UCLH Board directors announced
26 March 2021
Publish date: 19 March 2021
University College London Hospitals NHS Foundation Trust (UCLH) and Microsoft have announced a new strategic research and innovation collaboration to develop and apply advances in artificial intelligence (AI) and digital technologies for the future of healthcare delivery in the NHS.
The collaboration will see Microsoft Research work alongside UCLH research and clinical teams to develop AI algorithms and applications which will be tested and deployed at UCLH to assess data driven improvements in efficiency, patient safety, experience and outcomes (subject to the appropriate regulatory approvals). Until now, most AI-driven algorithms have been developed remote from the environment in which they are designed to be deployed.
The teams will evaluate the use of machine learning models to help with the allocation of hospital resources, manage patient demand, adapt to emerging care pressures, and augment and accelerate clinicians’ ability to perform radiotherapy planning for certain types of cancer.
The collaboration will see UCLH and Microsoft teams work on a range of projects. Each research programme will have a specific agreement that will govern the use of patient data which will be in accordance with current regulatory guidance. It is expected that most projects will not use identifiable data. Where programmes need to use identifiable data, it will only be done with patients’ explicit consent and will follow Health Research Authority (HRA) and other applicable guidance.
Professor Bryan Williams, UCLH Director of Research and Director of the Biomedical Research Centre at UCLH said: “This is a very exciting development and there is so much opportunity. The best way to realise the real potential of AI for healthcare and develop the right solutions for the NHS, is to create these solutions working together on the ground, in the hospital where these innovations need to be deployed. In this way we move beyond talking about the potential of AI and make it work for patients in the NHS.”
Professor Marcel Levi, UCLH chief executive, says: “Recent experience of the pandemic has demonstrated the importance of being able to quickly adapt to ensure that our clinical expertise remains focused on providing the best possible patient care. The collaboration with Microsoft provides UCLH with the opportunity to harness the most innovative technologies to develop and enhance what we do. Machines will never replace clinicians and staff, but the use of data, expertise and technology can radically change how we manage our services – for the better.”
Chris Bishop, Lab Director, Microsoft Research Cambridge, said: “We are excited to embark on this work with UCLH to co-create research technology that has the potential to have a significant impact on healthcare services in the NHS. Partnering to design principled AI technologies to be deployed in clinical practice is vital to guarantee that they meet the real needs of clinicians and their patients. UCLH, and our other clinical partners in the UK, are ensuring that the NHS remains at the forefront when it comes to digitally transforming healthcare.”
Improving patient flow
In one of the first projects, UCLH and Microsoft will explore how AI can be used to predict hour by hour and day by day fluctuations in demand for acute hospital care in UCLH’s hospitals. In doing this, researchers aim to reduce hospital strain, improve patient outcomes and look after staff wellbeing.
Working closely with hospital staff to understand their everyday practices and their challenges, the team will explore where AI can provide real time insights to help hospital managers and frontline staff to allocate their resources and adapt to ever changing care pressures – pressures which have only been heightened during the pandemic.
UCLH will leverage Microsoft’s open-source InnerEye technology to build and refine its own proprietary medical imaging AI models designed to automate time consuming radiotherapy preparations and speed up cancer treatment.
Project InnerEye develops machine learning techniques to support clinicians and help enable hospitals to meet growing demands on healthcare; help deliver precision medicine for better patient outcomes; and understand how we can combine medical imaging features with other types of data to change the way we do medicine today.
Approach to the collaboration
UCLH and Microsoft’s collaboration combines teams of researchers, software engineers, clinical staff and hospital managers, UCLH’s Digital Healthcare team and patients, to work together on a range of projects.
Established principles of the collaboration include:
UCLH combines excellent specialist healthcare with ground-breaking biomedical and clinical research and is working on translating that research, through partnership, into real world impact. The collaboration with Microsoft is a major step towards this.
While UCLH clinicians and analysts have expertise in patient data systems, working with Microsoft will give UCLH unprecedented access to AI technologies and expertise.
Until now, most AI-driven algorithms have been developed away from the environment in which they are designed to be deployed.
To realise the potential of AI for healthcare and develop the right solutions for the NHS, we believe the best approach is to create solutions working together on the ground, in the hospital where these innovations need to be deployed, subject to the appropriate regulatory approvals.
Over the last 3 years, UCLH has been working closely with Microsoft on specific projects. Some of UCLH’s IT systems are also Microsoft-based.
Microsoft Research Ltd brings extensive experience in computing, AI and cloud technologies and Microsoft’s cloud platform, Azure, offers a particularly broad and powerful range of capabilities for UCLH.
Microsoft researchers involved in the research programmes will be UCLH honorary employees and certified to UCLH data management and governance standards. Microsoft researchers will undertake collaborative research on UCLH premises and on UCLH research platforms. Each research programme will have a specific agreement that will set the appropriate legal frameworks, governance and controls required for Microsoft researchers to access any data for research purposes.
Data will be used to train technology to use large datasets to develop and test models with the potential to aid the diagnosis and treatment of common and complex medical conditions, as well as improving the overall efficiency of hospital operations, quality of care and systems (subject to the appropriate regulatory approvals).
Data will be securely stored in an environment governed by UCLH. Access to data will be managed by access only being granted to UCLH accredited research team members working on the projects. These research teams will involve clinical, managerial and academic staff at UCLH as well as researchers from Microsoft Research Labs. The research team members will have UCLH contracts and will all be working according to UCLH policies for the safe use of clinical data. No personal or patient data will be transferred outside of the environments governed by UCLH. Data may be aggregated for reporting on the research.
Data is securely stored on Microsoft Azure, with data being encrypted at rest and in transit. Access is managed by the principle of least privilege to only provide access to data where required based on a person’s accredited role. Microsoft has corporate and local policies and procedures to manage access, due diligence and if needed, reporting issues.
No. UCLH will own and retain full rights, Microsoft will only be granted access to the data for the duration of the research projects.
The research programmes will have specific agreements that will set the appropriate legal frameworks, governance and controls required for Microsoft researchers to access any data for research purposes. These agreements will govern consent to use of patient data. Where programmes need to use identifiable data, it will only be done with patients’ explicit consent, and Health Research Authority (HRA) and other guidance on use of patient data will be followed.
Dr Junaid Bajwa is both a non-executive director at UCLH and Chief Medical Scientist at Microsoft Research. Arrangements were in place, and remain in place, to ensure that Dr Bajwa has not been, and will not be involved in any of the planning, negotiations, agreements or future research agreed under this collaboration between UCLH and Microsoft, whilst he holds the position of non-executive director at UCLH.
It has ensured governance of the process through the UCLH Microsoft Oversight Committee. This is a subcommittee of the UCLH Board of Directors and is co-chaired by two non-executive directors.