Six decades of longitudinal health knowledge production: a systematic review on Nordic birth cohort studies

ABSTRACT This systematic review (a) identifies birth cohort studies (BCSs) established in the Nordic countries, (b) describes their basic characteristics, and (c) explores how these characteristics have evolved over time, discussing their implications to knowledge production. To identify Nordic BCSs, cohort databases and relevant scientific articles were systematically searched and screened. The review shows that since 1959, more than 600,000 index children have participated in the 79 Nordic BCSs (22 Danish, 20 Finnish, 12 Norwegian, 24 Swedish, one Icelandic), over half of them still ongoing. The Nordic BCSs cover a wide geographical area including the Nordic Arctic. The topics of BCSs have varied over time but most have focused on examining the developmental origins of diseases. A quarter of them had a general scope, while the rest started with a specific focus, commonly atopic diseases. All BCSs collected questionnaire and/or interview data and over 60% of the BCSs announced exclusion criteria for participants, typically insufficient language proficiency. NBCSs have produced crucial scientific knowledge for over six decades, but there are underutilised opportunities including systematic interdisciplinary collaboration, inclusion of children’s own views of their health and well-being, intergenerational data collection, and specific knowledge of Arctic indigenous peoples and other minorities.


Introduction
Longitudinal birth cohort studies (BCS) provide unique opportunities to produce scientific knowledge on the associations and causal relationships between early-life exposures with later health and wellbeing [1,2] extending all the way to old age [3].In a birth cohort study, longitudinal data is collected from a group of people born at a certain time in a defined geographical location by active measures such as clinical studies and surveys, and/or passive measures like hospital records and other registers, with follow-ups over a period of time [4].Recruitment for these studies happens during pregnancy, at birth [5], or later in life.A pregnancy cohort recruits pregnant women and follows them and their children prospectively whereas in a birth cohort study, recruitment happens at birth, and data collection concentrates on the children [6].The length of the study can vary from months to years or even decades, until the passing of the cohort member, the "index child" [7].In a third form of birth cohort study the "index child" is recruited later in life and is usually followed-up both retrospectively and prospectively [5].In this review, we will refer to all these three study types as birth cohort studies (see [5].
BCS as a research design, usually with long-term follow-up with multiple ways of data collection, provides insights into the development of different kinds of exposures for later development of disease, including genetic, socio-economic, environmental, and lifestyle factors [1,8].This way, BCSs enable the identification of key exposures and effective preventive methods of different types of diseases and sources of ill-being which benefits individuals and societies [9].
The UK is the pioneering country in establishing birth cohort studies [10], implementing the first large birth cohort study, The National Birth Cohort (MRC National Survey of Health and Development) in 1946, soon after World War II.Outside of the UK, the efforts of conducting birth cohort studies were sporadic and rare [11], but a dramatic rise in BCS was witnessed in the 1990s [4], especially in Northern Europe and in other high-income countries such as the US [12].This rise has happened in parallel with increased interest in the early-life developmental origins of diseases [1].From the turn of the millennium, genetic and epigenetic variables have been increasingly included in the studies (see [13], and many birth cohorts have created or have collaborated with biobanks to collect and store biological specimens to better investigate the interplay between genetic, lifestyle, and environmental factors in wellbeing and health [14,15].
The Nordic countries, Denmark, Finland, Norway, Sweden, and Iceland, have long traditions of conducting BCSs [5], some of them being among the oldest and largest in the world [1,3].Considering the exceptional prominence of BCSs in this geographical region and their scientific and health policy impact [16], it is timely to conduct a systematic review of them.In this review, we will present central characteristics of Nordic BCSs and examine how these characteristics have evolved over time and pay attention to the scientific knowledge production about and with BCS research participants.
Previous systematic reviews of birth cohort studies have focused on specific countries [15,17,18] or larger regions such as the Gulf Cooperation Council countries [19]; Sub-Saharan Africa [4] and Europe [1,5,6,20] or have had a specific focus, such as environmental exposures [21] or asthma and atopic diseases [22].Some of the Nordic cohorts are included in these previous review studies (e.g [5,22], but they are not systematically covered.Moreover, the Nordic countries include Arctic regions with indigenous populations of Sámi and Greenlandic Inuits, whose participation has not been discussed in previous BCS reviews. 1

