Cohort Profile: 1970 British Birth Cohort (BCS70)

How did the study come about?

The 1970 British Birth Cohort Study (BCS70) is an ongoing, multidisciplinary, longitudinal study. It takes as its subjects all those currently living in England, Scotland, and Wales who were born in a single week of 1970. To date there have been seven sweeps of the study, including the original birth survey.

BCS70 began as the British Births Survey, when data was collected about the births and social circumstances of over 17 000 babies born in England, Scotland, Wales, and Northern Ireland. Data were collected using a questionnaire completed by the midwife who had been present at the birth and, in addition, information was extracted from clinical records. The original study was sponsored by the National Birthday Trust Fund in collaboration with the Royal College of Obstetricians and Gynaecologists, under the directorship of Roma and Geoffrey Chamberlain. The study aimed to examine the social and biological characteristics of the mother in relation to neonatal morbidity, and to compare the results with those of the 1958 National Child Development Study.

When the cohort children were 3.5 yr the study transferred to the Department of Child Health at the University of Bristol where Neville Butler, Professor of Child Health, took over responsibility for the follow-ups at 5 and 10 yr, what came to be known as the Child Health and Education Study. In 1975 and 1980, parents of the children in the study were interviewed by health visitors, and information was gathered from the child's class teacher and head teacher, from the school health service, and from the children themselves. Cohort members who were born in Northern Ireland were included in the birth survey but dropped from the study in all subsequent sweeps. The main findings on health from the 5 and 10 yr surveys are published in a report by the original team.¹

When Neville Butler left the University of Bristol in 1985, he set up the International Centre for Child Studies (ICCS) to continue work on the 16 yr survey of the cohort members. This sweep was known as ‘Youthscan’ and comprised 16 separate survey instruments, including parental questionnaires, class teacher and head teacher questionnaires, and medical examinations. In addition to completing educational assessments, the cohort members themselves answered questionnaires on a wide range of different topics and were asked to keep two 4 day diaries, one on nutrition and one on general activity. It was originally planned to trace cohort members in time to interview them at 15.5 yr, well before the minimum school leaving age. Unfortunately, industrial action by the teachers, who were responsible for the educational tests, meant that the survey had to be delayed, and also resulted in incomplete data collection from schools.

In 1991 responsibility for the study was taken over by the Social Statistics Research Unit (SSRU), based at City University London. At 21 yr a 10% sample survey of the cohort was carried out to investigate the forces and circumstances which influence young people in their transition from full-time education to employment; and to examine the extent of literacy and numeracy problems in the cohort. In addition, this survey was designed to serve as a feasibility study for a full sweep of the cohort in their early 20s. This enabled the study to survive with no guaranteed further funding. Under the directorship of Prof. John Bynner, a further postal follow-up of cohort members was carried out at age 26 yr (between April and September 1996) by the SSRU. The survey was administered by MORI. In 1998, the SSRU moved to the Institute of Education, London, and became the Centre for Longitudinal Studies (CLS). CLS also houses the 1958 birth cohort (known as the National Child Development Study, NCDS), and in 1999/2000 a simultaneous survey of both cohorts was undertaken to facilitate comparisons between these two groups born 12 yr apart. This study was key both in restoring the BCS70 sample to over 11 000 and in establishing the scientific content of the adult surveys, ensuring that all major life domains are covered. The most recent follow-up of both BCS70 and NCDS was carried out in 2004/5. As in 1999/2000, this was managed by CLS, and fieldwork was carried out by the National Centre for Social Research. Information about sources of funding through the life of the study are provided in Table 1. The ESRC are committed to funding the next phase of follow-up. This is planned to be a 30 min telephone interview in 2008 when cohort members are aged 38.

What does it cover?

The 1970 study is a genuinely multi-purpose study which has collected detailed information from cohort members on many different aspects of their family circumstances, health, education, and social development as they have moved from childhood, through adolescence, and into adult life. In the most recent sweep to date, with data collected in 2004 and 2005, information has also been collected on and from the children of a 50% sample of cohort members. The focus has been on the health of cohort members' children as well as their performance on a number of ability scales. This means that in future it will be possible to carry out research on intergenerational relationships in health and health inequalities.

Who is in the sample, how often have they been followed-up, and what is attrition like?

