Biomedical data in longitudinal studies
Introduction
Until recently, measures of 'health' in social surveys have mainly been self-reported. Increasingly, however,
social surveys - and longitudinal studies in particular - have augmented or replaced these subjective responses to
survey questions with data derived from the collection of biological samples and/or from measures/tests conducted
by nurses and other professionals.
Cells and Surveys: Should Biological Measures Be Included in Social Science Research?, published
by the (United States) National Research Council in 2001, identified biomedical measures as an important tool for
understanding the associations between socio-economic status and health and mortality. Since then, biomedical
measures - or 'biomarkers' - have become an important feature in many large-scale social surveys.
A key reason for collecting biomarkers in population-based studies is to analyse the interaction of biological
systems with social factors, social behaviours and the environment, and thus to identify new risk factors and reduce
unexplained variance in analysis models more generally. Biomarkers are also used as an alternative to, or
as a validation measure of, self-reports. For example, serum cotinine concentration has been used to evaluate the
accuracy of self-reports of cigarette smoking (Perez-Stable et al., 1990). Perhaps most exciting is the likelihood that
controlling for biomarkers in social science analyses will open up new strands of research and challenge conventional
wisdom.
As with other data collected, the specific advantage of collecting biomedical data in longitudinal, as
opposed to cross-sectional, studies is the ability to identify causality rather than mere association. We can
expect biomarkers collected in the context of longitudinal studies to offer the potential for better explanations
of social and economic change.
An alternative to - or complementary to - collecting biomedical data as part of social surveys is the linkage with
increasingly available event data such as health registers. One example of this is the
ONS
Longitudinal Study which contains 1971, 1981, 1991 and 2001 Census data which have been linked with information
on events such as births, deaths and cancer registrations. Another example is the Millennium Cohort Study to which birth registration records, centrally
collected hospital records, and hospital of birth data may be linked to the survey data.
The
Administrative Data Liaison Service provides further information on the availability of administrative records
related to health and disability and how these might be matched to survey responses.
Types of biomedical measures
The terminology surrounding biomedical data varies and some of it will be novel to social scientists. Formally,
a biomarker is anything that can be used as an indicator of a particular disease state or some other biological
state of an organism. A National Institutes of Health (NIH) study group committed to the following definition in
1998: "a characteristic that is objectively measured and evaluated as an indicator of normal biologic processes,
pathogenic processes, or pharmacologic responses to a therapeutic intervention".
In practical terms, the collection of biomedical information from survey respondents will fall into one of
three categories:
- bio-material, like blood, hair, urine, saliva or a tooth; collected via sample taking
- anthropometrical measures, such as height and weight, demi-span, waist/hip circumference or body fat
- or physical and cognitive function tests, such as blood pressure, heart rate, grip strength or memory test
Social scientists are probably familiar with the latter two categories but, increasingly, the term biomarker is
used as a catch-all phrase to cover each of the above categories.
While the latter two data collection categories produce ready-made digitised, numerical data - for example,
a respondent's waist circumference is measured in centimetres and is easily recorded as an integer on a dataset -
biomaterial samples require laboratory analysis to produce numerical information (e.g. a measure of sodium
concencentration in urine in mmol).
In addition to the collection and production of
phenotypic data, genetic variants may also be extracted from biomaterial (e.g. DNA extracted from blood) and
increasingly are. Laboratory analysis allows genotyping of a sample, though access to genetic data is, for reasons
of sensitivity and confidentiality, subject to much more restrictive conditions.
Selected publications
(2008) American Journal of Sociology: Exploring Genetics and Social Structure, 114, Supplement.
Retrieved February 8, 2010, from http://www.journals.uchicago.edu/toc/ajs/114/s1.
Bearman, P. (2008) 'Introduction - Exploring genetics and social structure', American Journal of Sociology,
114, Supplement: v-x. Retrieved February 8, 2010, from http://www.journals.uchicago.edu/doi/full/10.1086/596596.
Finch, C.E., Vaupel J.W. and Kinsella, K. (eds.) (2001) Cells and surveys. Should biological measures be
included in social science research? Washington, D.C.: National Academies Press. Retrieved February 8, 2010,
from http://www.nap.edu/openbook.php?record_id=9995&page=1.
Freese, J., Li, J-C.A. and Wade, L.D. (2003) 'The potential relevances of biology to social enquiry',
Annual Review of Sociology, 29, pp.233-256. Retrieved February 8, 2010, from http://arjournals.annualreviews.org/doi/full/10.1146/annurev.soc.29.010202.100012.
Guo, G. (2006) 'The linking of Sociology and Biology', Social Forces, 85(1), pp.145-149. Retrieved
February 8, 2010, from http://socialforces.unc.edu/epub/folder.2007-02-09.8541500563/sept-2006-85-1.html.
