University of Groningen Healthy ageing in a comparative perspective Reus Pons, Matias

(1)

Healthy ageing in a comparative perspective

Reus Pons, Matias

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date: 2018

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Reus Pons, M. (2018). Healthy ageing in a comparative perspective: A study of the health of older migrants and non-migrants across Europe. University of Groningen.

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

(2)

chapter 5

Health differences between migrants and non-migrants aged

50 to 79 in Europe: the role of integration policies and public

attitudes towards migration and migrants (2004-2015)

This chapter is based on: Reus-Pons M, Vandenheede H, de Valk HAG (2018). Health differences between migrants and non-migrants aged 50 to 79 in Europe: the role of integration policies and public attitudes towards migration and migrants (2004–2015). Manuscript submitted for publication, current status: revise and resubmit.

(3)

Abstract: Integration policies and public attitudes towards migration and

mi-grants have received little attention in previous studies on the health differences between migrants and non-migrants. Our aim is to take these dimensions, which are likely to influence the position of migrants in society, into account in an examination of the health differences between migrants and non-migrants aged 50–79 across 10 European countries. To cover different dimensions of health, we use a variety of health indicators: self-rated health, diabetes, and depression. The data comes from the Survey of Health, Ageing, and Retirement in Europe (SHARE), and are enriched with data from the European Social Survey (ESS) and the Migrant Integration Policy Index (MIPEX). Results from multivariate logistic regression analyses showed that migrants, especially those of non-western origin, had higher odds than non-migrants of having poor self-rated health, diabetes, and depression. We furthermore found that less favourable public attitudes to-wards migration and migrants were associated with higher rates of poor self-rated health, diabetes, and depression among (non-western) migrants; however, the association between integration policies and migrant health was less clear. In light of these results, we conclude that favourable public attitudes towards migration and migrants are more important for migrant health inequalities at older ages than more inclusive integration policies. Our findings contribute to the scientific

literature by taking into account the role of the policy and societal context when assessing health inequalities between older migrants and non-migrants in Europe.

Keywords: migration, health, ageing, Europe, integration policies, attitudes

(4)

Chapter 5

5.1 InTroducTIon

While there are some studies on migrant health at older ages, most focused on specific health outcomes or on specific countries of origin or residence, while few considered the effect of contextual country of residence differences (Castañeda et al., 2015). Nevertheless, some researchers have argued that the context of reception is relevant for, among others, older migrants’ health outcomes (Bor-done & de Valk, 2016). Public attitudes towards migration and migrants, and policies that facilitate migrants’ integration into society have been suggested as factors that could influence migrants’ health (Castañeda et al., 2015). Including these dimensions in research that examines the health differences between older migrants and non-migrants in Europe is important for several reasons. First, the share of older migrants in European populations is increasing rapidly (Lanzieri, 2011). Second, health disparities merit attention because European health care systems are based on equity (Nørredam & Krasnik, 2011). In order to reduce po-tential health differences between migrants and non-migrants (Razum & Stronks, 2014), special attention must be paid to reducing socioeconomic inequalities and discrimination (Nazroo, 2003). The latter goal could be advanced through more inclusive integration policies and more favourable public attitudes towards migra-tion and migrants.

Evidence on the association between integration policies and health outcomes is however scarce. The existing studies have mainly focused on very specific (un-documented) migrant groups (Martinez et al., 2015). A few recent studies have focused on the association between integration policies and migrants’ self-rated health (Malmusi, 2015) and mortality (Ikram et al., 2015) in Europe. However, none of these studies specifically focused on older migrants, who had reached a life stage in which health issues tend to become more important.

For similar reasons, public attitudes towards migration and migrants are likely to be important when assessing health among migrants. Research that has fo-cused on the association between racism or discrimination and health has shown that there is a clear association between discrimination and poor self-rated and mental health, including depression (Karlsen & Nazroo, 2002; Williams et al., 2003; Paradies, 2006; Johnston & Lordan, 2012; Levecque & van Rossem, 2015). While discrimination measures individual experiences, we argue that the general climate regarding migration and migrants is also likely to contribute to migrants’ health. To our knowledge, only one previous study has analysed the association between more general attitudes towards migrants and migrant health (Huijts &

(5)

Kraaykamp, 2012), while there is no research on the association between attitudes towards migration and migrant health.

The health outcomes of both migrants and non-migrants in a given country are associated with the country’s broader societal and policy context (e.g. social and health care policies). However, the country’s integration policies and public attitudes towards migration and migrants are likely to affect the outcomes of migrants in particular. Accordingly, we believe that the relative (rather than the absolute) health differences between migrants and non-migrants are the most appropriate indicator for evaluating the association between policies and public attitudes, and migrant health outcomes. Although examining relative health dif-ferences does not allow us to identify in which countries migrants attain the best health outcomes, it can help us compare the health of migrants and non-migrants within a specific policy and societal context.

Therefore, the aim of this article is to study the differences in health between migrants and non-migrants aged 50–79 in ten European countries, and how these differences are affected by integration policies and public attitudes towards migration and migrants. We distinguished between migrants with a western and a non-western background. We decided to focus on the older population; since most of the older migrants in Europe immigrated many years ago, and have thus been exposed to the policy and societal context in the country of residence for many years. The length of residency is important because some chronic condi-tions (including diabetes) are not clinically diagnosed until after a long latency period after exposure to risk factors (Law & Wald, 1999; Razum & Twardella, 2002). Finally, since European societies are becoming older and increasingly multicultural (Lanzieri, 2011), our results could help us identify policies and at-titudes that could potentially reduce the health differences between migrants and non-migrants.

We included different dimensions of health: overall, physical, and mental health. Although poor self-rated health has been shown to be associated with more objective measures of poor health, its validity when comparing distinct ethnic groups has been contested (Chandola & Jenkinson, 2000). Therefore, we included additional health measures to strengthen the validity of our results and pinpoint the health dimensions where differences are potentially largest between migrants and non-migrants. Self-reports of chronic diseases subject to clinical diagnosis, and especially of diabetes, have been shown to be fairly accurate (Kriegsman et al., 1996). Finally, the EURO-D scale for measuring depression was developed

(6)

Chapter 5

by Prince et al. (1999), who demonstrated the scale’s reliability and validity for cross European comparisons. The EURO-D scale has also been used to analyse inequalities in mental health between migrants and non-migrants (Aichberger et al., 2010). The inclusion of various dimensions of health (i.e., self-rated health, diabetes, and depression) allows us to examine how differences in policies and attitudes might be related to diverse health outcomes.

We used pooled data from the Survey of Healthy Ageing and Retirement in Eu-rope (SHARE) from 2004 to 2015. SHARE collets demographic, socioeconomic, and health data on individuals aged 50 and older (Börsch-Supan et al., 2013). To study how policies and attitudes are associated with differences in health between older migrants and non-migrants, we enriched the dataset with country-level vari-ables. Information on integration policies was taken from the Migrant Integration Policy Index (Niessen et al., 2007), while information on public attitudes towards migration and migrants was derived from the European Social Survey (ESS, 2014). The relevant information was pooled with the SHARE data to enable us to study how and to what extent these factors were related to migrant and non-migrant health in Europe.

