Associate feature: Driving healthcare innovation
“The test bed function is essentially a model to accelerate innovation in the NHS,” explains Dr David Lowe, clinical lead at the West of Scotland (WoS) Innovation Hub.
“It is a way of being able to bring industry, the NHS and academia together to deliver productive innovation.”
Introducing new technology, services or pathways into the NHS requires evidence to show advantage over what is currently available. It is difficult to acquire this evidence without access to patients or clinical facilities to develop and test prototypes or new models.
Critical, Lowe adds, is “having people with expertise involved from the outset, be it clinicians, e-health, technical expertise or governance – people in the NHS who know how it operates and what is needed.”
The national network of innovation test beds, set up by the Chief Scientist Office of Scottish Government, allows innovative solutions to be tried out in a real-world environment and ensure they are fit for purpose – clinically, financially and operationally – before being brought into wider use. They cover different aspects of NHS care from managing long-term conditions to digitising services, to integrating artificial intelligence (AI) and machine learning.
Having this model in place has been vital in enabling the NHS to adapt and respond to the coronavirus pandemic and embrace new ways of working. For a public body as large as our health service, this is no mean feat.
Lowe’s project is one example: a new COVID-19 assessment app which allows doctors and nurses to complete a standardised assessment of patients. Based at Queen Elizabeth University Hospital, the project was led by the WoS Innovation Hub in collaboration with NHS Education for Scotland and Daysix, an industry partner. Not only does it help clinicians identify suspected COVID-19, it also ensures signs of other illness or disease are recorded. This can be used to improve individual patient care and inform the wider strategy for handling coronavirus.
Lowe says: “In the vast majority of emergency departments in Scotland, we still hand write our notes. Therefore, there’s variation in what’s recorded and definitely for me, legibility is an issue – as with many doctors.
“This is a way of structuring data that’s recorded and pushed into our electronic records system so it can be viewed by clinicians looking after the patient downstream.
“Critically, the other novel thing we built in was a business intelligence dashboard. This allows those running assessment centres to understand the number of patients presenting, their key features, be it their frailty, their home circumstances, right down to their physiology – heart rate, blood pressure – and start to understand the kind of gross characteristics of patients required to be admitted versus discharged.
“That’s helping to support how we think about our resource planning – how many beds we require to be available in the hospital and what the length of stay may be.”
Thanks to the test-bed model, within eight weeks the app went from idea to pilot to implementation in NHS Greater Glasgow and Clyde, with national rollout to follow. The data gathered is also being used to support clinical trials and to help inform patient care as the world deals with second and third waves of COVID.
Lessons learned also have wider application. This kind of technology could be used for any disease or condition for which it makes sense to collect national data, such as stroke, heart disease or COPD. Lowe believes doing this will help to “inform both patient-level care, the clinical pathway for that individual and then the strategic response.”
Innovation test beds mean teams have the resources, connections and time to clear these hurdles for new developments
Capitalising on the use of technology has never been more important for the NHS. It has helped support the response to COVID-19 and in many areas allowed routine care to continue virtually.
Dr Chris Carlin, a consultant respiratory physician, has spent the last two years building an app which supports patient self-management of COPD, a condition affecting approximately 120,000 people in Scotland and the second most common cause of emergency hospital admissions. This project, also led by the WoS Innovation Hub, allows patients to monitor their symptoms of COPD at home through a digital service.
Carlin explains: “They get a text or a Fitbit notification every day to interact with the app and fill out their symptom scores. That’s then visible to the clinicians to support routine care.
“We’ve got self-management resources in the app so they can see what to do if symptoms are worse. Some of that is generic advice, some specified by the clinicians – what antibiotics, what steroid tablets and other things to take if they’ve got a flare-up.
“They can message the clinical team through the app and we can message them – it provides reassurance and supports their self-management.”
Outcomes so far have been “really positive”, he says. Patients have continued to use the app throughout the first year and beyond. In addition, analysis indicates a reduction in hospital admissions and bed days, as well as requiring fewer face-to-face appointments.
The next step is looking at how to use the data to pursue a preventative model of care. Carlin explains: “What we’re doing with machine learning modelling is looking to predict which patients are at highest risk of dying over the next 12 months; coming back into hospital over the next three months; and then the next phase is looking at which patients are developing a flare-up and exacerbation over a 72-hour period.
“With this intelligence we can then take appropriate action – so patients who are at the highest risk of dying, we can target them for anticipatory care planning. Patients at highest risk of coming back into hospital, we can target them for more intensive community interventions.”
COVID-19 reached UK shores about halfway through this project, but the drive to increase remote management and reduce face-to-face interactions means the app felt almost “tailor-made” for the pandemic, Carlin says. “Conditions like COPD put patients at higher risk of becoming severely ill with COVID – reducing the risk of emergency admission for these patients has been crucial.”
Equally, the project has helped to drive forward Scotland’s response to COVID-19. He explains: “The architecture and the governance framework for the COPD support service are what test and trace and negative notifications are built on. It’s accelerated that.”
The need for greater efficiency in diagnostics has become more pressing as a result of COVID-19 too. A project using AI to examine mammogram images could provide valuable insight into AI-enabled clinical decision support.
Dr Gerald Lip, clinical director of the breast screening programme in the North East, is harnessing AI to tackle problems arising from recruitment challenges in radiology. His project is looking at 80,000 sets of mammograms taken between 2016 and 2019 to prove AI can successfully be deployed as a second pair of eyes.
“The position we see in the future, possibly, is that one of the reads [of a mammogram] will be done by the AI and the second read done by a human. There will always be a human doing the reading but using advanced technology can have benefits across the system. This research will go a long way to proving that the AI works in the Scottish population and may allow innovation and early adoption of AI in screening in Scotland,” Lip says.
The main strand of the first year of research was working with iCAIRD, setting up the infrastructure to allow AI to analyse the data. The aim is to ensure patients who come in for a breast screening – all women aged 50 to 70 – continue to receive high quality care, at the same time as freeing up clinicians to spend more time with patients and speed up the turnaround time for results. “That’s often something people worry about. They’ve had their mammogram and then they wait and wait, and it can take up to three weeks for the result to come back. If the AI does it straight away and a human does it the next day, then you’re going to get your report faster,” says Lip.
Three exemplar projects, all transforming the way care is delivered for staff, patients and carers, and harnessing the power of the test-bed model.
Carlin explains: “If you try to put through an individual project, you have e-health evaluations and approvals; you’ve got governance documentation; you’ve got research proposals, approvals, sponsorship, governance, evaluation; you have interactions with the SMEs; you have contracts, finances, programme management and so on and so forth.”
Innovation test beds mean teams have the resources, connections and time to clear these hurdles. “It just unblocks so many of the things that hold back service development and transforms research and innovation into routine care,” Carlin adds.
As the NHS responds to the most significant challenge it has ever faced, accelerating new approaches to care and creating an environment that supports innovation is crucial. Ultimately, it will lead to better care for people in Scotland and around the world.
This special feature was written in association with Scottish Health Innovations Ltd (SHIL), NHS Greater Glasgow and Clyde and NHS Grampian.
The projects referenced are made possible through the expertise, funding, support and collaboration of multiple partners including Chief Scientist Office (CSO), Innovate UK, Industrial Centre for Artificial Intelligence Research in Digital Diagnostics (iCAIRD), NHS Education Scotland, Digital Health and Care Innovation Centre, Technology Enabled Care Team, Scottish Government, ResMed, StormID and DaySix.
A virtual industry, NHS and Social Care Innovation Network event will take place on 26 November, 10am – 1pm. Follow @HSCInnoScot to stay up to date.