Principal Investigator: Professor Martin Rossor
Institution: University College London (UCL)Tags: 'cognitive footprint', 41519, cognition, impairment, Machine Learning, medication
Medications are commonly prescribed in the UK and are taken for a variety of illnesses. Some medications can affect mental processes and abilities, commonly referred to as cognition. For example, some medications taken for blood pressure can cause fatigue, dizziness or headaches but it is not clear how much these impact upon formal cognitive measures. Indeed, how certain medication affects cognition is rarely assessed in the general population and is, instead, limited to studies of people with pre-existing impairment i.e. those diagnosed with Alzheimers disease or other forms of dementia. Some medications may impact upon cognition to a small degree but this may still be clinically meaningful. For example, a headache may impact upon levels of concentration and, therefore, affect a persons ability in their employment. However, to date, no studies have formally examined the effect of commonly prescribed medications on different areas of cognition.
We will explore associations between certain medications and cognition at baseline, whilst adjusting for other variables that may impact on cognition (e.g. age, disease, other medications taken). We will then use causal modelling techniques to examine more complicated relationships where the condition can impact on both cognition and on medication use such as for blood pressure.
To complement the traditional statistical analysis, we will also use an innovative artificial intelligence technique developed at UCL. This approach uses machine learning to compute very complex models in which infinitely more covariates, including white matter damage, can be adjusted for in an analysis. This will provide further evidence that any impairment is attributable to a medication alone. We envisage that the project would take approximately 18 months from receipt of data to dissemination.
A successful project would provide evidence for a unique, cognitive effect of certain medications. Some effects may be minor and some may be more significant. However, the formal study of these effects means that clinicians and patients will be able to more comprehensively weigh the benefits against any risks, including a risk of cognitive impairment, before commencing a medication. This will also act as a health technology tool for which two drugs may have equal efficacy for a condition but one is implicated in cognitive impairment. The results of this analysis will also be particularly relevant for those already experiencing or at significant risk of cognitive impairment such as those diagnosed with Alzheimers disease or a dementia, ensuring their cognition is not worsened unnecessarily.