On the Cutting Edge: Safely Accelerating AI Solutions in the NHS

In this article, we delve into the crucial discussions surrounding AI implementation in healthcare, examining the intended use of AI solutions, the importance of trust, data standardisation, and the need for upskilling the healthcare workforce. This article is based on the HETT Show 2023 talk ‘Does the NHS have the Data and Digital Infrastructure to Support a Future of AI on the Frontline?’. The session was chaired by DigitalHealth.London’s Programme Director, Sara Nelson.

In 2019, ‘The Topol Review: Preparing the healthcare workforce to deliver the digital future’ was published as a component of the provisional health and care Workforce Strategy for England until 2027 – Facing the Facts, Shaping the Future. The Topol report encouraged the empowerment of NHS staff in maximising the benefits of cutting-edge technologies like genomics, digital medicine, artificial intelligence (AI), and robotics to enhance healthcare services and improve patient outcomes.

Four years on, the NHS’s pursuit of digital transformation continues to significantly revolutionise the way we work within the health and care system. However, as laid out in Health Education England’s AI Roadmap report, the impact of large language models and complex AI solutions raises equally complex questions about the readiness of the healthcare sector to embrace these transformative technologies: where do people fit in the process to bring AI to market?

Ensuring Safe AI Solutions in Healthcare

Usage of AI in the NHS broadly falls into two categories: Clinical and Administrative.

The clinical domain involves complex decision-making processes where AI can significantly impact patient outcomes. On the other hand, AI solutions on the administrative front have fewer associated risks as they mainly focus on automating and reducing workloads. Though some healthcare trusts have already begun integrating AI solutions into their workflows, the stakes are often high when clinicians rely solely on these technologies.

To mitigate risks, pilots should be utilised to ensure that AI solutions are safe, effective and controlled before they can be rolled out on a larger scale. Earlier this year, DigitalHealth.London Accelerator alumnus company Concentric and Chelsea and Westminister Hospital NHS Foundation Trust worked together on a highly successful digital consent pilot for patients undergoing surgery. Consultants were able to complete e-consent forms before surgery and share them with patients for signing either virtually or in person.

Around 7,000 patients and 250 NHS/social care staff benefitted from the pilot, improving the overall efficiency of the surgical patient pathway. While the pilot initially started in General Surgery and Colorectal Surgery in Chelsea and Westminster Hospital, it has since been rolled out to nine specialities, with over 3,800 consent forms used on the platform to date.

It is pivotal to engage with not just the direct users, but also staff that are indirectly associated with the workflow. It is important to ensure that information and recommendations are obtained from them, and they are also offered training – as their knowledge is invaluable.

Phoebe Allen
Quality and Improvement Manager for Planned Care at Chelsea and Westminster Hospital
Digital Pioneer Fellow on the DigitalHealth.London Digital Pioneer Fellowship

Transformative change does not happen in a vacuum. Similarly, AI cannot be scaled without the workforce’s trust and knowledge. Clinical and administrative users must build a fundamental understanding of the technology underpinning these solutions, as well as their benefits and limitations.

Open and robust discussions on the classification and usage of AI (e.g., Is it a medical device? What sort of skills or people do we need to implement these solutions? Is specialised training necessary?) should be encouraged to close the workforce skill gap and improve digital literacy. Public resources such as the British Standard Institution’s BS 30440 Validation framework for the use of AI within healthcare (on-demand webinar available here) can also be used to help the workforce in safely procuring and implementing these innovative technologies.

Building trust in AI among healthcare professionals and patients requires both time and perseverance, but is essential and should be an ongoing process that requires constant validation and feedback.

Accelerating Pathways to Better Outcomes

While data is a precious and abundant resource within the NHS, it often suffers from fragmentation and a lack of standardisation. A prime example is imaging data – crucial for diagnoses and treatments, particularly within Dermatology – which frequently exhibits variations in quality due to differences in angles and lighting in photographs. AI solutions can help implement standardised processes within clinical pathways and offer a substantial advantage, saving both staff and patients valuable time and effort.

In a recent Innovation Exchange run by NHS England and the Health Innovation Network, the Academic Health Science Network (AHSN) for south London, Ambient Voice Technology (AVT) was spotlighted as a solution that could capture patient interactions in real-time. This innovative technology enables products to, with the patient’s consent, listen in on patient-clinician interactions and subsequently generate a summary. The clinician can then review and add this summary to the patient’s records.

When applied to the vast landscape of the NHS – encompassing approximately 1.75 million daily patient appointments – the adoption of AVT could result in saving clinicians thousands of clinical hours. This transformative approach has the potential to significantly enhance capacity within the NHS, ultimately benefitting patients and healthcare professionals alike.

AI also has the potential to expedite the diagnosis and treatment of diseases. HN, an alumnus company of the DigitalHealth.London Accelerator, leverages AI to anticipate and prevent unplanned care events using existing healthcare data. Their technology can pinpoint 8 out of 10 patients who might need emergency admission, aiding them in enhancing their health outcomes, thus resulting in a 34% reduction in the probability of hospital admission.

The Road Ahead

For the past 75 years, the NHS has pioneered revolutionary change in the healthcare industry by continuously adapting to meet the needs of each generation. By harnessing innovative digital solutions safely and effectively, healthcare providers can ultimately accelerate pathways to meet demand and achieve better patient outcomes for all. As AI permeates our daily lives, understanding its implications and applications is essential. It’s not just a healthcare revolution; it’s a societal transformation.

Concentric and HN are alumni companies of the DigitalHealth.London Accelerator Programme.

The DigitalHealth.London Accelerator is a collaborative programme funded by the UK Government through the UK Shared Prosperity Fund, the Health Innovation Network South London, the Office of Life Sciences, CW+, the Mayor of London and the Levelling Up Fund.