Last month, members from the MAGENTA team attended the Advances in environmental health research: Platforms, linkages, and methodseventin London. The event brought together researchers, epidemiologists, statisticians, data scientists, and policymakers to explore the opportunities and challenges of using linked environmental and health data for research.
The MAGENTA Project uses linked health, administrative, temperature and environmental data to investigate how prolonged heat exposure affects pregnancy outcomes for women in Wales and London. Given its focus on environmental data linkage and population health research, the event provided a valuable opportunity for our team to connect with and learn from researchers working on similar environmental health data platforms, linkage projects and methodological developments.


MAGENTA team members Morgan Byatt, Prof Pia Hardelid (UCL), Dr Amal Rammah (UCL) and Dr Amy Mizen attended the event, with Amy presenting the MAGENTA Project as an example of a large-scale cohort study using environment–health data linkage.
In this blog, we hear from Morgan Byatt, Epidemiological Research Assistant, who attended the event, sharing his highlights from the day and key takeaways that will inspire his work at MAGENTA.
What did you find useful about connecting with projects and researchers tackling projects related to MAGENTA?
Learning about the large-scale environmental data linkage projects currently being done across the rest of the UK was probably one of the most useful aspects of the event. Air pollution and climate data science/epidemiology are rapidly growing fields in the UK with many interesting projects (such as MAGENTA) being conducted at the moment. Whilst it is amazing to be part of such a thriving research community, it can also be quite hard to navigate what projects of a particular type are out there. So, having a day to learn about some of the largest data linkage projects from people who are either leading or are directly involved in these projects was really valuable.
It was also great to catch up with other early career researchers and collaborators, as well as meet some people in-person for the first time!
Was there a presentation, discussion, or networking moment that stood out to you most?
One project I found particularly interesting was the implementation of pollution data (such as air and noise pollution) into the UK Longitudinal Linkage Collaboration. The UK has a well-established history in setting up longitudinal population studies to analyse how a whole array of exposures impact health. However, when many of these studies were set up, there were technological and computational constraints in linking individual-level health data to environmental exposures. Therefore, the opportunity to (as a result of this project) analyse the impact of environmental exposures within these historical cohorts is incredibly valuable. Such a project couldn’t be done without extensive collaboration between institutions (with Swansea University being an involved member!) and so I think this project is a testament to how powerful and productive a collaborative approach to research can be.
Were there any takeaways that you have applied to your work on the MAGNETA project?
There were a lot of takeaways from the event, though one that has stuck to me the most is the following:
Large-scale environmental data linkage is an incredibly valuable process that allows for some really detailed environmental epidemiological research to be produced. However, just because the capacity to conduct some analyses exists, doesn’t mean that we should just dive in and do so. At its core, epidemiology is ultimately concerned with what causes certain health outcomes and why, rather than precise predictive modelling. For example, rather than trying to predict some outcome Y based on an exposure X, epidemiology tries to understand both if and how X causes Y. Therefore, despite lots of information for individuals being linked, we may only really care about how a fraction of this information relates. In practice, this means conducting analyses grounded in hypotheses that are theory informed.
Myer Glickman from the ONS raised the point of theory-driven analysis in the panel discussion, further highlighting that every data set has its own biases and therefore every linkage can introduce some bias. Therefore, it’s really important that epidemiologists and medical statisticians understand how the data they are analysing can affect their results and the intended relationship they are seeking to understand. Furthermore, every linkage makes an individual more identifiable and so it’s also really important to remember to ensure individuals’ confidentiality is maintained. All of this may seem obvious, but it can be easily forgotten, especially when detailed large-scale data is available.
These were some points that really stuck with me, and they served as a great reminder of what the ultimate goal of research in this field is. I’ll definitely be remembering them during my work on the MAGENTA project and in the future.
For more information about the MAGENTA Project, visit: https://magentaresearch.org/