Nordic welfare states as research setting
The Nordic countries are technologically developed, politically stable democracies with gross domestic products above the European average [23] and share a lot of common history and culture.All five implement a version of the Nordic welfare model, which is characterised by universalism, high level of taxation and welfare expenditures, high employment rate, active labour market policies, and a high degree of equality [24,25].These countries provide free education and universal and predominantly publicly financed health-care services.
The Nordic countries are particularly well suited for conducting birth cohort studies.As welfare societies, they share a strong interest in monitoring and improving the health and well-being of their population, and they have invested substantially in research and research infrastructures.The high legitimacy of and sense of trust in the state and its health institutions [26] have likely contributed to high participation rates in birth cohort studies in the Nordic countries (e.g [27].But even more importantly, the data-gathering processes are greatly enhanced by the long traditions of maintaining nationwide registers and establishing personal identification numbers for citizens and permanent residents.This makes it possible to link data from several registers and other sources with registerbased information at an individual level [28,29].In addition, standardised maternity care practices produce large amounts of widely comparable medical data on every pregnancy and infant, and facilitate the recruitment of mothers and babies in a birth cohort study (e.g [30].
While the Nordic welfare states perform well in terms of life expectancy and overall population health, they also experience significant health inequalities.Besides inequalities related to socioeconomic differences, especially in Sweden and Finland, there are health disparities between southern and northern parts of the country and between urban and scattered settlement areas [31].Northern and Arctic areas of Nordic countries face specific conditions and challenges affecting population health and wellbeing, such as scarcely populated areas, long distances, strong female outmigration and decreasing population rates, higher unemployment [32,33], cold climate, high number of substance abuse and interpersonal violence, as well as suicides [34,35].
In the Nordic Arctic, two indigenous peoples are residing: Greenlandic Inuits in semi-autonomous area of Greenland, and Sámi people.According to two global reviews, indigenous people´s health and wellbeing are poorer in comparison to majority populations living in the same countries [35,36].Especially in Greenland, its Inuit population has more health problems and lower life expectancy than the rest of the Nordic region [37].With the Sámi in the Nordic countries, 2 only minor health differences compared to non-Sámi populations 1 Weihe et al. [73] provided an overview of ongoing cohort and dietary studies in the Arctic, but while describing some birth cohorts, some birth cohorts located in the Arctic areas in the Nordic countries were missing.

2
The traditional settlement area of Sámi people, Sápmi, stretches over the northern parts of Norway, Sweden, and Finland, and a small portion covers the Kola Peninsula in Russia, but a large part of the Sámi lives outside of this area [79].
have been found [38].Yet, there are special challenges and threats to Sámi people and their health, connected to, for instance, the cumulative effects of resource exploitation, climate change, and conflicts with the majority society [38].

Aim and objectives
Given the unique context of the Nordic countries, their long tradition of establishing birth cohorts, and the fact that they host some of the world's largest birth cohorts, it is surprising that to date, no systematic reviews of Nordic birth cohort studies have been conducted.Reviews of birth cohort studies are important as they allow us to recognise research trends and knowledge gaps and help in the identification of the strengths and weaknesses of the studies [20,39].Moreover, they offer information for future crosscohort collaboration, which is desirable as combined analysis of data may bring efficacy to research efforts, improve statistical power and reduce publication bias as well as improve contrast and diversity in exposure and outcome measures [1,20,21].
The aim of this systematic review is to identify and review birth cohort studies conducted within the five Nordic countries, Finland, Sweden, Norway, Iceland, and Denmark, including the autonomous areas of Faroese Islands and Greenland (Denmark) and Åland Islands (Finland).The objectives are to (a) identify existing birth cohort studies, (b) describe their basic characteristics including geographical locations, study focus, types of data collected, and exclusion criteria for study participation, and to (c) explore how these characteristics have evolved over time.These issues are discussed in terms of their potential implications for the scientific knowledge produced with the birth cohort data, especially considering research participation; the review takes into consideration the regional coverage, eligibility criteria for the participants and the respondents to the questionnaires and interviews.As such, we examine the participation in the knowledge production process in BCS, an aspect often neglected in birth cohort reviews.
Even though some of the Nordic birth cohort studies are included in earlier reviews, a lot of information about them is missing.We seek to identify Nordic pregnancy and birth cohort studies that have not been included in previous reviews and to reach a larger variety and representativeness, we also include BCSs where recruitment happened later in life.Moreover, we do not restrict the initiation date of BCSs, allowing the inclusion of the older birth cohorts.

Materials and methods
To ensure transparent reporting, the PRISMA guidelines (Preferred Reporting Items for Systematic reviews and Meta-Analyses for Protocols) for systematic reviews were followed [40] Detailed PRISMA checklists are found in Appendix 4 and Appendix 5.

