Primary care, whether in the United Kingdom or wider, is one of the most rewarding careers in the medical sector. It is the backbone of a health system and the gateway to the NHS. Although not a path I initially chose, it has been the best decision, and I can safely say it has allowed me to have a comprehensive and holistic overview of a patient’s life. However, UK general practice is at a tipping point. A stagnant or declining General Practice (GP) workforce, increasing patient numbers, high consultation volumes, and burdensome administrative duties have resulted in unsustainable conditions. Without intervention, these challenges jeopardise the quality and accessibility of primary care. This context opens the door for digital health and AI innovations as potential solutions to support GPs and improve patient care.

Challenges Facing UK General Practice and GPs

Workforce shortages and workload

General practice in the UK is under intense strain from a shrinking workforce and rising patient demand. GP numbers have stagnated over the past decade, failing to keep up with population growth. As of January 2025, England had about 28,235 full-time equivalent (FTE) GPs – 1,129 fewer FTE GPs than in 2015, even as patient registrations climbed by nearly 7 million. Fewer doctors care for more patients; the average GP now has 2,258 patients, 17% more than in 2015. Efforts to recruit GPs have lagged behind government targets, and many experienced GPs are leaving or reducing hours. Over 40% of GPs report considering leaving the profession within five years, reflecting a critical workforce retention problem. This shortfall directly impacts care continuity and access, especially in deprived areas.

Rising patient demand

In late 2024, GPs provide over 1.4 million consultations daily, totalling approximately 369 million appointments in the year leading to January 2025. This surge is driven largely by an ageing population with complex health conditions; patients over 70 have, on average, five times more GP visits than younger people. UK GPs manage a wider range of care, including paediatrics, maternity, and mental health, increasing their workload. Many full-time GPs see over 35 patients daily, surpassing the British Medical Association’s safe limit of 25, leading to stress and lowered consultation quality. The 10-minute appointment model, meant for efficiency, often results in unresolved issues being passed to other parts of the system.

Administrative burden and burnout

GPs face a significant administrative load, spending up to 40% of their working hours on tasks like managing correspondence, referrals, and documentation. This leaves less time for patients, often forcing them into unpaid overtime and contributing to burnout and poor work-life balance. Bureaucracy and “box-ticking” add to this stress, with 60% of clinicians identifying documentation as a key stressor. Many GPs have reduced their clinical hours yet still face unpaid work. These pressures result in longer wait times, difficulty accessing appointments, and decreased patient satisfaction.

Challenges in chronic disease management

As multimorbidity rises, managing chronic diseases has become more complicated. Each condition requires tailored follow-up schedules, care pathways, and performance metrics. This increase in parallel protocols places a considerable administrative and cognitive load on GPs who must navigate various guidelines while delivering comprehensive, patient-centred care. The Quality Outcomes Framework (QOF) is a UK-wide performance incentive scheme that rewards general practices for achieving evidence-based clinical and organisational targets to enhance patient outcomes. While intended to standardise and elevate clinical standards, it can sometimes focus too much on metrics, potentially overshadowing personalised care. Though it has improved certain aspects of chronic disease management, the strong emphasis on data collection and target indicators may distract GPs from having more holistic consultations. As a result, significant time is spent on documenting compliance instead of nurturing deeper, patient-focused interactions.

How Digital Health and AI Technologies Can Address These Challenges

Emerging digital health and AI tools offer promising opportunities to alleviate pressures on GPs and enhance primary care delivery. Policymakers regard digital tools and artificial intelligence as essential for transforming primary care and easing the strain on clinicians. However, there is broad recognition that genuine digital transformation involves more than merely digitising existing processes. Many have found that NHS digital innovations often replicate traditional workflows instead of fundamentally enhancing them. In this context, a variety of AI-enabled and digital tools have surfaced, each addressing specific pressure points in general practice.

AI-powered triage and online consultation systems

AI-driven triage systems act as a first point of contact, helping to manage patient demand before it reaches the GP. These “digital front doors” interact with patients using chatbots or questionnaires, then provide advice or direct them to the appropriate service by risk stratifying based on the symptom presentation and urgency. In theory, such tools can advise if self-care is suitable, if a pharmacist or physiotherapist can help, or if the patient truly needs a GP or emergency care. By filtering minor ailments and guiding patients appropriately, AI triage can reduce unnecessary GP appointments and ensures patients get the right care at the right time.

Ambient AI assistants

An emerging application is using AI in consultation rooms to automatically transcribe and summarize consultations. These ambient scribes (employing speech recognition and large language models) record the doctor-patient conversations and generate structured clinical notes or letters, thereby reducing the administrative burden.

Predictive analytics for risk stratification

AI and machine learning algorithms are being applied to primary care data to identify various cohorts, such as those at the highest risk of chronic disease complications, patients likely to miss follow-ups, and individuals at risk of hospital admission or certain diseases. Primary care data provide detailed medication histories and routine check-ups that help classify risk more granularly. This information then informs the types of targeted interventions that need to be applied across the system, rather than working in siloed, disease-specific pathways. Although still in its early stages, it holds promise for more proactive, preventive primary care, which should demonstrate greater cost savings further down the line. A healthy population is a less expensive population.

