The Health of Immigrant Youth in Denmark: Examining Immigrant Generations and Origin

First-generation migrants tend to have better health than those of native ancestry, while second-generation migrants often are affected by negative health assimilation. Less is known about immigrants arriving before their teens, the 1.5 generation, or those with one native and one immigrant parent, the 2.5 generation. We apply logistic regression models to study physical and mental health outcomes for immigrant youth across generations using Danish register data. We take heterogeneities into account studying gender and origin differences. Our analysis shows that first- and 1.5-generation immigrants have better physical health than natives, while the physical health of the second and 2.5 generations is the same or worse than natives. For mental health, the 2.5 generation is the only group with worse outcomes than natives, which we relate to identification struggles translating into poorer mental health. When studying potential heterogeneities, we find that males belonging to the second generation with an African background have a higher likelihood of hospitalization for a mental health issue.


Introduction
The migrant group in Western European immigrant-receiving countries has become increasingly heterogeneous over the last decades as a result of globalizing migration patterns. Denmark is no exception. The majority of migrants in Denmark originate from non-western countries and most minority groups by now, consist of more than one generation. Moreover, with increasing immigration, Denmark has seen rising intermarriage rates between native Danes and migrants, which means there is 1 3 a growing 2.5 immigrant generation. This calls for an analysis of Danish migrant health considering differences in both origin and generation. With the use of detailed longitudinal register data, we can compare health outcomes of different migrant generations and origin groups while also controlling for important sociodemographic variables. Studying immigrant mental and physical health from a multi-generational perspective adds to the broader literature on migrant health while also increasing the knowledge of health differences between different migrant groups.
Previous studies on migrant health find that it differs across immigrant generations. First-generation migrants tend to have better health, especially at the time of arrival, than those of native ancestry in the receiving country, and this is found for health outcomes such as self-rated health (Akresh & Frank, 2011;Bostean, 2012;Ichou & Wallace, 2019) or all-cause mortality (Anikeeva et al., 2014;Markides & Eschbach, 2011). This phenomenon is widely known and referred to as the Healthy Immigrant Effect (HIE) and is viewed as paradoxical since migrants typically have lower socioeconomic status and fewer resources than the native-born population (Urquia et al., 2012) and since migrants often originate from countries of the global south with higher mortality and lower life expectancy (Aldridge et al., 2018). HIE is generally explained by different selection processes such as receiving country policies 'selecting' healthier individuals, and self-selection by migrants in that healthier individuals are more likely to migrate and negative self-selection in return migration, where less healthy and less successful migrants return to the origin country (Vang et al., 2017).
The HIE tends to fade with time (Ng et al., 2021) and with the second (or subsequent) generations (Markides & Rote, 2019;Urquia et al., 2012). Second-generation migrants 1 are affected by a Negative Health Assimilation (NHA) process or health convergence, resulting in a health status more similar to that of native-born with native ancestry (E. R. Markides & Rote, 2019). The NHA is explained by an unhealthy adjustment to a western lifestyle and food, including risky health behaviors (Ru & Li, 2021), but also by an exposure to different stressors related to the experience of being a migrant or having a migrant background. The stressors are associated with failed economic or social integration or related to experiences of discrimination and racism and affects first-generation migrants cumulatively with increasing years since migration (Elshahat et al., 2021), but is also a valid explanation to the NHA experienced by second, and subsequent generations of migrants as the immigrant disadvantage can be intergenerational (E. R. Loi et al., 2021;Ramraj et al., 2015).
The HIE has been substantially less studied for mental health, and the available studies have indicated inconsistent evidence (Elshahat et al., 2021). Theoretically, it is expected that stress in relation to migration and integration would mean worse mental health for migrants, and this has been shown in some contexts ( see, e.g., Cavdar et al., 2021 for the UK and Di Napoli et al., 2021 for Italy.) 1 3 The Health of Immigrant Youth in Denmark: Examining Immigrant… Few studies have studied origin country differences and immigrants arriving in the host country before their early teens, the 1.5 generation, or those with one native and one immigrant parent, the 2.5 generation. The 1.5 generation spend a majority of their youth in the host country and are thus likely to be affected by negative health assimilation, although they belong to the first immigrant generation, which tends to be healthier in line with the HIE. 2.5-generation migrants, individuals with one native and one immigrant parent, are a growing group in many immigrant-receiving countries. They have more social contact with the receiving population, and they share values and culture, and tend to have better socio-economic positions than first-and second-generation migrants (Kalmijn, 2015b). Yet, they are exposed to the native lifestyle and health behavior through having a native parent. Parental conflict and separation are moreover more common in ethnic intermarriage (Dribe & Lundh, 2012;Zhang & Van Hook, 2009) which could affect children negatively in terms of mental health and wellbeing (Bernardi & Radl, 2014). Previous research has also shown that children of the 2.5 generation struggle with self-identification and may be subject to discrimination and prejudice, from both minority and majority groups, which also have the potential of affecting the mental health of the 2.5 generation (Kalmijn, 2015a;Lee & Bean, 2007;Rodríguez-García et al., 2018).
This study contributes to the literature on migrant health through three important aspects. First it studies both physical and mental health, the latter being less studied in previous research. Second, we study the health of immigrant youth across generations, also including the 1.5 and 2.5 generations, and finally, we take potential heterogeneities into account, as we study differences between groups based on origin.

