Q&A: Dr Erik Mayer, Digital Collaboration Space
The Digital Collaboration Space at Paddington Life Sciences is a new research hub that adds value to the vast amount of data routinely collected across Imperial College Healthcare NHS Trust.
The space houses the digital health team from the National Institute for Health and Care Research (NIHR) Imperial Biomedical Research Centre (BRC), a partnership between the Trust and Imperial College London. They provide state-of-the-art management and analysis of the huge amount of health data collected across Imperial College Healthcare’s five hospitals and collaborate with the wider North West London Integrated Care System on a complementary data set from a diverse population of 2.4 million people.
Following the formal launch of the space this month, we spoke to Dr Erik Mayer, NIHR Imperial BRC digital health team leader:
Q: Tell us a little about the Digital Collaboration Space
Every year at the Trust, we have about a million new patient clinical events. We went fully live with our electronic patient record system in 2014, so since then we’ve collected over 2 billion rows of structured data across a consolidated database, and 100s of millions of unstructured records. It’s a vast set of high quality, granular data that is really valuable for researchers.
The Digital Collaboration Space allows multidisciplinary teams to come together in one place and, with the right approvals, securely use anonymised data to inform research questions and deliver on research plans.
Our research informatics team previously supported this kind of activity, but the most important part of our new space is that we now have a central place to come together. Sitting within Paddington Life Sciences we’re co-located, not only with St Mary’s hospital, but with our industry partners. With one vision across healthcare professionals, academia and industry, we have everyone together, from doctors, nurses and community partners, to information governance, data scientists and data engineers.
Being co-located like this allows us to work much more effectively with all stakeholders right from the start of a project. We’re able to help inform the planning and design of projects and be an integral part of the end-to-end process to ensure the project’s success.
Q: How does it work in practice?
Essentially, what we’ve done is made sense of roughly 2,500 tables at the back end of our electronic patient record, so that we can curate patient records and transform them into research-ready data.
An example could be for some research on diabetes, where a researcher may need data on all patients with uncontrolled blood sugars, of a certain age, and perhaps other factors. We’ll pull together a data table for them that includes all the data points and allows them to complete that research.
Since the informatics function was first formed in 2019, we’ve created a standard data set which answers 80% of the research questions that come up. And of course, as more requests come in, we build on that further.
We have an agreed set of projects and research priorities set by the Trust, but we also have additional data requests that come in via our monthly data access committee. The committee has representation from all the core stakeholders: clinicians, academics, data protection, community partners and so on. The committee then decides on whether the research is in the public interest, and whether the data requested is appropriate and proportionate. A lay summary is written up and published on our website, so that it’s all very transparent.
Q: Can you explain a bit about your data security measures?
We’re classed as a secure data environment (SDE). SDEs follow the ‘Five safes’ framework: safe people, safe projects, safe settings (including safe compute), safe outputs, and safe data. We also support safe return to ensure that we can link research back to patients where appropriate and supported by direct care use cases and or patient consent.
As part of the process, we gained overall database ethics approval from the Health Research Authority. That means that anyone that uses the de-identified routinely collected health data will automatically get ethics approval for their work, once fully approved by the data access committee.
One of the key differences about the new approach with SDEs, is that an extract of data won’t just be given to someone to take away (as was often the case with data for research in the past). The data is accessed in a very controlled way, through our SDE, which sits within the NHS firewalls. Within the environment there is no way to access emails or the internet. If a researcher needs to take away aggregated tables or graphs, there is what’s called an ‘air lock’ system, which is another layer of checking to ensure there’s no patient-identifiable data. This is key to having public confidence for re-using patient data because it ensures we’re being good and fair custodians.
Q: Are there other benefits, outside of those to researchers?
One of our unique aspects, compared to other SDEs, is our real time data. Many SDEs work with quite static data sets that can be used for research and decision making such as commissioning. But our data has also been used to support frontline clinical decision making.
Because we take both structured and unstructured data from patient records and essentially ‘make sense’ of it, we’re then able to push better data back into the front end of the system that clinicians use.
An example is a project we did to support a smoking cessation programme. Smoking status is typically recorded in a very fragmented way and data can often be out of date. We were able to combine and cleanse data in a way that meant pharmacists were better able to identify smokers and proactively offer them support to quit. Another piece of work we did informed changes to a risk assessment for blood clots that meant clinicians were able to use the risk assessment much more effectively.
Q: What are your hopes for future developments in the way your team works?
One of the things I think we can do better in the future is being more proactive. What we’ve spoken about so far is people coming to us for support, but we also have a lot of knowledge of where the problems are and what we can do to address them. That’s where organisations like MedCity are really valuable. Although we have a good local, regional and even global network of our own, we don’t always know which are the best industry partners to work with on a particular problem. MedCity is very well placed to support that kind of activity as a middle partner to help identify industry partners for work we’ve identified. That’s the type of activity I hope we’ll be doing more of in the future.