- Split View
-
Views
-
Cite
Cite
Angelique Grosser, Oliver Razum, Tanja G.M. Vrijkotte, Ina-Merle Hinz, Jacob Spallek, Inclusion of migrants and ethnic minorities in European birth cohort studies—a scoping review, European Journal of Public Health, Volume 26, Issue 6, December 2016, Pages 984–991, https://doi.org/10.1093/eurpub/ckw068
- Share Icon Share
Background: Migrant and ethnic minority groups constitute substantial parts of European populations. They frequently experience health disadvantages relative to the respective majority populations. Birth cohort studies can help to disentangle social and biological factors producing these health inequalities over the life course. We investigated whether birth cohorts in European countries (i) assess migration history and ethnicity in the study design; and (ii) use this information in data analyses. Methods: A scoping review was performed in which European birth cohort studies were identified using dedicated web-based registries, MEDLINE and EMBASE. Two reviewers systematically assessed all identified birth cohorts and selected those fulfilling defined inclusion criteria (e.g. enrolment after 1980). Publications and websites were screened for information on the inclusion of migrants and ethnic minorities. To obtain more detailed information, researchers of enrolled birth cohorts were contacted individually. Results: Eighty-eight birth cohorts were identified in 20 European countries, with more than 486 250 children enrolled in total. Sixty-two studies (70.5%) reported collecting data about migration history or ethnic background. Twenty-three studies (26%) used information on migration history or ethnicity for data analyses or plan to do so in future. Conclusion: The majority of European birth cohorts assessed participants’ migration history or ethnic background; however, this information was seldom used for comparative analyses in trying to disentangle reasons for health inequalities. Also, heterogeneous indicators were used. Better use of data already available, as well as harmonization of data collection on migration history and ethnicity, could yield interesting insights into the production of health inequalities.
Introduction
Birth cohort studies can be a powerful tool to investigate biological, environmental and social influences on health.1 They help with understanding diseases from a life-course perspective as several aspects of health status in later life are associated with early-life exposures.2,3 They allow researchers to disentangle the contributions of individual and contextual factors to inequalities in health.2 Including diverse population groups in birth cohort studies enables researchers to study the respective influences of “nature” (genetics, with increased phenotypical richness) and “nurture” (environment and culture, including social determinants), as well as their interaction, on health.1,4. When birth cohort studies aim to investigate the production of health inequalities, it is particularly important that diverse and potentially disadvantaged population groups are enrolled. Examples of such groups are migrants and ethnic minorities.5
Throughout Europe, the diversity of populations is increasing, largely as a consequence of past and present migration movements. In 2011, ∼10.1% of the EU population were either foreign-born or citizen of a country other than that in which they resided.6 In Western European countries, a considerable part of new-borns have parents with a migration history or are part of an ethnic minority. This is of concern for public health as being a migrant or member of an ethnic minority can have an impact on health and has been associated with health inequalities.7,8 Even though migrants are often initially healthier (the so-called healthy migrant effect), they tend to experience health inequalities with increasing duration of residence in the host country. Socioeconomic disadvantages, language barriers and other access barriers to health services contribute to health inequalities;7,9,10 the effect of such exposures may accumulate over the life course.4 Elucidating the underlying mechanisms can help to identify starting points for interventions.11 Within the last decades, awareness has grown and researchers in Europe are increasingly enrolling migrants and ethnic minority populations in epidemiological studies. However, in many studies migrants and ethnic minorities are still excluded, underrepresented or their migration status is not clearly defined or assessed.5,12
The aim of this article is to evaluate how birth cohort studies in Europe are handling diversity in terms of migration history and ethnic background. The main research questions refer to whether birth cohorts in European countries (i) assess ethnicity and migration history in the study design; and (ii) use this information in data analyses. More specifically, we assessed what proportion of birth cohort studies in Europe considers migration history or ethnicity in their study design; how they define it; and whether this information, if available, has been (or will be) used for comparative analyses of health differentials.
Methods
We conducted a scoping review with the aim of identifying all European birth cohort studies and describing whether and how they have considered migrants and ethnic minorities; it was not our aim, however, to identify all publications reporting results from analyses of birth cohort studies.13,14
European birth cohort studies were identified using web-based registries of birth cohorts, and by a search in the scientific databases MEDLINE and EMBASE as well as in the reference sections of published papers that were retrieved. The aim of this search strategy was to identify all birth cohorts in Europe but not all publications of each cohort.
A birth cohort study enrolling or following up participants simultaneously in multiple locations was considered as one study, whereas studies set up by the same investigators at different time periods were counted separately. Cohort identification took place between June and August 2014 and in January 2016. Search terms were intentionally kept broad, with ‘birth cohort’ in the title, ‘Europe’ for place of study, and ‘1980–2014’ for publication date (c.f. table 1). While birth cohorts have been set up in European countries at least since 1946,1 interest in migrant/ethnic minority research grew only in the 1980s, hence the lower limit of publication years. The MEDLINE and EMBASE search resulted in 1888 titles that were screened to assess whether they indeed reported on a birth cohort study, and then checked for duplicates. Thus, 122 different birth cohort studies were identified.
Step . | Search strategy . | No. of results in MEDLINE, June 2014 . | No. of results in EMBASE, January 2016 . | No. of results in total . |
---|---|---|---|---|
1 | birth cohort.m_titl. | 1925 | 3161 | 5086 |
2 | exp Europe/ | 1 086 835 | 1 252 988 | 2 339 823 |
3 | 1 and 2 | 920 | 980 | 1900 |
4 | Limit 3 to yr= “1980–2014” | 911 | 977 | 1888 |
Step . | Search strategy . | No. of results in MEDLINE, June 2014 . | No. of results in EMBASE, January 2016 . | No. of results in total . |
---|---|---|---|---|
1 | birth cohort.m_titl. | 1925 | 3161 | 5086 |
2 | exp Europe/ | 1 086 835 | 1 252 988 | 2 339 823 |
3 | 1 and 2 | 920 | 980 | 1900 |
4 | Limit 3 to yr= “1980–2014” | 911 | 977 | 1888 |
Step . | Search strategy . | No. of results in MEDLINE, June 2014 . | No. of results in EMBASE, January 2016 . | No. of results in total . |
---|---|---|---|---|
1 | birth cohort.m_titl. | 1925 | 3161 | 5086 |
2 | exp Europe/ | 1 086 835 | 1 252 988 | 2 339 823 |
3 | 1 and 2 | 920 | 980 | 1900 |
4 | Limit 3 to yr= “1980–2014” | 911 | 977 | 1888 |
Step . | Search strategy . | No. of results in MEDLINE, June 2014 . | No. of results in EMBASE, January 2016 . | No. of results in total . |
---|---|---|---|---|
1 | birth cohort.m_titl. | 1925 | 3161 | 5086 |
2 | exp Europe/ | 1 086 835 | 1 252 988 | 2 339 823 |
3 | 1 and 2 | 920 | 980 | 1900 |
4 | Limit 3 to yr= “1980–2014” | 911 | 977 | 1888 |
In August 2014, all birth cohort studies listed in the following inventories were retrieved: www.birthcohorts.net (n = 66), www.chicos.eu (n = 77) and www.enrieco.eu (n = 46). The registries have been established in 2010 and 2011. The website www.birthcohorts.net is searchable and contains birth cohort studies worldwide.15 The webpage www.chicosproject.eu represents an inventory of European-based birth cohorts as well as mother–child cohorts covering a variety of topics.16 The webpage www.birthcohortsenrieco.net specifically focuses on birth cohort studies with environmental exposure data.17 Registration of birth cohort studies at these registries is voluntary, they may therefore not contain all existing birth cohort studies.