5.2 mIgrATIon To wEsTErn EuropE

Most of the older migrants who currently live in Europe immigrated after World War II. We can distinguish three main periods of migration to western Europe from the 1950s until today. Between the 1950s and the 1973 oil crisis, most migrants to western European countries were from former colonies, or were so called ‘guest workers’ from southern Europe, Turkey, Morocco, or Al-geria (van Mol & de Valk, 2016). Considerable labour migration also took place between neighbouring countries: e.g., from Ireland to the UK or from Finland to Sweden (Jennissen et al., 2006). After this period, the next phase until the end of the 1980s was characterised by a cessation of labour recruitment. While some migrants returned to their country of origin, others settled in Europe, and migration continued through family reunification or family formation (van Mol & de Valk, 2016). From the mid-1980s onwards, international migrants were also entering former emigration countries in southern Europe, especially Italy and Spain (Castles et al., 2014). A third period characterised by an increase in the number of asylum seekers and in intra-European mobility started in the 1990s (van Mol & de Valk, 2016).

(7)

Most of the older migrants living in western Europe today are first generation migrants who arrived during one of these migration waves (de Valk & Schans, 2008). Accordingly, most came as a labour migrant, or as a migrant from a neigh-bouring country or former colony (Lanzieri, 2011). Germany is different from other European countries in that many ethnic Germans (‘aussiedler’) migrated to Germany from the 1950s to the 2000s, first from central European countries, and later (after 1989) from the former Soviet Union (Jennissen et al., 2006; van Mol & de Valk, 2016). The ‘aussiedler’ are unlike other migrant groups because they were granted immediate citizenship, and were not officially considered migrants (Castles et al., 2014). However, their diverse backgrounds, their position in soci-ety, and their frequent lack of knowledge of the German language or of German society made them de facto migrants (Münz & Ohliger, 1998).

5.3 HEAlTH dIffErEncEs bETwEEn (oldEr) mIgrAnTs And non-mIgrAnTs

Most previous research on health differences between migrants and non-migrants focused on the total population or the working-age population only. These studies generally found that, despite their relatively low socioeconomic status, migrants in Europe tend to have lower mortality than non-migrants; a phenomenon that has been referred to as the ‘migrant mortality paradox’ (Razum et al., 1998; Deboosere & Gadeyne, 2005; Boulogne et al., 2012). However, this does not necessarily imply that migrants are healthier than non-migrants, since some chronic conditions may affect levels of disability and quality of life without being life threatening (Uiten-broek & Verhoeff, 2002). Indeed, the gaps in health outcomes between migrants and non-migrants seem to vary according to the health indicator in question, since different health indicators capture different aspects of health or different stages in the disablement process (Jagger et al., 2011).

Nevertheless, studies on older migrants (Carnein et al., 2014; Lariscy et al., 2015; Reus-Pons et al., 2016) have suggested that the migrant mortality paradox per-sists even at older ages. However, when other health dimensions are taken into account, older migrants appear to be in worse health than non-migrants. For example, compared to their non-migrant counterparts, older migrants tend to have more chronic conditions, lower levels of self-rated health and functioning, more limitations in daily activities, and higher rates of depression (Solé-Auró & Crimmins, 2008; Aichberger et al., 2010; Lanari & Bussini, 2012; Carnein et al., 2014; Reus-Pons et al., 2017).

(8)

Chapter 5

Previous attempts to explain the health differences between migrants and non-migrants have primarily focused on selection effects, such as the ‘healthy migrant effect’ and the ‘salmon bias’ hypotheses (Razum et al., 1998; Abraído-Lanza et al., 1999; Palloni & Arias, 2004; Norredam et al., 2014; Vandenheede et al., 2015); and on health related behaviours and their relationship with length of residence, such as the ‘acculturation’ hypothesis (Abraído-Lanza et al., 2005; Riosmena et al., 2015). These hypotheses have also been applied to the older population (Aguila et al., 2013; Riosmena et al., 2013; Thomson et al., 2013). While most of these studies found support for the assumption that migrants were selected for good health at the time of migration, they found less support for the assumption that unhealthy migrants returned to their home country, especially in Europe. Although support for the acculturation hypothesis is inconclusive, there is evidence that migrants’ health levels tend to worsen with increasing length of residence, and thus to converge with non-migrants’ health levels as the initial healthy migrant effect fades away (Norredam et al., 2014; Vandenheede et al., 2015).

Unlike selection and acculturation effects, contextual country of residence fac-tors have so far received little attention in migrant health research. Integration policies and attitudes towards migration and migrants could explain why health differences between migrants and non-migrants vary across destination countries (Agyemang et al., 2010; Castañeda et al., 2015).

5.4 socIAl dETErmInAnTs of HEAlTH

Our framework for analysing the association between integration policies and public attitudes towards migration and migrants follows the principles of the WHO model on the social determinants of health (Solar & Irwin, 2010), in line with previous studies (Ikram et al., 2015; Malmusi, 2015). This model assumes that the association between the structural determinants and health outcomes is mediated through three intermediate determinants or pathways: material, psy-chosocial, and behavioural factors (Solar & Irwin, 2010). In our study, we focus on specific structural determinants of health differences between older migrants and non-migrants: namely, the policy and societal context, and more specifically, integration policies and public attitudes towards migration and migrants.

5.4.1 Health and integration policies

Strictly speaking, integration policies are policies that focus on minorities as a group (Wright & Bloemraad, 2012). However, this definition excludes many

(9)

poli-cies that extend or deny rights to migrants as individuals (Koopmans, 2013). A broader definition of integration policies would include policies aimed at both individual migrants and groups of migrants (Ersanilli & Koopmans, 2011). As the full socioeconomic incorporation of individuals into the host society is impor-tant for improving health outcomes (Levecque & van Rossem, 2015), we use the broader definition of integration policies in the current study.

Given the relatively rapid changes in the composition of the migrant population, most western European countries have been introducing policies designed to help migrants integrate into the host society. Traditionally, there have been three integration models: differential exclusion, assimilation, and pluralism or multicul-turalism (Castles, 1995; Meuleman & Reeskens, 2008). The aim of the differential exclusionist model is to incorporate migrants into certain domains of society, mainly as workers and consumers; while denying them access to other domains of society (permanent residence, citizenship, welfare system, or political participa-tion). In the assimilationist model, by contrast, integration is conceptualised as occurring in a single direction, and migrants are expected to adopt the culture, traditions, and values of the destination society. Finally, in the pluralist model migrants are incorporated into society, but are allowed to maintain their language, culture, traditions, values, and religion (Castles, 1995). Within Europe, Germany, Switzerland, and Austria are considered exponents of the differential exclusionist model; France is often classified as an assimilationist country; and Sweden and the Netherlands are often classified as pluralist countries (Castles, 1995; Statham et al., 2005). In certain countries, such as Belgium and Spain, integration policies were not developed until the late 1990s or the 2000s (Phalet & Swyngedouw, 2003; Pasetti, 2014). Although these classifications may be less relevant nowadays, and the rights granted to migrants have indeed improved over time, the relative positioning of countries from those granting the most to those granting the few-est rights to migrants has remained largely unchanged (Koopmans et al., 2012). In recent decades, there have been new initiatives aimed at better capturing the diverse dimensions of integration policies. For example, the Migrant Integration Policy Index (MIPEX) is a widely used multidimensional indicator that measures institutional opportunities for migrants to participate in the host society, and is reassessed yearly. We use this Index to capture the diversity in integration policies across Europe.

Generally, more inclusive policies foster the social and economic integration of migrants in the destination country (Wright & Bloemraad, 2012). By facilitating

(10)

Chapter 5

migrants’ access to employment opportunities, higher income, or better housing, such policies may have beneficial effects on migrants’ health (Ikram et al., 2015). In contrast, policies that limit migrants’ rights may have the opposite effect, and could even induce stress (Gushulak et al., 2010; Agyemang et al., 2011). Stress may in turn cause physiological changes that, if prolonged, could result in a dysregula-tion of inflammatory responses of the body to a wide range of diseases, including diabetes and depression (Cohen et al., 2007; Kinzie et al., 2008; Agyemang et al., 2011; Johnston & Lordan, 2012; Iwata et al., 2013). Additionally, both poor living conditions and stress are likely to have a negative influence on health related practices, such as diet and exercise. These suboptimal practices could in turn lead to overweight and obesity, both of which are associated with diabetes and depression (Onyike et al., 2003; WHO, 2003; Kodjebacheva et al., 2015).

