Developing statistical longitudinal models to investigate the effects of life course events and decisions over the long term on the physical and mental health of the elderly

This study aims to develop statistical models to describe and investigate the effects of life course events and decisions over the years on well-being in later life. The project will take advantage of increasing accessibility of longitudinal datasets which has led to growing research interest in the statistical analysis and modelling of such data including the development of new statistical techniques and novel applications Britain is in the fortunate position of housing several cohort and panel studies with detailed information on health, economic and social circumstances. In particular, in the first instance, this study will use the National Child Development Study (NCDS) and the English Longitudinal Study of Ageing (ELSA) which provide measures of self- reported functioning, well-being and mental health, lifetime social conditions and health measures for those aged 50+. The current Government’s policy of extending the retirement age for both men and women and increased life expectancy has implications for the UK’s ageing population. Hence the mental and physical well-being of the elderly is important on a national scale. This project will analyse the information given in various longitudinal datasets to model, and to use such models to predict, how life course events and decisions may affect the physical and mental health of the elderly.

The aim of this project is to understand, develop and implement appropriate statistical methodologies for the analysis of longitudinal data with specific application in the broad medical field of ageing studies. There have been several cross-sectional studies looking at the benefits to the elderly of changes in lifestyle, but this project looks at the long term implications of life course events and decisions on the mental and physical well-being of the aged.

The UK has a prominent role in the development of longitudinal studies, with its wealth of publically accessible cohort studies which follow a group of people over time to gather repeated measures of life experiences and health status. One such dataset is the English Longitudinal Study of Ageing (ELSA) which has been providing information on 50+ year olds since 2002; another is the National Child Development Study (NCDS), whose participants were 50 in 2008. Although it is envisaged that these will be the two most important sources of information, other datasets will be investigated as common themes and factors emerge.

Longitudinal research is a broad term and hence the methods of analysis can vary substantially. This study will investigate several approaches to examine the effects of life course events and decisions on the health of the elderly. It is envisaged that this project will involve a number of techniques for the investigation and development of models for longitudinal data, including multiple and logistic regression techniques, multi-level modelling, event history analysis, survival analysis, repeated measures analysis and Structured Equation Modelling. Generally, standard statistical packages (e.g. SAS, SPSS, R) would be used, although some more specific software may be required for the more specialized methodology.

Overall the outcome of the project is to develop a statistical modelling framework to provide an understanding of the contributory factors and progressive nature of the effects of life course events and decisions over the long term, and to use such models to predict the mental and physical well-being of the elderly.