PhD Studentship What are the Biological Mechanisms Underlying the Association Between Smoking, Smoking Cessation and Mental Health: A Triangulation Approach using Machine Learning, Mendelian Randomisation, and G-methods Applied to Multiple Biological Cohort Studies

Background: Smoking is the world's leading cause of preventable illness and death. One in every two people who smoke will die from a smoking-related disease unless they quit. While smoking prevalence has decreased markedly in the general population, it remains disproportionately higher among people with mental illness, leading to significant disparities in morbidity and mortality.

There is evidence indicating that smoking causes mental ill health, and cessation improves mental health. However, little is known about the causal biological pathways underlying this association. Preliminary evidence suggests that smoking can damage neurological systems and oxidative stress pathways, and these systems may return to normal functioning after sustained smoking cessation.

All studies on this topic are likely to face issues with unmeasured and time-varying confounders. Therefore, we propose an innovative approach to: a) provide evidence for the biological basis of the association between smoking (and cessation) and mental health, and b) address potential biases from confounding and time-varying confounders using novel methods.

Aims and Method: The student will apply a triangulation approach across their program of work to strengthen causal inference by comparing different approaches with assumed unrelated sources of bias. The student will work with multiple data sources from different populations and compare effect estimates derived from traditional modelling to estimates derived from novel and more sophisticated approaches, including G-methods, machine learning, adjustment for genetic risk exposure, and Mendelian Randomization. This triangulation approach not only leads to more robust conclusions but also provides an excellent training opportunity for the student in a variety of methods.

The student will have the opportunity to take full ownership of both the methodological and theoretical aspects of the project. This type of project offers an extensive range of research decisions, such as selecting variables for biological data, selecting specific biomarkers, neurological systems, mental health outcomes, or time-varying confounders. The student will be encouraged to explore various modelling approaches and select methods based on their scientific rationale and training needs. They will be able to choose from a variety of techniques, including G-methods and machine learning, tailoring their approach to best fit the project’s objectives. Additionally, the student will have the opportunity to shape the patient and public involvement aspect of the project, including selecting platforms and crafting messaging that align with their personal interests and career objectives.

International Impact: These findings will significantly improve our understanding of the mechanisms through which tobacco harms mental health and whether these harms are reversible after smoking cessation. This information will have important implications for the development of bespoke individual-level interventions and public health interventions (e.g., specific messaging around the benefits of cessation on one’s mental health) to improve smoking cessation. It will also have important policy and healthcare implications in the UK and globally. The importance of targeted messaging was recently highlighted in the UK Government’s Khan Review, “Making Smoking Obsolete.”

£940 p/a Research Training Support

International student fees may be covered depending on successful application to funding for this purpose from the University of Bath.

Advert information