s country of origin. If the individual could not be linked to his/her parents in 2001, we checked whether the individual could be linked to his/her parents in the 1991 census, and then repeated the first two steps. If, however both parents had their origins in Belgium or if the parents’ nationality at birth was missing, the individual’s nationality at birth was used to define the individual’s country of origin. If this information was also lacking, current nationality was used to define the person’s country of origin. We did not distinguish first from second-generation migrants since previous research has already examined the different mortality patterns for both migrant generations in Belgium [18, 21, 22]. These studies generally observed a clear mortality advantage among first-generation migrants whereas the mortality patterns among second-generation migrants tended to converge towards mortality levels of native Belgians with similar SEP. Instead, in this paper we wanted to study the association with gender and SEP, two understudied subtopics in migrant mortality studies. We included three SEP indicators as they may represent different forms of disadvantage during different periods of life [23], and may have a different meaning in explaining mortality inequalities for men and women, migrant origin or COD. Educational attainment reflects chances early in life and may therefore not be the best suitable indicator among migrants [24, 25]. Employment status on the other hand is an SEP indicator at middle-age. Yet, given the different migration motivations of men and women, this indicator may be more informative for men than for women. That is why we also included home ownership, which refers to the economic assets at a household level. Educational attainment was categorized according to the International Standard Classification of Education (ISCED): primary education or no diploma (ISCED 0–1), lower secondary education (ISCED 2), upper secondary education (ISCED 3–4), and tertiary education (ISCED 5–6). Home ownership differentiates between tenants and owners of a dwelling. Finally, employment status contrasts employed people, unemployed people who are actively looking for a job, and unemployed people who are not looking for a job (e.g. students or retired people). The associations between migrant origin, gender and SEP were studied for all-cause mortality as well as for the five most common COD among Belgian adults, as classified according to the International Statistical Classification of Diseases and Related Health Problems (ICD - tenth revision): all cancers (C00-C99); circulatory diseases (I00-I99); lung cancer (C33–34); injuries (S00-T99); and respiratory diseases (J00-J99). Age-standardized mortality rates were calculated by gender and migrant origin as well as age-adjusted mortality rate ratios using age-adjusted Poisson models with native Belgians as the reference group. In these Poisson models, the interaction between migrant origin and gender was introduced. Furthermore, to assess the role of various indicators of SEP in these mortality differences, we adjusted the Poisson models for the three indicators of SEP: educational attainment (model 1), home ownership (model 2) and employment status (model 3). All analyses were performed using Stata/MP 14.2.

In order to provide high-quality, accessible and adequate healthcare, it is necessary continuously to monitor the mortality patterns of the population, including the ageing migrant population [1, 18, 23, 30]. Future studies should also take morbidity differences into account between the migrant and host population, since lower mortality not necessarily implies better health. Moreover, longitudinal studies, including information on life-long exposures, health behaviours and healthcare use might allow for a better understanding of the nature of health inequalities [15, 30].

We learned that migrants generally had lower mortality than the native Belgian population, for some groups even after adjustment for SEP. This suggests that there is potential room for improvement to protect our health [1, 18]. Primary prevention aiming at the adaptation of the typical western lifestyle with high levels of consumption of tobacco, alcohol and meat and a sedentary lifestyle could have a huge effect on the mortality pattern in countries such as Belgium [8, 18, 23, 30, 38], as well as in other European countries, taking the example of the high lifestyle-related mortality among French migrants.

This study stressed the importance of integrating a gender perspective in studies assessing mortality and morbidity patterns within the migrant and host population [26]. We observed larger migrant mortality variation among men. Yet, family healthcare is often organized by the women in immigrant families, and these women are more often unemployed and hence at risk of being socially isolated. Therefore, it is essential to eliminate structural, social, financial and cultural barriers for healthcare use by migrant men and women [26, 39, 40]. Since SEP had quite some impact on the mortality differentials, organising preventative and curative healthcare that is equally accessible across social strata is also important [26]. Our results suggest that one’s SEP may contribute as a protective factor and vice versa. It is thus essential to stimulate good housing and employment chances among all social strata and migrant groups [4, 18], not only for their well-being but also in terms of physical health.