Statistical approaches for life course studies: developmental trajectories and adult health

Lead Research Organisation: University College London
Department Name: Institute of Child Health

Abstract

Circumstances in early life are important for health in later life. They might act by influencing a child‘s physical, intellectual, social and emotional development. These pathways to later health and illness can best be studied using information collected on the same individuals from birth onwards (so-called cohort studies) and taking a ‘life-course‘ perspective. However, statistical analyses of such information can be very challenging as each of these different aspects of a child‘s development are inter-related, are usually measured several times at different ages and may be missing. Better statistical methods to deal with these issues are needed to make reliable assessments of how early life circumstances influence later health and assess whether some ages are more important than others.

I will use information from different cohort studies to develop statistical methods appropriate for repeated inter-related measures and for situations where data are missing. Using these methods I will examine the influence of childhood development on mental health and cardiovascular disease risks in mid-life. An award of this fellowship will provide me with an important opportunity to develop my research expertise and skills, and to establish myself as an independent statistical researcher with experience in supervising students and obtaining grant income.

Technical Summary

Background: Life course studies are interested in how different developmental domains, physical, cognitive, and socio-emotional, influence adult chronic disease. But few studies have considered trajectories of development using adequate statistical approaches for repeated developmental measures as exposures at different life stages. Moreover, developmental domains have been examined separately rather than as inter-related phenomenon. Further advancement in life course research requires applications of more advanced models.

Aims: First to bring together recently developed statistical methods and life course conceptual models to investigate multiple domains of development and their impact on adult health outcomes; and second, to tackle the important issue of non-response in life course data.

Methodology: I propose two statistical approaches, each with two stages. First approach, (i) fit a random effect growth model to repeated development measures and estimate individual-level parameters, which indicates how each individual changes across time, and (ii) for each health outcome, fit a regression model which includes estimated growth parameters. Second approach, (i) fit a latent class growth model, which classifies distinct groups of trajectories, and (ii) the distinct groups are then related to the outcome of interest.

Because physical, cognitive, and socio-emotional development are inter-related, I will extend the two approaches and analyse all three developmental trajectories simultaneously by including an additional level to the random effects growth models, to incorporate the multiple aspects of development. Individuals are treated as level-3 units, multivariate responses (physical, cognitive, and socio-emotional development) as level-2 units, and repeated measurements as level-1 units. The growth curve for each individual can then be estimated. For some physical and cognitive measures trajectories will be extended to adulthood. For the second approach, joint trajectory models, an extension of latent class growth models will be applied to analyse all three developmental trajectories. These models will provide the capability to study distinct, but related domains of development.

I will further extend this work to consider multivariate response models for multiple health outcomes, as some outcomes are likely to be clustered (e.g. cardiovascular disease risks).

Life course data raise an important issue of missing data. Weighting and multiple imputation methods will be developed and implemented to adjust for the non-response.

Scientific opportunities: The proposed work will offer me the opportunity to explore statistical methods that can be applied in life course research. It will also allow me to take specialised training in statistics and advance my research expertise.

Publications

10 25 50
publication icon
Adamson A (2007) The changing social patterning of obesity: an analysis to inform practice and policy development in Peer reviewed report to the Policy Research Programme, DOH

 
Description Expert referee for NIHR HTA Programme
Geographic Reach National 
Policy Influence Type Participation in advisory committee
 
Description DOH supported Public Health Research Consortium (Co-I)
Amount £48,023 (GBP)
Organisation Department of Health (DH) 
Sector Public
Country United Kingdom
Start 01/2007 
End 01/2008
 
Description DOH supported Public Health Research Consortium (CoI)
Amount £404,647 (GBP)
Organisation Department of Health (DH) 
Sector Public
Country United Kingdom
Start 03/2012 
End 02/2015
 
Description ICH/GOSH Obesity Initiative Grant (PI)
Amount £49,956 (GBP)
Organisation University College London 
Department Institute of Child Health
Sector Academic/University
Country United Kingdom
Start 01/2007 
End 01/2011
 
