Author: Azeem Majeed

I am Professor of Primary Care and Public Health, and Head of the Department of Primary Care & Public Health at Imperial College London. I am also involved in postgraduate education and training in both general practice and public health, and I am the Course Director of the Imperial College Master of Public Health (MPH) programme.

Managing Complications of Overseas Medical Procedures

General practitioners and emergency medicine doctors in the UK are increasingly encountering patients who return from overseas with complications following medical procedures. These cases can often be challenging to manage. Frequently, there is limited or no access to operative notes, discharge summaries, or detailed information about the techniques and materials used, making clinical assessment and safe follow-up difficult.

The complications themselves are often complex, including serious infections, wound breakdown, thromboembolic events, and implant-related problems. Many require urgent specialist input. GPs are typically the first point of contact and must manage patient distress, clinical uncertainty, and risk, while navigating referral decisions in the absence of clear guidance or established care pathways.

For patients, the lower cost of surgical procedures overseas can be an appealing alternative to private care in the UK. However, this often comes at the expense of structured follow-up, continuity of care, and access to the original operating team. Once complications arise, patients may find themselves without support from the healthcare provider who delivered the procedure.

The impact of these cases extends beyond general practice and is increasingly felt in NHS emergency departments. Patients may present acutely with sepsis, bleeding, wound failure, or suspected thromboembolic disease, often without any reliable documentation of the original procedure. This creates significant diagnostic and risk-management challenges for clinicians working in emergency departments, leading extensive investigations, senior clinician input, and sometimes to precautionary admission to an NHS hospital. These presentations add to the pressures on already overstretched emergency services.

These presentations place a significant additional burden on general practice. Consultations are typically longer and more complex, requiring careful documentation, risk management, and coordination with secondary care. This work is undertaken within already stretched services and is often compounded by medico-legal uncertainty. More broadly, the NHS absorbs the cost and workload of managing complications from procedures that were neither planned nor delivered within the UK healthcare system.

Together, these challenges highlight the need for improved patient awareness of the risks associated with overseas medical treatment, clearer clinical pathways for managing post-procedure complications, and greater recognition of the pressure placed on NHS services; including general practice, emergency departments, and specialist teams. Addressing these issues will be increasingly important as international medical travel continues to grow.

Embedding AI Error Detection Into Primary Care Safety Culture

As artificial intelligence (AI) become increasingly embedded in routine healthcare – supporting tasks such as triage, documentation, interpretation of investigation, diagnosis and patient communication – it introduces new patient safety risks through incorrect outputs (“hallucinations”) that should be treated as safety errors rather than technical glitches. In our article in the Journal of Patient Safety, we argue that primary care must extend its established safety culture to AI by systematically detecting, classifying, reporting, and learning from AI-related errors using principles already applied to human error, such as audit, governance, and incident reporting.

We highlight evidence that AI-generated clinical text can contain omissions, fabrications, or unsafe recommendations that may not be apparent to clinicians and patients and that risk becoming “silent errors” in electronic health records. These errors can then contribute to cognitive offloading if clinicians over-trust AI outputs. To mitigate these risks, we call for routine AI oversight in practice (including review, sampling, and escalation), explicit clinician accountability for AI-influenced outputs, patient engagement in spotting discrepancies, and closer collaboration with AI developers.

Ultimately, AI errors are inevitable, and that embedding AI safety as a core, proactive design feature – rather than an afterthought – is essential to ensure AI enhances rather than compromises patient safety in primary care.

The role of vaccination, infection control measures and early treatment in curbing the impact of flu

Influenza remains a major cause of preventable illness each winter and continues to place significant pressure on NHS general practices, urgent care services, and hospitals. This has been particularly evident this winter, with flu rates much higher than we would normally expect for this time of year.

As of mid-December 2025, UKHSA surveillance shows influenza positivity in primary-care sentinel samples running well above most pre-COVID seasons, and hospital and ICU admissions for confirmed influenza are rising sharply — especially among adults aged 65 and over and those with long-term medical conditions. In general practice, we see first-hand how flu can lead to severe complications, particularly in older adults, people with underlying conditions, and those who are immunocompromised.