Seven additional birth cohort studies were found through hand searches, e.g. from reference lists of screened literature, or through conference attendance. Combining all search strategies and removing obvious duplicate studies resulted in 202 birth cohort studies.
The final set comprised 88 birth cohort studies. In February 2015 and February 2016, the investigators of these studies were e-mailed a short questionnaire to validate and complement the publicly available information about the study. Reminders were sent 1–2 weeks later. Seventy-four of the 88 studies (84.1%) responded. All 88 birth cohort studies were included in the qualitative synthesis.
For the qualitative synthesis, information on the inclusion of migrants and ethnic minorities in study design and data analysis was collated. It needs to be noted that in the context of birth cohorts, “migration status” always relates to the parents or grandparents (most frequently to the mothers), as new-borns usually have not migrated. In recent years, substantial efforts have been made to develop standardized operational definitions for the group of migrants and their descendants as well as for ethnic minority groups.18–20 However, these definitions were published only in the last years of the study period, and—given the heterogeneity of these target groups—are not undisputed. Hence, we tried to be as inclusive as possible, screening the studies on information on (i) translation of study materials into a language other than the official one(s) of the host country; (ii) data collection on place of birth, nationality and ethnicity of mother, father and/or grandparents of children enrolled; (iii) other measurements to identify ethnic or migrant groups (e.g. language spoken at home) and (iv) use of these data for analyses (already performed or planned) comparing migrants/ethnic minorities to the (ethnic) majority population.
Results
General properties of the birth cohort studies
We included 88 European birth cohorts from 21 different European countries. These studies have enrolled, or are planning to enrol, a total of 486 250 children. The median size of the birth cohorts is 1409 children, ranging from 100 (the lower limit for inclusion) to 100 000. The majority of studies are located in Northern, Western and Central Europe. Most studies started in the 2000s, and eight studies are still in the recruitment phase (see table 2).
Start of enrolment . | Name of study (reference number, only one reference listed) . | Country of study . | No. of children . | Proportion of migrants/ ethnic minorities of mothers in sample . | Concept(s) of migration history/ethnicity measured . | ||||
---|---|---|---|---|---|---|---|---|---|
(X = concept used to depict proportion of migrants/ ethnic minorities among mothers in study) . | |||||||||
Place of birth . | Nationality . | Self-assigned ethnicity . | Other, e.g. mother tongue . | No. of concept(s) measured . | |||||
1982 | Merthyr Allergy Study15 | UK | 497 | 0 | X | 1 | |||
1983b | Greek birth cohort24 | Greece | 10 859 | Unknown | |||||
1985b | NFBC25 | Finland | 9479 | Unknown | |||||
1986 | CHEF 115 | Faroe Island | 1022 | Unknown | x | x | 2 | ||
1989 | Isle of Wight birth cohort study16 | UK | 1456 | Unknown | 0 | ||||
1990c | MAS 5 (Cities birth cohort)15 | Germany | 1314 | 3 | X | x | 2 | ||
1990 | MEFAB16 | Netherlands | 1200 | Unknown | 0 | ||||
1991 | ALSPAC15 | UK | 15 390 | 2.2 | x | X | 2 | ||
1992 | ECA birth cohort in Oslo26 | Finland | 3754 | 2.5 | X | 1 | |||
1992b,c | FCOU16 | Ukraine | 1148 | Unknown | |||||
1993 | Düsseldorf (1993–2000)27 | Germany | 126 | Unknown | 0 | ||||
1994 | Tyrol birth cohort7 | Austria | 33 808 | Unknown | 0 | ||||
1994 | CHEF 228 | Faroe Island | 182 | Unknown | x | x | 2 | ||
1994b | DIPP (1994)29 | Finland | 31 526 | Unknown | x | 1 | |||
1996 | BAMSE (Birthcohorts.net 2011)1 | Sweden | 4093 | 13.6 | X | 1 | |||
1996 | DNBC15 | Denmark | 100 000 | Unknown | 0 | ||||
1996 | PIAMA15 | Netherlands | 3963 | 8.2 | X | x | 2 | ||
1996 | Asthma Multicenter Infant Cohorts Study30 | Spain | 487 | 25.1 | x | x | X | 3 | |
1996b | Swedish birth cohort31 | Sweden | 1228 | Unknown | |||||
1997 | ABIS15 | Sweden | 17 055 | 10.7 | X | 1 | |||
1997 | CHEF 332 | Faroe Island | 656 | Unknown | x | x | 2 | ||
1997 | INMA-Environment and Childhood15 | Spain | 3768 | 9.5 | x | X | 2 | ||
1997 | LISA15 | Germany | 3097 | Unknown | 0 | ||||
1998 | COPSAC—200033 | Denmark | 411 | 5 | X | 1 | |||
1998 | DARC15 | Denmark | 562 | 0 | X | x | 2 | ||
1998 | SEATON15 | UK | 1400 | Unknown | 0 | ||||
1998 | SWS15 | UK | 2567 | 4 | x | X | 2 | ||
1999 | German birth cohort study34 | Germany | 3132 | 16 | X | x | 2 | ||
1999 | GMS (Gatesehead Millenium Cohort Study)15 | UK | 1029 | 2 | X | 1 | |||
1999 | IVAAQ35 | Greenland | 403 | 0 | X | x | 2 | ||
1999 | MoBa15 | Norway | 85 176 | 10.8 | X | 1 | |||
