The new data-analysis centre brings together South East Scotland’s health and social care data for the first time, driving forward data-driven approaches to improving care
The importance of data in directing health and social care delivery has been highlighted by the coronavirus pandemic. We spoke to Colan Mehaffey, Chief Operating Officer for the Health and Social Care Sector at the Usher Institute, and Dr. Atul Anand, senior clinical research fellow and consultant geriatrician at the University of Edinburgh, about the health and social care data repository they are creating: DataLoch.
The health and care sector is facing a variety of long-term challenges which have been compounded by the coronavirus pandemic: an ageing population, more people with long term conditions, and increasing drug costs, to name a few. Data driven decision making has been key in navigating the return to ‘normal’ life in the context of a pandemic, and the creators of DataLoch believe data will be vital for making all kinds of decisions in the health and social care sector in future.
The goal of DataLoch’s partners, including NHS Boards, Local Authorities and the University of Edinburgh, is to create a data repository linking health and social care data for Edinburgh and South East Scotland. The venture, started in 2019, predates the pandemic and aims to bring data driven approaches to a wide range of ordinary functions of the healthcare system including prevention, treatment, and service provision. These approaches, fuelled by DataLoch, aim to improve outcomes for people and reduce health inequalities.
We already have it all
The data that will populate DataLoch are ‘routine’ data – most often they are scattered and fragmented. As Mehaffey explains, “the best way to put it is that a lot of these data would never have met each other before. Take the two big categories, health care data and social care data, we don’t routinely link those two data sources.”
Even within these two overarching categories there is a lot of fragmentation. Under the umbrella of ‘healthcare data’, there is primary data, collected by GPs, secondary care data, collected by hospitals, and disease specific datasets. As Mehaffey elaborates, even “primary healthcare data and secondary healthcare data are not being routinely linked, which seems shocking to most people.” Dr. Anand goes on to explain that “we have an estimate of over 200 different data sources that we are looking to combine within health care.” Adding the layer of fractured social care data, which can be held by councils, community centres, care homes, and many other organisations, demonstrates the complexity of the problem DataLoch hopes to solve.
The disjointed nature of data in health and social care directly affects patients’ experiences. This is illustrated by Mehaffey: “Say you were severely asthmatic. Chances are most of your interactions would be with your GP, with your pharmacist and you would be given occasional referrals to a hospital environment, to see a specialist. You might be admitted to hospital because you had a severe asthma attack. We wouldn’t routinely link all those data together to know how we could manage your case better or how environmental factors might be playing a role. All the information is there, it is just not routinely linked.”
The notion that “all the information is there, it is just not routinely linked” underpins the power of what researchers call routine data, or more mysteriously, dark data. It is these data that researchers such as Dr. Anand want to mobilise to improve outcomes for patients: “the idea behind dark data is that there are lots of data that we routinely collect but never look at. That we never try to learn from, improve practice, or gain insights from. By systematically making data available to researchers and academics there may be opportunities to improve care without collecting additional data, meaning no additional effort on the part of the NHS, just using what was already sitting there.”
Benefits of linking data
Creating a DataLoch that links these data sources will have various advantages. Dr. Anand discusses how this would improve the research process: “having a single resource that brings together health and social care data would be a powerful way of doing research a lot quicker and in a much more agile way. For future projects there wouldn’t be a need to individually link lots of data sources because everything would be pre linked and ready to go.”
Beyond the practical advantages of being able to conduct research more rapidly, the quality of the research could also improve as data from all types of people becomes available. Discussing his PhD research which explored heart attacks, Dr. Anand explains the importance of inclusivity in research: “Rather than using a traditional trial method and recruiting individual patients, which is quite time consuming and costly, we did a cluster randomised trial where we included everybody attending ten hospitals in Scotland with suspected heart attacks. That is quite a powerful way of doing clinical research because it means you can include everybody. I am a geriatrician, and often the people that get excluded from normal trials are older people, frailer people, people with dementia, people who can’t consent to trials. A lot of our evidence in medicine comes from trials with quite selective people who can include themselves in research.”
The DataLoch will also be integral for the improvement of service delivery within the healthcare system. The goal is to create a learning healthcare system in Scotland, which Mehaffey describes as “a system where knowledge generation processes are part of your daily practice and produce continual improvements in care.” Data will be indispensable in these knowledge generation processes.
Finally, Mehaffey discusses how data can be used to train a new generation of health care workers and innovators: “having good quality synthetic or learning data sets that people can work on is a big benefit. People who want to innovate can use synthetic data sets to develop innovative health and social care products. Because we are embedded in the NHS and with local authorities, we would hope that they can then be productionised quickly. A lot of innovation will fall by the wayside, that is its nature, but you give it a better chance of survival by engaging directly with it.”
Local to global?
Mehaffey emphasizes that DataLoch is first and foremost a local project aimed at benefiting local people: “We have to focus on what we do for the city region first. That is number one – what is delivered for the city region and its citizens.” Although the local area is the primary priority, he also discusses the long-term scalability of the work being done: “Across Scotland we have 14 health boards and 32 local authorities, so we hope to feed into bigger initiatives. But that takes a long time to do at that scale, so we saw the opportunity to do this in a very scalable way, to develop the best practice technologies or approaches that could be scaled up.”
The impact of the research DataLoch enables is already being felt locally. Dr. Anand discusses the project’s rapid response to the Covid-19 pandemic in the city region: “what we have done is created a single repository of Covid-19 data and that comes from lots of different sources and is growing day by day. It is hospital records, community records, swab test results from hospitals and community centres and it has been augmented by a clinical review panel. The clinical panel has reviewed every patient that has been admitted to hospital with Covid-19 and gone through the hospital records to provide even more clarity around the diagnosis.” There are already around 20 different research projects, whose findings will surely influence local policy in the coming months, using this repository to answer a wide variety of questions about the virus.
As Mehaffey highlights, while the focus remains local, there is potential to “engender and enable international collaborations” to benefit the local and global community. For example, Dr. Anand discusses the European wide Ehden Project: “The idea is that within different Covid-19 datasets across Europe there are likely to be very similar data points collected, they are probably called something different, their structure is probably different, but they can be harmonised. By participating in these international common data model collaborations, we get better insights from larger data sets. It gives local academics access to a much bigger community, allowing them to ask much larger questions.
The vitality of data driven decision making in improving health and care provision has come to the fore in 2020 and the creators of DataLoch hope that their project can help with the continuous improvement of this sector for years to come.
Beyond the practical advantages of being able to conduct research more rapidly, the quality of the research could also improve as data from all types of people becomes available
Read the latest Case studies
Many people with neurodivergent conditions such as autism, ADHD, dyslexia and dyspraxia face barriers to…
A world-first data science qualification for school-aged learners has been developed in Scotland. As part…