Transformation Directorate

AI Skunkworks projects

The NHS AI Lab Skunkworks team ran short-term projects to investigate the use of AI for improving efficiency and accuracy in health and care.

The Skunkworks team looked to help the health and care system, through practical experience, to understand, build, buy, deploy, support, and challenge AI solutions.

Multiple NHS health icons

Here’s a quick summary of some of the ways we worked on this vision with colleagues in the health and care sector:

Using AI to find optimal placement schedules for nursing students

This project explored whether AI could automatically generate student nurse placement schedules that met the requirements needed by the hospital, while also providing a range of experiences for students.

Read more about the nursing placement schedules project

Arranging a student nursing schedule is complex. There are a range of needs which require consideration before a student nursing schedule can be developed. For example, the universities will have set dates when placements should take place. Some universities will require students to visit specific types of wards over the course of their placements, while others may be more flexible.

The Trust has requirements too. Each ward at a hospital has a maximum student hosting capacity, and this can vary depending on the student’s current year of study. Wards must also keep to internal Education Audit Standards to ensure they can provide necessary support to students.

We worked with Imperial College Healthcare NHS Trust, in conjunction with North West London CCGs, to develop an AI-driven solution to this problem.

Deep dive AI workshops - sharing skills and experience

This series of 5 workshops aims to provide organisations with the relevant knowledge and tools to understand how to safely launch an AI experiment in healthcare.

Read more about an AI deep dive workshop with the Digital team at University Hospital Southampton

These cost-free sessions are aimed at individuals within public sector health and care organisations who are interested in understanding and implementing AI. We welcome everyone from clinicians and technology teams to operations and senior stakeholders.

The Digital team at University Hospital Southampton (UHS) approached us with a request to explore the ethical and safety considerations of applying AI in their work. It is especially critical with AI in health and care that the people affected by its use are confident the tools are robust, and any support for decisions is fair for all patients.


Read the full deep dive workshop case study.

Exploring how to create mock patient data (synthetic data)

This project aimed to provide others with a simple, re-usable way of generating safe and effective synthetic data to be used in technologies that improve health and social care.

Read more about the synthetic data project

One of the challenges for AI development is that without suitable test data it is not possible to properly demonstrate AI tools, preventing users without data access from being able to see the tool in action. However, using real patient data for research and development carries with it safety and privacy concerns about the anonymity of the people behind the information.

In a partnership project, the NHS Transformation Directorate’s Analytics Unit and the NHS AI Lab Skunkworks team sought to further improve an existing synthetic data generation model (called SynthVAE) and develop a framework for generating synthetic data that could be shared for others to re-use.


Read the full synthetic data generation case study.

Identifying features in CT scans with George Eliot Hospital

This challenge was suggested by a team at George Eliot Hospital who wanted to speed up the analysis of computerised tomography (CT) scans to free up radiologists’ time and help identify tissue growth.

Read more about the CT scan project

We worked with the team at the George Eliot Hospital to research how to identify features in a computerised tomography (CT) scan and automatically align their scan "slices" to enable early detection, diagnosis and, later, the treatment of lesions (tissue growth).

Comparing computed tomography (CT) scans is a labour-intensive task and there are no satisfactory automation tools that can speed up the process. Difficulties in aligning scans 100% accurately make it hard for radiologists to accurately measure changes.

The project aimed to support radiologists by comparing two CT scans taken at different dates, to see if a patient has improved or deteriorated. We explored how AI might identify organs and lesions, report the change in size, and highlight areas of concern to the radiologist.


Read the full CT scan alignment case study.

Predicting long hospital stays with Gloucestershire Hospitals

In this project we used admission data and pseudonymised patient records to investigate whether artificial intelligence machine learning methods could identify the patients most at risk of becoming “longstayers” (hospital stays of more than 21 days).

Read more about the long stayers project

Many patients who stay in hospital for extended periods experience worse health outcomes than patients whose stays are shorter. Long stays also create problems for busy hospitals because beds stay occupied for longer and require extra resources to manage. In this context, a long stay is regarded as any stay longer than 21 days.

Long stayers at the Gloucestershire Hospitals Trust occupy an average of 278 beds per day, which is around 4% of all admissions but accounts for 34% of bed use. It is possible that identifying and intervening early could make a real difference to these patients.