Immigrants in Denmark
Denmark has a relatively modest foreign-born population of 13.8% of the total population, along with a rather short history of immigration (Statistics Denmark, 2018). A rising labor demand in the 1960s led to an increased inflow of guest workers, mainly from Turkey, Yugoslavia, and Pakistan (Liebig, 2007). Danish immigration has since the oil crisis of the 1970s consisted mainly of refugees and family reunification migrants.
The majority of the foreign-born in Denmark of today originate from non-western countries with dominating groups from Iran, Iraq, Turkey, Pakistan, and the Balkans. Foreign-born in Denmark are not fully integrated into the Danish labor market as there are large immigrant-native wage-and employment gaps (Statistics Denmark, 2018). This is potentially explained by a shift in the composition of the foreign-born, with an increasing share of refugees from the 1990s onward, but it could also be explained by a changing labor market structure which increased the importance of country-specific skills (such as language proficiency) resulting in a decreasing demand for immigrant employees (Rosholm et al., 2006). Moreover, the relatively compressed wage structure with high effective minimum wages in Denmark has been suggested as an explanation to the employment gap (Tranaes, 2014). This means that it is difficult for migrants to get employment, as wages are relatively high and migrants often lack the country specific skills needed. The employment gap between non-Western immigrants and natives can, moreover, be explained by the incidence of discrimination against non-Western immigrants in the Danish labor market (Sauer & Siim, 2019) or by the specific institutional structure of Denmark, as discussed extensively by (Brown, 2020). Especially immigrant from the Middle East and North Africa fall behind native Danes and other migrant groups in terms of labor market integration (Statistics Denmark, 2018).

Health across different immigrant generations
Different selection processes are proposed as main explanations to the HIE (see Kennedy et al., 2015, for an overview). First, selective health screening by receiving governments has been presented as a possible explanation, although far from all immigrant-receiving countries use health as a screening mechanism. Moreover, Laroche (2000) found in a study of Canada, which is a country where the health status of the applicant potentially could play a role for the outcome, that the number of individuals being turned down on their request to immigrate on the basis of health is minimal. Another more convincing explanation of the HIE is the one in which immigrants self-select into migration and that healthier and resource-strong individuals migrate while leaving weaker and older individuals behind. In other words, migrants are at the high end of the income and health distribution in their home country. Because of this selectivity, immigrants perform better than nativeborn individuals in various health outcomes. 2 Whether or not the favorable health of the first immigrant generation is shared by following generations has been investigated in previous studies. An individuals' socioeconomic status, health, and demographic outcomes are strongly influenced by intergenerational processes. (Becker & Tomes, 1986;Chetty et al., 2014). Across contexts and time, the correlation in attainment between biological parents and their children has proven to be robust reflecting both the transmission of genetically determined ability as well as a range of different resources, used as inputs in the health status attainment process. One common argument is that the selective process of migrants leads to better health also for their offspring. This is seen comparing birth weights of natives and second-generation migrants, with second-generation migrants having better health as infants (Andrasfay & Goldman, 2020;Cebolla-Boado & Salazar, 2016). This phenomenon is sometimes referred to as a paradox, due to migrants' socio-economic disadvantage, but is, however, less documented later during the life course (E. R. . Some studies instead find that the HIE disappears with the second generation (Stirbu et al., 2006;Tarnutzer & Bopp, 2012). This