Data sources and study selection
To identify relevant original research articles, EBSCO, Scopus, and ProQuest citation databases were systematically searched 26 th -27 th May 2020 using a strategy where the search term "birth cohort" was combined with terms describing the Nordic countries (Finland, Finnish, Sweden, Swedish, Norway, Norwegian, Denmark, Danish, Iceland, Icelandic).No language or publication status restrictions were imposed at this stage, but the search was directed to titles, abstracts, and keywords and restricted to peer-reviewed research articles.The search was updated with a similar search strategy on the 16 th of May 2022.
To be included in the review, the articles needed to meet the following criteria: they had to a) be peerreviewed and published in English, Finnish, Swedish, Norwegian, Icelandic or Danish; b) report empirical research on a birth cohort study with a study population consisting of a group of people born in specified time and place; including c) two or more types of data, for example, surveys/interviews, clinical data, biological specimen collection, or data from psychological tests; d) more than two follow-ups over a one year period, and e) a sample of at least 200 participants.
An article was excluded if it a) was not an original empirical research article, b) did not report a birth cohort study (was e.g. a case-control study, randomised clinical trial), reported a BCS including c) only passive types of data (registers, hospital or other types of records, national statistical data), d) only one type of data (e.g.only survey data), or e) participants outside the Nordic countries or more than one country as, for the purposes of this review, we wanted to centre on individual Nordic cohorts only.
The search resulted in altogether 10,309 literature references (8,901 from the original and 1,408 from the update search), and after the removal of duplicates 4,226 unique articles were identified (3,562 from the original, and 664 from the update search, see Figure 1).The publications were used to detect birth cohort studies and study programmes meeting the predetermined inclusion criteria.In other words, the aim was to identify individual birth cohort studies from the published empirical articles, rather than examining the publications as such.
Information based on the abstracts (n = 4,226) of studies included in the review were inspected independently by the authors (Authors 1 and 3 after the first search, and Authors 1, 2, and 3 after the update search).If the abstract did not yield sufficient information on the birth cohort study or the reviewer was unsure about its inclusion, the full article was retrieved and read.Based on abstract screening, 1,172 articles were excluded and 2,237 included.Moreover, 817 full papers were reviewed and 292 of them were excluded.In total, 2,762 articles were included for further analysis.Based on this screening, 74 BCSs were identified.
Moreover, web-based databases [41,42] which have previously mapped birth cohorts globally were explored, following the search protocol of Larsen et al. [1].In addition, potentially relevant birth cohort studies were identified from other sources through citation chaining, that is, seeking relevant birth cohorts from the retrieved publications.This resulted in the identification of five additional BCSs.
Descriptive information of the total of 79 BCSs was gathered from individual articles, cohort webpages, and cohort profile papers, commonly published in International Journal of Epidemiology, and cross-checked by all authors.In all phases of study selection, a consensus was reached in discussions among all authors.

Data extraction
Once a birth cohort study had been included, information was extracted from the articles and other sources using a data extraction sheet outlined for the purposes of this review.Guided by the objectives of the review, the following information was sought from each birth cohort study (see Appendix 1): Country; BCS name; Timing; Region of the BCS; Number of index children; Initial topic; Whether the topic of the cohort study was general or specific; Age of participants at recruitment (pregnancy, at birth, later in life/years); Age at follow-ups; Types of data; Respondents on questionnaires (including surveys and diaries) or interviews; and Exclusion criteria or specific inclusion criteria.

Results
In total, 79 birth cohort studies met our inclusion criteria.Of the cohorts, 20 were Finnish (F1-F20 4 ), 12 Norwegian (N1-N12), 24 Swedish (S1-S24), 22 Danish (D1-D22), of which one from Faroese Island and two from Greenland, and one Icelandic (I1).Eight of the cohorts were geographically located in the Arctic areas.In addition, there were altogether seven national cohorts which included also Arctic areas.All the included BCSs are listed in Appendix 1, which depicts their central characteristics.

