RSS Merseyside - Mental Health Statistics - 13th May 2019
On the 13th May 2019, the RSS Merseyside Local Group will welcome speakers Professor Richard Emsley (Kings College London), and Dr Anais Rouanet (MRC Biostatistics Unit, Cambridge University) for our event "Mental Health Statistics: Showcasing the analysis of mental health data", which coincides with the UK Mental Health Awareness Week. The meeting will take place in Room 107, Brodie Tower, University of Liverpool Campus (Building 233, in gridsquare C8 on campus map, https://www.liverpool.ac.uk/files/docs/maps/liverpool-university-campus-map.pdf)
14:00-14:50 Dr Anais Rouanet (MRC Biostatistics Unit, Cambridge University): Clustering methods for the study of cognitive decline and dementia for precision medicine: applications to French and US cohorts.
Dementia is one of the major causes of dependency amongst the elderly. No treatments have proved efficient to slow down the progression of this syndrome as they are probably administered too late. Identifying subjects at high risk of dementia is thus essential to start patient management earlier and improve individual health care, in a precision medicine context.
Clustering methods are a powerful tool to identify and profile the at-risk population over time. A first application will be based on the North American ADNI cohort, using a nonparametric Bayesian clustering method to identify patterns of cognitive declines, associated with brain imaging profiles. In a risk-stratified approach, a joint latent class model will be considered to analyze simultaneously the time-to-dementia in the French Paquid study, to associate different risks of dementia to each cognitive pattern, giving a more clinical interpretation to the clusters. This method accounts in particular for the interval censoring of dementia diagnosis as well as the competing risk of death. Finally, a dynamic prediction tool will be derived, allowing to compute for a given subject, the probability to develop dementia within the next 5 years, based on cognitive tests. This probability can then be updated at each new marker value collected.
These statistical tools are thus promising to refine the stratification of subjects with high risk of dementia and ultimately develop individual dynamic predictive tools for dementia, to improve individual care and/or for recruitment in clinical trials.
14:50-15:10 Tea and Coffee break
15:10-16:00 Professor Richard Emsley (Kings College London): How do mental health interventions work and who do they work for?
Well-designed randomised trials should involve examining the intermediate pathways through which the intervention under study affects the outcome. Similarly, in observational studies understanding how exposures impact on later outcomes is of crucial importance. Both of these questions can be answered using an approach called mediation analysis. This talk will discuss some of our new statistical methodology in mediation analysis arising from our research programme in efficacy and mechanisms evaluation. We illustrate these methods with examples from the field of mental health, examining the mediating role of various cognitive processes in the effect of childhood trauma on symptoms of psychosis and the role of sleep in psychological well-being.
All are welcome to attend the meeting, however we ask that people register in advance so that we can organise sufficient refreshments. More information, and registration links can be found at https://sites.google.com/site/rssmerseyside/research-meetings/mentalhealthstatistics .
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|