Advancing Compositional data analysis to harness the dynamic interplay of lifestyle behaviours for public health.

Ref: S2017SHLS007

Glasgow Caledonian University and Biomathematics and Statistics Scotland present an exciting PhD opportunity to work at the cutting edge of public health science, developing new statistical techniques and applying them to large datasets addressing a global health issue.

Background
Lifestyle behaviours throughout the day and physical activity have a significant impact on health and healthy ageing. Public health and clinical interventions seek to modify these behaviours to prevent and manage chronic disease and promote healthy independent ageing. As time is limited throughout the day, time spent in each behaviour necessarily impacts time spent in the others. Nowadays it is possible to record high volumes of high-precision continuous behavioural data thanks to modern body worn sensor and mobile health technologies. However, one key issue is dealing with them consistently and using them to understand the dynamics of the interplay between behaviours and how it can affect the effectiveness of interventions. Current conceptual and analytical approaches in physical activity epidemiology essentially consider lifestyle behaviours as isolated actions, ignoring the interplay between them and the intrinsic multivariate relative nature of the data generated. The introduction of compositional analysis for relative data in physical activity research was recently pioneered by a collaboration between GCU and BioSS as a major paradigm shift and progress.

Aims
This PhD project will focus on developing, implementing and applying compositional methodology for the analysis of longitudinal and randomised control trials to inform public health interventions targeting lifestyle behaviour synergistically. The first year will focus on familiarisation with the technical fundamentals, the application field and the development of models and methods. In the second year these methods will be implemented, put to the test, contrasted and refined using real data to study the life course determinants and the effects on health and healthy ageing of the dynamics of physical activity behaviours. The third year will be devoted to complete scientific publications, elaborate guidance to inform intervention and public health policy about the most effective ways to gain health benefits when targeting multiple lifestyle behaviour synergistically and thesis writing.

Specific requirements of the project:
- Degree or Master in a quantitative discipline (statistics, mathematics or related are preferable).
- Demonstrable training in multivariate data analysis and models for longitudinal data would be advantageous. Familiarity with compositional data analysis methods would be valuable.
- Previous experience with at least one programming language and statistical programming experience using R would be desirable.
- Experience in research would be desirable.
- Experience in dealing with health and physical activity data would be desirable.
- Keen interest in developing a research career in the interface between quantitative sciences and scientific applications, and acquiring widely transferable skills.
- Interest and ability to work in a multidisciplinary team.