B3612 - Using detailed cohort data to investigate collider bias in mental health outcomes - 08/09/2020

B number: 
B3612
Principal applicant name: 
Gareth Griffith | University of Bristol MRC-IEU (United Kingdom)
Co-applicants: 
Dan Smith, Matt Tudball, Dr Tim Morris, Dr Hannah Sallis, Professor Kate Tilling, Professor George Davey Smith, Professor Marcus Munafo
Title of project: 
Using detailed cohort data to investigate collider bias in mental health outcomes.
Proposal summary: 

The need for comprehensive and representative data collection on populations for epidemiological research has been brought into sharp focus by the COVID-19 pandemic. This has resulted in the generation of many COVID-specific modules within existing cohorts and datasets (e.g. Henderson et al, 2020, Kwong et al, 2020). These datasets are going to be invaluable in understanding the mental health response of individuals to the COVID-19 pandemic. These studies reflect individuals responding under unique circumstances, presenting unique selection effects, which have the capacity to substantially bias results (Griffith et al. 2020). These selection effects are likely to be particularly stark with respect to mental health, which is known to be associated with non-response (Kwong, 2019). The data and analysis will be carried out by GG and DS, and data stored on the University of Bristol RDSF.

Impact of research: 
Elucidate potential impacts of non-random dropout, selection and collider bias. Increase understanding of the importance of representation in a mental health context.
Date proposal received: 
Monday, 7 September, 2020
Date proposal approved: 
Tuesday, 8 September, 2020
Keywords: 
Mental health - Psychology, Psychiatry, Cognition, Mental health, Statistical methods, Statistical methods