Vaccination remains the single most effective way to reduce the risk of severe illness, hospitalisation, and death from flu. Interim data for the 2025–26 season suggest that vaccination is already reducing the risk of GP consultations and hospital admissions in older adults. In primary care, we also see the wider benefit of vaccination in helping to protect the NHS by reducing demand on GP appointments, out-of-hours services, and emergency care during the winter months.

General practices — working closely with community pharmacies — play a crucial role in delivering the flu vaccination programme, identifying eligible patients, reaching underserved groups, and offering trusted, personalised advice. Many patients value the opportunity to discuss vaccination with their GP or practice nurse, and these trusted relationships are central to addressing concerns, tackling vaccine hesitancy, and improving uptake. It is absolutely still worth vaccinating eligible patients now; protection begins within 10–14 days.

Alongside vaccination, simple practical measures remain important: good hand hygiene, staying at home when unwell, and seeking timely advice if symptoms worsen. For higher-risk patients who develop suspected influenza, antiviral treatment with oseltamivir or zanamivir is recommended if started within 48 hours of symptom onset — or later in severe illness or immunocompromised patients. Those eligible for antiviral treatment include people aged 65 and over, pregnant women, and individuals with long-term conditions or immunosuppression.

Flu should not be underestimated. For some patients it can be life-threatening, but many cases are preventable through vaccination, early treatment in high-risk groups, and sensible infection-prevention measures.

Relevance Over Recall: Rethinking How AI Uses Clinical Data

Our article in the Journal of the Royal Society of Medicine argues that safe and effective AI in healthcare must incorporate mechanisms that emulate human judgement – down-weighting old, inaccurate or superseded information and prioritising what is recent, clinically relevant and reaffirmed – so that AI supports, rather than disrupts, high-quality patient care.

Clinicians constantly revise, reinterpret and filter past information so that only what is relevant, accurate and timely shapes present-day management decisions; medical records function as dynamic “working tools” rather than fixed archives. By contrast, many AI systems lack this capacity for selective forgetting and often treat all historical data as equally meaningful.

This can lead to outdated or low-confidence diagnoses being repeatedly resurfaced, persistent labels influencing clinical expectations, and irrelevant, long-resolved events cluttering summaries and decision-support outputs. Such indiscriminate recall not only risks misdirecting clinical care, but also adds to information overload, exacerbates cognitive burden and contributes to clinician burnout. Importantly, it can also undermine patient trust when obsolete or stigmatising terms continue to shape interactions with the clinicians and the healthcare system.

Getting mental health diagnoses right without undermining access to care and disability rights

The UK government’s forthcoming review of mental health and neurodevelopmental diagnoses presents an opportunity to improve the healthcare and benefits system if the potential risks are averted. Rising rates of conditions such as ADHD, autism, and anxiety disorders have raised questions about whether we are seeing a genuine increase in need or greater awareness and possible over-diagnosis. A thoughtful, evidence-based review could help bring much-needed clarity. But if mishandled, it could deepen inequalities and undermine support for those who need it most.

Done well, the review could improve diagnostic quality and reduce the postcode lottery that too often defines access to assessment and treatment. Clearer clinical standards and properly funded services would allow professionals to make more accurate diagnoses, shorten long waiting lists, and better match interventions to individuals’ needs. This is an outcome everyone should welcome.

But the review must not become a vehicle for restricting access to treatment, reasonable adjustments, or disability benefits. The increase in diagnosed mental-health and neurodevelopmental conditions reflects, in large part, years of unmet need and increasing public willingness to seek help. Tightening diagnostic thresholds or narrowing eligibility criteria risks penalising individuals who already face significant barriers to care, particularly those from disadvantaged communities. It also risks pushing desperate families towards unregulated private providers.

Framing this issue as one of “too many diagnoses” is unhelpful. It risks stigmatising people who are already struggling and undermines recent progress in public understanding of neurodiversity and mental health. The question we should be asking is not “How do we reduce the numbers?” but “How do we ensure people receive timely, appropriate and equitable support?”