Juerges, H. (2007) 'True health vs. response styles: exploring cross-country differences in self-reported
health', Health Economics, 16(2), pp.163-178. Retrieved February 8, 2010, from
http://www3.interscience.wiley.com/journal/112770972/abstract.
(2009) ko’hört: Special Issue on Biomedical Data Collection, Spring 2009. Retrieved February 14, 2012, from
http://www.cls.ioe.ac.uk/downloads/kohort_spring_final%20web.pdf.
Lillard, D. and Wagner, G.G. (2006) 'The value added of biomarkers in household panel studies', DIW Data
Documentation, 14, pp.1-12. Retrieved February 8, 2010, from http://www.diw.de/documents/publikationen/73/diw_01.c.44684.de/diw_datadoc_2006-014.pdf.
Lindau, S.T. and McDade, T.W. (2008) 'Minimally invasive and innovative methods for biomeasure collection in
population based research', in National Research Council (ed.) Biosocial Surveys, Washington, D.C.: National
Academy Press, pp.251-277. Retrieved February 8, 2010, from http://www.nap.edu/openbook.php?record_id=11939&page=251.
McDade, T. W., Williams, S. and Snodgrass, J.J. (2007) 'What a drop can do. Dried blood spots as a minimally invasive method for
integrating biomarkers into population-based research', Demography, 44(4), pp.899-925. Retrieved February 8,
2010, from http://muse.jhu.edu/journals/demography/v044/44.4mcdade.pdf.
National Research Council (ed.) (2008) Biosocial Surveys, Washington, D.C.: National Academy Press.
Retrieved February 8, 2010, from http://books.nap.edu/openbook.php?record_id=11939&page=1.
Pérez-Stable E.J., et al. (1990) 'Apparent underreporting of cigarette consumption among Mexican American
smokers', American Journal of Public Health, 80(9), pp.1057-1061. Retrieved February 8, 2010, from
http://ajph.aphapublications.org/cgi/reprint/80/9/1057.
Schnell, R. (2010) 'Biological variables in social surveys', RATSWD Working Paper, 138. Retrieved August 31, 2010, from
http://www.ratswd.de/download/RatSWD_WP_2010/RatSWD_WP_138.pdf.
(2008) Sociological Methods & Research: Special Issue on Society and Genetics, 37(2). Retrieved February
8, 2010, from http://smr.sagepub.com/content/vol37/issue2/.
ESDS Longitudinal studies: Biomedical and health events data collected
In Wave 2, in addition to the main interview and the self-completion questionnaire, biomedical
and physical performance measures were collected from respondents by a trained nurse. The nurse visit included
the taking of several measures and samples including:
- blood pressure
- blood sample (later analysed for genetic information)
- lung function
- saliva sample
- anthropometric measures – weight, sitting height, standing height, and waist and hip measures
- physical performance measures - grip strength, chair rises, balance and leg raises
- gait speed (timed walk)
More details on the main interview, nurse visit, blood sample and self-completion questionnaire can be found in
the Wave 2 technical report.
Access to survey/phenotypic data via ESDS (special conditions
apply).
Access to genetic data
(available from ELSA DNA Repository (EDNAR)).
Where the mother consented, the following records have been added to the main MCS survey data:
- birth registration records (personal and demographic information about baby and parents at birth
registration);
- centrally collected hospital records (information about the mother's stay in hospital, including details of
any operations performed or any diagnoses that were made during the hospital stay and the length of stay in
hospital);
- coded hospital of birth data. (Note that the corresponding uncoded hospital of birth data are held under
SN 5724 and are subject to stringent
Special Licence access conditions).
More details on the information collected can be found in the
user guide
prepared by the Centre for Longitudinal Studies.
Access to health records data via ESDS (special conditions
apply).
The NCDS biomedical survey was funded under the MRC 'Health of the Public' initiative, and was carried out in
collaboration with the Institute of Child Health, St George's Hospital Medical School, and NatCen.
The CAPI interview included the following elements:
- vision: measures of near vision in right and left eyes (using appropriate visual correction), with and without
pinhole viewer; stereo vision; distance vision (using appropriate visual correction)
- blood pressure and pulse: three measures of systolic and diastolic blood pressure and resting pulse
- prescription drugs: all prescribed drugs taken, by name and BNF code
- hearing: thresholds of hearing in right and left ears at 1kHz and 4kHz
- standing height, sitting height, weight, waist circumference, hip circumference
- lung function: three measures of forced vital capacity (FVC), forced expiratory volume (FEV1) and peak flow (PF)
- eye measurements using autorefractor: sphere, cylinder and axis of right and left eyes
- non-fasting blood sample
The CASI interview covered audit and supplementary questions about drinking alcohol, adverse childhood
experiences; and the CIS-R interview modules on appetite, fatigue, concentration and forgetfulness, sleep problems,
irritability, depression, depressive ideas, anxiety, phobias, and panic.