Study designs and topics
The majority of the birth cohort studies included in the review were prospective mother/parent-child BCSs.In them, a parent, commonly the mother, and more rarely mother together with the father/partner, was recruited during pregnancy (n = 30) or soon after the birth of their child (n = 26).The data collection started during pregnancy with the mother or upon the birth of the index child with both mother and the child and continued in follow-ups.In 29% (n = 23) of the birth cohort studies recruitment happened years or decades after the birth of the index child.Most of these studies looked also back in time and included retrospective data collection.
Nordic birth cohort studies have predominantly had a medical focus.According to their initial aims, a third (n = 16) of the reviewed mother/parent-child BCSs had a general developmental origins of health and disease, "DOHaD", (see [13; 15]) approach, covering a broad range of prenatal and/or early life exposures on later child development, health, and wellbeing, while the rest (n = 40) centred on more specific exposures and/ or outcomes (see Table 1 and Appendix 2).Outcomespecific studies focused on the risk factors for and the development of specific diseases, 13 of them on atopic diseases.Exposure-specific studies concentrated on the impact of maternal medical conditions, lifestyles, and pollutants, for example, on the later development of Each cohort was assigned a code with a letter referring to the country and a running number in ascending order.
INTERNATIONAL JOURNAL OF CIRCUMPOLAR HEALTH the index child.Moreover, six BCSs specified on mental health including psychiatric problems of the index child or the mother.Of the BCSs commenced later in life, four had general focus on the ageing process or health at an older age while 19 had more specific focus.Of the specific BCSs, six focused on cardiovascular diseases and eight on mental health or behavioural/social matters including the effects of inequalities and risk factors for later mental and behavioural problems.
It should be noted that many of the specific birth cohort studies extended their initial focus to cover other subjects.For example, The GLAKU birth cohort (F11) was initially designed to study maternal intake of glycyrrhizin in liquorice, as glycyrrhizin is an inhibitor of cortisol metabolism and might be related in the aetiology of low birth weight [43] but, later, the collected data has been utilised for studying a variety of topics including early life origins of mental health [44].
The typical study designs, characteristics, and topics of the BCSs varied to some extent both between the countries and over time.The recruitment of older children or adults as index children was common in the first two decades of Nordic birth cohort studies.However, the very first Nordic birth cohort study, the Danish Copenhagen Perinatal Birth Cohort (D1), initiated in 1959, was a prospective study with general DOHaD aims and recruited mothers during pregnancy (see Appendix 2).The 1980s seems to have started the golden age of birth cohort studies in Denmark (six new studies) and Finland (five new studies).The 1990s was a prolific decade in the other Nordic countries as well (Figure 2), as 18 birth cohort studies were established, comprising over 290,000 new index children.In the first decade of the 2000s, as many as 19 new birth cohorts were established but with approximately only 55,000 new index children.During these decades, the topics and study designs diversified.However, the 1990s and the early 2000s were especially marked by the number of new cohort studies on atopic diseases -altogether over 10 new studies were launched (e.g.D14, D15, D19, S10, S12, S13, S15, S18, N4).During the same decades, mental health and social issues (S11, D17, D18, N9, N10, S21) and the general health of elderly people (D10, F10, S16) were popular topics.Moreover, studies established in that era addressed the effects of diet on specific diseases (F9) and the effects of specific foods on general child development (F11) as well as gut microbiota development (N8).The effects of preterm birth (F13, F15) and pollutants (N7, S4) were also studied.Some studies focused on the risk factors of specific conditions such as diabetes and other immune-mediated diseases (N6, S14, S15), and cardiovascular diseases (S9).
In the 2010s, only three large scale general prospective birth cohort studies were initiated (D20, F18, S24).Otherwise, studies focused on specific topics including sleep (F16) and postpartum mental health (N10), the effects of stress on harmful behavioural outcomes among adolescents (I1), brain development (F17), epigenetic alterations in children conceived with assisted reproductive technology (S22), the effects on maternal H1N1 infection (N11), and development of allergy/ asthma (S23).
Finland and Denmark have had many longitudinal birth cohort programmes with general aims and multiple follow-ups.Sweden and Norway, in contrast, have had fewer programmes with general aims.The high number of allergy and/or asthma cohorts in Sweden is particularly notable.Norway differs from Sweden, Finland, and Denmark in having fewer birth cohorts and a proportional emphasis on examining the effects of pollutants (N4, N7).The only Icelandic birth cohort (I1), a study with 15/16year-olds focusing on the influence of stress on harmful behaviour, was established as late as 2016.In Finland and Sweden, the latest birth cohort studies were established in 2016 (F20, S24), in Denmark in 2017 (D21), and in Norway in 2011 (N12).In recent years, no new birth cohort studies have been initiated in the Nordic countries.

Sample sizes and geographical distribution of the cohorts
The sample sizes between the BCSs varied considerably, ranging from 205 (F5) to 114,000 (N5).Sweden has established the largest number of birth cohort studies (n = 24) among the Nordic countries, but it has enrolled fewer participants (approximately 100,000) in its cohorts than Denmark and Norway (see Figure 3).In Denmark alone, over 260,000 index children have been recruited, and on top of them, mothers and fathers.Three Nordic cohorts (Aarhus Birth Cohort, Denmark; Danish National Birth Cohort DNBC, Denmark; and Norwegian Mother, Father and Child Cohort Study MoBa, Norway) have had 100,000 or more participants each.Birth cohort studies are established in specified geographical locations, ranging from hospital districts and cities to counties, larger regions, and whole countries.Only seven of the reviewed studies were nationwide, among them the DNBC and MoBa.MoBa is one of the most extensive Nordic BCS with 12 follow-ups since its initiation.
Since the acreage of the Nordic countries is relatively large, and the average population density very low (ranging from 0,24 inhabitants/km2 in Greenland and 3,33 inhabitants/km2 in Iceland to 137 inhabitants/km2 in Denmark), it is not surprising that different regions have not been equally represented in the Nordic birth cohort studies.In Finland, Norway, and Sweden, there has clearly been an overrepresentation of BCSs in the southern urban areas of the country, especially around university hospitals.On the other hand, a large part of the total population lives in these geographically limited but more densely populated areas; for instance, over 1/3 of the Danish population live in the Copenhagen metropolitan area, which is the best represented region in the Nordic BCSs.The populations of the large metropolitan areas are usually more varied in terms of socio-economic conditions and ethnicity than rural and small-town populations.While there are some areas in Finland (such as South-Eastern Finland) and Sweden (in Northern Sweden) that are not covered by any cohorts, the studies represent both urban and rural, and wealthy and disadvantaged areas.The same applies to the other Nordic countries, including Iceland, whose only study is nationwide.Especially in Sweden, the geographical scope of the birth cohort studies has clearly become more varied after the first decades.
Eight BCSs were from Arctic regions, and their topics reflect the specificities of these areas (Figure 4).The first Finnish birth cohort study (F1) focused on the disadvantaged Northern and Arctic areas, and specifically addressed the regional health inequalities in the country, initially in relation to perinatal health and mortality.Two other Finnish birth cohorts (F7, F15) are located in the same area.The only Arctic birth cohort of Sweden (S23) took place in Luleå, focusing on environmental exposures.Among Danish cohorts, D7 is located in Faroe Islands and focuses on Arctic livelihoods, particularly on seafood diet and marine contaminants.Two Greenlandic BCs focus on Greenlandic Inuit health.Indigenous people also participate in birth cohort studies outside their traditional settlement area.For instance, there are large populations of Sámi people in the Southern and especially in the capital areas of Norway, Sweden, and Finland.

