Digital Health

Communities in north Westminster, north Kensington, and south Brent are amongst the 20% most deprived areas in the country, with significant health outcome disparities and an 18-year  life expectancy gap between the most and least deprived. Alongside these public health challenges, the pandemic has increased the demand for innovative digital solutions that can enhance the resilience of the healthcare system and support patient engagement. Our research is responding to these challenges by generating data-driven insights and digital solutions to address the most pressing health priorities and inequalities – for example, in diabetes, mental health, and multi-morbidity.

Three programmes of research aim to improve healthcare delivery using digital innovation and AI technologies:

• The Translational Data Analytics workstream brings together information across different electronic health record systems to provide clinical decision support to front-line healthcare staff – ensuring they have the right information, in the right place, at the right time. Robust evaluation of our innovative decision support tools will include assessing the impacts on clinician behaviours, productivity metrics, patient experience, health inequalities, and clinical outcomes.
• The AI Testbed for Digital Health programme will create the digital research infrastructure for safe, effective and ethical adoption of AI-driven technologies in health and social care. We will apply AI to image and omics data in routine care for early diagnosis and precision medicine, and develop new-generation Interpretable AI technologies that are understood and trusted by patients and clinicians.
• The Real-World Evidence workstream aims to develop new approaches to increasing the efficiency of clinical trials and improving recruitment in under-represented population groups. De-identified routinely-collected data, captured during the patient journey through the health system, will enable us to evaluate real-world adoption of interventions to inform complex healthcare delivery problems.

Inclusive Recruitment

Diverse NHS staff are essential to the delivery of high-quality and person-centred services, however, the COVID-19 pandemic has shown that Black and Minority Ethnic (BME) staff have very different and unequal experiences of the NHS as a workplace, with less representation in more senior roles. To address this, Imperial College Healthcare NHS Trust has implemented an inclusive recruitment programme which includes mandatory diverse interview panels and requires all hiring managers to write a letter to the Trust Chief Executive Officer (CEO) to justify their chosen candidate. Working with the Trust’s workforce team, we are using advanced analytics and natural language processing to analyse routine recruitment data and the Letters to the CEO, providing real-time insights into the effectiveness of this inclusive recruitment programme to support the Trust in achieving its EDI objectives.

Falls

Every day the NHS spends around £6 million on falls; falls occurring in hospitals account for 25% of this expenditure. The human cost of falls ranges from distress and loss of confidence to head injuries, bone fractures, and even death. Our research will use routine data from electronic patient records and incident reporting systems to better understand why patients fall in hospitals, with a particular emphasis on under-represented groups (e.g. patients with English as a second language). We will develop and evaluate an integrated informatics platform that applies Natural Language Processing to falls documentation to provide semi-automated, near-real-time reports to Imperial College Healthcare NHS Trust for coordinated safety monitoring and improvement.

Clinical Decision Support in Ovarian Cancer

Ovarian cancer is a devastating disease with most patients presenting at an advanced, incurable stage. There is a wealth of data held across NHS systems that could help improve outcomes for such patients but it is often written in clinical notes across the electronic health record, and not held in a single location. This makes finding all the relevant information quickly challenging, with the potential for some aspects to be missed.

Our team of healthcare professionals, researchers, and data scientists have built a new system using advanced text analysis called natural language processing bringing important and relevant clinical facts together into a single location such that it can be easily accessed by the multi-disciplinary cancer team. We are now evaluating the impact of the system on delivering benefits for patients and their outcomes, improving data quality, and saving money through improved processes and time savings.

Increasing real-world clinical trial capacity

The use of electronic health records (EHRs) can support the conduct of randomised clinical trials (RCTs) in conditions closer to usual clinical practice. This project will expand consultation with patients and citizens on the acceptable use of their healthcare records for clinical trials within a safe and secure data environment and deploy clinical trial technologies for the design, recruitment, and conduct of studies embedded within digital primary and secondary care clinical records.

Real World follow-up of interventions

The project aims to use linked healthcare data and additional regional/national data sources to supplement clinical trial data to provide evidence of effectiveness in interventions where good data are lacking. The goal is to understand health and digital inequalities, unintended consequences, and the holistic impact of healthcare interventions by taking a population approach. One exemplar project will evaluate the impact of a digital intervention for a mental health condition on adverse events and improved outcomes for a patient group in the community.