1999b | NAC I36 | Italy | 242 | Unknown | |||||
2000 | CHEF 432 | Faroe Island | 100 | Unknown | x | x | 2 | ||
2000 | German; Ulm (2000–2001)37 | Germany | 1090 | 14.7 | x | X | x | 3 | |
2000 | KOALA Birth Cohort Study16 | Netherlands | 2834 | 10 | X | x | 2 | ||
2000 | Ukraine European Collaborative38 | Ukraine | 4759 | 5 | X | x | 2 | ||
2001c | Lifeways Cross-GenerationCohort15 | Ireland | 1082 | Unknown | 0 | ||||
2001 | WHISTLER15 | Netherlands | 2000 | 20 | X | x | 2 | ||
2002 | Berlin Pregnancy Cohort39 | Germany | 473 | 5 | X | 1 | |||
2002 | Duisburg40 | Germany | 234 | 12.5 | x | X | x | 3 | |
2002 | EDEN15 | France | 1418 | Unknown | x | 1 | |||
2002 | Generation R41,42 | Netherlands | 9749 | 41 | x | x | X | x | 4 |
2002 | INUENDO43 | Greenland, Poland, Ukraine | 1322 | Unknown | 0 | ||||
2002 | LUKAS16 | Finland | 396 | 0 | X | 1 | |||
2002 | Slovak PCB Cohort44 | Slovakia | 1200 | 21 | x | X | x | 3 | |
2002c | PELAGIE43 | France | 2398 | 1.5 | X | 1 | |||
2003c | ABCD45 | Netherlands | 8266 | 38.5 | X | x | x | x | 4 |
2003b | EUROPEAN, MOSAIC-study46 | EU-Project | 7222 | Unknown | |||||
2003b | GASPII47,48 | Italy | 719 | 4.1 | X | 1 | |||
2003b | HUMIS49,50 | Sweden | 6000 | Unknown | x | 1 | |||
2003 | Krakow cohort43 | Poland | 400 | 0 | X | x | 2 | ||
2003b | PARIS51 | France | 4115 | 13.5 | X | 1 | |||
2003 | SNiP43 | Germany | 4840 | 3 | X | x | 2 | ||
2004 | Co.N.ER (Bologna birth cohort)15 | Italy | 654 | 0 | x | X | 2 | ||
2004 | Danish birth cohort52 | Denmark | 228 | 4 | X | x | 2 | ||
2004b | TI-MOUN53 | France | 141 | 8.5 | X | 1 | |||
2005 | EuroPrevall15 | EU-Project | 12 049 | 7 | X | 1 | |||
2005 | G21 (Generation XXI)54,55 | Portugal | 8647 | 9 | X | 1 | |||
2005a | NINFEA56 | Italy | 1105 | 4.2 | X | x | 2 | ||
2006a,b | LucKi15 | Netherlands | 5000 | Unknown | |||||
2006 | STEPS study15 | Finland | 1827 | Unknown | 0 | ||||
2006 | UPPSAT57 | Sweden | 2493 | Unknown | x | x | 2 | ||
2006 | Dutch birth cohort58 | Netherlands | 156 | 32 | X | 1 | |||
2006 | Aladdin birth cohort59 | Finland | 552 | 15.4 | X | 1 | |||
2007 | BiB (Born in Bradford)22 | UK | 13 818 | 60.6 | x | X | x | 3 | |
2007 | CHEF 532 | Faroe Island | 475 | Unknown | x | x | 2 | ||
2007c | KANC (Kaunas Cohort)60 | Lithuania | 1919 | 3 | X | 1 | |||
2007 | NAC II and EAC61,62 | Italy | 632 | 0 | X | x | 2 | ||
2007 | REPRO_L15 | Poland | 1800 | 0 | X | x | 2 | ||
2007c | RHEA15 | Greece | 1317 | 8.9 | x | X | x | 3 | |
2008 | Cork BASELINE15 | Ireland | 2137 | 16.7 | X | x | x | 3 | |
2008 | RISK63 | Netherlands | 2703 | Unknown | x | 1 | |||
2009 | BILD15 | Switzerland | 364 | Unknown | 0 | ||||
2009 | Italian birth cohort64 | Italy | 1064 | Unknown | 0 | ||||
2009a | MUBICOS15 | Italy | 1000 | 0 | X | x | 2 | ||
2010 | NorFlu16 | Norway | 3200 | 8.5 | X | 1 | |||
2010a | Odense Child Cohort15 | Denmark | 4000 | 8 | x | X | 2 | ||
2011 | Piccolipiù15 | Italy | 3000 | Unknown | x | x | 2 | ||
2011a | PRIDE65 | Netherlands | 459 | Unknown | x | x | 2 | ||
2011c | Child Sleep Cohort66 | Finland | 1678 | 0 | x | 1 | |||
2011b | EPIPAGE 267 | France | 4290 | Unknown | 0 | ||||
2011a,b | LIFE Child15 | Germany | 2000 | Unknown | |||||
2012 | SPATZ (Ulm)68 | Germany | 934 | 12.5 | x | X | 2 | ||
2012 | LeuBiCo16 | Germany | 100 | Unknown | 0 | ||||
2013a | BaBi birth cohort study23 | Germany | 1000 | Unknown | x | x | x | 3 | |
Unknown | PASTURE69,70 | EU-Project | 938 | 3.5 | X | 1 | |||
Unknown | SELMA71 | Sweden | 2582 | Unknown | x | x | 2 | ||
Unknowna | Go Child birth cohort72 | UK | 341 | Unknown | 0 |
Start of enrolment . | Name of study (reference number, only one reference listed) . | Country of study . | No. of children . | Proportion of migrants/ ethnic minorities of mothers in sample . | Concept(s) of migration history/ethnicity measured . | ||||
---|---|---|---|---|---|---|---|---|---|
(X = concept used to depict proportion of migrants/ ethnic minorities among mothers in study) . | |||||||||
Place of birth . | Nationality . | Self-assigned ethnicity . | Other, e.g. mother tongue . | No. of concept(s) measured . | |||||
1982 | Merthyr Allergy Study15 | UK | 497 | 0 | X | 1 | |||
1983b | Greek birth cohort24 | Greece | 10 859 | Unknown | |||||
1985b | NFBC25 | Finland | 9479 | Unknown | |||||
1986 | CHEF 115 | Faroe Island | 1022 | Unknown | x | x | 2 | ||
1989 | Isle of Wight birth cohort study16 | UK | 1456 | Unknown | 0 | ||||
1990c | MAS 5 (Cities birth cohort)15 | Germany | 1314 | 3 | X | x | 2 | ||
1990 | MEFAB16 | Netherlands | 1200 | Unknown | 0 | ||||
1991 | ALSPAC15 | UK | 15 390 | 2.2 | x | X | 2 | ||
1992 | ECA birth cohort in Oslo26 | Finland | 3754 | 2.5 | X | 1 | |||
1992b,c | FCOU16 | Ukraine | 1148 | Unknown | |||||
1993 | Düsseldorf (1993–2000)27 | Germany | 126 | Unknown | 0 | ||||
1994 | Tyrol birth cohort7 | Austria | 33 808 | Unknown | 0 | ||||
