Pediatric Clinic( LS) عيادة الأطفال

04/09/2026

Your patients are being targeted by doctors who don’t exist

AI-generated “medical experts” are spreading misinformation rampantly on social media, with real-world consequences for health.

Scroll through social media and you may encounter a doctor speaking confidently to camera about carcinogens, natural antibiotics or pain-relief remedies. They may appear knowledgeable, dressed in scrubs with framed diplomas behind them. They may even talk about what they recommend to their own patients.

Except many of them are not doctors at all. They are AI-generated avatars, created to look and sound like real healthcare professionals.

And they are flooding feeds with misinformation that can frighten patients, distort treatment decisions and put their health at risk.

According to an analysis done by the digital marketing firm Hallam and reported on by The Guardian, an alarming 40% of the top health-related videos on TikTok contain AI-generated material, much of it from a number of synthetic influencer accounts made to look like doctors. The review included 1,198 highly ranked videos across searches including “health tips”, “fitness” and “doctor advice”. For “health tips”, the prevalence of AI-generated material soared to 84%.

the leading AI-generated doctor and health videos amassed a staggering 2.5 million views. Among the more popular were warnings that microwaving food in plastic, using deodorant and sleeping next to a mobile phone could cause cancer.

Some AI avatars are easy to spot as artificial. But as the technology evolves, others are becoming disturbingly realistic, and telling the difference is nearly impossible. A 2024 systematic review covering 56 studies and more than 86,000 participants found that they detected deepfakes only about 55 per cent of the time – barely better than flipping a coin.

This is especially troubling when such bogus medical content reaches patients with debilitating conditions searching for answers. These fake videos borrow the credibility of medicine and exploit the trust we place in healthcare professionals, making misinformation appear far more authoritative and reliable than it is.

A New York Times investigation found hundreds of AI-generated doctors, healers and wellness influencers promoting supplements to older women and people living with health problems. In one example, an AI-generated avatar claimed that a herbal supplement could treat stage 3 kidney disease better than medication.


To fight medical misinformation, more doctors are becoming influencers
Frustrated by the rise and rapid spread of ill-founded health advice on social media, a growing number of doctors are fighting back with facts – one video at a time

The investigation also showed how vulnerable patients can be drawn in. One 71-year-old woman with autoimmune-related health issues described being drawn to Facebook advertisements for a moringa supplement because she hoped it might help with her joint problems, inflammation and fatigue. After taking it daily for months, she discovered that it had been urgently recalled after the FDA and CDC linked the supplement to salmonella infections.

It is also very worrying just how cheap this fakery is to produce: one Chinese AI advertising agency said it generates around 1,200 AI videos a day, for as little as $10 each. At that price and volume, counterfeit medical authority can now be manufactured at industrial scale.

TikTok’s policies say it removes health misinformation that could lead to serious harm and prohibits “false endorsements or impersonations of health professionals”, while Meta says it applies “AI info” labels to detected AI-generated content.

Still, removal is mostly reactive, can come too late and is not always guaranteed. Doctors may not even know they have been deepfaked until someone recognises them. And getting impersonations removed can itself be an ordeal.

That’s what happened with public-health doctor Professor David Taylor-Robinson. He didn’t know he had become a victim of impersonation until someone emailed him about it. TikTok initially found no violation, and the deepfake remained live despite being repeatedly flagged. By the time the malicious account was eventually removed, the video exploiting his likeness had attracted more than 365,000 views.

In the clinic

Beyond reputational harm, the burden of medical misinformation is spilling over from patients’ feeds into healthcare appointments. A 2026 World Cancer Research Fund survey found that a striking 44% of patient-facing NHS staff encounter inaccurate or misleading nutrition and supplement claims at least weekly.

Unfortunately, pointing out that these supposed experts are not what they seem doesn’t necessarily undo the harm. In two experiments involving 1,410 people exposed to cancer misinformation, simply correcting the false claim was often more effective in changing the person’s views than challenging the supposed expert’s credentials.