Description MRC Centre Small Project Grant (PI)
Amount £36,500 (GBP)
Funding ID NA 
Organisation University College London 
Department MRC Centre for Epidemiology for Child Health
Sector Academic/University
Country United Kingdom
Start 09/2011 
End 10/2012
 
Description MRC Centre of Epidemiology for Child Health PhD Studentship (PI)
Amount £56,195 (GBP)
Organisation Medical Research Council (MRC) 
Sector Academic/University
Country United Kingdom
Start 09/2009 
End 10/2012
 
Description UCL Graduate Study Travel Grants (2009)
Amount £500 (GBP)
Organisation University College London 
Sector Academic/University
Country United Kingdom
Start 05/2009 
End 06/2009
 
Description UCL IMPACT Collaborative PhD Studentship (PI)
Amount £32,535 (GBP)
Organisation University College London 
Sector Academic/University
Country United Kingdom
Start 10/2012 
End 09/2015
 
Description BMI trajectories and CVD risk factors -a cross cohort comparison 
Organisation University College London
Department MRC Unit for Lifelong Health and Ageing
Country United Kingdom 
Sector Public 
PI Contribution Development and application of statistical methods. Develop and draft research papers.
Collaborator Contribution Data sets. Publications. Networking opportunities.
Impact Publications: one in AJE (doi: 10.1093/aje/kwn227), a second paper has been revised for IJE. Presentations at Joint Statistical Meetings, Miami, Florida (2011), HALCyon/IALSA Methods Conference, Sussex (invited, 2011), SLLS Amsterdam (2013), UCL MRC Unit for Lifelong Health and Ageing (invited, 2013). A third paper will be submitted shortly. I am also a collaborator on a work package on harmonise measures of body function across cohorts (part of a £5 million ESRC funded initiative on Cohort and Longitudinal Studies Enhancement Resources. PI: J Elliott, IOE).
Start Year 2007
 
Description CLOSER project 
Organisation University College London
Department MRC Unit for Lifelong Health and Ageing
Country United Kingdom 
Sector Public 
PI Contribution My link with the UCL LHA MRC Unit has been strengthened by my role as a collaborator on the ESRC CLOSER project on data harmonisation of measures of biological function across the cohorts. The collaboration has led to a paper (as a first author) on estimation of cross-cohort comparisons of life-course BMI associations with adult BP using a novel approach and a second paper (as a co-author) on co-ordinated analyses in five UK birth cohorts.
Collaborator Contribution Research collaboration, data harmonization, writing papers
Impact Published two papers in peer reviewed journals
Start Year 2012
 
Description Clean delivery kit use and neonatal/maternal mortality in rural South Asia 
Organisation University College London
Department Centre for International Health and Development
Country United Kingdom 
Sector Academic/University 
PI Contribution Statistical expertise. PhD supervision. Contribution to manuscripts.
Collaborator Contribution Data collection. Analysis. Paper drafting. Influence on policy and practice. Networking opportunities.
Impact Publications: one in PLOS MED (doi:10.1371/journal.pmed.1001180) and a second paper revised for PLOS ONE.
Start Year 2010
 
Description Cross-classified random effects modelling of cognitive outcomes in British Columbia children 
Organisation University of British Columbia
Department Human Early Learning Partnership (HELP)
Country Canada 
Sector Academic/University 
PI Contribution Statistical expertise. Contribution to manuscript
Collaborator Contribution Database on cognitive assessment for Canadian children. Publication. Networking opportunities
Impact Publication: Lloyd J, Li L, Hertzman C (2010). Early experiences matter: lasting effect of concentrated disadvantage on children's language and cognitive outcomes. Health & Place. 16: 371-380. Academic visit to University of British Columbia.
Start Year 2008
 
Description Growth in Autism 
Organisation Norwegian Institute of Public Health
Country Norway 
Sector Academic/University 
PI Contribution Supervision of a Research Council of Norway Oversea Fellow. Provide statistical/epidemiological expertise.
Collaborator Contribution Data of Norwegian MoBa cohort. Expertise on ASD. Networking opportunities.
Impact Two papers have been published (one by Pal Suren in JAMA doi:10.1001/jama.2012.155925; one with Pal in Epidemiology doi: 10.1097/EDE.0b013e31829e1d45).
Start Year 2010
 