If the government wants to improve outcomes, the path forward is clear: strengthen diagnostic pathways, expand clinical capacity, invest in multidisciplinary teams, and safeguard the rights and protections that enable people to live well. Any attempt to use the review as justification for cuts, gatekeeping, or narrowing access would be a profound step backwards.

A better mental health system is one that supports people — not one that seeks to reduce the number who qualify for support. The review should be a catalyst for improvement, not an excuse for limiting care.

Balancing Innovation and Affordability: The New UK Approach to Drug Pricing

The announcement of a new UK-US pharmaceuticals deal is an important change in the UK’s approach to how new medicines are evaluated, priced and adopted. Faster access to innovative treatments for conditions such as cancer will be welcomed by patients and professionals. The increased investment in medicines may also help the UK attract more clinical research.

However, the impact of the proposals will depend on implementation. Raising NICE’s cost-effectiveness thresholds will increase overall NHS spending on medicines. Without a corresponding investment in areas such as workforce, diagnostics and primary care, there is a risk that higher drug spending could divert resources from other parts of the NHS. A more flexible pricing environment could also reduce the UK’s future negotiating leverage with industry. Maintaining NICE’s independence will be essential to maintain both public and professional confidence in its decision-making.

The changes have could benefit patients and strengthen the UK’s life sciences sector. But they will require careful monitoring to ensure that the promised improvements in access, equity, and health outcomes are realised; whilst also managing the financial challenges for an already pressured NHS.

Teaching medical trainees to see societal infrastructure as a clinical issue

In an article published in the journal Frontiers in Medicine, we argue that medical education must broaden its focus: rather than treating infrastructure such as housing, transportation, energy, water supply as only a public-health or social background issue, trainees should regard infrastructural deficiencies as direct clinical determinants of patient health.

We highlight concrete examples (e.g., missed appointments due to transport failures, disrupted dialysis from electricity outages, contaminated water causing toxicity) showing how infrastructure can precipitate or worsen clinical problems. We propose educational innovations: embedding infrastructure-related history-taking, case-based learning driven by infrastructural triggers, community placements in underserved areas, and interdisciplinary learning (with urban planners, engineers, public health) to equip future clinicians with “systems-citizen” skills and advocacy capability.

The goal is to reframe clinical practice to include infrastructure as a proximal driver of disease, thereby reducing health inequities and enabling clinicians to engage meaningfully in structural interventions beyond traditional biomedical care.

From Lloyd George Envelopes to Artificial Intelligence: The Evolution of Medical Records in Primary Care

I spoke to GP Registrars on the Imperial College GP Training Scheme about the evolution of medical records in primary care. This is a journey that mirrors the broader transformation of healthcare itself.

The story begins in 1911, with the introduction of the Lloyd George Envelope following the National Insurance Act. These brown paper envelopes (named after the then Chancellor and future Prime Minister, David Lloyd George), each containing a patient’s handwritten medical notes and printed correspondence, became the standard for decades. They were simple, portable, and remarkably durable but also limited by their physical nature. Searching for information meant literally leafing through these paper records, and continuity of care relied on legibility and the clinician’s diligence in recording.

The late 20th century brought a revolution: the computerisation of general practice. Early adopters in the 1980s and 1990s began using systems like EMIS and Vision, digitising the record and transforming how we document, code, and retrieve information on people’s healthcare. Over time, these systems became essential clinical tools by enabling prescribing safety checks, audit, population health management as well as research and quality improvement.

Today, almost every consultation, prescription, and referral is logged electronically. The electronic health record (EHR) has become the backbone of primary care. Yet, despite these advances, challenges remain: data fragmentation across systems, burdensome data entry, and limited interoperability between sectors.

Looking ahead, I believe we are on the verge of another major transformation. Artificial intelligence (AI) and machine learning have the potential to redefine how we record, understand, and act on health information. In the coming years, we can expect:

  • AI-assisted consultation notes, automatically generated from clinician–patient conversations.