Across Europe, migrants come from a wide range of countries, and can be roughly classified as being of non-western or western origin. Most western migrants come from another European Union country, and thus have residency and political rights that many non-western migrants lack (European Commission, 2015). The association between integration policies and health is therefore especially relevant for non-western migrants.

Thus, we hypothesise that more inclusive integration policies are associated with better health outcomes among migrants in Europe, especially among those of non-western origin (H1).

5.4.2 Health and public attitudes towards migration and migrants

Public attitudes towards migration should be distinguished from public attitudes towards migrants. The former set of attitudes reflects public opinion on, for instance, the numbers of migrants who should be admitted to the country and under which conditions; while the latter set of attitudes reflects public percep-tions of migrants who have already entered the country, and of their impact on society (Ceobanu & Escandell, 2010; Gorodzeisky & Richards, 2016). Recent studies conducted in Europe showed that public attitudes towards migration and migrants were very positive in Sweden and relatively negative in Austria (Goro-dzeisky & Semyonov, 2009; Meuleman et al., 2009). The relative positioning of countries in terms of public attitudes towards migration was generally stable between 2002 and 2007 (Meuleman et al., 2009).

Negative attitudes towards migrants, which are often measured by levels of discrimination, are associated with poor labour market outcomes and access to

(11)

social and health services (Johnston & Lordan, 2012). Repeated exposure to these health damaging conditions can lead to poor health. Discrimination may also induce stress among migrants. A study of Latina migrants in the United States showed that racism and discrimination were among the factors associated with acculturative stress, and increased the likelihood that acculturative stress would develop into mental health problems (Bekteshi & van Hook, 2015). Furthermore, discrimination is associated with health damaging behaviours such as smoking and poor eating and sleeping habits (Sims et al., 2016). In contrast, positive at-titudes towards migration and migrants provide a more favourable context for the social and economic integration of migrants into society (Berry, 1997).

Given that most western migrants are of European origin, we would expect to find that they are less excluded than non-western migrants, whose culture and behaviour are more distant from those of the host society (Gorodzeisky & Semyonov, 2009). Indeed, attitudes towards migrants tend to be more negative with increasing cultural (Schneider, 2008) and religious (Coenders et al., 2008) distance.

Accordingly, we hypothesise that positive public attitudes towards migration and migrants are associated with better health outcomes among migrants in Europe, especially among those of non-western origin (H2).

5.4.3 Additional control variables

Low socioeconomic status is strongly associated with poor physical and mental health outcomes (Nazroo, 2003; Kunst et al., 2005; Mackenbach et al., 2008), and it is very likely that the association between health and policies and attitudes is mediated indirectly via socioeconomic inequalities. Therefore, we include socio-economic status indicators (education and job situation) in our analyses as control variables. By considering education, we also partially capture socioeconomic status during childhood and youth (Bhopal et al., 2002). In addition, we control for age to adjust for potentially different age structures in our populations, and for length of residence in the country of destination to account for the possibility that a migrant’s health will worsen with increasing length of residence (Norredam et al., 2014; Vandenheede et al., 2015). Finally, given the strong association between obesity and obesity related behaviours and diabetes and depression, we included body mass index (BMI) (Onyike et al., 2003; WHO, 2003; Kodjebacheva et al., 2015).

(12)

Chapter 5

5.5 dATA And mETHods 5.5.1 Data sources

Our study sample consists of respondents of migrant and non-migrant origin from the Survey of Health, Ageing and Retirement in Europe (SHARE). SHARE collects data on the health status, socioeconomic status, and social networks of individuals aged 50 and older in several European countries and in Israel (Börsch-Supan et al., 2013). We pooled the data from waves 1, 2, 4, 5, and 6 (2004–2015) (Börsch-Supan, 2017a; 2017b; 2017c; 2017d; 2017e). We selected only those countries for which data was available in all waves: Austria, Belgium, Denmark, France, Germany, Italy, the Netherlands, Spain, Sweden, and Switzerland. Fur-thermore, eastern European countries have very different migration histories and most of them remain up to this date mainly emigration countries (Castles et al., 2014), which makes it difficult to compare the health situation of migrants in these countries with those in western and southern Europe. Although SHARE is a panel survey, at each wave refreshment samples were drawn to increase the sample size and to compensate for panel attrition (Börsch-Supan et al., 2013). For each respondent, we considered information at baseline only. Respondents whose origin was unknown were deleted from our study. Due to the small sample sizes after age 80 for non-western migrants, we restricted our analysis to ages 50–79. 5.5.2 Dependent variables

We included three dimensions of health in our study. First, we used self-rated health as a subjective measure of the respondent’s overall health. Answers to the question: Would you say your health is…? were dichotomised as good (excellent, very good, or good) or poor (fair or poor).

Second, we included diabetes as a measure of physical health. Respondents could select diabetes as a response to the question: Has a doctor ever told you that you had any of the conditions on this card? The original categories were per se dichotomous, indicating whether the respondent did or did not have the specified condition.

Third, we included depression as an indicator of mental health using the EURO-D scale (Prince et al., 1999), which consists of 12 items: depression, pessimism, death wish, guilt, sleep, interest, irritability, appetite, fatigue, concentration, enjoy-ment, and tearfulness. We classified individuals with a EURO-D scale score of more than three as suffering from depression (Dewey & Prince, 2005).

(13)

5.5.3 Independent variables

Migrants were defined as individuals who were not born in the country where they were currently living. While this implies that we included only first gen-eration migrants in our study sample, it should be noted that the vast majority of migrants aged 50+ in Europe belong to the first generation. Furthermore, following the definition of Statistics Netherlands, we classified migrants as being of western or non-western origin. According to this definition, a western migrant is a migrant who was born in Europe (except Turkey), North America, Oceania, Japan, or Indonesia (CBS, 2014).

The SHARE data was enriched with two other data sources to test our hypotheses on the importance of integration policies and of public attitudes towards migra-tion and migrants. Data on migramigra-tion policies was retrieved from the Migrant Integration Policy Index (MIPEX, 2017). MIPEX is a multidimensional measure of institutional opportunities for migrants to participate in the host society. It is based on more than 140 policy indicators in six policy areas (labour market access, family reunification, long term residence, political participation, access to nationality, and antidiscrimination) (Niessen et al., 2007). We chose MIPEX (2007) over other indicators of multicultural policies, including the Multicultural-ism Policy Index (MPI) (Banting & Kymlicka, 2006). While these two indices are highly correlated (Helbling, 2013), we selected MIPEX because of its robustness (based on more than 140 indicators) and its multidimensionality, and especially for its coverage of labour market integration, a policy domain with potentially large effects on migrants’ health. The final MIPEX score, which is an average of the index for each of the six policy domains, ranges from zero to 100.