Optimising health and social care systems

This project aims to use integrated machine learning and optimisation approaches to improve the delivery of high-quality, safe care and reduce inequalities for patients in North West London. The project will focus on analysing patient journeys through the health and social care system, characterising delays and inequalities, and developing an open-source framework to optimise pathways based on routinely collected integrated health data.

Giving Parents a Voice: A Digital Tool to Capture the Stress of Having a Baby in Intensive Care

What is this project about?

When a baby is born prematurely or becomes seriously unwell, they may need specialist care in a Neonatal Intensive Care Unit (NICU). This experience can be one of the most stressful and emotionally overwhelming events a parent can face. Yet hospitals do not routinely record how parents are feeling in any consistent or structured way, meaning the emotional and practical challenges families experience are largely invisible to researchers and care teams. This project, developed together with Pregnancy and Prematurity theme, tested a secure digital questionnaire that parents can complete on their phone, tablet, or computer, allowing them to share how stressful their NICU experience has been and linking those responses directly to their baby’s medical records in the UK’s National Neonatal Research Database (NNRD).

Why does it matter?

Around one in ten babies born in the UK are admitted to a neonatal unit, meaning this issue touches a huge number of families every year. Black and Asian families are disproportionately more likely to have a premature baby, and families from disadvantaged backgrounds often face additional burdens — including travel costs, language barriers, and limited support networks. By capturing parental stress in a standardised, digital way and linking it to babies’ clinical data, this project creates a powerful new resource that could help researchers understand how the NICU experience affects families, identify which groups need more support, and ultimately improve the care and wellbeing of both babies and their parents.

What are the outputs of the project?

A validated digital version of the Parental Stressor Scale: NICU (PSS:NICU) was successfully developed in collaboration with the Imperial Clinical Trials Unit (ICTU) Data Systems Team, clinicians, and psychologists, with embedded electronic consent and automated reminders built in. Three NHS neonatal units in North West London implemented the tool, and over a five-month period, 34 parents registered interest, all provided consent, and 29 completed the questionnaire — with minimal missing data and near-universal agreement to share responses with the NNRD.

How were patients and the public involved?

Parents with lived experience of neonatal care were involved from the very beginning and played a central role in shaping every aspect of the project. A parent co-applicant contributed to the study design, ethical considerations, and interpretation of results. The parent advisory group reviewed all study documents, refined the digital questionnaire to ensure it was emotionally sensitive and accessible, and provided practical advice on how and when to approach families directly influencing the decision to invite parents a week before discharge rather than on the day itself. Based on parent feedback, one distressing question was removed from the questionnaire, an open-text box was added so parents could share experiences not captured by structured questions, and follow-up reminders were limited to just one to avoid placing pressure on families. Parents reported feeling valued, respected, and empowered by contributing to improvements in neonatal care, describing participation as meaningful and emotionally validating.

Improving How We Test Digital Health Tools: Understanding What Gets in the Way

What is this project about?

Digital health interventions, such as apps, online programmes, and wearable devices designed to help people manage their health, are increasingly being tested in clinical trials. However, the methods used to run these trials have not always kept pace with the technology itself. Newer, smarter ways of designing trials exist. Adaptive trials can be adjusted as results come in, while decentralised trials allow participants to take part from home rather than travelling to hospital. However, these approaches are not widely used. This project surveyed researchers across the UK to find out what is stopping them from using these more efficient methods and explored why clinical trials of digital health tools often fail to include a diverse range of participants.

Why does it matter?

Traditional clinical trials are often slow, expensive, and rigid. They cannot be changed once started, even if early results suggest something is not working. This means patients may continue receiving ineffective treatments for longer than necessary, and the evidence needed to bring better digital health tools into routine NHS care takes far longer to generate. By identifying the specific barriers that prevent researchers from using smarter trial designs, this project lays the groundwork for practical solutions that could make digital health trials faster, cheaper, and more inclusive — ultimately getting effective tools to patients sooner.

What are the outputs of the project?

A national survey received 138 responses over four months, with 97 eligible respondents who had direct experience of running clinical trials of digital health interventions. Respondents came from 42% of UK Clinical Research Collaboration registered Clinical Trials Units, spanning all nine regions of England, Scotland, and Wales providing a strong national picture. The key barriers identified included the complexity of calculating costs for decentralised trials, a lack of practical training in newer trial designs, and limited experience of knowing when these approaches are most appropriate. The survey was developed and piloted with support from staff at the Imperial Clinical Trials Unit (ICTU).

