Cohort Profile Update: The 1970 British Cohort Study (BCS70)

The 1970 British Cohort Study (BCS70) began in 1970 with data collection on the births and social circumstances of over 17 000 babies born in the UK. Cohort members who were born in Northern Ireland were included in the birth survey but dropped from the study in all subsequent sweeps. At the time of writing, the cohort members are in their early fifties. The initial BCS70 birth study had a particular focus on perinatal mortality. The focus of the study has broadened over time, reflecting the interest of both social science and health disciplines in each life stage, as the cohort has moved through childhood into adolescence, adulthood and mid-life. Sullivan, Brown and Bann describe the history and context of the cohort from birth to mid-life. BCS70 is the third of a series of UK national birth cohorts which began with the 1946 National Survey of Health and Development (NSHD), followed by a second cohort in 1958, the National Child Development Study (NCDS). The Millennium Cohort Study began 30 years after BCS70. The Key Features

Centre for Longitudinal Studies at University College London houses the 1958, 1970 and 2000 cohorts.
Elliott and Shepherd 7 describe the BCS70 study up to 2004, when the cohort members were aged 34. BCS70 began as the British Births Survey, under the directorship of Roma and Geoffrey Chamberlain. Data collection at birth (1970) was conducted by midwives. Subsequent surveys at the ages of 5 (1975) and 10 years (1980) took place under the auspices of the Department of Child Health at the University of Bristol, led by Neville Butler. Butler set up the International Centre for Child Studies (ICCS) which carried out the age- 16 (1986) survey. Following a 10-year hiatus, the next major wave of data collection was at age 26 (1996), led by John Bynner at the Social Statistics Research Unit (SSRU), which subsequently became the Centre for Longitudinal Studies now based at University College London. This was followed by a face-to-face survey at age 30 (2000) carried out simultaneously with data collection at age 42 for the 1958 cohort (NCDS), with 90% of questions shared between the two cohorts. Heather Joshi directed the age-34 (2004) survey, a face-to-face interview including adult literacy and numeracy assessments. Jane Elliott become Principal Investigator of the study in 2005, followed by Alice Sullivan in 2010 and George Ploubidis in 2021. Mark Hamer was joint Principal Investigator with Alice Sullivan of the age-46 biomedical study.
What is the reason for the new data collection? Subsequent surveys have been carried out at regular 4-yearly intervals in order to track the lives of the study members into mid-life, with the intention of continuing to follow them into old age. This is designed to provide the opportunity to understand the precursors of beneficial and adverse circumstances in mid-life and their consequences in subsequent older age, maximizing the value of the earlier data collections. A summary of new data collection is shown in Table 1. Planned surveys of the whole cohort have been completed at ages 38, 42 and 46.
The Age 38 Survey consisted of a 25-min telephone interview and sought to establish changes in circumstances since the previous interview. The topics covered included: household situation, housing, relationships, children and wider family, family income and wealth, employment, lifelong learning, health and health behaviour.
The Age 42 Sweep involved a 60-min interview and a paper self-completion questionnaire. It aimed to provide rich data on the cohort members' lives across a wide range of domains. Topics covered included relationships, children, parents, place of residence, economic activity, income, qualifications and training, physical and mental health, smoking, drinking, diet, exercise, identity, attitudes and values, religion and leisure activities. A vocabulary test was also administered.
The Age 46 Survey included a full range of bio-measures, including for the first time collection of blood samples administered by a nurse. The inclusion of objective measures of health was designed to allow researchers to assess the longitudinal predictors of health in mid-life. Many of the measures were designed to allow for cross-cohort comparisons with the bio-measures administered to NCDS in mid-life.
The Age 50 Survey, due to commence in 2020, was delayed by the pandemic and started in September 2021. The pandemic has prompted new data collection in order to understand its impact on study members' lives. Three surveys collecting information about the impact of the pandemic were conducted in May 2020, October 2020 and February 2021. The COVID-19 Surveys were also completed by participants in NSHD, NCDS, the Millennium Cohort Study and Next Steps-the first simultaneous data collection across these UK cohorts. In March 2021, participants provided blood samples which were tested for COVID-19 antibodies.