Data on public attitudes towards migration and migrants was derived from the first round of the European Social Survey (ESS, 2002). ESS is a survey on public attitudes and values in several European countries (ESS, 2014). The first round of the survey included more questions regarding public attitudes towards migra-tion and migrants than the successive rounds of the survey. ESS (2002) captures information on 19,423 respondents in the ten countries considered in this study. We constructed our index on attitudes towards migration and migrants based on 12 items, which can be grouped into two broad domains. The first group of items covers attitudes towards migration, including openness to immigration and requirements for migration, as developed by Davidov et al. (2008). The second group of items covers attitudes towards migrants, including perceived ethnic threat, as developed by Schneider (2008). The specific questions of the 12 items

(14)

Chapter 5

are shown in Table 5.1. Other studies on attitudes towards migration or migrants used similar indicators (Gorodzeisky & Semyonov, 2009; Meuleman et al., 2009; Schlueter et al., 2013). We calculated the index of attitudes towards migration and the index of attitudes towards migrants from the original response categories in a similar way as the perceived group threat index in the study by Schlueter et al. (2013). Like MIPEX, which covers different policy domains, we averaged the index of attitudes towards migration and the index of attitudes towards migrants for each country to derive the final index of public attitudes towards migration and migrants.

Table 5.1. Questions of the 12 items used to measure public attitudes towards migration and migrants

1) To what extent do you think [country] should allow people of a different race or ethnic group from most [country] people to come and live here?

2) To what extent do you think [country] should allow people from the poorer countries in Europe to come and live here?

3) To what extent do you think [country] should allow people from the richer countries outside Europe to come and live here?

4) To what extent do you think [country] should allow people from the poorer countries outside Europe to come and live here?

5) Please tell me how important you think having good educational qualifications should be in deciding whether someone born, brought up and living outside [country] should be able to come and live here.

6) Please tell me how important you think having work skills that [country] needs should be in deciding whether someone born, brought up and living outside [country] should be able to come and live here.

7) Would you say that people who come to live here generally take jobs away from workers in [country], or generally help to create new jobs?

8) Most people who come to live here work and pay taxes. They also use health and welfare services. On balance, do you think people who come here take out more than they put in or put in more than they take out?

9) Would you say it is generally bad or good for [country]’s economy that people come to live here from other countries?

10) Would you say that [country]’s cultural life is generally undermined or enriched by people coming to live here from other countries?

11) Is [country] made a worse or a better place to live by people coming to live here from other countries?

12) Are [country]’s crime problems made worse or better by people coming to live here from other countries?

(15)

We then classified the ten countries in the study according to their relative posi-tioning on MIPEX, and on the index of public attitudes towards migration and migrants. The relative positioning of countries in terms of integration policies and public attitudes towards migration and migrants has been shown to be relatively stable over time (Meuleman et al., 2009; Koopmans et al., 2012). Based on their integration policies, we classified each country as more inclusive, intermediate, or less inclusive. We used a similar procedure to label each country’s public attitudes towards migration and migrants as more favourable, intermediate, or less favour-able. The final classification of countries according to their integration policies and attitudes towards migration and migrants is shown in Table 5.2.

5.5.4 Control variables

Age was recoded into five-year age groups. Education ISCED 1997 codes were recoded as primary education or lower (ISCED codes 0 and 1), secondary education (codes 2 and 3), higher education (codes 4, 5, and 6), and other (‘still in education’, and ‘other’). The current job situation was recoded as retired, active (‘employed’ or ‘self-employed’), non-active (‘unemployed’, ‘permanently sick or disabled’, or ‘homemaker’), and other (‘other’). We used the original BMI coding: underweight (< 18.5), normal weight (18.5–24.9), overweight (25–29.9), and obese (> 30). Length of residence (in years) was derived from the year of migration and the year when the interview took place, and was recoded into two categories: up to ten years, and ten years and over. 5.5.5 Descriptive statistics of the sample

Out of a total sample of 64,966 individuals, 59,056 (90.9 %) were non-migrants, 4,208 (6.5 %) were western migrants, and 1,702 (2.6 %) were non-western migrants. The detailed counts and percentages by country of origin and of residence are shown in Table 5.3. Compared with Eurostat data for 2013 (data not shown), mi-grants were underrepresented in the sample, except in Germany, possibly due to the fact that we considered ‘aussiedler’ - who are not officially classified as migrants - to be migrants as they were born outside of the current Germany (Castles et al., 2014).

Table 5.2. Types of countries according to policies and public attitudes Integration policies

More inclusive Intermediate Less inclusive

A

ttitudes

More favourable Sweden Italy Switzerland

Intermediate Netherlands Spain Denmark

Less favourable Belgium Germany Austria, France

(16)

Chapter 5

The age composition and job situation of western migrants were similar to those of non-migrants (Table 5.4). Non-western migrants were, however, younger and less likely to be retired. While western migrants tended to have a higher educational level than non-migrants, the opposite was the case for non-western migrants. Although non-western migrants were less likely than non-migrants and western migrants to be overweight, they were more likely to be obese.

5.5.6 Methods

To assess the health differences between migrants and non-migrants at older ages, and whether the health outcomes of migrants were associated with integration policies and with public attitudes towards migration and migrants, we applied multivariate logistic regression models for each of the health indicators consid-ered, separately by sex. We calculated robust standard errors to account for the structure of the data, in particular the fact that respondents are clustered within countries (Huber, 1967; White, 1980). In model 1, we included migrant origin (non-migrant, western migrant, or non-western migrant) and controlled for age and wave. In model 2, we introduced the effects of the country of residence together with the effects of length of residence in the host country, highest level of education, current job status, and BMI as control variables. Model 3 assessed the effect of integration policies (H1) and public attitudes towards migration and migrants (H2), keeping all control variables except for country of residence,

Table 5.3. Number and proportions of non-migrants, western migrants, and non-western migrants in the sample by country (2004–2015)

Country

Non-migrants Western migrants Non-western migrants

N % N % N % Austria 5,119 91.17 417 7.43 79 1.41 Germany 6,663 84.78 1,083 13.78 113 1.44 Sweden 5,273 90.82 443 7.63 90 1.55 Netherlands 5,396 93.28 218 3.77 171 2.96 Spain 6,871 94.71 128 1.76 256 3.53 Italy 7,540 98.59 65 0.85 43 0.56 France 6,230 87.29 379 5.31 528 7.40 Denmark 4,981 95.86 154 2.96 61 1.17 Switzerland 3,296 81.81 651 16.16 82 2.04 Belgium 7,687 89.01 670 7.76 279 3.23 Total 59,056 90.90 4,208 6.48 1,702 2.62

(17)

since policies and attitudes are defined at the country level. Finally, in models 4 and 5, we introduced interaction effects between migrant origin and policies and attitudes, respectively, which allows us to answer to what extent the health differences between migrants and non-migrants in Europe are associated with integration policies, and with public attitudes towards migration and migrants.

Table 5.4. Descriptive statistics on age, education, current job situation, BMI, and length of residence according to migrant origin (2004–2015)

Non-migrants Western migrants Non-western migrants

N % N % N % Age (years) 50-54 14,782 25.0 1,048 24.9 726 42.7 55-59 11,405 19.3 762 18.1 369 21.7 60-64 10,610 18.0 732 17.4 254 14.9 65-69 9,109 15.4 697 16.6 174 10.2 70-74 7,483 12.7 562 13.4 107 6.3 75-79 5,667 9.6 407 9.7 72 4.2 Education Primary or lower 13,956 23.7 644 15.3 560 33.2 Secondary 29,593 50.3 2,015 48.0 613 36.4 Higher 15,064 25.6 1,467 34.9 468 27.8 Other 230 0.4 72 1.7 44 2.6 Job status Retired 25,284 43.1 1,850 44.3 390 23.3 Active 21,501 36.7 1,487 35.6 682 40.7 Non-active 11,128 19.0 789 18.9 568 33.9 Other 726 1.2 51 1.2 35 2.1 BMI Underweight 664 1.2 67 1.6 21 1.3 Normal weight 22,966 39.8 1,554 37.7 649 39.4 Overweight 23,943 41.5 1,738 42.2 644 39.1 Obese 10,156 17.6 763 18.5 334 20.3

Length of residence (years)

0-9 - - 272 6.5 136 8.0

10+ 59,056 100.0 3,921 93.5 1,559 92.0

(18)

Chapter 5

5.6 rEsulTs

We start with a description of the main health outcomes among migrants and non-migrants (Table 5.5). Both male and female migrants were more likely to have poor self-rated health than non-migrants. In contrast, diabetes prevalence was very similar among western migrants and non-migrants, but was higher among non-western migrants. In terms of mental health, migrants, especially those of non-western origin, were more likely to suffer from depression than non-migrants. Females were more likely to report poor self-rated health and depression than males, while the opposite was true for diabetes.