3
The Health of Immigrant Youth in Denmark: Examining Immigrant… is often explained by Negative Health Assimilation (NHA) that successive generations of migrants adjust to native lifestyles and health behaviors of natives, leading to poorer health. Economic assimilation theory, expect migrants and their descendants to become economically indistinct from the majority group through an upward mobility with time and with subsequent generations (Alba & Nee, 1997). Negative health assimilation is instead a pattern of downward adjustment, where migrants with time have worse health over time, more in line with the majority group (E. R. . Several factors are in the literature brought forward as potential explanations to the NHA. First, health assimilation can be influenced by ethnic and racial identification that can change over time and across generations. An individual's identification is in turn dependent on factors, such as socio-economic status, education and intermarriage (Alba & Islam, 2009). Second, exposure to racism and discrimination as well as other negative social and economic settings can have a undesirable effect on the health of migrants and their descendants (T. G. . Social and institutional factors in both the origin and receiving country are also important determinants of health as shown for Somali migrants in Finland (Tiilikainen & Koehn, 2011). Differences in terms of negative health assimilation according to origin have, moreover, been shown (Vik et al., 2019). E. R.  showed for the USA that the process of health assimilation was segmented along racial and ethnic lines for later generations migrants, mainly explained by institutional structures, racial segmentation, discrimination, and racism in contemporary society.
The health of the second generation is likely also affected by family background, and environmental factors growing up. Immigrants tend to fall behind natives economically and socially in many western European countries, and Denmark is no exception. They have higher levels of unemployment, lower incomes (Brell et al., 2020), and live in more socially disadvantaged areas (Damm, 2007). With socialization into the youth culture of the receiving country, second-generation immigrants increasingly adopt health behavioral norms that could lead to a lower health status and higher health risks (Alidu & Grunfeld, 2017). Second-generation immigrant youth that belong to more marginalized groups in the host country may display worse health than the native population due to a downward assimilation process (Alidu & Grunfeld, 2017;Portes & Zhou, 1993).
Few studies have investigated the health of immigrant 1.5 and 2.5 generations. This is unfortunate, as they constitute a growing group in many western European countries due to an increasingly multi-ethnic population with growing intermarriage rates. Their health status is interesting to examine also from a theoretical perspective, and we need to draw upon several theories within the social sciences to explain their outcomes.
1.5-generation migrants arrive in the host countries as small children (here defined as before the age of 13) and thus spend the dominant part of their life in the host country. Most previous studies do not focus on this group in particular, although (Roshania et al., 2008) found that migrants that arrive in the USA below the age of 20 have a higher likelihood of being overweight or obese than immigrants who arrive at later ages. The authors explain their finding by arguing that immigrants who migrate during childhood and adolescence are more exposed to the lifestyles 1 3 of their native-born counterparts. Older immigrants are less likely to acculturate or build new social networks and are more likely to join family members already in the receiving country.
The health status of the 2.5 generation is affected by both a native and an immigrant parent. Through the native parent, they have access to native networks and information sharing, as well as native language proficiency. This can lead to positive health outcomes due to increased health literacy (Ward et al., 2018), but also negative health outcomes through exposure to native lifestyles (see above). The gains that the immigrant parent obtains through intermarriage will also affect child outcomes through intergenerational transmissions. Immigrants married to natives tend to have a higher socioeconomic position than endogenously married immigrants as a result of spousal spillover and network sharing (Elwert & Tegunimataka, 2016;Furtado & Theodoropoulos, 2010;Meng & Gregory, 2005). The positive effects of intermarriage, such as income and employment, are then transmitted from generation to generation. Hence, children of intermarriage are less likely to suffer from obstacles in economic and social integration facing immigrants and their descendants. This is, for example, seen comparing grades of 2.5-generation migrants to the grades of immigrants or second-generation immigrants (Kalmijn, 2015b) (although origin differences exist, see Tegunimataka, 2020). The overall better socioeconomic position of the children of the 2.5 generation would translate into positive health outcomes since health strongly correlates with socio-economic status (Smith, 1999).

