MIREDA (Mother and Infant Research Electronic Data Analysis) Partnership

Lead Research Organisation: Swansea University
Department Name: Institute of Life Science Medical School


The MIREDA partnership aims to improve maternal and infant health, particularly among disadvantaged groups, by developing new resources and tools for research that uses routinely collected data. It will do so by bringing together people and datasets from existing UK research programmes that are addressing infant and maternal health.

Poverty, disadvantage and associated poor health frequently start at the earliest stages of life. Poor parental health and adverse health behaviours such as drug and alcohol use affect development of the baby before and after birth, resulting in foetal growth restriction/small-at-birth, preterm birth, or being large for gestational age. The implications of these can last a lifetime, affecting health (brain and lung development, hearing/sight impairment), educational outcomes and subsequent life chances. This means that addressing inequalities in society starts with improving maternal and neonatal health. However, given a backdrop of rapidly-emerging and growing societal and economic challenges that include rising obesity, escalating living costs, inequality and large sections of society - including many migrants and asylum seekers - facing poverty and deprivation, it has never been more important to be able to understand how best to protect future generations from the long-term disadvantages that arise from adverse exposures before birth, including those linked to changing societal challenges. This partnership aims to facilitate such understanding, to inform evidence-based preventative interventions.

MIREDA will capitalise on and add value to recent investments and developments in health-data research. Due to the expansion of big data, there are many maternal and infant datasets around the UK that could be harmonised, analysed and compared for data science in population health. But access, governance, computational capacity and the analyst skills required are all serious (but addressable) barriers. Recent years have seen rapid developments in the infrastructure and expertise needed to safely bring together diverse data for public benefit. The urgent response to the COVID-19 crisis accelerated our capabilities, and we now know it is possible to enable researchers UK-wide to access and compare datasets across the four nations.

What the partnership will do: MIREDA will create a UK resource including harmonised maternal and infant birth-cohort health data, linked to local datasets including those in public health, neonatal health, imaging, primary care and hospitals. It will establish a multidisciplinary collaboration to provide the tools and supporting expertise for undertaking analysis in each of the cohorts without need to move the data. In addition, it will develop methods for data standardisation and common data management across datasets to ensure comparable analysis,and implement software for automating epidemiological study methods for ease of obtaining analysable datasets. It will work with others to build research capacity and networks in the field, using online and face-to-face workshops, seminars, conferences, and research development group meetings to share knowledge and skills within the UK and beyond. Finally, it will provide 'pump-priming' funding to support rising-star researchers and international collaborations in maternal and infant health, and to leverage additional funds for research to improve maternal care and infant outcomes.

Why this partnership is needed now: In the UK we have been very successful in reducing maternal and infant mortality, but this means more infants and mothers - predominantly from disadvantaged groups - living with chronic conditions. As disadvantage and deprivation become ever-more pressing societal issues, it is becoming increasingly urgent to mitigate the risks to maternal and infant health that are strongly associated with them, so as to improve life-course outcomes for successive generations and help break the cycle of poverty.

Technical Summary

To enable partnership working we will harmonise and recruit UK birth cohorts that use electronic linked data (core dataset and common data model, WP1). We will use a modified OMOP CDM (Observational Medical Outcomes Partnership Common Data Model) data standard for pregnancy and maternal/infant health data to transform the data to a common format. This will be undertaken using the Convenient And Reusable Rapid OMOP Transformer (CaRROT: HDR UK CaRROT Mapper on GitHib), which can help to automate the rules to convert data to OMOP standardised format. The OMOP CDM is patient level, with data mapped in domains and metadata therefore standardised across cohorts. Data from both the electronic linked data and questionnaires (where collected) will be mapped to different linkable domains. There will be a federated approach to analysis (WP2), whereby standard queries can be run on each cohort's core datasets. This will include developing common access approvals and common scripts that can be run on the harmonised data in each cohort, so that data remains with its local provider but meta-analysis of findings in each region can be used to examine outcomes for specific groups and for those with rare exposures.
Each core dataset will be linked locally with existing linkable datasets (WP3) and information from local linkages can be extrapolated to impute and estimate these variables in the core of the other cohort studies.
To build capacity and expertise across the partnership, MIREDA will work directly with the HDRUK training group and ADR training. We will join existing initiatives to contribute to delivering online and face-to-face workshops, seminars, conferences, and research-development group meetings in order to establish multidisciplinary collaborative national and international consortia and encourage sharing of knowledge and skills. This work will provide researchers, including early career researchers, with access to and training in using the MIREDA partnership data.