The subsequent multivariate analyses show that at older ages, migrants, especially those of non-western origin, had higher odds than non-migrants of having poor self-rated health (Table 5.6), diabetes (Table 5.7), and depression (Table 5.8). The odds ratios were very robust, and changed very little with the inclusion of the effects of integration policies, public attitudes towards migration and migrants, and individual characteristics for each of the health outcomes.

In terms of differences between countries of settlement we found the worst health outcomes in Germany, Italy, and Spain in terms of poor self-rated health and diabetes; and in France, Italy, and Belgium in terms of depression. The best health outcomes were found in Switzerland (all three domains), Sweden (good self-rated health and low prevalence of depression), and Denmark and Austria (low prevalence of depression). The picture became more complex when

look-Table 5.5. Prevalence of poor self-rated health, diabetes (ever diagnosed), and depres-sion according to sex and migrant origin (2004–2015)

Non-migrants Western migrants Non-western migrants

N % N % N %

Males

Poor self-rated health 7,263 26.6 560 30.4 258 31.0

Diabetes 2,918 10.7 220 12.0 135 16.3

Depression 4,315 16.1 338 18.8 219 28.0

Females

Poor self-rated health 9,145 29.0 790 33.5 306 35.6

Diabetes 2,539 8.1 207 8.8 97 11.3

Depression 9,185 29.6 783 34.0 335 41.7

(19)

ing only at the health outcomes of migrants in different countries (results not shown); for instance, the Netherlands would be in the group of countries with the poorest health outcomes, and Spain among those with the best health outcomes. This may not be surprising as migration to Spain is much more recent than it is in the Netherlands, and thus we may have captured the healthy migrant effect here where migrants are found to be healthier at arrival and shortly after.

We found some evidence that less inclusive integration policies were associated with poorer self-rated health (Table 5.6, model 3) and diabetes (Table 5.7, model 3), (H1). However, especially in countries with intermediate integration policies self-rated health was more likely reported to be poorer (males, self-rated health OR: 1.74, diabetes OR: 1.28; females, OR: 1.74, 1.19), while the less inclusive group of countries did not differ significantly from the most inclusive group (reference category). However, when including the interaction effects between origin and integration policies, non-western migrants were found to have the poorest self-rated health outcomes in countries with less inclusive policies, but also in countries with more inclusive policies (Table 5.6, model 4). Thus, health outcomes seemed to be poorer in countries with intermediate levels of integra-tion policies mainly for non-migrants and western migrants.

The odds of having poor self-rated health or diabetes were higher in countries with less favourable attitudes towards migration and migrants (H2). As expected, the best health outcomes were in countries where more favourable attitudes prevailed (reference category), followed by in the intermediate (males OR: 1.22; 1.02; females OR: 1.20, 1.16), and the less favourable country groups (males OR: 1.73; 1.12; females OR: 1.36, 1.32). The association between public attitudes and health also seemed to be stronger among non-western migrants. When looking at the interaction effect (Tables 5.6 and 5.7, model 5), migrants had higher odds to suffer from poor health and diabetes as compared to non-migrants in countries with less favourable attitudes towards migration and migrants. Although the odds ratios tended to be higher among non-western migrants than among western migrants in all typologies of countries according to public attitudes, the negative effects of intermediate and less favourable attitudes on self-rated health were rather similar for both migrant groups.

Although we did not find significant differences in the likelihood of having de-pression based on integration policies or public attitudes towards migration and migrants (Table 5.8, model 3), the general patterns remained similar, with

(20)

depres-Chapter 5

Table 5.6. Odds ra

tio of ha

ving poor self-ra

ted heal

th by sex (2004–2015)

Log pseudolikelihood Pseudo R

2 M ales Females M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 -16696 -15160 -15228 -15222 -15226 -19675 -18041 -18085 -18072 -18084 0.022 0.112 0.108 0.109 0.108 0.028 0.108 0.106 0.107 0.106 N on-migrant (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) W estern migrant 1.20* 1.21*** 1.20*** 1.30*** 1.32** 1.26 1.49*** 1.44*** 1.35*** 1.45** N on-w estern migrant 1.45** 1.17 1.17 1.54 1.45*** 1.64*** 1.37** 1.39** 2.05*** 1.61 50-54 (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) 55-59 1.25*** 1.08* 1.08* 1.08* 1.08* 1.33*** 1.10 1.09 1.10 1.09 60-64 1.46*** 1.01 1.01 1.01 1.01 1.44*** 0.90 0.89 0.90 0.89 65-69 1.69*** 1.10 1.09 1.09 1.09 1.74*** 0.96 0.94 0.95 0.94 70-74 2.27*** 1.41*** 1.40*** 1.40*** 1.39*** 2.55*** 1.33 1.3 1.31 1.30 75-79 3.21*** 2.00*** 1.97*** 1.97*** 1.97*** 3.55*** 1.82*** 1.78*** 1.79*** 1.78*** W av e 1 (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) W av e 2 1.07 1.18 1.19* 1.19* 1.19* 1.11 1.22* 1.24* 1.24* 1.24* W av e 4 1.07 1.20*** 1.13* 1.13* 1.13* 1.04 1.32*** 1.25** 1.25** 1.25** W av e 5 1.11 1.17 1.15 1.15 1.15 1.04 1.15 1.14 1.15 1.15 W av e 6 1.24 1.31*** 1.31*** 1.30*** 1.31*** 1.12 1.21*** 1.25*** 1.25*** 1.25*** France … (r ef ) … … … … (r ef ) … … … Austria … 0.95** … … … … 0.74*** … … … G ermany … 1.71*** … … … … 1.57*** … … … Sw eden … 0.43*** … … … … 0.61*** … … … N etherlands … 0.90*** … … … … 0.88*** … … …

(21)

Table 5.6. (c

on

tin

ued)

2 M ales Females M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 -16696 -15160 -15228 -15222 -15226 -19675 -18041 -18085 -18072 -18084 0.022 0.112 0.108 0.109 0.108 0.028 0.108 0.106 0.107 0.106 Spain … 0.84*** … … … … 1.17*** … … … Italy … 1.03 … … … … 1.27*** … … … D enmar k … 0.73*** … … … … 0.87*** … … … Switz erland … 0.47*** … … … … 0.49*** … … … Belgium … 0.79*** … … … … 0.84*** … … …

Integration policies: mor

e inclusiv e … … (r ef ) (r ef ) (r ef ) … … (r ef ) (r ef ) (r ef )

Integration policies: intermediate

… … 1.74*** 1.78*** 1.75*** … … 1.74*** 1.77*** 1.74***

Integration policies: less I

nclusiv e … … 1.09 1.10 1.08 … … 0.99 0.99 0.99

Public attitudes: mor

e fav ourable … … (r ef ) (r ef ) (r ef ) … … (r ef ) (r ef ) (r ef )