Eligibility criteria for participants
The general inclusion criterion for a birth cohort participation is being born or being a pregnant mother at a certain time in a certain geographical area.However, not all the children all mothers meeting these criteria participate in the studies either because they or their parents decline or because they are not eligible.Of the reviewed studies, 61% announced one or more specific exclusion criteria (see Table 1), insufficient language skills (n = 26, a third of all studies) being the most common criterion.In these studies, a parent, usually the mother, or parents needed to be able to understand the official language of the country well enough to answer questionnaires or participate in interviews.Another common criterion was the health of the mother, the baby, or both.Preterm babies, and babies with severe malfunctions or other severe diseases were often excluded from the study.On the other hand, we also identified 12 at-risk cohorts with a special focus on children with genetic or other risk factors, such as preterm birth (F15).Healthy babies without the specified risk factor(s) were excluded from these studies.
Five studies excluded potential participants because of their ethnic background.The Copenhagen mother-child cohort (D11), established in 1996, required that both the parents and grandparents of the unborn child were born and raised in Denmark; the Ivaaq cohort (D16) and ACCEPT cohort (D21) examined only ethnic Greenlanders; the GLAKU BCS (F11) excluded participants of non-Finnish origin; and the MIDIA BCS (N6) excluded participants of non-Caucasian origin, because the examined risk gene for diabetes is lacking in non-Caucasians.
All the cohorts examining childhood diseases and conditions included children to the study regardless of gender whereas some of the retrospective cohort studies recruiting adults were single-gender cohorts, including only women (S4) or only men (D3, S1, S6, S9).One cohort (S21) studying perinatal depression used an age criterion and excluded mothers under the age of 18.Some studies have excluded twin pregnancies or subsequent babies from the same mother (e.g.D8, D16, D22, F11, F20, S3, S10).

Types of data collected
In our review, we included studies which have collected two or more types of data within a study programme.Questionnaires, surveys, or interviews were included in all studies.Questionnaires were included in 76, and interviews in 32 studies.Biological data such as blood samples were the second most common data types across the birth cohort studies (included in 73% of the studies).Register data was also commonly utilised (in 65% of the studies).Clinical data, which often includes multiple different tests, such as rest-ECG, skin allergy tests, spirometry, physical activity tests, eye examinations, dental examinations, or body composition tests (see for instance [45,46] were collected in 51 (65%) of the cohort studies.Psychological, psychiatric, neurological, or cognitive evaluations were less common but were included altogether in one third of the studies (n = 28, 35%).It should be noted that many of the study programmes also initiated sub-studies with smaller sub-samples of research participants, which we have not systematically reviewed.In these sub-studies, more varied data collection methods were commonly utilised.
Fifteen of the reviewed BCSs collected exceptionally versatile data, representing five out of the six data type categories of this review (types 1-5, see Appendix 2).These studies were mostly medium-sized general mother-child birth cohorts with multiple follow-ups over the years.However, environmental data (type 6), such as house dust, was not included in these studies, and was rarely collected in the birth cohort studies in general.

Participants and knowledge producers
The active participants of the birth cohort studies were commonly the mother and her child, i.e. the index child.Fathers also participated in many studies, but not to the same extent as mothers.In all the traditional BCSs where the participants were recruited either during pregnancy or right after birth, mothers responded to questionnaires/ interviews, and fathers/partners were given the possibility to also respond in little over half of them (66%, n = 37).In these types of studies, the index child rarely responded to the questionnaires/interviews themselves (33%, n = 19), and parent(s) responded about their own and their child's health.Naturally, this has to do with the age of the index children; they usually started to respond to the surveys only as teenagers, and many Nordic birth cohorts are not that mature yet.In those BCSs where recruitment happened years or decades after the birth, "index children" responded to questionnaires in nearly all studies (96%, n = 22).Recently, grandparents have been invited to participate in birth cohort studies that have been established decades earlier (for example in S13 initiated in 1996, F1 initiated in 1965), although their participation is still rare in Nordic BCSs.Healthcare professionals were respondents in one fourth of all BCSs (n = 25, 32%), and school teachers responded in 9 studies.No clear temporal trends regarding participant groups were observed.That is, for instance, fathers were as often included in the studies established in the 1960s and in 2000s.

