Redesigning Primary Care in the Age of AI

Artificial intelligence is reshaping the fundamental delivery model of primary care to support the NHS 10-Year Plan's transformation agenda

To start with, a warm welcome to so many new subscribers who are part of this growing community. Thank you, I'm glad these reflections are resonating with increasing numbers of readers. My hope is that these articles help to stimulate us all to consider how we navigate healthcare in this time of transition, where digital, economic, and social pressures are reshaping not just the systems and structures, but the meaning of care, in a way that is fair and equitable.

This latest article explores one of the most significant shifts underway in UK healthcare: the redesign of primary care as the foundation of a more preventive, digitally enabled, and community-based NHS

Central to this shift is increasing use of AI tools as we reshaping who receives care, where, and when. As neighbourhood health services take shape, we need to ask: what does it really mean to shift care upstream, and how we can ensure this transformation serves equitable care, not just efficiency?

Delivering the NHS 10-Year Plan requires three fundamental shifts: shift care out of hospitals and into local communities, shift care from being largely reactive and treatment based to one focussed on prevention, and transition healthcare from analogue to being digitally driven. These shifts can't be delivered through policies alone, they fundamentally require a rethink; new models of care, new infrastructure, new roles for primary care and new tools.

The design of general practice has remained largely static for decades: small, independent practices offering reactive consultations. Walk into many GP surgeries today, and you'll still see a model shaped in the 1960s, individual clinicians working in professional silos, responding to whoever walks through the door. In this traditional setup, the system waits for the patient to present.

The government's vision for 2035 could not be more different. In a digitally-enabled neighbourhood model, the system will actively identify risk, initiate contact, and intervene before conditions escalate. This requires more than a change in workflow. It requires a fundamental shift in power and responsibility. And it demands more than infrastructure and funding, it requires trust.

To navigate this shift we need clarity about what we mean by 'community,' 'prevention,' and 'digital.' Transformation means rethinking not only what primary care does, but where it happens, how it happens, and who it happens with. It's not enough to modernise general practice, we must fundamentally redesign it. This transformation must be systemic, not just technological. It involves redistributing power, reallocating resources, and redefining relationships across sectors. It means equipping primary care to take on more risk, manage greater complexity, and deliver more care, earlier and closer to where people live.

The plans for neighbourhood health services provide a window into the future. These hubs, typically serving 30,000 to 50,000 people, will not just be bigger GP practices, but entirely new ecosystems. Bringing together primary, community, social care, mental health, and voluntary sector professionals, to function as integrated care teams.

By way of example, the Birmingham and Solihull's Community Care Collaborative, primary care sits alongside acute and mental health services, social care, and ambulance teams. Shared digital records, care coordinators, and population health dashboards identify patients requiring targeted early interventions, and proactive planning. It fundamentally shifts care from "treating who walks through the door" to "reaching those who need care but haven't sought it."

In this changing landscape, AI-powered diagnostic tools are beginning to redefine the scope, remit and scale of primary care. These 'system-shifting' technologies, are already being tested, trialled, and deployed in primary care and community settings. They promise to detect risk earlier, reduce the burden on clinicians, and support faster, more accurate decisions, enabling the shift of primary care towards the NHS's future direction. Here are some examples:

• AI-enabled stethoscopes, now deployed in over 200 GP practices, more than doubled diagnosis rates for heart failure, AF, and valve disease, and saving an estimated £2,400 per case. These patients remained in primary care, with no need for referral into secondary care services. The technology enabled a shift in agency, empowering GPs to do more, catch deterioration earlier, and with greater confidence.
• EchoGo, developed by Ultromics, is reshaping heart failure diagnosis. The technology provides automated echocardiogram analysis with over 90% accuracy, enabling earlier, more effective interventions. Previously such complex forms of heart failure that often went undiagnosed until patients suffered serious complications and diagnosis required invasive hospital procedures.
• Cancer detection remains one of the most critical time-sensitive challenges for GPs. AI tools like C the Signs integrate with GP systems to triage patients based on symptoms and history, increasing early detection by over 12% without increasing clinician workload.
• Diabetic retinopathy affects one in three people with diabetes and is a leading cause of preventable blindness. AI is transforming the screening pathway. EyeArt replaces hospital eye appointments with AI-powered retinal scans in community settings, offering same-day results and predicting disease years in advance.
• Generative models like Foresight uses primary care data to forecast disease progression. Validated at over 95% accuracy, they support early, targeted interventions for conditions like heart failure, CKD, and diabetes.

Technology alone can't deliver a future neighbourhood health service. We will require new workforce models, roles that don't yet exist in traditional primary care. As digital tools become embedded in everyday practice, the workforce around them must evolve too.

Clinicians of the future must have the skills to blend AI outputs with clinical reasoning, interpreting algorithms as part of a broader clinical decision-making process. We will need data health workers to use population-level insights, using analytics to spot patterns, anticipate needs, and drive the delivery of targeted interventions. We will need digital care coordinators to guide patients through hybrid care journeys as they navigate online and face-to-face care. These examples of essential roles that will be required to ensure we digitise care in a way that is safe, personal and inclusive.

We will also need a financial shift. Over the next few years, we'll see investment shift from hospitals to neighbourhood care. The traditional GP partnership structure, optimised for activity and volume will be insufficient to meet the needs of population health. Instead, greater outcome-focused funding frameworks will be required that reward collaboration, equity, and prevention. Predictive analytics will guide resource allocation toward communities with the greatest burden of need, aligning money not just with demand, but with impact.

However exciting this tech enabled future is, this transformation also carries significant risk.

I've talked frequently about how AI use in healthcare risks perpetuating existing biases and inequalities, disproportionately impacting marginalised populations when these disparities aren't considered during development stages. With the growth in the use of AI, this will be an increasingly critical patient safety concern. A 2024 study by Oxford University found that 10% of NHS patients lacked any ethnicity record, and 12% had conflicting data, AI can perpetuate systemic biases if built on incomplete or skewed data, undermining model fairness from the outset.

Digital exclusion is another well known barrier. Those with the highest health risks are often the least digitally equipped. As neighbourhood models increasingly rely on digital triage, remote monitoring, and online portals, we risk creating a two-tier system where the digitally confident receive proactive, preventive care, while others only receive care when they are in crisis.

The redesign of primary care is not just about technology or efficiency. It's about reimagining the relationship between people and the system that serves them. Done well, AI-enabled neighbourhood health services could democratise early detection, reduce health inequalities, and create a more responsive, preventive model of care. Done poorly, they risk entrenching digital divides, eroding trust, and perpetuating biases.

The choice is ours. The time to shape that choice is now.