DOI: https://doi.org/10.58248/HS55

Overview

The government, healthcare professionals and researchers have highlighted the transformative role that technology innovations may play in:[1],[2],[3],[4],[5]

  • diagnosing diseases and mental health conditions earlier and improving patient outcomes
  • reducing NHS waiting lists
  • predicting health outcomes, such as heart attacks,
  • reducing A&E pressure
  • discovering and developing drugs
  • reducing healthcare costs
  • improving health literacy and personalised information that could enable people to be more engaged in managing their own health
  • supporting people with disability

There is increasing focus within the NHS on developing better diagnostic technologies and disease prediction tools (Developments in the diagnosis of disease for medicine and public health). For example, in October 2023 the NHS announced a £21 million investment to roll out AI diagnostic tools covering 64 trusts in England for faster and more accurate diagnosis compared to a clinician.4 Some MIT researchers have shown that if trained properly with good quality data, AI algorithms could compare brain scans and other 3-D images up to 1000 times faster than humans.[6]

The government, NHS, charities, research and industry stakeholders play key roles in funding, supporting, researching, developing and testing new technologies.[7],[8],[9],[10] Many emerging technologies are still in development and trial stages and are yet to be deployed and implemented at scale in the NHS.6

Many healthcare professionals have highlighted a need for policy makers to also focus on using non-clinical technologies that can help with administrative and operational tasks, such as note taking, communication and scheduling.1 Many healthcare professionals highlight how these operational uses of technology could free up staff time, improve NHS productivity, carry less clinical risk, be more cost-efficient and be implemented more quickly.1

Challenges and opportunities

Artificial Intelligence

To better classify diseases and understand their genetic origins, researchers can use AI to unravel patterns in large volumes of data on:[11]

  • complete sets of DNA (known as genomics)
  • patient health, lifestyle and demographics and thus risk of disease
  • proteins and their structures for drug discovery and development

Various agreements exist in the UK for researchers to access genomic and biomedical data, such as through the UK Biobank or Genomics England.[12],[13]

To help treat mental health conditions, AI chatbots could provide cheaper and accessible therapy alternatives.[14] AI wearable devices could collect and interpret biodata, such as sleeping patterns.[15] Computer vision, where algorithms interpret images, could help to understand non-verbal cues, such as facial expressions. 24 These devices could help to assess risks and predict and diagnose mental health conditions.24

Digital Twins

Although still in research and development stages, virtual data replications of a patient’s organs or entire body, known as ‘digital twins,’ have the potential to:[16],[17],[18]

  • assess the safety and efficacy of drugs
  • simulate treatments for patients
  • monitor patients health trajectory and allow for the early intervention and prevention of diseases
  • allow surgeons to practice complex medical procedures

Medical and surgical technologies

Medical technical devices such as surgical implants and wearable sensors can collect data on, analyse and alter the nervous system, known as ‘neurotech’.[19] This technology can assist patients who have lost motor and sensory functions.[20],[21] They are being developed to treat a range of chronic illnesses or injuries, including epilepsy, strokes, Parkinson’s disease, chronic pain, depression and more.

In June 2024, the NHS started its first clinical trials using neurotech to treat children with epilepsy. The trial has shown that one patients’ daytime seizures were found to decrease by 80%.[22],[23] Successful trials may result in a significant future uptake.18,[24]

The NHS use robots performing independent or semi-independent actions to assist with minimally invasive surgery techniques that improve clinical outcomes such as removing prostate glands.[25] 3D printing, such as anatomical models of feet, hands and a pelvis, has significantly increased in the past decade, particularly in Head and Neck and Cardiology departments.[26] Models have been used to:

  • help patients understand their injuries and make informed decisions about treatments
  • assist surgeons in planning accurate surgeries6,[27]

In 2022, the NHS launched trials that give lab-grown red blood cells to people and assess their recovery, survival and safety.[28],[29] If proven effective, lab-made blood could revolutionise treatments for people with rare blood types, blood disorders such as sickle cells or people who are not able to have normal blood transfusions.13

Rehabilitation technologies

There has been a growing use of technologies, including sensors, robots, gaming, virtual reality, wearable structures and mobile apps, to support rehabilitation within hospital and home settings for people with disability and long-term conditions including stroke, dementia, respiratory conditions and cancer care.[30],[31],[32],[33],[34]

For example, some studies show virtual reality has helped patients to adhere to rehabilitation exercises and improved training intensity.32 Translation of technologies into rehabilitation settings remains limited.5

Challenges

While emerging technologies offer promising advancements in healthcare, some researchers caution against overhyping their capabilities.[35],[36]

A 2019 study estimated that it takes approximately ten years for a new medical diagnostic technology to be routinely adopted in the NHS.15 Reasons include:15,[37]

