A University of Edinburgh report into the NHS AI Lab has identified “promising financial and patient care benefits” while also pointing to barriers to widespread artificial intelligence (AI) implementation and adoption, including a lack of alignment with other NHS systems.
A team of public health, social science, informatics and business researchers from the university’s Usher Institute assessed the NHS AI Lab’s work so far. The team was aided by the NHS Arden & GEM Commissioning Support Unit (CSU).
The NHS AI Lab was established in 2019 by the Department of Health and Social Care (DHSC) and NHS England, with an initial outlay of £143.5m. It is said, by DHSC, to have supported AI research, development and implementation to improve patient outcomes and operational efficiency within the NHS.
The CSU that supported the report has 1,300 staff and 70 clients, which include Integrated Care Boards, NHS England, Primary Care Networks, NHS provider trusts and local authorities. It provided support in the area of health economics for the Edinburgh study.
The research was conducted between March and December 2024, using evaluation techniques that included document reviews, interviews, observations, analytics and outputs measurement.
Kathrin Cresswell, lead researcher on the evaluation, said: “The NHS AI Lab has been instrumental in positioning the UK at the forefront of delivering system-based change to promote AI-driven healthcare. This evaluation provides real-world empirical evidence and learning that can help to shape future efforts in the UK and internationally.”
One AI project the team examined implemented a diagnostic tool in a non-elective care setting across a regional NHS. According to the report, the technology helped clinical staff “make time-critical treatment decisions”, which led to a cost saving of over £44m across a population of 150,000 patients.
The report’s 11 authors said the NHS AI Lab has helped to advance AI development and scaling in healthcare, generating valuable insights and lessons that can help to shape future AI strategies for the NHS.
The findings from this report will inform the ongoing development of AI strategies and approaches that can help the NHS … shape a future where AI will enhance patient care, operational efficiency and overall healthcare outcomes Dom Cushnan, NHS England
As well as Cresswell, the authors include Robin Williams, professor of social research on technology at the University of Edinburgh; Miguel O Bernabeu, professor of computational medicine at the Usher Institute; and Sheena Dungey, senior health economist at the NHS Arden & GEM CSU.
They found early evidence of financial and patient care benefits and of “returns on investment for taxpayers”. But they also identified barriers to “widespread AI implementation and adoption, including procurement processes, integration with existing infrastructures and processes, and the need for stronger alignment with NHS system needs”.
In the report, the authors cautioned: “Research in advanced AI has shown good results. But not all results are available yet, and it is hard to predict how valuable these new technologies will be in the future. Outcomes are also still being reported.”
They went on to say: “Expanding the use of AI has been tough, partly because it is unclear how to buy and use these technologies within the NHS. National plans sometimes did not match what the health and care system specifically needed. For example, many AI projects focused on popular areas like diagnostics but did not always address bigger system needs, such as helping with resource planning.”
They also added the caveat that “progress has been slowed by big changes, such as the Covid-19 pandemic, changes in leadership, government spending reviews, and restructuring within government. Many groups were involved in coordinating the AI Lab, often with conflicting goals, which made things more complicated”.
They concluded that their report “emphasises the need for sustained national support, strategic leadership and evidence-based decision-making to ensure AI’s full potential is realised in healthcare” and “highlights the importance of fostering positive collaboration between AI developers, policymakers and frontline healthcare providers”.
Dom Cushnan, director of AI, imaging and deployment at NHS England, said: “The findings from this report will inform the ongoing development of AI strategies and approaches that can help the NHS to make the strategic shift from analogue to digital in health and care, helping to shape a future where AI will enhance patient care, operational efficiency and overall healthcare outcomes.”
Rose Taylor, executive director of health and care transformation at NHS Arden & GEM CSU, commented: “This evaluation demonstrates the important role that AI can play in the transformation of NHS services. The health economics approach taken in the review has enabled systems to demonstrate that AI technologies can deliver benefits for patients while simultaneously providing productivity and efficiency gains.”