1994 | CHEF 228 | Faroe Island | 182 | Unknown | x | x | 2 | ||
1994b | DIPP (1994)29 | Finland | 31 526 | Unknown | x | 1 | |||
1996 | BAMSE (Birthcohorts.net 2011)1 | Sweden | 4093 | 13.6 | X | 1 | |||
1996 | DNBC15 | Denmark | 100 000 | Unknown | 0 | ||||
1996 | PIAMA15 | Netherlands | 3963 | 8.2 | X | x | 2 | ||
1996 | Asthma Multicenter Infant Cohorts Study30 | Spain | 487 | 25.1 | x | x | X | 3 | |
1996b | Swedish birth cohort31 | Sweden | 1228 | Unknown | |||||
1997 | ABIS15 | Sweden | 17 055 | 10.7 | X | 1 | |||
1997 | CHEF 332 | Faroe Island | 656 | Unknown | x | x | 2 | ||
1997 | INMA-Environment and Childhood15 | Spain | 3768 | 9.5 | x | X | 2 | ||
1997 | LISA15 | Germany | 3097 | Unknown | 0 | ||||
1998 | COPSAC—200033 | Denmark | 411 | 5 | X | 1 | |||
1998 | DARC15 | Denmark | 562 | 0 | X | x | 2 | ||
1998 | SEATON15 | UK | 1400 | Unknown | 0 | ||||
1998 | SWS15 | UK | 2567 | 4 | x | X | 2 | ||
1999 | German birth cohort study34 | Germany | 3132 | 16 | X | x | 2 | ||
1999 | GMS (Gatesehead Millenium Cohort Study)15 | UK | 1029 | 2 | X | 1 | |||
1999 | IVAAQ35 | Greenland | 403 | 0 | X | x | 2 | ||
1999 | MoBa15 | Norway | 85 176 | 10.8 | X | 1 | |||
1999b | NAC I36 | Italy | 242 | Unknown | |||||
2000 | CHEF 432 | Faroe Island | 100 | Unknown | x | x | 2 | ||
2000 | German; Ulm (2000–2001)37 | Germany | 1090 | 14.7 | x | X | x | 3 | |
2000 | KOALA Birth Cohort Study16 | Netherlands | 2834 | 10 | X | x | 2 | ||
2000 | Ukraine European Collaborative38 | Ukraine | 4759 | 5 | X | x | 2 | ||
2001c | Lifeways Cross-GenerationCohort15 | Ireland | 1082 | Unknown | 0 | ||||
2001 | WHISTLER15 | Netherlands | 2000 | 20 | X | x | 2 | ||
2002 | Berlin Pregnancy Cohort39 | Germany | 473 | 5 | X | 1 | |||
2002 | Duisburg40 | Germany | 234 | 12.5 | x | X | x | 3 | |
2002 | EDEN15 | France | 1418 | Unknown | x | 1 | |||
2002 | Generation R41,42 | Netherlands | 9749 | 41 | x | x | X | x | 4 |
2002 | INUENDO43 | Greenland, Poland, Ukraine | 1322 | Unknown | 0 | ||||
2002 | LUKAS16 | Finland | 396 | 0 | X | 1 | |||
2002 | Slovak PCB Cohort44 | Slovakia | 1200 | 21 | x | X | x | 3 | |
2002c | PELAGIE43 | France | 2398 | 1.5 | X | 1 | |||
2003c | ABCD45 | Netherlands | 8266 | 38.5 | X | x | x | x | 4 |
2003b | EUROPEAN, MOSAIC-study46 | EU-Project | 7222 | Unknown | |||||
2003b | GASPII47,48 | Italy | 719 | 4.1 | X | 1 | |||
2003b | HUMIS49,50 | Sweden | 6000 | Unknown | x | 1 | |||
2003 | Krakow cohort43 | Poland | 400 | 0 | X | x | 2 | ||
2003b | PARIS51 | France | 4115 | 13.5 | X | 1 | |||
2003 | SNiP43 | Germany | 4840 | 3 | X | x | 2 | ||
2004 | Co.N.ER (Bologna birth cohort)15 | Italy | 654 | 0 | x | X | 2 | ||
2004 | Danish birth cohort52 | Denmark | 228 | 4 | X | x | 2 | ||
2004b | TI-MOUN53 | France | 141 | 8.5 | X | 1 | |||
2005 | EuroPrevall15 | EU-Project | 12 049 | 7 | X | 1 | |||
2005 | G21 (Generation XXI)54,55 | Portugal | 8647 | 9 | X | 1 | |||
2005a | NINFEA56 | Italy | 1105 | 4.2 | X | x | 2 | ||
2006a,b | LucKi15 | Netherlands | 5000 | Unknown | |||||
2006 | STEPS study15 | Finland | 1827 | Unknown | 0 | ||||
2006 | UPPSAT57 | Sweden | 2493 | Unknown | x | x | 2 | ||
2006 | Dutch birth cohort58 | Netherlands | 156 | 32 | X | 1 | |||
2006 | Aladdin birth cohort59 | Finland | 552 | 15.4 | X | 1 | |||
2007 | BiB (Born in Bradford)22 | UK | 13 818 | 60.6 | x | X | x | 3 | |
2007 | CHEF 532 | Faroe Island | 475 | Unknown | x | x | 2 | ||
2007c | KANC (Kaunas Cohort)60 | Lithuania | 1919 | 3 | X | 1 | |||
2007 | NAC II and EAC61,62 | Italy | 632 | 0 | X | x | 2 | ||
2007 | REPRO_L15 | Poland | 1800 | 0 | X | x | 2 | ||
2007c | RHEA15 | Greece | 1317 | 8.9 | x | X | x | 3 | |
2008 | Cork BASELINE15 | Ireland | 2137 | 16.7 | X | x | x | 3 | |
2008 | RISK63 | Netherlands | 2703 | Unknown | x | 1 | |||
2009 | BILD15 | Switzerland | 364 | Unknown | 0 | ||||
2009 | Italian birth cohort64 | Italy | 1064 | Unknown | 0 | ||||
2009a | MUBICOS15 | Italy | 1000 | 0 | X | x | 2 | ||
2010 | NorFlu16 | Norway | 3200 | 8.5 | X | 1 | |||
2010a | Odense Child Cohort15 | Denmark | 4000 | 8 | x | X | 2 | ||
2011 | Piccolipiù15 | Italy | 3000 | Unknown | x | x | 2 | ||
2011a | PRIDE65 | Netherlands | 459 | Unknown | x | x | 2 | ||
2011c | Child Sleep Cohort66 | Finland | 1678 | 0 | x | 1 | |||
2011b | EPIPAGE 267 | France | 4290 | Unknown | 0 | ||||
2011a,b | LIFE Child15 | Germany | 2000 | Unknown | |||||
2012 | SPATZ (Ulm)68 | Germany | 934 | 12.5 | x | X | 2 | ||
2012 | LeuBiCo16 | Germany | 100 | Unknown | 0 | ||||
2013a | BaBi birth cohort study23 | Germany | 1000 | Unknown | x | x | x | 3 | |
Unknown | PASTURE69,70 | EU-Project | 938 | 3.5 | X | 1 | |||
Unknown | SELMA71 | Sweden | 2582 | Unknown | x | x | 2 | ||
Unknowna | Go Child birth cohort72 | UK | 341 | Unknown | 0 |
References 33–72 can be found in Supplementary Appendix S3.
Recruitment ongoing.
Information based on publicly available material only.
Sample size mothers.