Discussion
This review is the first to summarise and describe the basic characteristics of birth cohort studies established in Nordic countries.Over the course of several decades, from the 1950s to the 2020s, 79 Nordic birth cohort studies have produced important knowledge especially on the effects of early life exposures on health later in life, but also on other topics such as the ageing process, mental health, and behavioural and social issues.Even though no new birth cohort studies have been established in recent years, more than half of the reviewed birth cohort studies are still ongoing.Our review was able to detect a considerably greater number of Nordic mother/parent-child birth cohorts than previous reviews on European BCSs [5,6], and also included BCSs where recruitment happened later in life.

Hundreds of thousands of Nordic 'index children'
Over 600,000 Nordic inhabitants have participated in the 79 birth cohort studies as index children, and, on top of that, several hundred thousand parents have also done so.
Recently, grandparents of the index children have been included in some of the Nordic birth cohort studies as well.The participant numbers are high with respect to the total population of the five Nordic countries, which is approximately 27 million.For instance, in Australia and New Zealand, an area with a comparable population, Townsend et al. [15], identified 23 birth cohort studies with less than 40,000 index children.
The general societal conditions and practices in the Nordic countries provide explanations for the prominence of BCSs in this area.As wealthy welfare societies they have had sufficient funds to facilitate large scale longitudinal study programmes [47][48][49].They also have a long tradition of monitoring the health of pregnant women and children as well as giving birth in hospital settings.These facilitate contacting prospective participants and gathering data.Moreover, tracking down cohort members to follow-ups is enabled by existing national registers, citizen databases, and personal identification numbers [28].

Regional and ethnic representativeness
Despite the large surface area of the Nordic countries and some overrepresentation of large cities, different geographical areas are quite well represented in the Nordic birth cohort studies, at least in proportion to their population.Only seven nationwide birth cohort studies (none in Finland or Sweden) have been conducted, but wide geographical areas are still covered by many regional birth cohorts as well.These areas include both wealthy and disadvantaged and urban and rural areas.
There have also been birth cohort studies in areas that are inhabited with substantial minority populations, such as the traditional settlement areas of the Sámi people.As there might be some specific health and wellbeing issues among indigenous peoples, it is valuable that BC data is produced among them as well.However, since the Nordic countries do not collect data on ethnicity in large-scale or population based surveys [38,50], the health problems of the minorities do not become visible even when the minorities are included in these studies.Sometimes, this problem is partially circumvented by using measures such as the place of birth, mother tongue, or nationality in data sweeps [49].Several Nordic birth cohorts include health information on the indigenous peoples, and this information could benefit health research and promotion.
Even though only a few Nordic birth cohort studies announced ethnic exclusion criteria, the more common language criterion has reduced the possibilities of other ethnic minorities, especially newly arrived immigrants, to participate in birth cohort studies.While the language criterion is understandable for practical reasons, it easily leaves disadvantaged parts of the population out of the study scope.Birth cohort studies could offer an opportunity to disentangle the underlying mechanisms of health inequalities regarding ethnic, migration or minority background, if this information was collected.Grosser and colleagues [49] have reviewed European birth cohort studies from this perspective and criticised them for wasting the opportunity to address health inequalities.Moreover, based on earlier research, ethnic minorities, people in disadvantaged socio-economic positions and those at greater risk of developing health problems tend to participate in this type of studies less often and are more likely to drop out [51].This type of attrition can cause systematic biasin the data collection and limit the generalisability of findings [4,52].
Oversampling of ethnic minority groups has been suggested as a way to address the mentioned systematic biases [49].Even though this is not simple because of the lack of data on ethnicity, some Nordic birth cohort studies have addressed these questions.For example, the Danish Longitudinal Survey of Children Born in 1995 (DALSC) and the Odense Child Cohort have made special efforts to recruit participants with non-western backgrounds.Since the Nordic countries face relatively large health inequalities despite their welfare model [37,53], both socio-economic and ethnic variables should preferably be addressed.However, as this is a sensitive and complex matter, it requires taking various and conflicting ethical aspects into consideration.