Public attitudes: intermediate

… … 1.22 1.23 1.22 … … 1.20 1.21* 1.21*

Public attitudes: less fav

ourable … … 1.73*** 1.73*** 1.76*** … … 1.36*** 1.36*** 1.36*** Intermediate IP* W estern … … … 0.87*** … … … … 1.13 … Intermediate IP * N on-w estern … … … 0.46** … … … … 0.35*** … Less inclusiv e IP * W estern … … … 0.92 … … … … 1.07 … Less inclusiv e IP * N on-w estern … … … 0.82 … … … … 0.75 … Intermediate P A* W estern … … … … 1.04 … … … … 0.84 Intermediate P A* N on-w estern … … … … 0.88 … … … … 0.71 Less fav ourable P A* W estern … … … … 0.86 … … … … 1.02 Less fav ourable P A * N on-w estern … … … … 0.75* … … … … 0.92

(22)

Chapter 5

Table 5.6. (c

on

tin

ued)

2 M ales Females M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 -16696 -15160 -15228 -15222 -15226 -19675 -18041 -18085 -18072 -18084 0.022 0.112 0.108 0.109 0.108 0.028 0.108 0.106 0.107 0.106 Length of r esidence: 10+ y ears … (r ef ) (r ef ) (r ef ) (r ef ) … (r ef ) (r ef ) (r ef ) (r ef ) Length of r esidence: 0-9 y ears … 0.51** 0.49*** 0.51** 0.49** … 0.77** 0.74** 0.78*** 0.78** Secondar y education … (r ef ) (r ef ) (r ef ) (r ef ) … (r ef ) (r ef ) (r ef ) (r ef ) Primar y education or lo w er … 1.55*** 1.43*** 1.43*** 1.44*** … 1.55*** 1.55*** 1.55*** 1.55*** H igher education … 0.63*** 0.62*** 0.62*** 0.62*** … 0.69*** 0.69*** 0.69*** 0.69*** O ther … 0.90 0.89 0.88 0.89 … 1.21 1.20 1.19 1.19 Retir ed … (r ef ) (r ef ) (r ef ) (r ef ) … (r ef ) (r ef ) (r ef ) (r ef ) Activ e … 0.47*** 0.46*** 0.46*** 0.46*** … 0.44*** 0.44*** 0.44*** 0.44*** N on-activ e … 2.64*** 2.61*** 2.62*** 2.61*** … 1.17*** 1.17*** 1.17*** 1.16*** O ther … 0.94 1.00 1.01 1.00 … 0.92 0.92 0.92 0.916 N ormal w eight … (r ef ) (r ef ) (r ef ) (r ef ) … (r ef ) (r ef ) (r ef ) (r ef ) U nder w eight … 3.79*** 3.66*** 3.66*** 3.65*** … 2.19*** 2.19*** 2.19*** 2.19*** Ov er w eight … 1.08** 1.07** 1.07** 1.07** … 1.35*** 1.34*** 1.34*** 1.34*** O bese … 1.89*** 1.86*** 1.85*** 1.86*** … 2.56*** 2.53*** 2.53*** 2.53*** Source: o

wn calculation based on SHARE data (2004-2015)

* p < 0.10, ** p < 0.05, *** p < 0.01; IP (integ ration policies), P A (public attitudes to w ards mig

ration and mig

(23)

sion being more prevalent in countries with intermediate integration policies and in countries with less favourable attitudes.

Finally, all control variables were in the expected direction. The odds of having poor self-rated health, diabetes, or depression generally increased with age and length of residence in the host country, and diminished with increasing levels of education and among the active population. Respondents who were underweight or obese were much more likely than their normal weight counterparts to have poor self-rated health or depression.

5.7 DISCUSSION

In this article, we studied the health differences between migrants and non-migrants aged 50–79 living in ten European countries using data from the Survey of Health, Ageing and Retirement in Europe (SHARE). As we were particularly interested in investigating the association between health and integration poli-cies and public attitudes, we enriched the SHARE data with country level data on integration policies (MIPEX) and on public attitudes towards migration and migrants (ESS). In addition to the specific focus on migrant health at older ages, our study contributes to the existing literature (Huijts & Kraaykamp, 2012; Ikram et al., 2015; Malmusi, 2015) in two ways. First, we considered the role of both integration policies and public attitudes towards migration and migrants as social determinants of migrant health inequalities. Second, we covered different health measures in our analyses.

Our results showed that older (western) migrants were more likely than non-migrants to suffer from poor self-rated health, diabetes, and depression. These findings are in line with those of previous studies (Solé-Auró & Crimmins, 2008; Aichberger et al., 2010; Kunst et al., 2011). However, our findings also suggest that integration policies and public attitudes towards migration and migrants play a role in explaining the migrant health inequality gap in different countries. Regarding the association between migrant health inequalities and integration policies and public attitudes towards migration and migrants, we had hypothesised that both more inclusive policies (H1) and more favourable public attitudes (H2) would be associated with better health outcomes among migrants, and especially among those of non-western origin. Our results partially supported the assump-tion that integraassump-tion policies are associated with differences between migrants

(24)

Chapter 5

Table 5.

7. Odds ra

tio of ha

ving diabetes by sex (2004–2015)

Log pseudolikelihood Pseudo R2

M ales Females M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 -9855 -9399 -9403 -9398 -9402 -9157 -8427 -8432 -8426 -8428 0.024 0.070 0.069 0.070 0.069 0.025 0.103 0.103 0.103 0.103 N on-migrant (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) W estern migrant 1.13* 1.12 1.12 1.13 1.19 1.11 1.17* 1.17* 1.21 0.96 N on-w estern migrant 2.03*** 2.01*** 2.00*** 2.16*** 1.57*** 1.91*** 1.60** 1.58** 2.66*** 1.48 50-54 (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) 55-59 1.61*** 1.45*** 1.45*** 1.45*** 1.45*** 1.63*** 1.38*** 1.38*** 1.38*** 1.38*** 60-64 2.12*** 1.71*** 1.72*** 1.73*** 1.72*** 1.96*** 1.38*** 1.38*** 1.39*** 1.39*** 65-69 2.84*** 2.18*** 2.19*** 2.20*** 2.19*** 2.67*** 1.73*** 1.73*** 1.74*** 1.73*** 70-74 2.79*** 2.14*** 2.15*** 2.16*** 2.15*** 2.97*** 1.82*** 1.81*** 1.82*** 1.82*** 75-79 3.27*** 2.53*** 2.54*** 2.56*** 2.54*** 3.69*** 2.27*** 2.26*** 2.27*** 2.28*** W av e 1 (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) W av e 2 0.96 0.98 0.98 0.98 0.98 0.90 1.02 1.03 1.02 1.03 W av e 4 1.26** 1.30*** 1.28*** 1.27*** 1.28*** 1.03 1.23*** 1.18*** 1.17*** 1.18*** W av e 5 1.20 1.18 1.19 1.20 1.19 0.96 1.02 1.03 1.03 1.03 W av e 6 1.09 1.09 1.06 1.05 1.06 0.90 1.00 0.98 0.98 0.98 France … (r ef ) … … … … (r ef ) … … … Austria … 1.03 … … … … 1.06 … … … G ermany … 1.30*** … … … … 1.62*** … … … Sw eden … 1.00 … … … … 1.05 … … … N etherlands … 0.90** … … … … 1.15*** … … …

(25)

Table 5.