  • Predictive analytics, helping us identify patients at risk of deterioration or with unmet health needs.

  • Natural language processing, allowing clinicians to query records using plain English.

  • Integrated patient data, linking information from hospitals, social care, and personal devices.

The shift from the Lloyd George envelope to intelligent digital systems represents more than just technological progress. It also reflects an ongoing effort to improve care, enhance safety, and make information work for both clinicians and patients.

As we look to the future, the challenge will be ensuring that these innovations support, rather than replace, the human connection that is at the heart of general practice. The tools may change rapidly, but our purpose remains the same: delivering compassionate, person-centred care grounded in good records and good professional relationships.

Rethinking NICE Cost-Effectiveness Thresholds: Implications for the NHS and UK Industrial Strategy

There has been recent discussion about the need to revise drug pricing frameworks within the United Kingdom’s National Health Service (NHS), particularly amid the ongoing transatlantic trade frictions involving potential tariffs from the United States administration.

Elevating the cost-effectiveness threshold applied by the National Institute for Health and Care Excellence (NICE) by 25 percent from its established range of £20,000 to £30,000 per quality-adjusted life year (QALY) would increase access for NHS patients to innovative treatments that were previously excluded on grounds of excessive cost relative to their clinical benefits.

However, this change would also put increased pressure on the NHS budget. It is difficult to quantify the extra spending that might result from a wider range of drugs becoming available for use in the NHS through this change but any extra spending on these treatments would have to be matched by reductions in spending on other health services. Effective implementation would therefore require not only additional funding but also robust mechanisms for monitoring real-world effectiveness and ensuring that new treatments deliver value commensurate with their higher costs.

The government would also need to consider any benefits that might occur from increased investment in research and development in the UK by global pharmaceutical companies. From a trade and industrial policy perspective, revising the NICE cost-effectiveness threshold would have broader implications for the UK’s position in global pharmaceutical markets. A more permissive pricing environment may enhance the attractiveness of the UK as a destination for clinical trials, research, and early market access, reinforcing the government’s ambition to establish the UK as a “science superpower.”

But it could also be perceived internationally as a shift towards higher healthcare costs, potentially complicating trade negotiations that involve intellectual property protections, market access, and pricing transparency. The Department for Business and Trade would need to balance these considerations carefully, ensuring that any change supports the competitiveness of the UK life sciences sector while maintaining affordability and equity within the NHS.

Hence, this is not a straightforward issue and the Department of Health and Social Care, the Department for Business and Trade, and the Treasury may all have differing views about the relative costs and benefits of the change. Ultimately, the decision will depend on the government’s political and economic priorities and its assessment of the relative importance of the competing costs and benefits.

The importance of coding Long Covid in electronic medical records

As the world continues to grapple with the aftermath of the COVID-19 pandemic, Long Covid has emerged as a significant public health challenge. Characterised by persistent symptoms like fatigue, brain fog, shortness of breath, and joint pain lasting weeks, months or even years after an infection, Long Covid affects millions globally. Yet, one major hurdle in understanding and addressing this condition is its under-recording in electronic medical records (EMRs).

Accurate coding of Long Covid in EMRs is essential for studying its epidemiology, improving patient care, and managing its impact on healthcare systems and on societies. Electronic medical records are at the core of modern health systems and have largely replaced the more traditional paper-based records used by healthcare providers for many decades. Electronic medical records are used to track patient diagnoses, treatments, and clinical outcomes. When Long Covid is not properly coded, it becomes difficult to use this data to carry out these tasks. Accurate coding using standardised systems such as SNOMED enables researchers and healthcare providers to:

  • Track Prevalence and Trends: Consistent coding allows us to estimate how many people are affected by Long Covid, identify demographic patterns, and monitor changes over time.
  • Study Risk Factors and Outcomes: Properly coded data helps researchers pinpoint who is most at risk and how Long Covid impacts long-term health.
  • Optimise Healthcare Resources: Understanding the true burden of Long Covid helps health systems allocate resources, plan specialised clinics, and train clinicians.
  • Support Policy and Funding: Reliable data informs public health policies and justifies funding for Long Covid research and treatment programs.