Two paper self-completion questionnaires cover the topics: sun exposure; physical activity connected with work;
hearing; eyesight; pain; working conditions; household circumstances; social support as well as general health and
diet; leisure exercise; employment; partnership status and children; life events; and (women only) contraception
and HRT.
More details can be found in the Technical Report and in the survey documentation.
Access to survey/phenotypic data via
ESDS (special conditions apply).
Access to genetic data (available from St. George's,
University of London).
The study intends to collect phenotypic information related to people's health such as height, weight, blood
pressure and cholesterol levels. Dependent on funding, the study might also include the analysis of blood and
saliva samples. For more information see the Understanding
Society web page on biomedical research or a Longview
report outlining the options for collecting biomedical information.
Access to data: The survey data collection started in 2009, though a biomedical data collection is not expected
before Wave 3.
Other ESDS studies: Biomedical and health events data collected
Health and Lifestyle Survey
The first Health and Lifestyle Survey (HALS1) was conducted in 1984-1985 as an attempt to measure self-reported health, attitudes
towards health and beliefs about the causes of illness. Initially no further survey was foreseen but a follow-up study (HALS2) was
conducted in 1991-1992 which included as many respondents of HALS1 as possible. HALS data have been linked to deaths and cancer records.
HALS1 and HALS2 contain a number of biomedical variables:
- physiological measures (including anthropometry, blood pressure, respiratory function, exhaled carbon monoxide)
- cognitive function tests (including memory and reasoning)
HALS2 also collected data on salivary cotinine.
Access to HALS data via
ESDS
Health Survey for England | Scottish Health Survey | Welsh Health Survey
The Health Survey for England (HSE) is a series of annual surveys sponsored by the Department of Health. Starting in 1991, these surveys are
designed to measure health and health related behaviours in adults and children, in England. The surveys consists of core elements that are
included every year and special topics that are included in selected years. Core topics include: general health; smoking, drinking and fruit and
vegetable consumption; height; weight; blood pressure measurements and blood and saliva samples. Special topics include: cardiovascular disease;
physical activity; accidents; lung function measurement and certain blood analytes.
Access to HSE data via
ESDS
Health Survey for England (HSE) Bloodbank:
The HSE collects blood and saliva samples in addition to data on items like height, weight and blood pressure. The blood samples have been
collected since 1994 as part of the Health Survey for England Bloodbank project.
Bloodbank project
The Scottish Health Survey (SHS) is commissioned by the Scottish Executive Health Department and is closely modelled on the Health Survey for
England. Surveys were conducted in 1995, 1998, 2003 and 2008. The survey is currently running continuously from 2008-2011. The SHS collects
samples of urine, blood and saliva. Other variables include measures on lung function and blood pressure, as well as anthropometric data on
height, weight, waist and hip circumference, and demi-span.
Access to SHS data via
ESDS
The current Welsh Health Survey (WHS) series was conducted for the first time in 2003-2004. It is commissioned by the Welsh Assembly
Government (WAG), and carried out by the National Centre for Social Research. The WHS collects data on height and weight.
Access to WHS data via
ESDS
Growing Up in Scotland
The Growing Up in Scotland (GUS) study is a large-scale longitudinal social survey which follows the lives of groups of Scotland's
children from infancy through to their teens, and aims to provide important new information on young children and their families in Scotland.
Height and weight measurements were taken in Sweeps 2 and 4. Derived BMI measures are available in the datasets deposited with the UK Data
Archive.
Access to GUS data via
ESDS
National Diet and Nutrition Surveys
In 2008 a new rolling cross-sectional survey was established as part of a continuation of the National Diet and Nutrition Survey (NDNS).
The rolling programme will provide the detailed food consumption data essential to support risk assessments for food chemicals and will
also benefit a wide range of Government activities related to diet and health.
The survey includes components to:
- measure blood and urine indices that give evidence of nutritional status or dietary biomarkers and to relate these to dietary,
physiological and social data
- provide height, weight and other measurements of body size on a representative sample of individuals and examine their relationship
to social, dietary, health and anthropometric data as well as data from blood analyses
Users should note that blood and urine analytes data will become available when the NDNS year 2 data are archived during 2011.
Access to NDNS data via
ESDS
Non-ESDS studies: Biomedical and health events data collected
Avon Longitudinal Study of Parents and Children
National Survey of Health and Development
Whitehall II
UK Biobank
Print-friendly page