  • insufficient robust clinical evaluations in real-world contexts
  • resource constraints
  • institutional barriers within the NHS, including a lack of incentives for innovation adoption
  • fragmented national and regional systems
  • lack of co-design with stakeholders involved in the supply and demand of technologies
  • a lack of education and training to support confidence in staff and patients using technologies

Ethical challenges exist, particularly around AI and personal data:

  • Health data is highly vulnerable to misuse and security breaches. A 2023 report found 8 in 10 UK health organisations have had a security breach since 2021.[38] There have been high-profile incidents of NHS data breaches such as the June 2024 exposure of 300 million pieces of blood test patient data from two trusts.[39], [40]­
  • Multiple studies have shown how inaccurate, biased or incomplete datasets directly impacts the effectiveness, safety and fairness of AI technologies in healthcare (PN 637).[41],[42],[43] Inaccurate data can lead to AI diagnosing symptoms incorrectly and potentially harmful treatments.21 A 2019 study found that an algorithm used to allocate healthcare resources in US hospitals showed racial bias, being less likely to refer Black patients to healthcare programmes compared to equally ill White people (PN 708).[44]
  • The use of technologies could exacerbate inequalities in access to and quality of healthcare for certain regions and demographics. The independent cancer task force noted that the NHS has not historically adopted technologies equitably across England in a review from 2015-2020.[45]
  • Stakeholders have raised concerns about AI dehumanising the healthcare system and some have expressed that doctors are able to make more holistic judgements about diagnosis or treatments than AI systems (PN 637).
  • Some reports have highlighted the importance of assessing trustworthiness and confidence in healthcare technologies for them to be used in the NHS.[46],[47]

The Health Foundation has highlighted a need for greater focus, supportive policy and investment to maximise the impact of technologies that could help with operational and administrative tasks.1 It emphasised the importance of effective skills training, infrastructure, efficient internet connectivity, and skilled change management to decide how to repurpose freed-up time.1

Key uncertainties/unknowns

There is insufficient robust data measuring the cost and health impacts of AI innovations currently being used in the NHS. There is also a lack of evidence on the effectiveness of both emerging and incremental technologies to treat health conditions, and particularly mental ill health.

The Health Foundation said there is a “lack of an overarching strategy for AI in healthcare”, which limits the ability to assess its potential in improving quality and efficiency of healthcare.1

It can often be impossible, even in principle, to understand how complex AI models have generated their output (PB 57). A lack of transparency in how the AI model calculated its output can make it difficult for someone adversely affected by an automated decision to know what went wrong, who is liable and how to seek reimbursement (PN 708).[48] Medical ethicists have stated that AI in diagnosis requires transparency, human oversight, and for regulation and governance to be clear on how liability is defined (PN 708).[49],[50]

Key questions for Parliament

  • What are the government’s plans for measuring the impact of AI and other innovations in the NHS?
  • Should there be more of a focus on how to maximise the use of technologies that can help with administrative and operational tasks in the NHS?
  • Is there adequate funding and resources for implementing technology innovations and upskilling staff and patients to use them?
  • How can policymakers help overcome institutional barriers towards adopting innovations in the NHS?
  • How can Parliament adequately consider and regulate ethical concerns, including patient privacy and the impacts of biased algorithms?
  • How can personal data in the health system be securely managed and safeguarded against increasing cyberattacks?

Related documents

References

[1] Horton, T et al. (2024). Can technology and AI ‘save the NHS’? A look at the main party manifestos. The Health Foundation.

[2] Asaria, M. (2024). How AI could revolutionise NHS healthcare. LSE.

[3] Jones, W. (2024). AI-powered ‘deep medicine’ could transform healthcare in the NHS and reconnect staff with their patients. The Conversation.

[4] UK Government (2023). AI to speed up lung cancer diagnosis deployed in NHS hospitals.

[5] Mitchell, J., et al. (2023). Factors that influence the adoption of rehabilitation technologies: a multi-disciplinary qualitative exploration. Journal of Neuroengineering and rehabilitation, Volume 20, 1.

[6] Matheson, R. (2018). Faster analysis of medical images. MIT News.

[7] NHS Health Education England (2019). The Topol Review: Preparing the healthcare workforce

to deliver the digital future.

[8] NHS England (online). The NHS AI Lab. Accessed 26 July 2024.

[9] Department of Health and Social Care (2023). Call for evidence outcome – 10-Year Cancer Plan: call for evidence.

[10] Cancer Research (online). Detect cancer earlier by interrogating medical and non-medical data sets using machine and deep learning. Accessed 26 July 2024.

[11] Agrawal, R., et al. (2020).  Big data in digital healthcare: lessons learnt and recommendations for general practice. Heredity, Vol 124, pages 525–534.

[12] uk biobank (online). Background.

[13] Genomics England (online). 100,000 Genomes Project.

