Cohorts as Platforms for Mental Health research (CaP:MH)

This research programme aims to enhance two of the most richly characterised population based resources in the research area of early life course and inter-generational influences on the aetiology of mental disorder. Our focus reflects evidence from cohort studies, highlighted in the MRC Strategy for Lifelong Mental Health Research, suggesting that at least 75% of mental disorders are determined by factors acting before the age of 18 years. Considerations of statistical power and external validity mean that many questions on the epidemiology of mental disorder require multi-cohort resources. In addition to enhancement of the mental health capabilities of these premier cohort resources individually, we will also build the foundation of a future multi-cohort platform for mental health research.

CaP:MH aims to address the following challenges:

  • Developing methodologies for linkage and integration of biological, psychological, clinical, behavioural, lifestyle-related, social, educational or other datasets
  • Enabling interoperable datasets, ethical access and data sharing
  • Harmonisation of mental disorders measurement and symptoms in cohorts
  • Enhancement of existing cohort and longitudinal studies
  • Incorporating a broader range of physical and mental health outcome measures in population research
  • Demonstrate how activities will utilise and link to existing informatics infrastructure and resources

This will be achieved through the following indicative projects:

  • Project 1: Psychological risk and resilience in vulnerable children
  • Project 2: Using social media linkage for high-resolution longitudinal measurement of mental health
  • Project 3: Predicting risk of transition to psychotic disorder in high risk individuals
  • Project 4: Using routine data to explore risk and resilience for longitudinal change in maternal mental health and the association with child outcomes

Project 4 will develop processes for ‘translating’ routinely collected data into usable variables for research. It will provide tools and processes that can be applied to future research using routinely collected mental health data. The project will enhance the existing cohort studies, by ‘translating’ routine data and accelerating progress towards making routinely collected mental health data accessible for research. 

The work will be divided into 4 workpackages (WP):

WP1: Data mapping and pathways - In WP1 we will undertake a thorough mapping exercise of routinely collected mental health data that are linked, will be linked, or could be linked in each of the cohorts. We will map where data are held, who holds it, when it is collected and the kinds of data that are held e.g. Read codes, health care visits, free-text notes, prescriptions, as well as how this has changed over time. This may include interviews with key informants. Analyses of currently held data will inform this work.

WP2: Processes - Qualitative work with health professionals and patients (WP2) will develop our understanding of human factors that are associated with data collection locally. These include decision-making associated with entering data such as whether, and what, data are entered, Read code selection, completion patterns and assumptions made by health professionals or non-health professionals when entering data. As part of this work package we will also review the evidence on human and healthcare factors in mental health medical record keeping in the UK more generally.

WP3: Variable and analysis definition - In WP3 we will characterise the presence and timing of relevant mental health data. We will review the statistical, health service and epidemiological literature for least-bias approaches to classifying and analysing routine health data, with a focus on mental health data. We will then be in a position to develop algorithms that best predict cohort collected measures of ‘caseness’ during pregnancy and postnatally from routine data. Factors associated with unidentified mental illness will be modelled and methods to account for bias in the medical records explored. We will apply our work to characterise mental health in the postnatal period where only routinely collected data are available

WP4: Perinatal mental health analysis - Finally, in WP4 we will apply the understanding generated in WP1 and 2 and variables derived in WP3 to model the relationship between maternal mental ill health and selected child outcomes, and risk and protective factors.

Funding

Funder: MRC National Institute for Medical Research
Start Date: March 2018
Expiry Date: March 2020

Members

Internal Staff

External Collaborators

  • John Wright, Bradford Institute for Health Research
  • Pippa Bird, Bradford Institute for Health Research
  • Andy Boyd, University of Bristol
  • Oliver Davis, University of Bristol
  • Claire Haworth, University of Bristol
  • Mark Mon-Williams, University of Leeds
  • Marcus Munafo, University of Bristol
  • Kate Tilling, University of Bristol
  • Nic Timpson, University of Bristol
  • Stan Zammit, University of Bristol

Public Health and Society Research in the Department of Health Sciences