7. (c

on

tin

ued)

M ales Females M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 -9855 -9399 -9403 -9398 -9402 -9157 -8427 -8432 -8426 -8428 0.024 0.070 0.069 0.070 0.069 0.025 0.103 0.103 0.103 0.103 Spain … 1.17*** … … … … 1.24*** … … … Italy … 1.10*** … … … … 1.15*** … … … D enmar k … 0.89** … … … … 1.04 … … … Switz erland … 0.78*** … … … … 0.63*** … … … Belgium … 0.91*** … … … … 1.20*** … … …

… … 1.28*** 1.32*** 1.28*** … … 1.19*** 1.23*** 1.19***

nclusiv e … … 1.00 0.98 1.00 … … 0.82*** 0.85** 0.82***

… … 1.02 1.03 1.02 … … 1.16*** 1.16*** 1.12*

ourable … … 1.12 1.13* 1.11 … … 1.32*** 1.33*** 1.31*** Intermediate IP* W estern … … … 0.80 … … … … 0.95 … Intermediate IP * N on-w estern … … … 0.61*** … … … … 0.41*** … Less inclusiv e IP * W estern … … … 1.22 … … … … 0.97 … Less inclusiv e IP * N on-w estern … … … 1.11 … … … … 0.47** … Intermediate P A* W estern … … … … 0.82 … … … … 1.71 Intermediate P A* N on-w estern … … … … 1.14 … … … … 1.62 Less fav ourable P A* W estern … … … … 0.94 … … … … 1.23 Less fav ourable P A * N on-w estern … … … … 1.41* … … … … 0.88

(26)

Chapter 5 Table 5. 7. (c on tin ued)

M ales Females M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 -9855 -9399 -9403 -9398 -9402 -9157 -8427 -8432 -8426 -8428 0.024 0.070 0.069 0.070 0.069 0.025 0.103 0.103 0.103 0.103 Length of r esidence: 10+ y ears … (r ef ) (r ef ) (r ef ) (r ef ) … (r ef ) (r ef ) (r ef ) (r ef ) Length of r esidence: 0-9 y ears … 0.76 0.76 0.78 0.78 … 0.74 0.74 0.75 0.65* Secondar y education … (r ef ) (r ef ) (r ef ) (r ef ) … (r ef ) (r ef ) (r ef ) (r ef ) Primar y education or lo w er … 1.20** 1.20** 1.19** 1.20** … 1.36*** 1.33*** 1.33*** 1.33*** H igher education … 0.87*** 0.87*** 0.87*** 0.87*** … 0.77*** 0.78*** 0.78*** 0.78*** O ther … 1.24 1.25 1.23 1.24 … 1.37 1.36 1.34 1.39 Retir ed … (r ef ) (r ef ) (r ef ) (r ef ) … (r ef ) (r ef ) (r ef ) (r ef ) Activ e … 0.69*** 0.68*** 0.69*** 0.68*** … 0.58*** 0.58*** 0.58*** 0.58*** N on-activ e … 1.45*** 1.44*** 1.44*** 1.44*** … 1.12 1.10 1.10 1.10 O ther … 0.95 0.94 0.95 0.95 … 1.05 1.04 1.05 1.05 N ormal w eight … (r ef ) (r ef ) (r ef ) (r ef ) … (r ef ) (r ef ) (r ef ) (r ef ) U nder w eight … 1.43 1.43 1.45 1.44 … 0.81 0.81 0.81 0.81 Ov er w eight … 1.71*** 1.71*** 1.71*** 1.71*** … 2.23*** 2.24*** 2.24*** 2.24*** O bese … 3.77*** 3.77*** 3.77*** 3.77*** … 5.19*** 5.21*** 5.21*** 5.23*** Source: o

wn calculation based on SHARE data (2004-2015).

ration and mig

(27)

and non-migrants in terms of self-rated health and diabetes (H1). The worst health outcomes (in terms of SRH and diabetes) were found in countries with intermediate integration policies, which may reflect non-observed country level characteristics such as the health or social policy systems leading to the poor-est health outcomes in countries with intermediate levels of integration policies (Germany, Italy, and Spain). For non-western migrants we however found the poorest self-rated health and diabetes outcomes in countries with less inclusive integration policies, but also in countries with more inclusive integration policies. This is an important finding, that reflects that, in this case, the relative health disadvantage of non-western migrants as compared to non-migrants does not necessarily take place in those countries with the overall poorer general health outcomes, and neither follows the expected distribution according to integration policy levels (H1). A study by Koopmans (2010) suggests that more inclusive integration policies in the form of multiculturalism may not be the most efficient way to integrate non-western migrants into society and into the labour force, which may, on the long run, explain the relatively poor health situation of older non-western migrants in these countries. Another study suggests that egalitarian policies, with the notable exception of anti-discrimination policies, may have the opposite as intended effect since they may increase minorities’ sensitivity towards existing inequalities (Ziller, 2017).

Furthermore, the results confirmed our hypotheses regarding the association between public attitudes and health (H2) for both self-rated health and diabetes. The odds of having poor self-rated health or diabetes were higher in countries with less favourable public attitudes towards migration and migrants. While the negative effect of less favourable attitudes seemed to be greater on non-western migrants in terms of diabetes prevalence, this effect seemed to be rather similar for western and non-western migrants in terms of self-rated health. These find-ings confirm our assumption that less favourable attitudes towards migration and migrants are an important predictor of migrant health inequalities and are impor-tant for migrants regardless of their origin. This has imporimpor-tant implications, since as opposed to the case of integration policies, a more positive societal general view and attitudes towards migration and migrants is directly associated with bet-ter migrants’ health outcomes via the provision of a more favourable context for the social and economic integration of migrants into society (Berry, 1997); while negative perceptions of migration and migrants, and especially events of racism or discrimination, have a very negative impact on migrants’ health (Karlsen & Nazroo, 2002; Williams et al., 2003; Paradies, 2006; Johnston & Lordan, 2012; Levecque & van Rossem, 2015).

(28)

Chapter 5

Table 5.8. Odds ra

tio of ha

ving depression by sex (2004–2015)

2 M ales Females M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 -12819 -12259 -12321 -12314 -12314 -20080 -19418 -19593 -19586 -19592 0.006 0.050 0.045 0.045 0.045 0.005 0.038 0.029 0.029 0.029 N on-migrant (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) W estern migrant 1.21** 1.26*** 1.17* 1.38 1.12 1.25** 1.46*** 1.34*** 1.43*** 1.46*** N on-w estern migrant 2.04*** 1.61*** 1.68*** 2.04*** 1.84** 1.79*** 1.50*** 1.62*** 1.85*** 1.85** 50-54 (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) 55-59 0.92** 0.81*** 0.81*** 0.81*** 0.81*** 0.90*** 0.82*** 0.80*** 0.80*** 0.80*** 60-64 0.85** 0.70*** 0.67*** 0.67*** 0.67*** 0.90 0.73*** 0.71*** 0.71*** 0.71*** 65-69 0.89 0.74** 0.70*** 0.70*** 0.70*** 0.90 0.69*** 0.67*** 0.67*** 0.67*** 70-74 1.03 0.83* 0.77** 0.78** 0.77** 1.01 0.74** 0.70** 0.70** 0.70** 75-79 1.26** 0.98 0.92 0.92 0.92 1.27** 0.89 0.85 0.85 0.85 W av e 1 (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) (r ef ) W av e 2 0.89* 0.96 0.98 0.98 0.98 0.93 0.99 1.00 1.00 1.00 W av e 4 1.14 1.15 1.20 1.19 1.20 1.05 1.10 1.13 1.13 1.13 W av e 5 1.03 1.18 1.11 1.11 1.11 0.93 1.09 1.04 1.04 1.04 W av e 6 1.36*** 1.21** 1.40*** 1.39*** 1.40*** 1.29** 1.21* 1.38*** 1.38** 1.38*** France … (r ef ) … … … … (r ef ) … … … Austria … 0.59*** … … … … 0.46*** … … … G ermany … 0.67*** … … … … 0.57*** … … … Sw eden … 0.52*** … … … … 0.51*** … … … N etherlands … 0.61*** … … … … 0.48*** … … …