Without accurate coding, Long Covid remains under-represented in the data that shapes healthcare decisions.

The Challenges of Coding Long Covid

Despite its importance, coding Long Covid in EMRs faces many challenges:

  1. Lack of Standardised Coding: The ICD-10 and SNOMED codes exist but their use is inconsistent. Many clinicians may not apply it, either due to unfamiliarity or because symptoms don’t neatly fit into a code’s description.
  2. Symptom Overlap and Complexity: Long Covid presents with a wide range of symptoms that overlap with other conditions like chronic fatigue syndrome or fibromyalgia. Clinicians may code individual symptoms (e.g., fatigue) rather than linking them to Long Covid, fragmenting the data.
  3. Confusion with other medical problems. Where patients have other long-term medical problems – such as chronic lung disease or cognitive impairment – new symptoms may be be assumed to be of these conditions rather than from Long Covid.
  4. Limited Clinician Awareness: In busy clinical settings, Long Covid may not be recognised or prioritised, especially if patients present with vague or multisystem symptoms. Lack of training on Long Covid’s diagnostic criteria exacerbates this issue.
  5. Under-reporting by Patients: Some patients may not seek care for persistent symptoms, or their concerns may be dismissed, leading to no record of Long Covid in EMRs.

These barriers lead to underestimation of Long Covid’s impact, which in turn limits research and resources needed to improve diagnosis and treatment.

The Consequences of Under-Coding

When Long Covid is not properly documented, the consequences ripple across patients, providers, and systems:

  • For Patients: Misdiagnosis or lack of a formal Long Covid diagnosis can delay treatment, leaving patients struggling without validation or access to specialised care.
  • For Research: Incomplete data hinders epidemiological studies, making it harder to understand Long Covid’s long-term effects or develop evidence-based treatments.
  • For Healthcare Systems: Without clear data, hospitals and clinics can’t plan for the growing demand for Long Covid care, leading to strained resources and inequities in access.
  • For Public Health: Policymakers rely on EMR data to justify funding and programs. Under-coding obscures the scale of the problem, potentially stalling progress.

Solutions to Improve Coding

Addressing the under-coding of Long Covid requires a multi-pronged approach:

  1. Standardised Coding Protocols: Healthcare systems should promote the consistent use of ICD-10 and SNOMED codes. Clear guidelines are also needed for when and how to apply these codes. Future iterations of coding systems could include more specific Long Covid codes to capture its diverse presentations.
  2. Clinician Training: Educating healthcare providers about Long Covid’s symptoms, diagnostic criteria, and coding practices is essential. Continuing medical education (CME) programs can bridge knowledge gaps and encourage proactive documentation.
  3. Technology and AI: Natural language processing (NLP) tools can analyse unstructured EMR data, such as clinician notes, to flag potential Long Covid cases that might otherwise go uncoded. Integrating these tools into EMR systems could improve case identification.
  4. Patient Awareness: Public health campaigns can encourage patients to report persistent symptoms and advocate for themselves, ensuring their conditions are documented.
  5. Research and Collaboration: Partnerships between health systems, researchers, and policymakers can drive the development of better diagnostic and coding frameworks, informed by real-world data.

Conclusions

The under-recording of Long Covid in electronic medical records is more than a technical issue. It is a barrier to understanding and addressing a condition that affects millions of people globally. Accurate coding is the foundation for robust research, effective patient care, and informed public health strategies. By prioritising standardised coding, clinician education, and innovative technologies, we can shine a light on Long Covid’s true impact and pave the way for better outcomes.

For healthcare providers and clinicians, the message is clear: document Long Covid deliberately and consistently. For patients, it is about advocating for your health and ensuring your symptoms are recorded. And for health systems, it’s about investing in the tools and training needed to make electronic medical records an effective tool in managing Long Covid at patient, healthcare provider and national levels.