Description Handling missing data in cohort studies 
Organisation University of Cambridge
Department MRC Biostatistics Unit
Country United Kingdom 
Sector Public 
PI Contribution Knowledge of life-course epidemiology on the topics of interest. Expertise of the measures in the 1958 British birth cohort. Contribution to two papers.
Collaborator Contribution Methodological development. Networking opportunities.
Impact Publications: one method paper (DOI: 10.1111/j.1541-0420.2011.01666.x) and a second paper (Handling Missing Data in Observational Cohort Studies) will be submitted.
Start Year 2008
 
Description MRC Centre of Epidemiology for Child Health 
Organisation University College London
Department MRC Centre for Epidemiology for Child Health
Country United Kingdom 
Sector Academic/University 
PI Contribution Participation in Centre scientific and academic seminars, workshops, symposia and masterclasses, and multi-disciplinary team working
Collaborator Contribution • Access to a critical mass of epidemiological and statistical expertise applied to children's health creating a positive working environment that contributes to researcher capacity, scientific excellence, methodological expertise and impact in this field. The mission of the MRC Centre of Epidemiology for Child Health is to improve the health and well being of children by conducting, and contributing to the application of, high quality interdisciplinary research. The MRC Centre adds value by: exploiting opportunities for electronic health record linkage; developing and applying advanced statistical methods; capacity building in genetic epidemiology; and developing approaches to policy research in child health. • Provision of Epilab: The epiLab is an advanced computing environment and data management service that has been developed at the MRC Centre to support researchers through all aspects of the research data management life-cycle. The service provides core data storage and backup provision in conjunction with a virtualised desktop and backend-server infrastructure, designed to provide high levels of resilience and information security. Additional services provide users with tools, training and assistance with data transfer, metadata management, data collection, archival and destruction and pseudonymisation. • Statistical support: several statisticians within the Centre provide statistical support for ongoing projects, in terms of study design, e.g. power calculations, or analysis, advising on and helping to interpret the analyses, or else doing the analyses in collaboration. • A strong track record in delivering excellent training and early career development support with opportunities to access supervision and supervisory experience for epidemiologists, statisticians and other population health scientists as well as clinicians from public health, paediatrics and paediatric sub-specialties reflecting the focus of the Centre's work on longitudinal epidemiological studies of population samples or patient cohorts. • Awarded two peer reviewed grants from the MRC Centre - MRC Centre PhD studentship and MRC Centre Small Project Grant.
Impact Publications, other collaborations and partnerships, influence on policy and practice, networking opportunities.
Start Year 2007
 
Description Maternal factors and childhood CVD risk factors in Portugal 
Organisation University of Coimbra
Department Department of Life Sciences
Country Portugal 
Sector Academic/University 
PI Contribution Initiated research project. Design statistical approaches. Supervise data analysis. Drafting papers.
Collaborator Contribution Provided access to data of the 2004 and 2010 national surveys of primary schools children in Portugal. Expertise on the topic and data.
Impact A paper has been published in Pediatric Obesity. Academic visit and invited talk (University of Coimbra). A conference presentation (DoHaD 2013).
Start Year 2010
 
Description Secular trends in CVD risk factors in UK children 
Organisation St George's University of London
Department Division of Population Health Sciences and Education
Country United Kingdom 
Sector Academic/University 
PI Contribution PI of the collaborative project. Line manage a research assistant. Secured funding for further work (MRC Centre Small Project Grant) .
Collaborator Contribution Access to multiple data sets of child CVD. Publications. Expertise on child CVD.
Impact Two published papers - in J Hypertens (doi:10.1097/HJH.0b013e3283562a6b) and in Arch Dis Child (to appear, selected for press release). MRC Centre Small Project Grant, £36,500.
Start Year 2008
 
Description Well-being of left behind children in China 
Organisation University College London
Department Institute For Global Health
Country United Kingdom 
Sector Academic/University 
PI Contribution Research expertise in childhood adversity and development and statistical expertise in large/complex data sets.
Collaborator Contribution Advise on study design, aims, data analysis, joint supervise a PhD student, writing papers.
Impact Joint supervision of a PhD student. Awarded an ESRC project grant (as a co-applicant).
Start Year 2012
 
Description Changes of BMI trajectories -MRC news (2009) 
Form Of Engagement Activity A magazine, newsletter or online publication
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Media (as a channel to the public)
Results and Impact Research findings on changes of life-course BMI trajectories were reported on the MRC News website (2009).