Mental health across immigrant generations
There are many theoretical reasons to expect worse mental health for migrants (especially for refugees) compared to natives. Stressors in connection to migration itself, such as exposure to war, terrorism, poverty, and natural disaster in their origin, as well as separation from a family member, or difficult journeys, can lead to increased risk for emotional imbalance for recent migrants. Migrants are, in addition, exposed to potential stressors after arriving in the receiving country. Unemployment, poverty, unsafe living arrangements, discrimination, and prejudice are all factors that can have potentially negative effects on the mental health of migrants. Yet, empirical research has been indecisive in this regard, with some studies showing that immigrant's mental health is better than the mental health of the native populations, (Breslau & Chang, 2006;Marks et al., 2014), (which is sometimes referred to as paradoxical), while other studies instead have shown indications of poorer mental health for migrants (Cavdar et al., 2021;Di Napoli et al., 2021;Klein et al., 2020). Migrant mental health also differs according to origin (Lindert et al., 2008) and intersects with other factors such as gender, socio-economic status (SES), race, and ethnicity (Alegría et al., 2017). The 1.5 generation sometimes also displays a higher likelihood of mental health issues compared to immigrants that arrive later in life (Alegria et al., 2007;Takeuchi et al., 2007), which is explained by negative health assimilation.

3
The Health of Immigrant Youth in Denmark: Examining Immigrant… Research focusing on the mental health of second-generation migrants has shown a lower mental health status compared to the first generation (Ampadu, 2011), and this is (again) often explained by negative health assimilation. Second-generation youth assimilate into the society of the host country while maintaining strong ties to the country and culture of the origin and are therefore affected by possible tension and divided loyalties that may play out in worse mental health (Giguère et al., 2010). Second-generation migrants with a more distant cultural background also face difficulties in terms of host country assimilation and may instead experience stagnating or downward assimilation, leading to negative health behavior and worse mental health outcomes (Portes & Zhou, 1993). Furthermore, they may experience discrimination and stigmatization from the majority society (Kalmijn, 2015b).
Children of the 2.5 generation may face difficulties having multiple ethnic and/ or racial identities. American literature on biracial children helps guide our expectations for the outcomes of the 2.5 generation in Europe. Biracial individuals might be unable to assimilate fully into single-race groups, as they may be rejected by their members since they do not fit clearly into any single race category. Being rejected by both single-race groups could lead to poor mental health (Campbell & Eggerling-Boeck, 2006). Another source of stress for biracial children is when there is a conflict between an individual's own identification and the identity assigned by others. Biracial individuals may choose racial labels for themselves that are not recognized by others or do not match their appearance (Campbell & Eggerling-Boeck, 2006). In other instances, the biracial individual may struggle to find a fitting label and selfidentity, which can lead to increased mental health issues. Although racial identity may change through life, adolescence is a period when issues of identity are emphasized (Hitlin et al., 2006).
Children with an immigrant parent originating from a geographically and/or culturally more distant country may experience greater identification struggles, both regarding own identification and in terms of the identification from the majority society (Kalmijn, 2015b). Along these lines, we would also expect that other generations of immigrants from more culturally or geographically distant countries would be subject to potential discrimination from the majority of society. In other words, origin from a more distant place may have negative mental health effects for all individuals with an immigrant background.
Another aspect that may affect the mental health of the 2.5 generation, in particular, is parental conflict and separation, which is more common in intermarriages than in other marriages (Dribe & Lundh, 2012;Zhang & Van Hook, 2009). Intermarried individuals may differ in terms of religion, education, and ethnicity which reproduce differences in norms, values, and ways of communication that may lead to increased conflict and possible dissolution. Intermarried families may, moreover, obtain less support from relatives and external family members (Kalmijn et al., 2005), as some may object to the marriage. Parental country of origin matters also in this regard as it is expected that the greater the cultural distance between spouses, the greater the risk of union dissolution (Dribe & Lundh, 2012). Divorce and separation mean a potential loss of financial resources and enhanced emotional stress of parents and children (Bernardi & Radl, 2014) and could potentially also lead to mental health problems (Strohschein, 2005). Furthermore, the mental health and well-being of parents affect the mental health of their children (Olfson et al., 2003). This is explained by mechanisms such as genes, negative parental practices, lack of emotional connection between parents and children. Children of the 2.5 generation are accordingly exposed to parental stress, both from the increased risk of parental divorce and conflict in intermarriages, but also since the immigrant parent could be experiencing stress in connection with the processes of migration and integration. A recent Finnish study (Loi et al., 2021) analyzed the physical and mental health across migrant generations in Finland, including 2.5-generation migrants and they found indication of negative health assimilation for first-and second-generation migrants as they poorer physical and mental compared to native children. They also found that the 2.5 generation had an elevated risk of mental health problems.

