Researchers propose new ways to assess and integrate AI health technologies
Imperial researchers suggest reforms in how digital health technologies are evaluated and integrated into the UK’s healthcare system.
“By modernising the NICE ESF, we can fast-track safe and effective technologies into the NHS, particularly in critical areas like heart disease detection, remote care, and chronic disease management.” Dr. Shayan Bahadori Research Associate, Department of Surgery and Cancer
A new paper by researchers from Imperial College London and the NIHR HealthTech Research Centre in In Vitro Diagnostic, titled “Evolving Digital Health Technologies: Aligning with and Enhancing the NICE Evidence Standards Framework – A Viewpoint Piece” outlines a roadmap to modernise the National Institute for Health and Care Excellence (NICE)’s Evidence Standards Framework (ESF).
There are growing concerns that that regulatory systems risk falling behind the pace of technological development, especially in light of rapidly advancing AI diagnostics and wearable health innovations.
Dr. Shayan Bahadori, lead author and a Research Associate at Imperial’s Department of Surgery and Cancer, said: “We’re entering a new era of healthcare where AI and wearable devices can detect disease earlier and more accurately than ever before but our health system risks falling behind if we don’t modernise how we evaluate these tools. Our paper sets out a blueprint for smarter, faster, and fairer regulation of digital health innovation.”
Bridging the innovation-regulation gap
The paper, published in JMIR Mhealth Uhealth, highlights a growing mismatch between emerging technologies, such as portable ECG monitors like KardiaMobile, which continuously refine their algorithms as they collect more user data, and outdated regulatory frameworks. These technologies, which learn and adapt through machine learning and real-time feedback, are often assessed using static, one-size-fits-all models that were never designed for such dynamic systems.
Dr. Bahadori added: “This creates delays, uncertainty, and significant barriers to adoption, especially for start-ups and smaller innovators.”
To address this, the authors propose that the NICE ESF undergo a comprehensive transformation into a dynamic, iterative framework that evolves in real time with technological progress. Among the specific proposals, the paper calls for:
- Modular Evaluation: Introducing modular processes that allow NICE to assess different aspects of a digital health technology (DHT) as it matures. For example, an early-stage AI tool could undergo an initial review based on preliminary data, with its efficacy and safety updated continuously as real-world evidence accumulates.
- Regulatory Sandboxing: Establishing controlled, real-world testing environments where developers, clinicians, and patients can trial and refine technologies in practice. This approach, inspired by successful models in the financial sector, would enable faster and more flexible approval pathways.
- Shared Data Repositories: Developing collaborative platforms that pool anonymised health data from wearables, records, and surveys, to train machine learning models while upholding strict data protection standards. Such repositories would improve the robustness of AI tools and enable fairer, more transparent oversight.
- Real-World Evidence Strategies: Embedding evidence generation that focuses on actual clinical outcomes rather than theoretical models alone, co-designed with industry partners to ensure relevance and timeliness.
Supporting innovation: next steps and impact
The researchers argue that by integrating these measures into NICE’s ESF, the UK can create a more agile and transparent regulatory framework that not only protects patients but also accelerates the safe adoption of new technologies.
“By modernising the NICE ESF, we can fast-track safe and effective technologies into the NHS, particularly in critical areas like heart disease detection, remote care, and chronic disease management.” said Dr Bahadori, who is also part of the NIHR HealthTech Research Centre in In Vitro Diagnostic.
As a viewpoint article, the paper does not present new empirical data but synthesises emerging challenges and offers actionable policy recommendations. The authors now plan to pilot their proposals, including regulatory sandboxes, in collaboration with digital health companies, regulators, and NHS organisations.
Article text (excluding photos or graphics) © Imperial College London.
Photos and graphics subject to third party copyright used with permission or © Imperial College London.
Reporter
Benjie Coleman
Department of Surgery & Cancer