[14] Haque M. D. R., et al. (2023). An Overview of Chatbot-Based Mobile Mental Health Apps: Insights From App Description and User Reviews. JMIR Mhealth Uhealth, Vol 11.

[15] Lavrentyeva, Y. (2024). The big promise AI holds for mental health.

[16] Katsoulakis, E., et al. (2024). Digital twins for health: a scoping review. npj Digit. Med., Vol 7, 77.

[17] Balasubramanian, S. (2023). Digital Twin Technology Has The Potential To Radically Disrupt Healthcare. Forbes.

[18] Geddes, L. (2023). How digital twins may enable personalised health treatment. The Guardian.

[19] Department for Business, Energy & Industrial Strategy (2022). Regulatory Horizons Council (RHC) publishes independent recommendations on the future regulation of neurotechnology and AI as a medical device.

[20] Information Commissioner’s Office (2024). Tech Horizons Report.

[21] Berger, S., et al. (2023). AI and Neurotechnology: Learning from AI Ethics to Address an Expanded Ethics Landscape. Communications of the ACM.

[22] University of Oxford (2024). First UK trial of Deep Brain Stimulation for children with epilepsy begins.

[23] NHS (2024). Clinical trial using new device to treat children with epilepsy.

[24] Leng, G., et al. (2018). Uptake of medical devices approved by NICE. BMJ Innov, Vol 4, pages 178-184.

[25] Maynou, L. et al. (2022). The diffusion of robotic surgery: Examining technology use in the English NHS. Health Policy, Vol 126, 4, pages 325-336.

[26] HETTSHOW (2024). 3D Planning and Printing in NHS Hospitals – The Future of Surgical Planning.

[27] NHS (2015). New 3D printer sparks ideas for patient care.

[28] NHS (online). Recovery and survival of stem cell originated red cells.

[29] National Institute for Health and Care Research (2022). Groundbreaking clinical trial gives lab-grown red blood cells to people for the first time.

[30] Arntz, A., et al. (2023). Technologies in home-based digital rehabilitation: scoping review. JMIR Rehabilitation and Assistive Technologies, Volume 10.

[31] Gebreheat, G., et al. (2023). The use of home-based digital technology to support post-stroke upper limb rehabilitation: A scoping review. Clinical Rehabilitation, Volume 38, 1, pages 60-71.

[32] Gebreheat, G., et al. (2024). Application of immersive virtual reality mirror therapy for upper limb rehabilitation following stroke: a scoping review. Neurological Sciences, Volume 45, pages 4173-4184.

[33] Roberts, N.J., et al. (2023). A scoping review exploring adoption of digital stress management resources for long-term health conditions – just useful for respiratory conditions? Thorax, Volume 78, 4.

[34] Melilo, A., et al. (2022). Virtual Reality Rehabilitation Systems for Cancer Survivors: A Narrative Review of the Literature. Cancers (Basel), Volume 14, 13.

[35] Janus, J. (2019). Innovation in the NHS: making the most of new technologies. PHG Foundation

[36] Mesko, B. (2023). The Most Overhyped Technologies in Healthcare. The Medical Futurist.

[37] Parliamentary Office of Science and Technology (2021). Developments in the diagnosis of disease for medicine and public health.

[38] Chipman, A. (2023). Eight in ten UK health orgs have had a security breach since 2021. digitalhealth.

[39] Commons Library research briefing CBP-9821, Cyber security in the UK

[40] The Guardian, Records on 300m patient interactions with NHS stolen in Russian hack (accessed 12 July 2024)

[41] Linder, C. (2024). The role of data representation for the design, development and evaluation of (medical imaging) AI innovations in healthcare. British Orthopaedic Association.

[42] Arora et al. (2023). The value of standards for health datasets in artificial intelligence-based applications. Nat Med, Vol 29, 11.

[43] Ali Chherawalla, M. (2024). High-quality data is critical to Improving AI in Healthcare.

[44] Obermeyer, Z. et al. (2019). Dissecting racial bias in an algorithm used to manage the health of populations. Science, Vol 366, 6464, pages 447-453.

[45] Independent Cancer Taskforce (2015). Achieving world class outcomes: A strategy for England 2015-2020.

[46] NHS (2023). Moving from trust to appropriate confidence.

[47] Steerling et al. (2023). Implementing AI in healthcare—the relevance of trust: a scoping review. Frontiers.

[48] Davies, M. et al. (2023). Regulating AI in the UK. Ada Lovelace Institute.

[49] World Health Organisation (2021). Ethics and governance of artificial intelligence for health.

[50] Zhang, J., et al. (2023). Ethics and governance of trustworthy medical artificial intelligence. BMC Med Inform Decis Mak, Vol 23, 7.


Photo by: Master Video via Adobe Stock

Horizon Scan 2024

Emerging policy issues for the next five years.