Start of enrolment . | Name of study (reference number, only one reference listed) . | Country of study . | No. of children . | Proportion of migrants/ ethnic minorities of mothers in sample . | Concept(s) of migration history/ethnicity measured . | ||||
---|---|---|---|---|---|---|---|---|---|
(X = concept used to depict proportion of migrants/ ethnic minorities among mothers in study) . | |||||||||
Place of birth . | Nationality . | Self-assigned ethnicity . | Other, e.g. mother tongue . | No. of concept(s) measured . | |||||
1982 | Merthyr Allergy Study15 | UK | 497 | 0 | X | 1 | |||
1983b | Greek birth cohort24 | Greece | 10 859 | Unknown | |||||
1985b | NFBC25 | Finland | 9479 | Unknown | |||||
1986 | CHEF 115 | Faroe Island | 1022 | Unknown | x | x | 2 | ||
1989 | Isle of Wight birth cohort study16 | UK | 1456 | Unknown | 0 | ||||
1990c | MAS 5 (Cities birth cohort)15 | Germany | 1314 | 3 | X | x | 2 | ||
1990 | MEFAB16 | Netherlands | 1200 | Unknown | 0 | ||||
1991 | ALSPAC15 | UK | 15 390 | 2.2 | x | X | 2 | ||
1992 | ECA birth cohort in Oslo26 | Finland | 3754 | 2.5 | X | 1 | |||
1992b,c | FCOU16 | Ukraine | 1148 | Unknown | |||||
1993 | Düsseldorf (1993–2000)27 | Germany | 126 | Unknown | 0 | ||||
1994 | Tyrol birth cohort7 | Austria | 33 808 | Unknown | 0 | ||||
1994 | CHEF 228 | Faroe Island | 182 | Unknown | x | x | 2 | ||
1994b | DIPP (1994)29 | Finland | 31 526 | Unknown | x | 1 | |||
1996 | BAMSE (Birthcohorts.net 2011)1 | Sweden | 4093 | 13.6 | X | 1 | |||
1996 | DNBC15 | Denmark | 100 000 | Unknown | 0 | ||||
1996 | PIAMA15 | Netherlands | 3963 | 8.2 | X | x | 2 | ||
1996 | Asthma Multicenter Infant Cohorts Study30 | Spain | 487 | 25.1 | x | x | X | 3 | |
1996b | Swedish birth cohort31 | Sweden | 1228 | Unknown | |||||
1997 | ABIS15 | Sweden | 17 055 | 10.7 | X | 1 | |||
1997 | CHEF 332 | Faroe Island | 656 | Unknown | x | x | 2 | ||
1997 | INMA-Environment and Childhood15 | Spain | 3768 | 9.5 | x | X | 2 | ||
1997 | LISA15 | Germany | 3097 | Unknown | 0 | ||||
1998 | COPSAC—200033 | Denmark | 411 | 5 | X | 1 | |||
1998 | DARC15 | Denmark | 562 | 0 | X | x | 2 | ||
1998 | SEATON15 | UK | 1400 | Unknown | 0 | ||||
1998 | SWS15 | UK | 2567 | 4 | x | X | 2 | ||
1999 | German birth cohort study34 | Germany | 3132 | 16 | X | x | 2 | ||
1999 | GMS (Gatesehead Millenium Cohort Study)15 | UK | 1029 | 2 | X | 1 | |||
1999 | IVAAQ35 | Greenland | 403 | 0 | X | x | 2 | ||
1999 | MoBa15 | Norway | 85 176 | 10.8 | X | 1 | |||
1999b | NAC I36 | Italy | 242 | Unknown | |||||
2000 | CHEF 432 | Faroe Island | 100 | Unknown | x | x | 2 | ||
2000 | German; Ulm (2000–2001)37 | Germany | 1090 | 14.7 | x | X | x | 3 | |
2000 | KOALA Birth Cohort Study16 | Netherlands | 2834 | 10 | X | x | 2 | ||
2000 | Ukraine European Collaborative38 | Ukraine | 4759 | 5 | X | x | 2 | ||
2001c | Lifeways Cross-GenerationCohort15 | Ireland | 1082 | Unknown | 0 | ||||
2001 | WHISTLER15 | Netherlands | 2000 | 20 | X | x | 2 | ||
2002 | Berlin Pregnancy Cohort39 | Germany | 473 | 5 | X | 1 | |||
2002 | Duisburg40 | Germany | 234 | 12.5 | x | X | x | 3 | |
2002 | EDEN15 | France | 1418 | Unknown | x | 1 | |||
2002 | Generation R41,42 | Netherlands | 9749 | 41 | x | x | X | x | 4 |
2002 | INUENDO43 | Greenland, Poland, Ukraine | 1322 | Unknown | 0 | ||||
2002 | LUKAS16 | Finland | 396 | 0 | X | 1 | |||
2002 | Slovak PCB Cohort44 | Slovakia | 1200 | 21 | x | X | x | 3 | |
2002c | PELAGIE43 | France | 2398 | 1.5 | X | 1 | |||
2003c | ABCD45 | Netherlands | 8266 | 38.5 | X | x | x | x | 4 |
2003b | EUROPEAN, MOSAIC-study46 | EU-Project | 7222 | Unknown | |||||
2003b | GASPII47,48 | Italy | 719 | 4.1 | X | 1 | |||
2003b | HUMIS49,50 | Sweden | 6000 | Unknown | x | 1 | |||
2003 | Krakow cohort43 | Poland | 400 | 0 | X | x | 2 | ||
2003b | PARIS51 | France | 4115 | 13.5 | X | 1 | |||
2003 | SNiP43 | Germany | 4840 | 3 | X | x | 2 | ||
2004 | Co.N.ER (Bologna birth cohort)15 | Italy | 654 | 0 | x | X | 2 | ||
2004 | Danish birth cohort52 | Denmark | 228 | 4 | X | x | 2 | ||
2004b | TI-MOUN53 | France | 141 | 8.5 | X | 1 | |||
2005 | EuroPrevall15 | EU-Project | 12 049 | 7 | X | 1 | |||
2005 | G21 (Generation XXI)54,55 | Portugal | 8647 | 9 | X | 1 | |||
2005a | NINFEA56 | Italy | 1105 | 4.2 | X | x | 2 | ||
2006a,b | LucKi15 | Netherlands | 5000 | Unknown | |||||
2006 | STEPS study15 | Finland | 1827 | Unknown | 0 | ||||
2006 | UPPSAT57 | Sweden | 2493 | Unknown | x | x | 2 | ||
2006 | Dutch birth cohort58 | Netherlands | 156 | 32 | X | 1 | |||
2006 | Aladdin birth cohort59 | Finland | 552 | 15.4 | X | 1 | |||
2007 | BiB (Born in Bradford)22 | UK | 13 818 | 60.6 | x | X | x | 3 | |
2007 | CHEF 532 | Faroe Island | 475 | Unknown | x | x | 2 | ||
2007c | KANC (Kaunas Cohort)60 | Lithuania | 1919 | 3 | X | 1 | |||
2007 | NAC II and EAC61,62 | Italy | 632 | 0 | X | x | 2 | ||
2007 | REPRO_L15 | Poland | 1800 | 0 | X | x | 2 | ||
2007c | RHEA15 | Greece | 1317 | 8.9 | x | X | x | 3 | |
2008 | Cork BASELINE15 | Ireland | 2137 | 16.7 | X | x | x | 3 | |
2008 | RISK63 | Netherlands | 2703 | Unknown | x | 1 | |||
2009 | BILD15 | Switzerland | 364 | Unknown | 0 | ||||
2009 | Italian birth cohort64 | Italy | 1064 | Unknown | 0 | ||||
2009a | MUBICOS15 | Italy | 1000 | 0 | X | x | 2 | ||