Hypotheses
Immigrants' physical health tends to be better than the native population in the receiving country in accordance with the HIE. A better physical health than natives is also expected for the 1.5 generation, although some negative health assimilation takes place for this group, and thus, we expect: H1: Immigrant youth and immigrant youth arriving in Denmark before the age of 13 have better physical health than native Danes in the same age groups, although the 1.5 generation's health is somewhat closer to the native population due to negative health assimilation.
The health of second-generation migrants is converging with the health of natives by adjusting to native norms and health behavior, and previous studies have even shown worse health than natives (Stirbu et al., 2006;Tarnutzer & Bopp, 2012) thus we expect: H2: Second-generation migrants have health that is poorer than the health of first-and 1.5-generation migrants. Especially second-generation immigrant youth that belong to more marginalized groups in the Danish society have a physical health that is poorer than to the health of native Danes in the same age groups.
For the 2.5 generation, we instead expect:

H3: 2.5 generation migrants have a physical health that is not significantly different from natives.
Previous literature has shown a paradoxical pattern of migrant mental health with better mental health compared to the native population, despite many theoretical reasons to expect the opposite. There is, however, no reason to expect any differences in outcomes in the Danish case, and thus we expect:

3
The Health of Immigrant Youth in Denmark: Examining Immigrant… H4: Immigrant youth and immigrant youth arriving in Denmark before the age of 13 have better mental health than native Danes in the same age groups, although the 1.5 generation's mental health is somewhat closer to the native population due to negative health assimilation.
For the second generation, we expect a negative health assimilation, also for mental health. Second-generation youth may experience tension and divided loyalties that could lead to mental health issues. There are many reasons to expect poorer mental health of the 2.5 generation compared to native Danes. 2.5-generation migrants often experience issues with identification, both regarding own identification and in terms of the identification from the majority society. They may also experience family conflict to a higher extent due to differences in norms, values, and ways of communication between the parents that may lead to increased tension and possible family dissolution. We thus expect: H5: The mental health of second and 2.5-generation youth in Denmark is poorer than the mental health of native Danes in the same age groups.
Youth with geographically and/or culturally more distant background may experience greater identification struggles and may be subject to potential discrimination from the majority society. Thus, we expect: H6: second-generation and 2.5-generation immigrant youth that belong to more marginalized groups in the Danish society have poorer mental health than native Danes in the same age groups.

Data and Methods
We use Danish longitudinal register data from several administrative registers available through the Danish Civil Registration System (CRS) administered by Statistics Denmark. All individuals in Denmark are required by law to register with the (CRS) and are thereby given a personal identification number, which gives full coverage of the Danish population. Parents register their children with the CRS at the time of birth and if the parents are married, the spouse of the mother will be automatically assigned as the father of the child. Basic demographic information in this study is obtained from a register covering the years 1991 onwards, and through an anonymized individual identifier, multigenerational links are obtained. We study hospitalization and data on this are obtained from the Danish Health Data Authority. Information on inpatient care has been collected since 1991 and can be linked to other register data through the CRS. Medical conditions are classified in accordance with the IDC 3 -10 categorization.
Our main explanatory variable denotes immigrant generation (1). 2.5 generation with one native and one immigrant parent, 2). second-generation migrants, and 3). 1.5-generation migrants and 4). Immigrants. 4 Native Danes serve as a reference category in all models. We control for demographic variables the number of siblings and to account for the socioeconomic position of the family, parental education is defined as the highest education obtained for both the mother and the father at age 12 of the child. The variable has three categories: 1) primary education, 2) secondary education, and 3) tertiary education.
We furthermore control for parental separation by age 12. Marriages are registered by the CRS and thereby definable in our data, whereas the information on cohabitation needs to be constructed. The Danish Family Registry (FAIN) obtains exact information on the house or apartment in which an individual lives, which enables the identification of cohabitation. We are using statistics Denmark's definition of cohabiting couples as two unrelated individuals of the opposite sex registered at the same address, with an age difference of fewer than 15 years and with no other adult person registered at the same address. Having information on cohabitation allows us to account for not only divorce, but also parental separation. We moreover control for region of residence, which consist of Denmark's five regions. 5 We perform subgroup analyses, first separating our analysis by gender and then by region of origin. We separate the data into five origin groups: 1). Europe and Western Countries, 2). Middle Eastern countries, 3). African countries, 4). Asian countries and 5) Latin American countries. 6 Our data on hospitalization are available for the years 1991-2015, and we include 90% of all individuals registered in Denmark aged 12-25 these years. In order to construct our main explanatory variable (immigrant generation), individuals must be linked to both biological parents. Thus, individuals with only one registered parent are excluded. We also drop those individuals with missing information on parental education and individuals of the 2.5 generation whose native parent is a secondgeneration immigrant.
We conduct logistic regressions with two main outcome variables using STATA16. The first is a dummy variable that denotes hospitalization for any physical conditions (0 = no hospitalization, 1 = ever hospitalized for physical conditions before the age of 25) The second one is constructed in the same way as a dummy variable, but here we are interested in hospitalization due to a mental condition (0 = no hospitalization, 1 = ever hospitalized for mental conditions before the age of 25). Table 1 includes the descriptive statistics of the analysis. We see that Native Danes are slightly more likely of being hospitalized for physical conditions than the 1 3 The Health of Immigrant Youth in Denmark: Examining Immigrant… other groups. For mental health, two percent of the 2.5 generation is ever hospitalized during the period of observation, and this number is slightly lower for the other groups. Next, looking at parental education, we see that the 2.5 generation have the highest educated parents, 44 percent of mothers have a university degree, compared to only 20 percent of mothers of the second generation. Our descriptive table also tell us that immigrants, second-generation immigrants and the 1.5 generation belong to families with more children than the other groups and that the 2.5 generation and second generation tend to live in the Copenhagen area.