Current participants are survivors from an original sample of over 17 000 births, all born in England, Wales, and Scotland, during 1 week in 1970. Cohort members were followed-up by parental interview and examination at ages 5, 10, and 16 yr. At age 26 a postal questionnaire was sent to cohort members and they were subsequently followed-up by interview at ages 30 and 34 (Table 1). During childhood, cohort members were traced through schools and immigrants found through this process, and born in the reference week, were added to the target sample for the current and later follow-ups. This approach was no longer possible once cohort members left school, and the success of subsequent follow-ups has relied to a great extent on efforts to maintain contact with cohort members through the annual mailing of a birthday card and other feedback materials. These efforts are supplemented by work to trace lost cohort members with reference to study records and telephone and other databases. It can be seen that the samples reached at age 26 yr and subsequently are somewhat smaller than at age 16 yr and earlier. The main reasons for sample loss over time are individuals moving to a new address and not being subsequently traced. Refusal rates are relatively low but also contribute to sample loss over time. For example, in the face-to-ace adult surveys, refusal rates have been 7.3% at 30 yr; and 7.6% at 34 yr (based on provisional figures provided by survey contractors). Further details of attrition in the study are provided by Plewis et al. (2004) in a technical report on the CLS website.²

What has been measured?

The main health and medical data collected from birth to age 34 is listed in Table 2. Several topics have repeated measures, while others are only appropriate for a specific life stage. Whereas during childhood health data were mostly obtained from the school medical examination or by parental report, during adulthood information has mostly been obtained directly from the cohort member. At age 34 (and for a 10% sample at age 21) detailed assessments were made of the cohort members' numeracy and literacy skills. The cohort is also flagged for mortality. Table 3 summarizes the socio-demographic information collected in each sweep of the study. It can be seen that topics include socio-economic circumstances, family background, cognitive development, educational achievement, employment, partnership histories, and health-related behaviour. The strength of the study is its multidisciplinary nature resulting in the ability to link data across different life domains.

What has it found? Key findings and publications

Research using data from the cohort has had an important impact on policy and practice and has improved our understanding of human development and the predictors of health outcomes and health inequalities. To date there have been over 300 publications based on analysis of data from this cohort study. Limitations of space inhibit a full review of all the different research topics covered by the published literature. However, several major themes are discussed here, including prenatal and perinatal antecedents of health problems; social inequalities and health outcomes; health-related behaviours; and cross-cohort comparisons.

Prenatal and perinatal antecedents of health problems

The initial focus of the 1970 cohort study was on the medical management of pregnancy and birth and early research findings therefore reflect this theme. In particular, a number of studies have looked at the impact of maternal smoking during pregnancy. In common with the results of research on earlier cohorts, maternal cigarette smoking during pregnancy was found significantly to decrease birthweight, and in addition to increase the risk of perinatal mortality among the offspring of smokers in the manual social classes.³ Longer term impacts of maternal smoking have been found to include an increase in the incidence of respiratory illnesses in children^4–6; an increase in offspring smoking at age 16 and at age 30⁷ and an increase in psychological and somatic distress at age 26.⁸ However, research on febrile convulsions and afebrile seizures in children up to age 10 suggests that low birthweight is a risk factor rather than maternal smoking during pregnancy.⁹

Other studies have specifically focussed on the impact of breast feeding on subsequent health outcomes including eczema and hay fever;^10,11 bronchitis, lower respiratory illness, and gastro-enteritis.¹² Associations between breast feeding and developmental outcomes for children have also been examined.^13,14

Social circumstances and health outcomes

One of the strengths of BCS70 is that its multi-purpose design makes it possible to examine links between health and social circumstances. Indeed a central concern of the original research team was the link between socio-economic status and infant health.^3,15 A number of early studies focussed on the social circumstances of the cohort member's birth and subsequent health outcomes. While some research has demonstrated an increased risk of perinatal mortality and low birthweight for babies born to unmarried mothers,¹⁶ other analyses have shown no statistically significant differences between outcomes for the group of babies conceived pre-maritally and those conceived within marriage, once maternal age, parity, and maternal smoking were taken into account.¹⁷ Paternal unemployment at the child's birth was found to be associated with maternal health behaviour but not with low birthweight or pre-term delivery, once controlling for social class.¹⁸ More recent research has examined the long-term impact of socio-economic disadvantage on psychological adjustment¹⁹ and has also shown a strong link between low levels of qualifications and depression among young adults.^20,21

Adult outcomes of childhood disease and health status, and predictors of adult health status

The longitudinal design of BCS70 makes it possible both to examine long-term correlates of health conditions and disease in childhood, and also to investigate which risk factors from childhood are the best predictors of adult health conditions. For example, studies have focussed on behaviour and cognitive outcomes from middle ear disease²²; the outcomes of childhood status epilepticus and lengthy febrile convulsions²³; pertussis infection in childhood and subsequent type 1 diabetes mellitus²⁴; adult outcomes of attention-deficit disorder²⁵; and adult socioeconomic, social, and psychological outcomes of childhood obesity.²⁶ The longitudinal nature of the BCS70 has also been exploited in examining childhood predictors of self-reported chronic fatigue syndrome in adults.²⁷ It has also been demonstrated that the initial handling of newborn infants (as a proxy for increased exposure to micro-organisms) has a significant impact on individuals developing hay fever by age 26.²⁸