2010 | NorFlu16 | Norway | 3200 | 8.5 | X | 1 | |||
2010a | Odense Child Cohort15 | Denmark | 4000 | 8 | x | X | 2 | ||
2011 | Piccolipiù15 | Italy | 3000 | Unknown | x | x | 2 | ||
2011a | PRIDE65 | Netherlands | 459 | Unknown | x | x | 2 | ||
2011c | Child Sleep Cohort66 | Finland | 1678 | 0 | x | 1 | |||
2011b | EPIPAGE 267 | France | 4290 | Unknown | 0 | ||||
2011a,b | LIFE Child15 | Germany | 2000 | Unknown | |||||
2012 | SPATZ (Ulm)68 | Germany | 934 | 12.5 | x | X | 2 | ||
2012 | LeuBiCo16 | Germany | 100 | Unknown | 0 | ||||
2013a | BaBi birth cohort study23 | Germany | 1000 | Unknown | x | x | x | 3 | |
Unknown | PASTURE69,70 | EU-Project | 938 | 3.5 | X | 1 | |||
Unknown | SELMA71 | Sweden | 2582 | Unknown | x | x | 2 | ||
Unknowna | Go Child birth cohort72 | UK | 341 | Unknown | 0 |
Start of enrolment . | Name of study (reference number, only one reference listed) . | Country of study . | No. of children . | Proportion of migrants/ ethnic minorities of mothers in sample . | Concept(s) of migration history/ethnicity measured . | ||||
---|---|---|---|---|---|---|---|---|---|
(X = concept used to depict proportion of migrants/ ethnic minorities among mothers in study) . | |||||||||
Place of birth . | Nationality . | Self-assigned ethnicity . | Other, e.g. mother tongue . | No. of concept(s) measured . | |||||
1982 | Merthyr Allergy Study15 | UK | 497 | 0 | X | 1 | |||
1983b | Greek birth cohort24 | Greece | 10 859 | Unknown | |||||
1985b | NFBC25 | Finland | 9479 | Unknown | |||||
1986 | CHEF 115 | Faroe Island | 1022 | Unknown | x | x | 2 | ||
1989 | Isle of Wight birth cohort study16 | UK | 1456 | Unknown | 0 | ||||
1990c | MAS 5 (Cities birth cohort)15 | Germany | 1314 | 3 | X | x | 2 | ||
1990 | MEFAB16 | Netherlands | 1200 | Unknown | 0 | ||||
1991 | ALSPAC15 | UK | 15 390 | 2.2 | x | X | 2 | ||
1992 | ECA birth cohort in Oslo26 | Finland | 3754 | 2.5 | X | 1 | |||
1992b,c | FCOU16 | Ukraine | 1148 | Unknown | |||||
1993 | Düsseldorf (1993–2000)27 | Germany | 126 | Unknown | 0 | ||||
1994 | Tyrol birth cohort7 | Austria | 33 808 | Unknown | 0 | ||||
1994 | CHEF 228 | Faroe Island | 182 | Unknown | x | x | 2 | ||
1994b | DIPP (1994)29 | Finland | 31 526 | Unknown | x | 1 | |||
1996 | BAMSE (Birthcohorts.net 2011)1 | Sweden | 4093 | 13.6 | X | 1 | |||
1996 | DNBC15 | Denmark | 100 000 | Unknown | 0 | ||||
1996 | PIAMA15 | Netherlands | 3963 | 8.2 | X | x | 2 | ||
1996 | Asthma Multicenter Infant Cohorts Study30 | Spain | 487 | 25.1 | x | x | X | 3 | |
1996b | Swedish birth cohort31 | Sweden | 1228 | Unknown | |||||
1997 | ABIS15 | Sweden | 17 055 | 10.7 | X | 1 | |||
1997 | CHEF 332 | Faroe Island | 656 | Unknown | x | x | 2 | ||
1997 | INMA-Environment and Childhood15 | Spain | 3768 | 9.5 | x | X | 2 | ||
1997 | LISA15 | Germany | 3097 | Unknown | 0 | ||||
1998 | COPSAC—200033 | Denmark | 411 | 5 | X | 1 | |||
1998 | DARC15 | Denmark | 562 | 0 | X | x | 2 | ||
1998 | SEATON15 | UK | 1400 | Unknown | 0 | ||||
1998 | SWS15 | UK | 2567 | 4 | x | X | 2 | ||
1999 | German birth cohort study34 | Germany | 3132 | 16 | X | x | 2 | ||
1999 | GMS (Gatesehead Millenium Cohort Study)15 | UK | 1029 | 2 | X | 1 | |||
1999 | IVAAQ35 | Greenland | 403 | 0 | X | x | 2 | ||
1999 | MoBa15 | Norway | 85 176 | 10.8 | X | 1 | |||
1999b | NAC I36 | Italy | 242 | Unknown | |||||
2000 | CHEF 432 | Faroe Island | 100 | Unknown | x | x | 2 | ||
2000 | German; Ulm (2000–2001)37 | Germany | 1090 | 14.7 | x | X | x | 3 | |
2000 | KOALA Birth Cohort Study16 | Netherlands | 2834 | 10 | X | x | 2 | ||
2000 | Ukraine European Collaborative38 | Ukraine | 4759 | 5 | X | x | 2 | ||
2001c | Lifeways Cross-GenerationCohort15 | Ireland | 1082 | Unknown | 0 | ||||
2001 | WHISTLER15 | Netherlands | 2000 | 20 | X | x | 2 | ||
2002 | Berlin Pregnancy Cohort39 | Germany | 473 | 5 | X | 1 | |||
2002 | Duisburg40 | Germany | 234 | 12.5 | x | X | x | 3 | |
2002 | EDEN15 | France | 1418 | Unknown | x | 1 | |||
2002 | Generation R41,42 | Netherlands | 9749 | 41 | x | x | X | x | 4 |
2002 | INUENDO43 | Greenland, Poland, Ukraine | 1322 | Unknown | 0 | ||||
2002 | LUKAS16 | Finland | 396 | 0 | X | 1 | |||
2002 | Slovak PCB Cohort44 | Slovakia | 1200 | 21 | x | X | x | 3 | |
2002c | PELAGIE43 | France | 2398 | 1.5 | X | 1 | |||
2003c | ABCD45 | Netherlands | 8266 | 38.5 | X | x | x | x | 4 |
2003b | EUROPEAN, MOSAIC-study46 | EU-Project | 7222 | Unknown | |||||
2003b | GASPII47,48 | Italy | 719 | 4.1 | X | 1 | |||
2003b | HUMIS49,50 | Sweden | 6000 | Unknown | x | 1 | |||
2003 | Krakow cohort43 | Poland | 400 | 0 | X | x | 2 | ||
2003b | PARIS51 | France | 4115 | 13.5 | X | 1 | |||
2003 | SNiP43 | Germany | 4840 | 3 | X | x | 2 | ||
2004 | Co.N.ER (Bologna birth cohort)15 | Italy | 654 | 0 | x | X | 2 | ||
2004 | Danish birth cohort52 | Denmark | 228 | 4 | X | x | 2 | ||
2004b | TI-MOUN53 | France | 141 | 8.5 | X | 1 | |||
2005 | EuroPrevall15 | EU-Project | 12 049 | 7 | X | 1 | |||
2005 | G21 (Generation XXI)54,55 | Portugal | 8647 | 9 | X | 1 | |||
2005a | NINFEA56 | Italy | 1105 | 4.2 | X | x | 2 | ||
2006a,b | LucKi15 | Netherlands | 5000 | Unknown | |||||
2006 | STEPS study15 | Finland | 1827 | Unknown | 0 | ||||
2006 | UPPSAT57 | Sweden | 2493 | Unknown | x | x | 2 | ||