Hospitalization Physical and Mental health issues
Our aim is to study the health of immigrant youth in Denmark, across generations. Our first outcome of interest denotes hospitalization for physical health problems. Figure 1 visually presents the odds ratios for the four generations of immigrants. Native Danes (individuals with two native Danish parents) serve as a reference category (OR = 1). Odds ratios for all variables are found in Table 2. We see that 2.5-generation migrants have the same odds of ever being hospitalized for physical health issues as the reference category, while the second-generation migrants are somewhat more likely to be hospitalized for a physical condition compared to the reference group (OR 1.088). For the first-generation and the 1.5-generation immigrants, we see a lower likelihood of hospitalization, which is in accordance with the HIE. We see a somewhat higher likelihood of hospitalization, due to a physical problem for the 1.5 generation, compared to immigrants, which confirms our first hypothesis as we expected the 1.5 generation's health being somewhat closer to the native population due to negative health assimilation. We further see a SES gradient with individuals having parents, all else equal, with primary education only being more likely to be hospitalized compared to individuals whose parents have higher levels of education. Females are slightly less likely to be hospitalized for physical conditions compared to males and individuals with siblings are more likely to be hospitalized compared to individuals who have no siblings. We also see that individuals with cohabiting parents are less likely to be hospitalized compared to individuals with separated parents (OR 0.750). There are only minor regional differences in hospitalization found which means that region of residence matters very little for an individual's hospitalization in Denmark (not shown), which can be expected from an Scandinavian welfare state with public healthcare. Separate analysis for men and women reveals an important gender difference for the second generation. The somewhat accentuated likelihood of being hospitalized for physical health issues that was found for the full sample turns out to be driven by males that have an OR of 1.229 displaying a higher likelihood of ever being hospitalized for physical conditions, while for women the OR is 0.961. For the other groups, there is no clear difference between men and women.

3
The Health of Immigrant Youth in Denmark: Examining Immigrant…   The Health of Immigrant Youth in Denmark: Examining Immigrant… Our second outcome of interest denotes hospitalization for mental health problems. We hypothesize that both the second generation and the 2.5 generation have poorer heath than native Danes, and we also expect origin differences.
In Fig. 1 and Table 3, we see that 2.5-generation migrants have a significantly higher likelihood of being hospitalized for mental health problems compared to the reference group (OR 1.118), which was according to our expectations. This is potentially explained by the 2.5-generation problematic self-identification and conflict between an individual's own identification and the identity assigned by others (Campbell & Eggerling-Boeck, 2006). Men belonging to the 2.5-generation immigrants have a somewhat stronger likelihood of hospitalization (OR 1.156) compared to women (OR 1.071). The remaining groups have a lower likelihood of hospitalization compared to the reference group native Danes. This was expected for immigrants and 1.5-generation migrants, however not what we expected for the second generation in accordance with the fifth hypothesis. For our control variables, we see similar patterns as for physical health.