The Business Intelligence team at Gloucestershire Hospital (GHFT), supported by GHFT's CIO and senior clinical leaders, developed an idea to use artificial intelligence to address the issue of “long stayers”, and applied to create a proof of concept with NHS AI Lab Skunkworks.

Bed occupancy management with Kettering General Hospital

This project with Kettering General Hospital will look at using AI to improve bed scheduling in hospitals. Using historic bed occupancy data we will apply optimisation methods to create a model that explores different ways of allocating patients to beds.

Read more about this bed occupancy project

This project investigated whether AI can support hospitals to manage bed occupancy more efficiently in order to benefit both patients and staff. The aim is to enable staff to schedule the best use of beds more quickly, by providing more information to the scheduler and guiding decision-making.

Fewer bed moves can mean a better experience for the patient as well as cost savings for the hospital. The optimisation method approach taken by this project will allow us to put patient welfare first.

This project created a proof of concept system that presents allocation options visually to the ward staff, alongside an explanation of the different factors and reasons involved, employing explainable AI (also known as XAI, which means ensuring a human can understand the path an AI system took to reach its decision). This also makes sure that the medical team are the final decision-makers, with AI playing a supporting role to keep the “human in the loop”.


Read the full bed allocation project case study.

Developing the health data search engine, Data Lens

In this project, we applied Natural Language Processing (NLP) and other AI technologies to test a prototype data search tool that would be a universal search engine for health and social care data catalogues and metadata.

Read more about Data Lens

Health data is held on numerous incompatible databases across different organisations. Analysts and researchers wanting to source relevant health data face a time-consuming and difficult task finding what they need, accessing and understanding it.

This project aimed to provide information for analysts and researchers from multiple sources across the health and care sector with one search.

The resulting tool, Data Lens, joins up data catalogues from NHS Digital, the Health Innovation Gateway, MDXCube, NHS Data Catalogue, PHE Fingertips and the Office for National Statistics. It gives improved access and supports increased collaboration by providing user-friendly access to separate data catalogues with one search, providing multilingual support.


Read the full Data Lens case study.

Predicting negligence claims with NHS Resolution

This project investigated whether it is possible to use machine learning AI to predict the number of claims a trust is likely to receive and learn what drives them in order to improve safety for patients.

Read more about the NHS Resolution project

NHS Resolution provides expertise to the NHS on resolving concerns and disputes. The organisation holds a wealth of historic data around claims, giving insight and valuable data around the causes and impacts of harm.

The NHS Resolution team wanted to understand whether AI methods could be applied to their data to better understand and identify risk, preventing harm and saving valuable resources.

We aimed to prove the value of machine learning in determining insights from the available data. Automated machine learning was used to run repeated processes on the available data in order to select the AI models that uncovered the most relevant information.


Read the full NHS Resolution case study.

Recruitment shortlisting with the NHS England London Talent team

This project looked at the issue of bias in using AI to help compare and review job descriptions and applications. It will test how algorithms perform while placing careful scrutiny on the issues of ethics, equality and inclusiveness.

Read more about this recruitment shortlisting project

Recruitment for an organisation as large as the NHS is a time-consuming and expensive operation. If artificial intelligence can successfully manage bias while improving the speed and efficiency of selection processes, this could lead to fairer opportunities, greater inclusivity and reductions in time and cost.

The project began some research to explore the various existing approaches to using AI to solve this problem, from chat bots to CV screening, and automated decision-making processes to decision-making support tools, looking at the advantages and disadvantages they offer.

This will allow the London Talent team to make an informed decision about what type of solution might be suitable for the NHS, and the possibilities for overcoming bias in this context. We will then use pseudonymised applications for closed job vacancies to train and test a model to see if a solution can be found that accounts for bias. Pseudonymisation separates data from direct identifiers (e.g. name, surname, NHS number) and replaces them with a pseudonym (for example, a reference number), so that identifying an individual from that data is not possible without additional information.


Read the full recruitment shortlisting case study.

Identifying and quantifying Parkinson’s Disease using AI on brain slices

This project looked at developing an approach to enhance the identification of biomarkers which are indicative of Parkinson’s Disease, and explore whether automated identification of Parkinson’s Disease in these slices is possible.

Read more about this quantifying Parkinson's Disease project

Identification of Parkinson’s Disease is carried out by neuropathologists who analyse post-mortem brain slices. This process is highly time intensive, and the neuropathologists are highly trained in their field. Being able to look at introducing automation to this process has potential to increase the speed at which Parkinson’s Disease can be diagnosed in a brain, as well as freeing up neuropathologists who are otherwise required to spend hours looking at the brain slices themselves.