Overcoming Key Barriers: Integration, Trust, and Regulation

Integration with existing clinical systems

Primary care IT infrastructure is often outdated and siloed. New AI tools frequently lack interoperability with GP software, complicating workflow integration. This disconnect results in AI solutions becoming isolated “point solutions” instead of integrating into care pathways and daily workflows, thus adding complexity and reducing efficiency. New solutions depend on established clinical systems, which are key to unlocking their APIs. A top-down mandate is needed for these systems to allow the integration of essential tools, removing bureaucratic hurdles that delay implementation.

Trust and reliability concerns

The key consideration here is establishing trust in AI to provide accurate results. However, current large language models still hallucinate, making it essential to involve a clinician in the loop. If the AI makes an error and the clinician does not identify it, the AI cannot be held liable. Ongoing research is identifying algorithms to mitigate hallucination. One approach involves using real-time references to validated medical data sources and prompting LLMs to highlight uncertainty, enabling clinicians to identify and correct potential hallucinations. Another approach is adopting a “clinician-in-the-loop” workflow, where final decisions remain anchored in professional expertise to reinforce trust and user confidence. However, a standardized method to validate LLMs is still needed, and further work is necessary to address this.

Data privacy, security, regulation, and clinical safety

Various regulatory frameworks govern AI in NHS primary care to ensure safety and efficacy. AI systems need patient data, raising concerns about confidentiality and data use. Information governance standards ensure compliance with NHS policies (GDPR, national data opt-out, etc). A Data Protection Impact Assessment (DPIA) is required whenever patient data is being accessed and used. The Medicines and Healthcare products Regulatory Agency (MHRA) regulates medical devices, including software and AI. The Digital Technology Assessment Criteria (DTAC) sets standards for clinical safety, data protection, cybersecurity, interoperability, and usability.

All digital tools being implemented need to be DCB0129 and DCB0160 compliant. DCB0129 lays out the mandatory processes and responsibilities for manufacturers to identify, assess, and mitigate clinical safety risks in health IT systems. DCB0160 shifts the focus to healthcare organizations themselves, guiding them in safely deploying, monitoring, and managing these systems within their local environments.

Bridging the Digital Divide: Empowering Patients and Clinicians for AI-Driven Care”

As primary care digitizes, we must address barriers preventing patient access to new services. Digital exclusion is a risk: not all patients have the means, skills, or desire to use online tools. Elderly patients, low-income individuals, those with disabilities, and residents in rural areas with poor connectivity may struggle with digital health services. Concerns exist that a reliance on apps and remote care might erode the valuable doctor-patient relationship, especially for older generations. If unaddressed, a digital-first approach could create a two-tier system benefiting the tech-savvy while leaving others behind. It's essential to support patients in developing digital skills; communities have introduced Digital Inclusion Champions to coach patients on using NHS applications. Libraries, community centers, and charities partner with the NHS to offer training and loan devices to those in need. By prioritizing inclusivity – co-designing with users lacking digital literacy and ensuring accessibility – primary care can prevent further inequalities.

Staff utilizing these tools also need to feel empowered and confident as adopting advanced AI tools necessitates new skills and training that many primary care staff have not received. GPs and nurses may not feel confident in using AI-driven software or interpreting its outputs without proper training. Building clinician digital literacy, data skills and understanding of the benefits and limitations of AI is essential. Without adequate training and change management, even the best tools could be underutilized or used incorrectly.

Conclusion:

UK primary care stands at a crossroads, confronting unsustainable workloads and escalating GP burnout. Against this backdrop, policymakers and healthcare leaders envision AI and digital tools as potential catalysts for a healthier, more efficient system. Each category of tool – from automated triage to ambient clinical documentation – is designed to offload burden from GPs, streamline administrative tasks, or enhance clinical decision-making. These pressures also present an opportunity for digital health and AI innovations to alleviate operational strains and enhance both patient care and clinician well-being. To capitalize on these solutions, clinicians and policymakers must integrate new technologies seamlessly into daily workflows, ensure robust data governance and privacy safeguards, implement within standardized frameworks, and invest in comprehensive training and inclusive design. By combining human clinical expertise with thoughtfully implemented digital tools, UK general practice can remain a cornerstone of holistic, patient-centred care and evolve to meet the complex needs of a rapidly changing healthcare landscape.

References

1. Baird B, Charles A, Honeyman M, Maguire D, Das P. Understanding pressures in general practice. London: The King’s Fund; 2023.
2. Marshall M, Patel A. The future of general practice: Workforce challenges and potential solutions. BMJ. 2022;377:e067874.
3. Royal College of General Practitioners (RCGP). General practice workload survey 2023. London: RCGP; 2023.
4. NHS England. GP Patient Survey 2023: National results. London: NHS England; 2023.
5. Nuffield Trust. Multimorbidity and long-term conditions in primary care: Challenges and responses. London: Nuffield Trust; 2022.
6. NHS England. 2023/24 Quality and Outcomes Framework guidance for General Medical Services contract [Internet]. London: NHS England; 2023 Available from: https://www.england.nhs.uk/publication/2023-24-qof-guidance-for-gms-contract/

--Issue 69--