How were patients and the public involved?

Before engaging their own public partners, the team consulted an established patient and public involvement group at Sheffield University with expertise in clinical trial methodology, whose input helped refine the discussion approach and confirmed that using real trial examples would be the most effective way to generate meaningful conversation. Four public partners then took part in an online discussion session, during which three real digital health trial examples were presented to illustrate common diversity and recruitment challenges. Partners identified important barriers to participation, including limited advertising reach, digital exclusion, and the failure to account for diverse lifestyles, and suggested practical solutions such as broader recruitment strategies and improved imagery and messaging. Their perspectives were incorporated into the resulting publication, and they were formally acknowledged as contributors and given the opportunity to review and comment on the manuscript before submission.

A Real-Time AI Tool to Spot Undiagnosed Heart Valve Disease During MRI Scans

What is this project about?

Heart valve disease is a common and potentially serious condition that often goes undetected until it has already caused significant damage. Cardiac MRI scans are increasingly used to assess the heart, but standard scans do not always include the specific images needed to detect valve problems — meaning patients can leave their scan with an undiagnosed condition. This project developed an AI tool called AVAI that analyses cardiac MRI images automatically and in real time, as the scan is being performed. If AVAI detects signs of aortic valve disease, it immediately alerts the radiographer, who can then decide whether to take additional images on the spot without the patient needing to return for another appointment.

Why does it matter?

Demand for cardiac MRI is growing faster than NHS capacity, leading to long waiting lists. Currently, many hospitals take a broad range of images for every patient just in case something is missed which is time-consuming and costly. AVAI offers a smarter approach: by flagging only those patients who show signs of valve disease, it allows radiographers to tailor each scan to the individual patient’s needs. This makes scans shorter and more comfortable for patients, frees up scanner time, and could allow significantly more patients to be seen each day reducing waiting lists and improving access for everyone.

What are the outputs of the project?

A working prototype of AVAI was successfully built and tested on retrospective cardiac MRI data from Hammersmith Hospital, demonstrating that the tool can detect aortic valve disease from standard scan images and highlight the specific regions of the image that informed its assessment. Although the lead researcher moved to another institution before prospective patient testing could be completed, the code and methodology developed for AVAI were subsequently used by MRI Physicist Dr Jan Sedlacik in related real-time cardiac MRI analysis projects, ensuring the work continued to generate value.

How were patients and the public involved?

Patient and public involvement shaped the fundamental design of the AVAI tool. A public involvement event held in 2020 brought together five patients with cardiac disease recruited through the British Heart Foundation, and five members of the Guy’s and St Thomas’ NHS Trust patient advisory team. Participants expressed a clear preference for AI tools that work alongside humans rather than replacing them, and for systems that explain how they reach their conclusions rather than operating as opaque “black boxes.” In direct response to this feedback, the team designed AVAI as a triage tool that flags concerns to a human radiographer, who retains full decision-making authority, and built in a feature that highlights the specific areas of the scan the AI used to reach its assessment, making the tool’s reasoning visible and transparent.

Harnessing 40 Years of Patient Data to Improve Care for Two Rare but Serious Conditions

What is this project about?

Pulmonary arteriovenous malformations (PAVMs) are abnormal connections between blood vessels in the lungs that allow blood to bypass the lungs’ natural filtering system. This puts people at high risk of stroke, brain abscess, and serious pregnancy complications. PAVMs are closely linked to an inherited condition called hereditary haemorrhagic telangiectasia (HHT), which also causes fragile blood vessels throughout the body leading to nosebleeds, gut bleeding, anaemia, and in some cases strokes or brain bleeds. Imperial College Healthcare NHS Trust has run national referral services for both conditions for over 40 years, accumulating one of the largest and most detailed patient datasets in the world. This project is building a suite of digital tools to unlock the value of that data and making it accessible to researchers, useful to clinicians who may never have encountered these conditions before, and informative for patients themselves.

Why does it matter?