2006 | Dutch birth cohort58 | Netherlands | 156 | 32 | X | 1 | |||
2006 | Aladdin birth cohort59 | Finland | 552 | 15.4 | X | 1 | |||
2007 | BiB (Born in Bradford)22 | UK | 13 818 | 60.6 | x | X | x | 3 | |
2007 | CHEF 532 | Faroe Island | 475 | Unknown | x | x | 2 | ||
2007c | KANC (Kaunas Cohort)60 | Lithuania | 1919 | 3 | X | 1 | |||
2007 | NAC II and EAC61,62 | Italy | 632 | 0 | X | x | 2 | ||
2007 | REPRO_L15 | Poland | 1800 | 0 | X | x | 2 | ||
2007c | RHEA15 | Greece | 1317 | 8.9 | x | X | x | 3 | |
2008 | Cork BASELINE15 | Ireland | 2137 | 16.7 | X | x | x | 3 | |
2008 | RISK63 | Netherlands | 2703 | Unknown | x | 1 | |||
2009 | BILD15 | Switzerland | 364 | Unknown | 0 | ||||
2009 | Italian birth cohort64 | Italy | 1064 | Unknown | 0 | ||||
2009a | MUBICOS15 | Italy | 1000 | 0 | X | x | 2 | ||
2010 | NorFlu16 | Norway | 3200 | 8.5 | X | 1 | |||
2010a | Odense Child Cohort15 | Denmark | 4000 | 8 | x | X | 2 | ||
2011 | Piccolipiù15 | Italy | 3000 | Unknown | x | x | 2 | ||
2011a | PRIDE65 | Netherlands | 459 | Unknown | x | x | 2 | ||
2011c | Child Sleep Cohort66 | Finland | 1678 | 0 | x | 1 | |||
2011b | EPIPAGE 267 | France | 4290 | Unknown | 0 | ||||
2011a,b | LIFE Child15 | Germany | 2000 | Unknown | |||||
2012 | SPATZ (Ulm)68 | Germany | 934 | 12.5 | x | X | 2 | ||
2012 | LeuBiCo16 | Germany | 100 | Unknown | 0 | ||||
2013a | BaBi birth cohort study23 | Germany | 1000 | Unknown | x | x | x | 3 | |
Unknown | PASTURE69,70 | EU-Project | 938 | 3.5 | X | 1 | |||
Unknown | SELMA71 | Sweden | 2582 | Unknown | x | x | 2 | ||
Unknowna | Go Child birth cohort72 | UK | 341 | Unknown | 0 |
References 33–72 can be found in Supplementary Appendix S3.
Recruitment ongoing.
Information based on publicly available material only.
Sample size mothers.
Outcomes of interest are mainly nutrition-related (comprising e.g. growth patterns, overweight, and obesity), neuro-psychological (e.g. mental health and neurodevelopment), or related to other specific diseases (e.g. respiratory conditions, infections and allergies). Exposures of interest relate to the physical environment, e.g. outdoor and indoor air pollution and toxins, to the social environment, and to lifestyle. Gene–environment interactions are investigated in the majority of the more recent studies. Some studies use comparative approaches with regard to maternal and perinatal health, either between different countries (e.g. INUENDO21) or between different populations within one country (e.g. ABCD,22 BiB,23 BaBi24).
Inclusion of migrants and ethnic minorities in the birth cohorts
Of the 88 birth cohorts, 62 (70.5%) have collected information on migration history or ethnic background of study participants using place of birth, nationality, self-assigned ethnicity or other measures. Table 2 shows the proportion of migrants or ethnic minorities and the different concepts, such as place of birth, nationality or ethnicity, that have been used to identify them. The number of concepts measured varied: 22 studies used just one concept (mainly place of birth) while two studies used four or more concepts (ABCD22 and Generation R25).
Maternal indicators were used most frequently to depict migration history or ethnicity. Information on maternal migration history or ethnic background was collected in 29–40 of the birth cohorts (32.9–45.5%), depending on which concept was used. On average, 10.1% of study participants in the respective birth cohort studies were classified as migrants or ethnic minorities based on these indicators, ranging from 0.0 up to 60.6%. Ten of the studies measuring maternal migration history or ethnicity reported no migrants or ethnic minorities among their study population. Paternal migration history/ethnicity was obtained in 18–27 studies (20.5–30.7%), and that of grandparents in 5–15 studies (5.7–17.0%), depending on the concept used.
Place of birth is the indicator used most frequently to establish migration history. Indicators other than place of birth, nationality and ethnicity of parents or grandparents are used in 17 birth cohorts (19.3%); they comprise language spoken at home (n = 7), years lived in country of study (n = 7), self-reported migrant or ethnic group, and nationality of the child.
Of 88 the birth cohort studies, 21 (23.9%) provided study materials in at least one language other than the national language(s), 14 of which translated their study materials to English, partly for purposes other than recruitment (e.g. international collaboration). Turkish study materials were provided in six studies. Other languages include Arabic, Russian and Urdu. One investigator from Portugal stated that there was no need to translate study materials into a foreign language as most migrants or ethnic minorities come from countries where Portuguese is spoken.
Information on migration history or ethnicity was (or is planned to be) used for analysis in 23 of the 88 studies (26.1%). The analyses range from adjusting for ethnic origin or migration history to conducting stratified analyses. Four birth cohorts (4.5%) performed comparative analyses trying to understand mechanisms leading to inequalities in health between migrant/ethnic minority populations and the majority population (ABCD,22 BiB,23 BaBi,24 Generation R25).