From preterm birth to allergy studies
The findings of this review evidence a strong medical focus and, more specifically, a "Developmental Origins of Health and Disease" approach (DOHaD, see [13; 15] in the Nordic birth cohort studies.Approximately a third of the mother/parent-child cohorts, and a fifth of the studies which recruited index children later in life had a general focus covering a range of exposures to child development, health, and well-being while the remaining studies addressed more specific topics, often still within the DOHaD frame.A similar trend between specific and general cohort types has been detected in previous birth cohort reviews [1,15].Many Nordic studies started with a specific focus, but extended it over the course of the study, a phenomenon also common in birth cohort studies elsewhere [15].
The cohort studies conducted in different geographical areas reflect, to some extent, their specific disease burden and conditions.For instance, the birth cohort studies conducted in South-East Asia and Eastern Mediterranean regions have focused on growth, mortality, nutrition, and infectious diseases, and less often on non-communicable diseases [54], whereas European birth cohort studies have predominantly concentrated on nutrition-related, neuropsychological, and disease-specific outcomes, and exposures related to lifestyles or the physical and social environments [49].Based on the findings of this review, Nordic birth cohorts with a specific focus have studied non-communicable diseases that are common in the region.Particularly common were birth cohort studies focusing on asthma, allergy, and other atopic diseases, which are among the most common chronic diseases among children in many industrialised countries (see [55].The incidence of different atopic diseases has increased in the last decades, and longitudinal epidemiological birth cohort study is an optimal research model for detecting cause-effect relationships in their development [22].In addition, data on allergies are not fully covered by available registers since they are health conditions that often do not result in hospital contact [56], which makes the collection of data in BCS especially valuable.Three of the Nordic BCSs also focus on type 1 diabetes, which has exceptionally high prevalence in Norway, Finland, and Sweden [57].
The foci of the study programmes have changed over time, which likely reflects the rise of the standard of living and changes in disease prevalence but also the cumulation of medical knowledge on certain topics and risk factors.Based on this review, many of the earliest Nordic birth cohort studies were centred on conditions with severe or even lethal consequences, such as risk factors of preterm birth and cardiovascular diseases.This review shows, similarly to other European birth cohort studies [22], that in the 1990s, atopic diseases became more prevalent topics.
When a study focuses purely on medical/clinical conditions, it may miss out the opportunity to study social determinants of diseases (see [49]) and wellbeing in general.Since BCSs with a medical scope commonly include questionnaires addressing a wide variety of issues, such as social situation, wealth, and occupation, there is clearly a lot of potential to use birth cohort data in other disciplines, too.Somewhat surprisingly, none of the Nordic birth cohorts had an educational focus, whereas for instance in the UK [58], long-term birth cohorts such as the Next Step study [59] specifically include variables exploring the educational realm.Education and schooling are instrumental for personal, social, political, and cultural development [60], and birth cohorts with a focus on educational settings, learning, school performance, educational transitions, and future adult health and wellbeing would be important to conduct in Nordic contexts as well.

Respondents and knowledge production
The birth cohort studies have been immensely valuable, as they focus on the underrepresented groups in medical research -mothers, infants, and young children [1,61].Medical treatment is largely based on knowledge produced on research on men, which is a pitfall and an obstacle also for gender equity in healthcare [62].On the other hand, men as fathers, in relation to their children, has not been a popular topic of medical research [63], and while birth cohort studies include fathers as participants, they have partially missed their opportunity to address the issue.According to our review, over half of the Nordic birth cohort studies collected data from fathers while all mothers were respondents, a finding similar to a previous Australian and New Zealand review [15].In addition, fathers are not as active as mothers in responding to surveys/questionnaires, even when given a chance (f.ex [64].The absence of fathers in birth cohort studies was discussed already in the 1990s [18], but no substantial progress has been seen regarding their involvement.The traditional family norm of two-parent families with mother and father has been a given in the older birth cohort studies.In future birth cohort studies, more inclusivity when addressing participants is called for.
Even though index children are the central participants in birth cohorts, they have seldom been active respondents in surveys or interviews regarding their own health and wellbeing for the first 10-15 years of the study.Clinical and biological data have been collected with young children, but self-reports have not.The tendency of adults to ignore children's views exists in many research fields and study types, and as seen, also in birth cohort studies.Often scientists explore children's experiences through the accounts provided by adults -parents, healthcare professionals, or teachers -rather than listen to children's own accounts of their life worlds [65][66][67].When it comes to medical topics, parents are likely to be able to provide information on their child's health issues when responding to surveys, but adult 'proxies' are not necessarily able to give valid accounts on children's social or emotional lives, which are often central to health and wellbeing (see [67].This is a shortcoming also from the perspective of children's rights and research ethics [68].

