¶¶Òõ¶ÌÊÓƵ

Dementia is the leading cause of disability in people aged over 65 years and the second leading cause of death in Australia. Preventing or delaying the onset of dementia is a critical health challenge and will have a major impact globally. Evidence suggests infections may play a pivotal role in triggering chronic inflammation in the brain, leading to neurodegeneration and dementia. However, the causal link between infections and dementia is not well established as current studies have failed to robustly address reverse causation (e.g. dementia increasing the risk of infection). 

Understanding the causal role infections play in the aetiology of dementia has the capacity to expand the horizon of dementia prevention and treatment. This project will provide real world evidence to estimate the population-level impact of infections on dementia and the potential for preventive interventions. The methods advanced in this work will have immediate application to other dementia risk factors and many other areas of health prevention and policy evaluation.

Aims

This project aims to advance both (i) the understanding of the causal role of infection in the aetiology of dementia, and (ii) the methodology used to investigate cause and effect within observational data. It will investigate whether hospital-diagnosed infections lead to subsequent dementia incidence in an Australian population and include a particular focus on persistent viral infections such as Herpes simplex virus 1, and Herpes zoster virus (Shingles). This project will also aim to develop tools and techniques which will provide a platform for robust evaluation of potential biases such as confounding or reverse causation - factors which often preclude drawing causal conclusions from observational studies. 

Design

The gold standard for establishing cause and effect in health is the randomised controlled trial - usually impossible when studying non-manipulable disease risk factors. This project will employ advanced epidemiological and statistical causal inference methods to try to mimic randomised experiments within observational data, such as through the emulation of target trials. There is also scope to develop novel approaches to studying cause and effect, e.g. by leveraging the rapidly developing fields of Causal Artificial Intelligence (AI) and network analysis to investigate and test the possibility that reverse causation can explain relationships in the data. It will use population-wide Australian data, including integrated Hospital, Aged Care, Medicare and Pharmaceutical data accessed through the National Health Data Hub. Dementia will be ascertained by using diagnosis codes combined across multiple datasets.