Publication of a paper in the American J of Epidemiology, influence on policy and practice
Year(s) Of Engagement Activity 2009
 
Description Changing influences on childhood obesity-press release (2010) 
Form Of Engagement Activity A press release, press conference or response to a media enquiry/interview
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Media (as a channel to the public)
Results and Impact A paper I was the senior author on ''Influences on childhood obesity in two generations of the 1958 British Birth Cohort'' was selected for press release by the American J of Epidemiology.

Influence on policy and practice through the contribution to a peer reviewed report to the Policy Research Programme, DOH.
Year(s) Of Engagement Activity 2010
 
Description Changing social patterns of obesity - DOH report (2007) 
Form Of Engagement Activity A magazine, newsletter or online publication
Part Of Official Scheme? Yes
Geographic Reach National
Primary Audience Policymakers/politicians
Results and Impact I contributed (co-authored) a peer-reviewed report on ''The changing social patterning of obesity: an analysis to inform practice and policy development'' to the Policy Research Programme, DOH.

Influence on policy and practice through the contribution to a peer reviewed report.
Year(s) Of Engagement Activity 2007
 
Description Childcare and risk of child obesity -BBC (2010) 
Form Of Engagement Activity A press release, press conference or response to a media enquiry/interview
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Media (as a channel to the public)
Results and Impact A paper I co-authored was covered by the BBC, Fox News, Scotsman, Daily Mail

Influence on policy and practice through the contribution to a peer reviewed report to the Policy Research Programme, DOH.
Year(s) Of Engagement Activity 2010
 
Description Childcare and unintentional injury- press release (2009) 
Form Of Engagement Activity A press release, press conference or response to a media enquiry/interview
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Media (as a channel to the public)
Results and Impact A paper I co-authored on ''childcare and unintentional injury in young children from deprived backgrounds'' was press released by the JECH.

Influence on policy and practice through the contribution to a peer reviewed report to the Policy Research Programme, DOH.
Year(s) Of Engagement Activity 2009
 
Description Intergenerational Influence Child Obesity -press release (2010) 
Form Of Engagement Activity A press release, press conference or response to a media enquiry/interview
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Media (as a channel to the public)
Results and Impact A paper which I was the first author on ''intergenerational influence child obesity'' was selected for press release by the American J of Clinical Nutrition.

Influence on policy and practice through the contribution to a peer reviewed report to the Policy Research Programme, DOH.
Year(s) Of Engagement Activity 2010
 
Description Media interest - trends in resting pulse rate in UK children 
Form Of Engagement Activity A press release, press conference or response to a media enquiry/interview
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Media (as a channel to the public)
Results and Impact A paper on secular trends in resting pulse rate in UK children (1980-2008) published in ADC was selected for press release .

The paper attracted media interests and appeared in Daily Mail, Mirror, Telegraph.
Year(s) Of Engagement Activity 2013
URL http://www.dailymail.co.uk/health/article-2506922/Warning-rising-pulse-rates-children-Average-restin...
 
Description Policies for the early years-DOH report (2011) 
Form Of Engagement Activity A magazine, newsletter or online publication
Part Of Official Scheme? Yes
Geographic Reach National
Primary Audience Policymakers/politicians
Results and Impact Contributed (co-authored) a peer-reviewed report on ''Will policies for the early years reduce inequalities in health? A synthesis of evidence to inform policy development, using the examples of unintentional injury and childcare'' to the Policy Research Programme, DOH.

Influence on policy and practice through the contribution to a peer reviewed report
Year(s) Of Engagement Activity 2011