Limitations
This review aimed to chart and review Nordic birth cohort studies from their first initiation until the spring of 2022.Our review was limited to information found in abstracts, published articles, cohort profile articles, official cohort webpages as well as cohort databases.It is possible that we have missed some birth cohort studies that would fulfil our inclusion criteria.We did not get in touch with cohort leads, like many previous birth cohort study reviews have done (e.g [1].Thus, the information we have collected from the birth cohort studies is somewhat limited, especially regarding most recent developments and the sub-study data.However, as our aim was to offer a basic description of the main data collection phases of the studies and how the cohort characteristics have evolved over time, the information found in the aforementioned sources can be considered sufficient.We excluded birth cohort studies which cover areas larger than a nation state.For instance, the Finnish study LUKAS, which is part of a European birth cohort PASTURE [69], was excluded, similarly to INUENDO, which include Greenlandic, Ukrainian, and Polish participants [70].Moreover, setting the limit to a minimum of 200 participants have left out some interesting and unique birth cohorts, such as the Arctic Sámi Birth Cohort [71] and Icelandic Citizens born in 1987 [72].

Conclusions
Nordic countries have hosted 79 longitudinal birth cohort studies with more than 600,000 index children.These studies have provided valuable scientific contributions in many fields, especially in epidemiology and medical sciences.Our review provided an extensive overall description and insights on the nature of the Nordic birth cohort studies, and the development of this research tradition over time.This type of information is useful when new birth cohorts are planned and established in Europe, and when cross-cohort collaboration is initiated.Indeed, many of the ongoing and some terminated studies offer possibilities for secondary analysis of data.This massive data pool could also be used more extensively in non-medical disciplines, such as social sciences, gender studies, education, and humanities, which in turn would widen the scope of knowledge production on birth cohort data.
In recent years, the Nordic countries have not been active in launching new birth cohort studies.Many BCSs have shown agility in adjusting and responding to the changing world; for instance, some BCSs have included COVID-19 related data collection to their data sweeps.Yet, the launch of new, generally aimed birth cohort studies would help understand the challenges babies born in the 2020s will be facing as they live in a very different environment than the older generations.In addition to commonly collected data from mothers and index children, we would encourage to include more versatile data including data from fathers or other primary caregivers and grandparents and questionnaire/interview data from young children.Furthermore, Nordic birth cohorts offer possibilities for indigenous and minority research, even though the laws regarding asking ethnicity in surveys in the Nordics limits these possibilities.These research directions could be utilised more in the existing Nordic birth cohort data, work that should be done in close collaboration with indigenous and other minority communities.Very few circumpolar birth cohorts exist outside Nordic countries [73].For instance, an inventory of Canadian birth cohorts reviewed 46 cohorts and only two of them (4.7%) were Arctic [17], the Nunavik Child Development Study (NCDS) being one of them [74].Cross-cohort comparisons between Arctic birth cohort studies and embarking new Arctic studies could greatly benefit knowledge production on health and wellbeing of the Arctic populations.
We recognise the challenges that the large birth cohort studies have faced in the first decades of the 2000s, such as the descendent willingness to participate in such studies and the stricter regulation of the data gathering and sharing processes, as well as related expenses.In the future, it would be worthwhile to study trends in science politics and financing to provide insights in how they direct the planning and execution of BCSs [75][76][77][78].

Appendices Appendix 1. Search strategy
To identify relevant original research articles, we conducted searches from EBSCO, ProQuest, and Scopus databases.EBSCO and ProQuest include several databases which are listed in Table A1.The following search strategies were used without imposing language or publication status restrictions.However, the search was restricted to peer-reviewed research articles in EBSCO and ProQuest.We have listed main references (e.g.article, cohort webpage) of each cohort to a Appendix 3. 2 Column "Respondents on questionnaires/interviews" refers to participants who responded to the questionnaires/surveys in order to illustrate who the active respondents were in contrast to, for instance, routinely collected anthropometric measures, such as height and weight.
Uppstart participants were adult couples.

Figure 1 .
Figure 1.Flow chart of the phases of the literature search and selection 3

Figure 3 .
Figure 3.The number of birth cohort studies by country (in green rectangles) and approximate total number of index children recruited in them.

Figure 4 . 6 6
Figure 4. Map of the Arctic administrative areas and the Nordic birth cohorts located in the Arctic area (marked as black).6

•
EBSCO (all databases) in 26th of May 2020 and update search in 16th of May 2022; abstract search [covers titles & keywords]: "birth cohort*" AND (Finland OR Finnish OR Sweden OR Swedish OR Norway OR Norwegian OR Denmark OR Danish OR Iceland OR Icelandic) • ProQuest (all databases) 27th of May 2020 and update search 16th of May 2022; abstract search [covers titles]: "birth cohort*" AND (Finland OR Finnish OR Sweden OR Swedish OR Norway OR Norwegian OR Denmark OR Danish OR Iceland OR Icelandic) • SCOPUS 27th of May 2020 and update search in 16th of May 2022; title, abstract and keyword search: "birth cohort*" AND (Finland OR Finnish OR Sweden OR Swedish OR Norway OR Norwegian OR Denmark OR Danish OR Iceland OR IcelandicThe review is not yet registered.The review protocol can be accessed from the authors.

Table 1 .
Characteristics of 79 Nordic birth cohort studies divided by age of enrolment.

Table A1 .
Databases included in EBSCO and ProQuest.