What will be the new areas of research?
A number of new areas of research have been generated by the new data: Who is in the cohort?
Participants are survivors from the original sample of over 17 000 births, all born in the UK during 1 week in 1970. During childhood, cohort members were traced through schools, and immigrants born in the reference week were added to the target sample. Efforts have been made to maintain participant engagement through feedback mailings, birthday cards, study websites and social media.
Efforts are made to trace lost participants through use of study records, internet searches, electoral records and administrative databases, but failure to trace individuals when they move is the main cause of attrition over time.
At the most recent major wave of data collection, in 2016, the cohort members were aged 46. Due to selective attrition, the sample includes more cohort members from an advantaged childhood socioeconomic background (27.1% non-manual paternal social class at birth, 32.2% at age 46) and more women (48.2% at birth, 52.0% at age 46). Nevertheless, recent work from the Centre for Longitudinal Studies has shown that capitalizing on observed variables from earlier waves allows researchers to replicate the original distribution of the baseline sample, reduce bias and restore sample representativeness. 8,9 What has been measured?
Measurements collected in the three main waves of data collection from age 38 onwards are listed in Table 2, and the COVID-19 survey content in Table 3.
The Age 38 Survey contained core measures only, due to time constraints.
Novel measures at age 42 included: sexuality (asked for the first time at this wave), reading behaviour and a vocabulary assessment (both included for the first time since age 16). Cohort members were asked for consent to link survey data with health and economic records held by the National Health Service, Department for Work and Pensions and Her Majesty's Revenue and Customs.
Hospital Episode Statistics have been linked, and are available via the UK Data Archive.
Novel measures at age 46 included: a range of bio-measures, cognitive assessments, thigh-worn accelerometry and a dietary diary. BCS70 is being genotyped at the time of writing.
The COVID-19 survey included assays for COVID-19 antibodies and consent for data linkage to the ZOE app. 10 What has it found? Key findings and publications

Overweight and obesity
Cross-cohort analysis has demonstrated that during childhood the 1970 generation were no more likely to be overweight than generations born in 1946 and 1958, but became overweight from their 1980s adolescence onwards. Later-born generations have higher body mass index, with a substantial increase in childhood overweight and obesity compared with the 1970 generation. 11 Exposure to obesity across adult life is related to cardiometabolic risk markers. 12

Mental health
Mid-life is a period of relatively high psychological distress, and individuals born in 1970 were more likely to experience psychological distress than those born in 1946 and 1958. 13,14 Psychological distress in adulthood accompanies a doubled risk of premature mortality. 15 Physical activity BCS70 self-reports of physical activity are complemented by a measure of sedentary time using a thigh-worn accelerometer at age 46. 16 Device-assessed physical activity and sitting-time were associated with physical and mental health in mid-life. 17,18 Biomarkers Blood collected at age 46 was used to measure a number of risk-markers. Secular trends in cholesterol have been Education, learning, social mobility and work Advantaged social origins and schooling have life course advantages in terms of educational and occupational attainment. 21,22 Reading for pleasure is linked to higher cognitive scores in adolescence and adulthood, controlling for earlier scores. 23,24 The pay gap between men and women has reduced for BCS70 compared with earlier generations. 25 What are the main strengths and weaknesses?
The strengths of the study include its large, nationally representative sample. It provides prospective, longitudinal data, with follow-up spanning a large portion of the life  Weaknesses include selective attrition, which is a consideration for all longitudinal studies. In addition, gaps in the series of national cohorts limit the scope for comparison across generations, as some generations are missing from the series. BCS70 was the third in a series of UK birth cohorts spaced 12 years apart (1946, 1958 and 1970). The intention was to continue the series at 12-year intervals, but a lack of political will and funding meant that 30 years passed before the next cohort emerged (the Millennium Cohort Study). The cancellation of a planned 2012 cohort led to a further gap in the series. However, the Early Life Cohort (ELC) and Children of the 2020s (COTS2020) are incipient additions to the birth cohorts series.
Can I get hold of the data? Where can I find out more? ). In addition, London-Central Research Ethics Committee have provided ethics approval for the ongoing activities of the study in between sweeps of data collection: keeping in touch and tracing study members; cleaning, documenting and providing access to the data for research; and linking data from administrative sources to survey data to increase the utility of the data for research (14/LO/0371).

Author contributions
All authors contributed to writing the paper.

Funding
We acknowledge funding from the Economic and Social Research Council, ES/M001660/1.