Because PAVMs and HHT are classified as rare diseases, most doctors are never taught about them during their training. This means patients frequently receive incorrect advice at general clinics sometimes with life-changing or even fatal consequences. One patient was told by their doctor after suffering a stroke that the condition is “usually only seen at post-mortem.” Yet these conditions are far more common than previously thought, affecting tens of thousands of people in the UK. By equipping general clinicians with digital decision support tools and making 40 years of expert patient data available for research, this project could prevent strokes, improve treatments, and transform the standard of care for a large and currently underserved patient population.

What are the outputs of the project?

A world-class patient database covering 1,785 patients, including 1,148 with confirmed PAVMs and 1,431 with confirmed HHT, spanning 40 years of data from 1984 to 2025 has been curated and is being uploaded to Imperial College Healthcare NHS Trust’s secure data environment. This is one of the largest PAVM datasets ever assembled anywhere in the world, and the only one with detailed, standardised physiological measurements. A new digital clinical tool, a PowerForm, integrated directly into the Trust’s Cerner electronic health record system has been approved by the Cerner Change Board and is awaiting technical go-live, enabling clinicians to safely assess and manage PAVM and HHT patients even without specialist training. The project directly underpinned Imperial’s successful application to lead a new NHS Rare Disease Collaborative Network (RDCN) for PAVMs, launched at the British Thoracic Society Winter Meeting in November 2025, with partner centres in Sheffield, Nottingham, Hull, Newcastle, and Manchester. Engagement with the National Disease Registration Service at NHS England is underway to transfer patient data into national NHS databases.

How were patients and the public involved?

Patients with lived experience of PAVMs and HHT were involved from the very beginning, co-partnering in the design of the digital tools, consent processes, and patient-facing information. They shaped what information should be prioritised on the Trust website, how it should be written, and how it should be delivered, insisting that both simple and detailed versions of information should be available, and that the public should be able to access as much detail as they feel is appropriate. Patient feedback directly changed the language used in information materials. For example, explaining circulation problems using the phrase “every drop of blood” after patients found this the clearest way to understand the condition. Patients also expressed a strong preference for explanatory videos from the expert doctors they know by name, rather than written information alone — a commitment the team has made to deliver in early 2026. A public-facing meeting was held in February 2024 to explore attitudes toward the use of AI in managing these conditions, and patients will be further involved in shaping how difficult findings — including data on strokes and gaps in care — are communicated when the database is released.

The theme has recruited four community partners that are all members of the public currently living and/or studying in Northwest London. These community partners will be actively involved in providing strategic input into the theme’s activities and meaningfully connecting researchers with community groups that are under-represented in healthcare research. Our current projects also involve lay partner representatives as part of steering groups and project teams and sit on our iCARE Data Access Committee to ensure all research use of iCARE’s routine health data is of public and patient benefit.

We will embed the EDI objectives by ensuring all new recruitments have a diverse interview panel and continuing to offer placements and apprenticeships, such as the HDRUK Black Internship Programme. We are also supporting Imperial College Healthcare NHS Trust to evaluate the effectiveness of their inclusive recruitment programme in increasing the ethnic diversity of senior roles within the Trust. Additionally, our diverse community partners will be invaluable in ensuring our research activities are inclusive and assess impacts on health inequalities. Finally, we are collaborating with local authorities, the voluntary sector, and Paddington Life Sciences industry partners to address digital exclusion in our Northwest London population.

In accordance with the Imperial BRC’s PPIEP Strategy , our theme has recruited a group of Community Partners to act as critical friends to our theme and share their valuable lived experience with our researchers and health professionals to help improve the relevance and quality of our research for the benefit of our North West London population.

“I am a multi-lingual seasoned leader with international expertise and experience of 25 years plus over 7 years as a Non-Executive focusing on governance, delivery and operations, driving sustainability and a culture of collaboration, continuous improvement & innovation with passion and empathy.” Rashmi Rungta

“I’m a 69 yr old Autistic person who is also working with the Integrated Care Board to redesign Autism Strategy delivery in Hammersmith and Fulham.” Suzanne Iwai

“I have been involved in the health sector since The Health & Social Care Act 2012. My particular interests are how best to implement emerging technologies into preventative and palliative medicine with all steps in between.” Nicola Brightman

“I am an undergraduate student interested in representing patients of South Asian descent and young people in patient care from a perspective that is underrepresented.” Lagsika Kugathas