The Networked Data Lab (NDL) is a Health Foundation initiative that’s building a collaborative network of analytical teams across the UK. Five groups of local partners are working with the Health Foundation to bring together analytical expertise in the use of linked datasets, which will help to address key issues for population health and care services.
We spoke to Kathryn Marszalek, Senior Analytical Manager at the Health Foundation, and Dr Jessica Butler from the Institute of Applied Health Sciences at the University of Aberdeen – one of the partners in the Networked Data Lab. They told us about what the programme is achieving in Scotland and more broadly.
Tell us about the Networked Data Lab and what it's aiming to achieve?
Kathryn: The aim of the Networked Data Lab is to produce insights on major challenges in health and care using linked datasets that can be used by both local and national policymakers to tackle pressing problems. For example, our first analysis explored the impact of the pandemic on the clinically extremely vulnerable population and provided insights into the support required for this population going forward.
A lack of joined up data is a big issue for people trying to improve health and care. It means that at a national level policymakers across the UK often don’t have the linked datasets needed to tackle the big health questions the country faces. Conversely, there are pockets of amazing linked datasets developed at a local level across the UK.
By creating the Networked Data Lab, the Health Foundation and our partners wanted to generate insights from these local linked datasets to help address major challenges that policymakers are facing. Another of our objectives is to showcase the value of linked data, encouraging further investment in the infrastructure and analytical capability needed, both locally and nationally.
Why is it so important to link up health data?
Kathryn: It’s important to point out that it’s not just about ‘health care’ data – it’s also about social care data, and the other information that local authorities capture that provides insights into the wider determinants of health. People’s health is looked after by multiple systems and providers, so when the data they hold isn’t linked, it’s not possible to see how those different services are working together to ensure that either an individual or a population remains healthy. Our recent report on the clinically extremely vulnerable population highlights the need for further investment in high quality linked data to ensure that, in the event of a future pandemic, people at particularly high risk can be quickly identified for additional support.
Jessica: That’s right, to better understand who is using care and why, it’s important to link data from across the whole health system. For example, in our Grampian NDL collaboration, we’re currently looking at whether children who live in deprived areas are more likely to reach crisis before they’re picked up by the mental health care system. To measure this, we need to understand which children are accessing care from GPs, out-patient clinics and emergency hospital admissions, as well as to understand their home area characteristics. This is a complicated data linkage problem, but it is possible and important.
What has working with the NDL partners taught you about successfully creating linked data systems across local health and care systems?
Kathryn: By working with five partners with advanced analytical capability who have invested in local linked datasets, we’re really demonstrating the art of the possible and learning some valuable things about what it takes to make this type of work happen. Having the right relationships with people who can facilitate conversations around data sharing is really important to getting things started. And having a good understanding of how data is going to be used, and why it’s important, is necessary to gain buy-in from the various teams that control the data.
Local expertise cannot be underestimated. Without a detailed understanding of each area’s nuances, it’s not possible to get to the full picture of what’s happening there – how their system works, how their data is captured, and how it can be analysed and interpreted to address key challenges across the health and care system.
Jessica: Those learnings are definitely echoed by us locally in Scotland. One big thing we’ve learned from our early work is that strong communities and personal relationships are key to making this type of difficult data analysis work. Our NDL group has made it a priority to work collaboratively with everyone from patients to clinicians to data managers to be sure the work we’re doing is safe, valuable, and well used.
How have you seen linked datasets being used during the pandemic?
Kathryn: The first area we looked at with the NDL was the clinically extremely vulnerable population shielding during the first wave of the pandemic (March to August 2020), and we’ve just published our analysis on that. That project really highlighted how valuable linked data was to shine a light on who the shielding population was, what their health care needs were, how the pandemic was affecting them, and how they should be prioritised. For local partners, this was key to being able to identify people and quickly put in place appropriate services.
Jessica: Our NDL analysis of the clinically extremely vulnerable has been important to NHS Grampian, and it has shaped how the NHS identify vulnerable populations. The work we did showed that the original method for detecting the vulnerable was very inefficient – almost half the vulnerable population was not detected by a screen of their medical records and, so, had to be added manually by clinicians. It led to us recommending that, going forward, NHS Grampian share GP records with local health boards to help make it easier to identify people considered clinically extremely vulnerable.
What’s next on the horizon for the NDL?
Kathryn: The next topics we’re looking at are really being informed by our partners, and what their priorities are at a local level for service planning and delivery. As Jessica mentioned, we’re currently looking at children and young people’s mental health. This is an area of care where pre-existing challenges have been further exacerbated by the pandemic and we’re interested in where there are inequalities and what’s happened, for example with referrals, over time. We plan to publish some of our research on this in the spring. Then, in the latter half of 2022 we’ll be looking at social care in more detail.
Jessica: At a local level, our NDL work has made it clear that, for this type of data analysis to be useful, the team has to be deeply embedded into the NHS. The Health Foundation’s support has allowed us to truly combine academic and NHS service work for the first time, and it’s been wonderfully successful.
What else can we expect to see in the future?
Jessica: We’re particularly keen that the analytical tools we make are openly available for all to use, and that the analysis we do always has the fundamental goal of improving health inequalities in the population.
Kathryn: Absolutely. We use the principles of team science and open analytics and try to keep everything we’re doing as open and transparent as possible. We share our code openly on GitHub, and present at conferences and different analytical huddles, so there’s a bigger ripple effect to the work we’re doing. That’s something we’d like to ensure we do more of.
As the NDL has progressed – and we’re around 18 months into the programme now – we’re seeing it become more of a network, with partners increasingly collaborating on projects and sharing insights. We’d like to see more of this happening, as well as continued sharing of learnings between like-minded analytical teams working on similar problems. By doing this we hope to showcase the value of analytics and linked data, as well as produce insights that can be used by local as well as national decision makers.
This content originally featured in our email newsletter, which explores perspectives and expert opinion on a different health or health care topic each month.
About 13 mins to read
Exploring the longstanding barriers to widespread use of data and data science in the NHS in...