Hospitalization Physical and Mental health issues, origin differences
Turing to origin differences shown in Figs. 2 and 3, 7 and Table 4 and 5. Starting with physical health issues shown in Fig. 2 Table 4. For the 2.5 generation, we find that both men and women with an immigrant parent originating from the Middle East have the highest likelihood of hospitalization (OR 1.147 for men and OR 1.126 for women). This is an unexpected result, and we can only speculate, that this is related to lower socio-economic status (SES) and unemployment of the foreign-born parent of the 2.5 generation that may lead to negative health outcomes for the children. For the second generation we see patterns that sustain our second hypothesis. Second-generation youth with a Middle Eastern background have higher odds of being hospitalized. For youth with an African background that belongs to the second generation, we also see accentuated odds (OR 1.179). Second-generation migrants from the Middle East and Africa are often children of refugees that struggle in the Danish labor market and fall behind native Danes and other immigrant groups to a larger extent (Statistics Denmark, 2018). Unemployment and poverty in the parental generation affects the health of children and youth. There are also clear gender differences, with males with a Middle Eastern and African background having higher odds of hospitalization compared to women. For youth with a Latin American or Asian background, we instead see a lower likelihood of hospitalization. Latin American and Asian migrants are less often refugees in Denmark and rather have other reasons for moving to Denmark, which may influence their economic integration and the likelihood of hospitalization for the second generation.

and in
First-and 1.5-generation migrants have a lower likelihood of hospitalization for physical conditions compared to native Danes for all origin groups and both men and women.

Table 3
Hospitalization Mental Health problems (Logistic Regression-Odds ratios), Immigrant generations. Denmark, 1991-2015 *** p < 0.01, ** p < 0.05, * p < 0.1, All models control for Region of residence. The Health of Immigrant Youth in Denmark: Examining Immigrant… In Fig. 3 and Table 5, we see origin differences for hospitalization due to a mental health. 2.5-generation immigrants with a parent from European/Western countries, Africa and Latin America have a higher likelihood of hospitalization for mental health issues compared to native Danes. This is especially visible for males with one African parent (OR 1.357) and males with one Latin American parent (OR 1.540). What our results here indicate is that youth of the 2.5 generation with an immigrant parent originating from a country more culturally distant from Denmark have substantially worse mental health compared to native Danes. This is an interesting finding in line with previous research on the 2.5-generation migrants in Demark (Tegunimataka, 2020). Youth belonging to the 2.5 generation may experience discrimination from the majority group or may experience greater identification struggles, both regarding their own identification and in terms of identification from the majority society (Kalmijn, 2015b). This would, however, not explain why we see an increased likelihood for hospitalization also for western/European 2.5-generation youth, and why we see lower odds ratios for those with a Middle Eastern and Asian background, which needs further investigation in future studies. For females with one Latin American parent, there is no significant difference from the reference group; however, females with an African parent have a higher likelihood compared to females with two Danish parents (OR 1.259).
Most origin groups of the second generation have a lower likelihood of hospitalization for mental health issues; however, second-generation male migrants with an African background stand out and instead have a higher likelihood (OR 1.232) compared to native Danes. To explain this pattern, we can refer to the marginalization of certain groups in the Danish society leading to a downward assimilation process. Most second-generation migrants with an African background in Denmark originate from Somalia, which is the dominant minority group of African descent in Denmark, and first-generation Somalis is one of the minority groups in Denmark with highest unemployment (Statistics Denmark, 2018). The higher likelihood of hospitalization for mental health for this group may also be related to a specific health behavior of this group; however, we lack such data. The higher likelihood of hospitalization is, however, not visible for women of the second generation with an African background. In general, the likelihood of hospitalization for mental health issues is lower than for natives for origin groups of the 1.5 generation. Results are, however, statistically insignificant for Latin American females and for Asian, African and Latin American men. For immigrants all groups (except those with Latin American background that are dropped due to small numbers), have a lower likelihood of hospitalization due to mental health issues than native Danes.

