#43. W/c 4th August '25
Alastair Allen
Partner, Head of Healthcare Technology at EY | Board Member at openEHR UK | Health and Social Care Council Member at TechUK
This week, my reading covers the widespread use of an 'Amazon-style' prescription tracker within the NHS App, a new AI co-pilot from OpenAI and Penda Health, and openEHR UK's view on the Single Patient Record. Additional topics include a new AI lifecycle management approach from the FDA and a new policy from the HSE on patient identity standards for IHI matching.
openEHR UK and the NHS SPR
openEHR UK recently responded to NHS England's Single Patient Record (SPR) Request for Information, emphasising its capability to enhance healthcare interoperability and digital transformation in England. The response highlighted openEHR’s strengths in delivering scalable, and semantically interoperable digital health solutions. It focused on three primary areas: clinical governance, interoperability, and innovation enablement.
Firstly, the submission recognised the importance of clinical governance and data quality assurance. Through clinical archetypes and templates developed collaboratively by clinicians and informaticians, openEHR supports reliable, shareable, and actionable patient information across care settings.
Secondly, openEHR’s strength in semantic interoperability was underscored, establishing a common language and set of rules that lets healthcare IT components work together. The response proposed building on the Shared Care Record Programme’s achievements and advocated for using HL7 FHIR to interface with an openEHR CDR, suggesting the International Patient Summary as a logical starting point.
Lastly, innovation was addressed, with openEHR facilitating integration with advanced analytics tools (FDP?), AI diagnostics, and patient-centred applications. The idea of a ‘data mesh’ of federated openEHR clinical data repositories was discussed, although personally I feel that idea requires more debate and is more of an implementation decision to consider among other options.
OpenAI and Penda Health Launch Clinical Copilot
OpenAI and Penda Health have partnered to develop AI Consult, an AI-powered “clinical copilot” designed to act as a safety net for clinicians during live patient consultations. Integrated into Penda’s electronic medical record system across 16 primary care clinics in Nairobi, Kenya, the tool runs silently in the background and intervenes only when it detects potential clinical errors. Using GPT-4o and adapted to Kenya’s clinical guidelines and health context, the system provides three levels of feedback: a green checkmark for no concerns, a yellow alert for moderate issues, and a red pop-up for safety-critical concerns requiring immediate review—ensuring clinicians remain in full control of decision-making.
The AI Consult system was rolled out in two phases: an induction period for testing and refinement, followed by a main deployment supported by peer champions, coaching, and incentives to encourage adoption. In a study of 39,849 patient visits—20,859 with AI support and 18,990 without—the AI reduced history-taking errors by 32%, investigation errors by 10%, diagnostic errors by 16%, and treatment errors by 13%. In cases where safety-critical red alerts were triggered, the improvements were even more significant, with diagnostic errors reduced by 31% and treatment errors by 18%. Clinicians reported strong approval, with 100% saying the tool improved care quality and three-quarters describing the improvement as substantial.
This initiative is one of the largest real-world trials of AI in a low-resource primary care setting, demonstrating how well-designed AI tools can enhance safety and quality without undermining clinician autonomy.
Hundreds of thousands use ‘Amazon-style’ prescription tracker in NHS App
The NHS App's new "Amazon-style" prescription tracker has been utilised by nearly 400,000 people in its initial ten weeks, significantly reducing unnecessary pharmacy visits and calls. This feature enables users to check the status of their prescriptions akin to tracking online purchases, with over 715,000 uses recorded since May. Around 1,650 pharmacies, including all Boots in England, offer this service. This integration allows patients to confirm if their medicines are ready for collection or dispatch.
The app has seen about 37.4 million registered users, with monthly logins averaging 11.4 million for managing healthcare. Over the past year, 61.5 million repeat prescriptions have been ordered via the NHS App, marking a 46% increase. Additionally, there were 87.4 million views of patient test results in the same period.
Dr. Vin Diwakar, Clinical Transformation Director at NHS England, highlighted that the tracker facilitates efficient pharmaceutical services, giving pharmacists more time to provide direct patient care by reducing call volumes about prescription readiness.
The government plans to expand this feature to nearly 5,000 more pharmacies within the next year, covering 60% of England’s pharmacies. The enhancement aligns with the government's broader digital transformation agenda under the 10 Year Health Plan, aiming for more user-friendly access to healthcare services including appointment management and instant non-urgent care advice through AI.
Lifecycle Management Approach toward Delivering AI-enabled Health Care
The U.S. Food and Drug Administration (FDA) addresses the increasing importance of AI in healthcare, focusing on the lifecycle management of AI-enabled medical devices to ensure safety, effectiveness, fairness, and trustworthiness. Recognising the challenges posed by AI’s adaptability and the risks of bias in data and algorithms, the Digital Health Center of Excellence emphasises the necessity of a robust AI lifecycle management approach.
The FDA outlines a structured framework known as the AI Lifecycle (AILC), which adapts traditional Software Development Lifecycles (SDLC) to AI applications. This approach aims to systematically manage the phases of data collection, model building, operation, and real-world performance evaluation. Central to AILC is the assurance of data quality, ethical standards, and transparency, guiding developers in maintaining high standards of AI practice.
The AILC model incorporates technical and procedural considerations across each phase of AI development, promoting a comprehensive understanding of lifecycle activities. Emphasising the role of standards, the FDA sees AILC as crucial for establishing interoperability and promoting ethical practices, with a community-driven approach to refine and enhance AI standards in healthcare.
National Policy on Patient Identity Standards for IHI Matching officially published on HSE Repository
The Health Service Executive (HSE) has released a national policy for Individual Health Identifier (IHI) Data Standards to enhance patient safety and system efficiency in Ireland. This new policy sets best-practice guidelines to ensure accurate patient identity data, crucial for matching patients with their unique IHI. The IHI is a unique identifier allotted to individuals in healthcare to facilitate precise health record connections. Prior inconsistencies in data entry impacted IHI matching, prompting this policy to address these gaps by establishing national standards.
The policy serves two primary objectives: it mandates essential data field standards for IHI matching in all new digital health systems and supports a national training programme via HSeLanD to upskill staff on proper data entry practices. This ensures a consistent approach to data registration, enhancing the overall reliability of IHI matching.
Developed by the HSE National HIDS Steering Committee and involving input from various stakeholders like clinical services and regulatory bodies, the policy aims to instil a standardised understanding of the IHI's significance to patient safety. The policy, officially released to the HSE 3PG National Central Repository, is bolstered by targeted communications and training initiatives to promote enduring improvements in data entry practices.
Integration Specialist in BSO ITS
1moThanks for sharing, Alastair. Just wondering who would have access to data? Would this type of AI be used to train other AI with real patient data?
Honorary Professor in Engineering & Applied Science at Aston University
1moThanks for sharing, Alastair. Fantastic to see your newsletter highlighting openEHR UK RFI response to the #NHS Single Patient Record (SPR) #openEHR
Project Manager
1moGreat round-up, Alastair — especially interesting to see the NHS App prescription tracker adoption and how it’s driving patient engagement. The AI lifecycle management work from the FDA is also a critical step in ensuring we deploy these tools responsibly in healthcare. At TherapyFast we’re seeing the importance of transparency and ethical AI in mental health too — particularly when scaling support across large populations. The Single Patient Record discussion also ties into one of the biggest challenges in mental health: ensuring that the right data is available to the right people at the right time without compromising privacy. If you’d like to see how we’re applying these principles in practice, you can book a free online demo of TherapyFast — AI-driven mental health support that cuts costs by up to 95% while maintaining quality. 📩 nicky@therapyfast.co.uk 🌐 www.therapyfast.app