Read the full quantifying Parkinson’s Disease case study.

Using deep learning to detect adrenal lesions in CT scans

This project explored whether applying AI and deep learning augment the detection of adrenal incidentalomas in patients’ CT scans.

Read more about the adrenal lesions in CT scans project

Many cases of adrenal lesions, known as adrenal incidentalomas, are discovered incidentally on CT scans performed for other medical conditions. These lesions can be malignant, and so early detection is crucial for patients to receive the correct treatment and allow the public health system to target resources efficiently. Traditionally, the detection of adrenal lesions on CT scans relies on manual analysis by radiologists, which can be time-consuming and unsystematic.

The main aim of this study was to examine whether or not using AI can improve the detection of adrenal incidentalomas in CT scans. Previous studies have suggested that AI has the potential in distinguishing different types of adrenal lesions. In this study, we specifically focused on detecting the presence of any type of adrenal lesion in CT scans. To demonstrate this proof-of-concept, we investigated the potential of applying deep learning techniques to predict the likelihood of a CT abdominal scan presenting as ‘normal’ or ‘abnormal’, the latter implying the presence of an adrenal lesion.


Read the full adrenal lesions in CT scans case study

Working with a Trusted Research Environment

This pilot examined using Trusted Research Environments to answer key and important questions for a defined project

Read more about about this project using a Trusted Research Environment

Trusted research environments (TREs) take the form of a secure data environment that allows analysts and researchers to undertake in-depth analysis on rich, joined-up datasets without them seeing any identifiable information. Data is held within a secure server and does not leave that server. Following a recommendation from the Goldacre review TREs form a key part of the Data saves lives: reshaping health and social care with data policy paper.

sing TREs for analytical projects is still a new concept to many analysts. The NHS AI Lab Skunkworks team and NHS England’s Health Inequalities and Evaluation Analytics Team embarked on a partnership project to gain experience of using a TRE to answer key and important questions for a defined project. They used a TRE called OpenSAFELY, an open-source software platform for analysis of electronic health care records data for COVID-19 related research.

Kidney deterioration prediction with University Hospitals Leicester

This project is looking at whether artificial intelligence can help to predict the suddenly deteriorating health of people with acute kidney injury, where your kidneys suddenly stop working properly.

Read more about this project with University Hospitals Leicester

People with acute kidney injury (AKI) sometimes experience sudden deterioration and need emergency dialysis or intensive care. Current electronic warning systems can alert healthcare staff shortly ahead of deterioration happening, but this project will explore whether artificial intelligence can be used to predict this further ahead.

University Hospitals Leicester (UHL) has access to a large collection of patient observation data that may allow us to find patterns indicating when sudden deterioration might happen, and the factors that enable its prediction. UHL currently use a team of nurses who specialise in evaluating patients for signs of sudden deterioration, but if this data can allow hospitals to automatically identify those patients most at risk at least 24 hours beforehand it will allow medical staff to focus their time on these patients, and potentially prevent the need for admission to the intensive care unit or emergency dialysis.

Clinical coding automation with the Royal Free and Kettering General

Data scientists in the AI Lab Skunkworks team and the NHS Transformation Directorate Analytics unit are supporting this project to investigate whether the process of clinical coding (applying standard code words to health records) can be supported by artificial intelligence.

Read more about this clinical coding project

When you visit your doctor or attend hospital a lot of information is collected about you on computers, including your symptoms, tests, investigations, diagnosis, and treatments. Across the NHS, this represents a huge amount of information that could be used to help us learn how to tailor treatments more accurately for individual patients and to offer them better and safer healthcare. The challenge is that most of the information held within these records is in written form that is difficult to use.

The process of reading health records and applying standardised codes based on particular words, conditions or treatments, is called "clinical coding". The process of clinical coding is time-consuming, expensive and carries the risk of mistakes.

We are providing data science capability to a joint project with the Royal Free Hospital and Kettering General Hospital. This project aims to understand which open source models are best to support clinical coders by automating part of the clinical coding process using natural language processing (NLP) to teach computers to ‘read’ electronic health records. The aim is for the technology to summarise and suggest the standardised codes that will then be checked by clinical coders.

NLP is a branch of AI used to interpret unstructured text data, such as free-text notes.