Conclusions
This paper studies mental and physical health of migrant youth in Denmark, while considering differences in terms of immigrant generation and origin. This is motivated by the growing heterogeneity of the immigrant group in many immigrant receiving western and northern European countries, which is a result of globalizing migration patterns. Moreover, many minority groups by now, consist of more     *** p < 0.01, ** p < 0.05, * p < 0.1, All models control for parental education, sex, number of children in household, parental cohabitation and region of residence.

3
The Health of Immigrant Youth in Denmark: Examining Immigrant…    *** p < 0.01, ** p < 0.05, * p < 0.1, All models control for parental education, sex, number of children in household, parental cohabitation and region of residence.

3
The Health of Immigrant Youth in Denmark: Examining Immigrant… than one generation, that together with increasing intermarriage rates have led to the migrant group consisting of more generations of immigrants. Despite these developments, very little is known about health differences between generations and origin groups. We apply Danish register data to study hospitalization during adolescence and young adulthood (ages 12-25) and set out to test five hypotheses while controlling for important socio-demographic variables. Our results sustain the first hypothesis and show that first-generation and the 1.5-generation immigrants, have a lower likelihood of hospitalization, which is in accordance with the HIE. As expected, we see a somewhat higher likelihood of hospitalization, due to a physical problem for the 1.5 generation, compared to immigrants, which would be due to negative health assimilation. The second generation has worse physical health than natives, as was expected according to previous literature (Stirbu et al., 2006;Tarnutzer & Bopp, 2012). Studying potential heterogeneities, we see that men, especially those with a Middle Eastern or African background belonging to the second generation, have worse physical health. We argue that the parental struggles on the Danish labor market can translate into worse health for children and youth and official statistics show that refuges originating from the Middle East and North Africa in particular, fall behind in the Danish society. For the overall 2.5 generation, there is no significant difference from the reference group, which was according to our expectations due to strong exposure to native lifestyle through the native parent.
Turning to mental health, we find interesting group differences. As expected, we see poorer mental health for the 2.5 generation as a whole. This was expected, as previous literature indicated that 2.5-generation migrants often experience issues with identification and may also experience family conflict and separation to a higher extent due to differences in norms, values, and ways of communication between the parents, which can lead to mental health issues for this group in particular. When considering origin differences, we find that youth of the 2.5 generation with an immigrant parent originating from Africa or Latin America have poorer mental health compared to native Danes. We relate this to identification issues, both in relation to own identification and identification by the majority society leading to adverse health outcomes. However, we also see an increased likelihood for hospitalization also for western/European 2.5-generation youth, but not for those with a Middle Eastern and Asian background, which needs further investigation in future studies.
For the second generation, we once again find an increased likelihood of mental health hospitalization when considering gender and origin differences, and we find worse mental health for males with an African background. An important limitation of this study is that we have no data on health behavior; thus, we cannot measure this specific association; however, we argue that this finding can be related to negative assimilation and health behavior and/or a marginalization in the labor market of the parental generation leading to lower family income and lower SES. Differences in health assimilation along racial lines have moreover been found in previous research (see, e.g., .
Another caveat is the inability to account for several important factors influencing the individual's home and surrounding environment, including the amount of conflict between the parents, the quality of relationships within the household, different types of health behavior such as smoking. Moreover, we are unable to account for income and instead use parental education as a proxy for family SES. These limitations are important to recognize when interpreting the results.
Despite obvious limitations, our findings shed important light on the association between mental and physical hospitalization and immigrant generation/origin. From the results, it is evident that physical health status differs according to immigrant generation for youth in Denmark. We see clear evidence for the HIE when studying physical health, as both the first immigrant generation and the 1.5 generation display significantly better physical health than native Danes. For mental health we see different patterns, with youth belonging to the 2.5 generation (if not all groups) experiencing poorer mental health.
This study highlights two important areas for further investigation. First, second-generation migrants, especially males from the Middle East and from Africa, have poorer physical health compared to native youth, but also compared to other immigrant origin groups. This is an important area of future research and should be highlighted in connection with policy making, to increase health equality for young people in Denmark. Second, the 2.5 generation struggle with mental health issues to a higher extent than other groups, and this is seen for 2.5-generation immigrants with an African and Latin American background, which we argue can be explained by identification struggles and stigmatization from the majority society. The mental health of the 2.5 generation in Denmark should thus also be an area of focus for policymakers and researchers. 1 3