(29)

Table 5.8. (c

on

tin

ued)

2 M ales Females M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 -12819 -12259 -12321 -12314 -12314 -20080 -19418 -19593 -19586 -19592 0.006 0.050 0.045 0.045 0.045 0.005 0.038 0.029 0.029 0.029 Spain … 0.64*** … … … … 0.76*** … … … Italy … 0.92*** … … … … 0.83*** … … … D enmar k … 0.53*** … … … … 0.50*** … … … Switz erland … 0.56*** … … … … 0.49*** … … … Belgium … 0.90*** … … … … 0.79*** … … …

… … 1.07 1.11 1.08 … … 1.16 1.19 1.16

nclusiv e … … 0.95 0.97 0.95 … … 0.97 0.97 0.97

… … 0.88 0.88 0.86 … … 0.93 0.94 0.94

ourable … … 1.19 1.20 1.21 … … 1.17 1.17 1.18 Intermediate IP* W estern … … … 0.76 … … … … 0.82 … Intermediate IP * N on-w estern … … … 0.48** … … … … 0.54*** … Less inclusiv e IP * W estern … … … 0.81 … … … … 0.98 … Less inclusiv e IP * N on-w estern … … … 0.92 … … … … 1.04 … Intermediate P A* W estern … … … … 1.96*** … … … … 0.97 Intermediate P A* N on-w estern … … … … 0.96 … … … … 0.79 Less fav ourable P A* W estern … … … … 0.94 … … … … 0.87 Less fav ourable P A * N on-w estern … … … … 0.89 … … … … 0.88

(30)

Chapter 5

Table 5.8. (c

on

tin

ued)

2 M ales Females M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 -12819 -12259 -12321 -12314 -12314 -20080 -19418 -19593 -19586 -19592 0.006 0.050 0.045 0.045 0.045 0.005 0.038 0.029 0.029 0.029 Length of r esidence: 10+ y ears … (r ef ) (r ef ) (r ef ) (r ef ) … (r ef ) (r ef ) (r ef ) (r ef ) Length of r esidence: 0-9 y ears … 0.75 0.75 0.78 0.72 … 0.59*** 0.59*** 0.63** 0.60** Secondar y education … (r ef ) (r ef ) (r ef ) (r ef ) … (r ef ) (r ef ) (r ef ) (r ef ) Primar y education or lo w er … 1.37*** 1.53*** 1.52*** 1.53*** … 1.27*** 1.53*** 1.53*** 1.53*** H igher education … 0.86** 0.83*** 0.83*** 0.83*** … 0.85*** 0.86*** 0.85*** 0.86*** O ther … 0.97 1.01 1.00 1.01 … 1.26 1.34 1.32 1.34 Retir ed … (r ef ) (r ef ) (r ef ) (r ef ) … (r ef ) (r ef ) (r ef ) (r ef ) Activ e … 0.67*** 0.65*** 0.66*** 0.65*** … 0.73*** 0.74*** 0.74*** 0.74*** N on-activ e … 2.44*** 2.36*** 2.38*** 2.37*** … 1.21*** 1.24*** 1.24*** 1.24*** O ther … 1.02 1.02 1.03 1.02 … 1.16 1.18 1.17 1.18 N ormal w eight … (r ef ) (r ef ) (r ef ) (r ef ) … (r ef ) (r ef ) (r ef ) (r ef ) U nder w eight … 2.72*** 2.66*** 2.66*** 2.62*** … 1.58*** 1.60*** 1.61*** 1.61*** Ov er w eight … 0.92* 0.92* 0.92* 0.92* … 1.11*** 1.09** 1.09** 1.09** O bese … 1.23*** 1.21*** 1.21*** 1.21*** … 1.47*** 1.43*** 1.43*** 1.43*** Source: o

wn calculation based on SHARE data (2004-2015).

ration and mig

(31)

While we found some evidence of an association between health and integration policies and attitudes for physical (diabetes) and overall (self-rated) health, the association was less clear for mental health (depression). Although the patterns remained similar to those found for self-rated health and diabetes, the main effect of mental health outcomes was not statistically significantly different depending on a country’s level of integration policies and public attitudes. It could well be that the experiences taking place before and during the migration process may be more important in determining migrants’ mental health than the context in the country of residence. Indeed, poor mental health among migrants is strongly associated with the context and conditions during the pre-departure and move-ment phases of the migration process, and the difficulties the individual faces in starting a new life in a new environment, including material deprivation and social isolation (Gushulak et al., 2010). More information about the conditions before and after migration and the selection of migrants is needed to assess how these factors influence mental health. Furthermore, our data did not allow us to conduct separate analyses for specific migrant groups, considering not only origin but also migrant status. This would be an important way ahead for future research since specific migrant groups, such as for instance refugees, may have a higher risk of depression and poor mental health outcomes due to previous traumatic experiences.

The associations found between health outcomes and the various control vari-ables were in line with the results of previous research. Health declined with the amount of time spent in the destination country, as the ‘healthy migrant effect’ faded away (Norredam et al., 2014; Vandenheede et al., 2015). Poor self-rated health and diabetes were positively associated with age, while the association between depression and age was U-shaped (Mirowsky & Ross, 1992; Wu et al., 2012). Poor socioeconomic status was strongly associated with both poor physical and mental health outcomes (Nazroo, 2003; Kunst et al., 2005; Mackenbach et al., 2008). Being underweight or being obese was also strongly associated with both poor physical and mental health (Jorm et al., 2003; Onyike et al., 2003; WHO, 2003; Kodjebacheva et al., 2015).

Although our analyses contribute to the literature on the effects of integration policies and public opinions on a range of health indicators for migrants across Europe, our study also has some limitations. First, the SHARE data was not specifically designed to sample the older migrant population in Europe. Indeed, as we mentioned above, migrants, and especially those of non-western origin, are underrepresented in SHARE. This meant that our country of origin categories

(32)

Chapter 5

had to be broad. In addition, the migrant population covered in SHARE is a selective group, since only migrants who could answer the questionnaire in the country of residence’s official language were included in the sample. This excludes poorly integrated migrants whose health status may have been more strongly as-sociated with both integration policies and public attitudes towards migration and migrants. Second, we considered western European destination countries only. Future studies should broaden this geographic scope to cover specific countries of origin and countries of destination beyond western Europe. Third, while MI-PEX is useful for assessing integration policies in a range of domains, it provides a cross sectional measure only. Although the relative positioning of countries according to integration policies over time has remained largely unchanged, it is also true that integration policies were much less developed before the 1990s, when most of the migrants in our study arrived. For future studies on migrants’ health outcomes, it would be useful to have more dynamic data in which policies are captured over the life course.

Despite these limitations, our study showed that integration policies, and especially public attitudes towards migration and migrants, were associated with health out-comes at older ages in Europe. Less favourable public attitudes towards migration and migrants were particularly strongly associated with poor health outcomes among non-western migrants. While governments seeking to reduce migrant health inequalities might want to develop more inclusive integration policies, tak-ing action to change public attitudes in favour of migration and migrants could be a more effective way to reduce such inequalities. These findings are especially relevant in the current European context, where anti-immigrant political parties are gaining support. We should also note, however, that even after controlling for a range of explanatory variables, we found persistent migrant health inequalities at older ages, with older non-western migrants having the worst health outcomes in all domains. Future research should focus more on the nexus between the country of origin and of destination, and on events that occur during the process of migration that could have long lasting effects on the individual. Cross national European comparative studies on these issues are needed to identify the policy and societal contexts that could reduce migrant health inequalities at a time when the older migrant population is growing across the continent.