Health-related behaviour

As summarized in Table 2, at age 16 cohort members were asked to complete detailed dietary diaries over a 4 day period and also provided information on smoking and alcohol consumption. This has led to analysis of the health behaviour of cohort members. Crawley and Summerbell²⁹ provide descriptive analyses of the nutrient and food intakes of boys aged 16–17, while Crawley³⁰ has found poorer quality diets (more soft drinks, chips, and white bread, and less non-processed vegetables and non-fried potatoes) reported by teenagers in Scotland compared with teenagers from elsewhere in Britain, even allowing for possible confounding factors such as family size, tenure, and smoking behaviour. An association was also found between parental smoking habits and the dietary behaviour of teenagers, with lower intakes of fibre, vitamin C, vitamin E, folates, and magnesium among both males and females where parents were smokers. Smoking in adults and passive smoking in children has also been found to be associated with acute appendicitis.³¹ Recent research has looked at links between behaviour at age 16 and BMI at age 30.³² Analyses suggest that 4 or more hours of sedentary behaviour each day at age 16 is linked to an increased BMI at age 30.

Cross-cohort comparisons

The BCS70 is the third in a series of four British Birth Cohort studies and therefore contributes to our ability to make comparisons of the health of successive cohorts of individuals. Studies focussing on cohort comparisons have examined a number of different health outcomes including hay fever and eczema^33,34; self-reported asthma and wheezy bronchitis,³⁵ psychological disorders,³⁶ psychological well-being²⁰; chickenpox³⁷; and health problems and illness in adulthood.³⁸ A helpful overview of cross-cohort changes in reports of physical illnesses in adulthood (including asthma, bronchitis, back pain, diabetes, high blood pressure, and cancer) is provided by Wadsworth et al.³⁹ This chapter also summarizes changes in birthweight and adult height and weight across the cohorts. Whereas some studies have explicitly focussed on the changing incidence and prevalence of health conditions across different cohorts^35,36,39 others have examined whether the risk factors for particular diseases remain similar through time.^37,40 For example, by looking at the increased association between maternal smoking and low birthweight for more recent cohorts (where mothers have higher awareness of the harms of prenatal smoking), Fertig⁴⁰ demonstrates that at least part of the association between maternal smoking and poor infant health outcomes is owing to selection effects. An overview of cross-cohort comparisons in health-related behaviour shows a downward trend in smoking behaviour of adults in their 30s particularly among women, but an increase in alcohol consumption and the use of illegal drugs.⁴¹

What are the main strengths and weaknesses?

The focus on all births in a single week in 1970 and the difficulty of recruiting subsequent immigrants to Britain mean that the cohort does not have the ethnic diversity of today's population. During the 1980s and 1990s, ad hoc funding inhibited the development of strategies for the timing and sometimes the content of each follow-up. Strengths include the large study sample; extensive data coverage; seven ages studied to date; information on cohort members and a sample of their children in 2004; and use of objective measures, standardized tests, or scales, especially in the earliest phases of follow-up (e.g. for height and cognition). An important facet of the childhood data is the simultaneous coverage of physical, cognitive, emotional, and behavioural development at ages 5, 10, and 16 yr. More recent sweeps of the study (i.e. in 2000 and 2004) include information about health behaviours such as smoking, alcohol consumption, diet, and exercise. These provide potential predictors of health status in later life.

Can I get hold of the data? Where can I find out more?

The 1970 cohort is conducted by the CLS at the Institute of Education, University of London, under the direction of Dr Jane Elliott. CLS is an ESRC Resource Centre, and offers support and advice to data users. The CLS website with documentation for the cohort and detailed information about current research and publications is http://www.cls.ioe.ac.uk/. Data from the 1970 cohort are held and distributed by the UK Data Archive, a service provider of the Economic and Social Data Service. Non-commercial users can download data free of charge. Further information about the cohort and other longitudinal studies can be obtained from ESDS longitudinal at http://www.esds.ac.uk/longitudinal/.

The Centre for Longitudinal Studies, with responsibility for the BCS70 is funded as a resource centre by the ESRC. Support is also received from the International Centre for Child Studies. The authors would like to thank Prof. John Bynner and Prof. David Blane for their helpful comments on an earlier draft of this paper. Particular thanks must also go to Prof. Chris Power whose profile of the NCDS (1958 cohort study) laid the foundations for this profile of the 1970 cohort.

References