Over the study period, the number of new birth cohort studies being set up increased, and so did the proportion of studies with at least one indicator for migrant or ethnic minority status (1982–1991: eight new studies; thereof measuring migrant/ethnic minority status: four [50%]; 1992–2001: 30; 21 [70%]; 2002 and later: 50; 40 [80%]).
Discussion
Migrants and ethnic minorities often experience health inequalities relative to the (ethnic) majority populations. Birth cohort studies can contribute toward disentangling the underlying mechanisms, provided information on migrant status and ethnicity is collected when the cohort is set up.
Eighty-eight birth cohort studies were identified in 21 different European countries, with more than 486 250 children enrolled in total. Many birth cohort studies that started in the 1980s or 1990s focused on clinical conditions, thereby missing the opportunity to study social determinants of disease. Sixty-two studies (70.5%) reported collecting data on migration history or ethnic background of study participants (mainly place of birth). The proportion of migrants or ethnic minorities among study participants ranged from 0.0 to 60.6%; this estimate varied depending on the indicator used for defining migration history or ethnic background. To date, 23 cohort studies (26.1%) have conducted, or are planning to conduct, comparative analyses using the respective indicators. Some studies used the information about migrant background or ethnicity for adjustment purposes only. Several investigators reported that they could not conduct comparative analyses because the number of migrant or ethnic minority groups in their study was too small. Our findings show small improvements relative to a study by Ranganathan and Bhopal5 who investigated whether ethnic minority populations were included in (adult) cardiovascular disease cohort studies in Europe and North America. They found no major cohort study focusing on ethnic group variations in Europe. Moreover, three European studies, all reporting on data collected in 1995 or later, were still explicitly restricted to “Caucasian” or “white” populations.
Having only a small proportion of migrants or ethnic minorities in the general population should not serve as an excuse for not enrolling them in cohort studies. There are obvious ethical reasons (it is a matter of equity), but also scientific arguments: Including even small numbers of ethnically diverse populations in cohorts will increase generalizability of results.5,26 One strategy to compensate for small numbers of migrants and ethnic minority group members in the general population is to oversample them when setting up the cohort. An alternative, or additional, strategy would be to pool data from several birth cohorts across Europe. Our review shows, however, that while the majority of European birth cohort studies measure ethnicity or migration history, they do so using widely differing indicators and concepts (Ranganathan and Bhopal5 observed the same problem in cardiovascular cohorts). Thus, a prerequisite for a European cooperation of birth cohorts with data pooling is the harmonization of broad indicators on migrant history and/or ethnicity. This could be achieved by agreeing on common definitions as well as a minimum set of such indicators, as has been suggested for other variables.1 Indicators for migrant/ethnic minority status need to go beyond the frequently collected but least informative variable “nationality”. “Country of birth” (including that of parents) can define migrant status but will not suffice to identify ethnic minorities. For this purpose, self-assigned ethnicity is favoured, but it has the disadvantage that it may change with time and context.18
Producing harmonized data across Europe would moreover allow comparing the effects of different health policies and health systems on migrants and ethnic minorities (given its variance in policies and systems, Europe constitutes a “natural laboratory” in this respect).27 So far, the few comparative studies of this type involving migrant populations have mostly been cross-sectional, rather than longitudinal, as would be possible when using birth cohort data.28 Finally, pooled datasets would facilitate the drawing of causal inferences from birth cohort data, thus helping to identify modifiable intra-uterine and early life risk factors.29
Enrolling migrants in cohort studies poses additional challenges: they are often reluctant to participate, requiring extra efforts to avoid under-representation and bias.30,31 Also, migrants tend to be more mobile than the majority population, so loss to follow-up may be larger and differential. A vision for the future is to set up a close-knit network of European birth cohort studies that could follow participants even when they move across national borders.32
Strengths and limitations of this study
To our knowledge, this scoping review is the first to summarize available information on strategies to include migrants and ethnically diverse population groups in birth cohort studies in Europe. A strength was that European birth cohort studies could be identified with the help of dedicated databases, thus reducing the chance that a cohort study would be missed in the literature search. In addition to analysing databases and published material, we sent questionnaires to the studies’ main investigators. This allowed us to ask which indicators of migration history and ethnic background were used. In this way our search was not restricted by the use of only the “obvious” indicators such as maternal place of birth, nationality or ethnic group; thus, we expect that we were able to identify all relevant cohorts within our sample.
There may, however, be birth cohort studies that have neither registered at one of the three web-based inventories nor published an article fulfilling our search criteria, so there is a small chance that our sample of European birth cohorts may not be complete. This could in particular apply to studies whose follow-up ended in the 1980s or early 1990s, and to those which just started. However, after checking conference abstracts and speaking to experts in the field we expect that the number of missed cohorts is low. Some cohorts may have continued to recruit and changed their recruitment strategy regarding migrants and ethnic minorities after publishing the paper that we analysed for our review. This may have led to misclassification. However, as most cohorts had defined (and not very extended) recruitment periods, this again is unlikely to have produced a major bias.
Conclusion
The majority of recent European birth cohorts assess migration history and/or ethnic background; however, this information is seldom used for comparative analyses trying to disentangle reasons for health inequalities faced by migrants and ethnic minorities. We propose improvements in the use of available data as well as in data collection of migration history and ethnicity (including Europe-wide harmonization of definitions and indicators). This could help to gain further insights and could help to identify starting points for interventions to reduce such inequalities. Birth cohort studies should routinely include, and possibly oversample, migrant and ethnic minority groups. This is an ethical obligation given the increasingly diverse populations of Europe. It is also a scientific necessity, as it will help to better understand the production of health inequalities and thus ultimately benefit all population groups living in Europe.
Acknowledgements
We are grateful to the investigators of birth cohort studies for providing additional information.
Funding
This work was supported by the Short Term Scientific Mission within the COST scientific programme on Adapting European health systems to diversity (ADAPT; grant number: COST-STSM-IS1103-12635 to A.G.) and the Federal Ministry of Education and Research as part of the study on ‘The production of health disparities in Germany’s population: disentangling the effects of individual and contextual factors in infancy among the offspring of immigrants’ (BaBi-Study; Grant number: 01ER1202; PIs: J.S. & O.R.).
Supplementary data
Supplementary data are available at EURPUB online.
Conflicts of interest: None declared.
Birth cohort studies can be a powerful tool to investigate the production of health inequalities in Europe, provided potentially disadvantaged population groups such as migrants/ethnic minorities are identifiable.
The majority of European birth cohorts assess participants’ migration history or ethnicity, but only few use this information for comparative analyses.
More widespread use of available data on migrants/ethnic minorities and harmonization of data could help to gain further insights into the production of health inequalities, which would ultimately benefit all population groups living in Europe.
New birth cohort studies should routinely include and possibly oversample migrant and ethnic minority groups, and plan comparative analyses.
References
Author notes
†First authorship
Comments