Expecting doctors to repeatedly unpick algorithmically amplified misinformation within already pressured appointments is hardly a scalable solution in any case – even if it is already becoming the norm. A systematic review of 22 studies recently found that doctors routinely have to serve as interpreters of patients’ online health information, with those who have done the most internet sleuthing placing greater demands on consultation time.

While medical professionals can be held professionally and legally accountable for the advice they give, who is liable when a fake AI doctor causes harm: the agency that created it, the account-holder profiting from it, the AI provider whose technology was used to generate it, the platform that amplified it, or all of them?

The tools to flag dangerous content are improving fast. Whether platforms choose to use them is another matter

Ofcom has proposed new measures under the Online Safety Act to tackle fraudulent advertising and misuse of AI advert-generation tools. But fake doctors and harmful health misinformation are not confined to paid ads. Platforms can push them into millions of feeds without a penny spent on promotion, regardless of the safety consequences.

Meanwhile, doctors are being left to deal with misinformation they did not create, distributed by systems they do not control. Worse, the longer counterfeit medical authority is allowed to proliferate, the greater the risk to patients and to the trust on which legitimate medical care depends.

But not all is lost, and there are promising ways forward. AI-content detection systems already exist, although identifying something as AI-generated is not the same as determining whether its advice could cause health harm. Harmful misinformation should be flagged, AI-generated or not.

Research at UCL is exploring how doctors and other health professionals can bring their expertise into misinformation risk assessment, helping define when online content crosses dangerous “red lines” and needs to be flagged, pushed down feeds or removed. The tools to do this are improving fast. Whether platforms choose to use them is another matter.

iStock. Explore the official iStock website for millions of exclusive, royalty-free, stock files. Find the perfect stock for your project, fast. Search now.

27/08/2026

FDA authorises first-of-a-kind robotic blood draw device
The US Food and Drug Administration (FDA) authorised a standalone robotic device, dubbed Aletta, the first of its kind, capable of drawing blood from a patient’s arm without the intervention of a hands-on operator.

The device is cleared for use in adults in outpatient settings under supervision of a trained phlebotomist. Each phlebotomist can supervise

“Blood draws are one of the most commonly performed medical procedures in the US, yet patients may face delays due to a growing shortage of trained phlebotomists,” said Michelle Tarver, Director of FDA’s Center for Devices and Radiological Health.

The robotic device employs near-infrared light and Doppler ultrasound to identify suitable veins and differentiate them from arteries. But if it cannot find a suitable vein, the device will not attempt the procedure. During an autonomous session, the device performs the application of a tourniquet, preparation of the skin, insertion and disposal of the needle, management of collection tubes and application of a bandage. The session is started by a supervising phlebotomist who checks the correct sequence of tube filling and is available to deal with any issues that arise.

The clinical data supporting the authorisation demonstrated the device achieved successful blood draw rates that were equal to or better than those of trained human phlebotomists. These findings were seen in a broad patient population that consisted of patients of various skin tones, health status and self-reporting difficult vein access. Adverse events related to the device were mild and infrequent.

Safety features include the continuous application of disinfectant during scanning, and sensors that detach the needle automatically if the patient moves too much. The approval was granted to Vitestro through the De Novo pathway, with special controls for performance and clinical testing to assure safety and effectiveness.

Healthcare professionals should report all serious adverse events suspected to be associated with the use of any medicine and device to FDA’s MedWatch Reporting System or by calling 1-800-FDA-1088.

Reference: https://www.fda.gov/news-events/press-announcements/fda-authorizes-first-its-kind-robotic-blood-draw-device

SOURCE: US Food and Drug Administration

The U.S. Food and Drug Administration today authorized the Aletta, the first standalone robotic device that can draw blood from a patient’s arm without hands-on operator intervention.

24/08/2026

A doctors’ guide to social media posts

A useful way to think about social media risk is to place a post into one of three zones: green, amber or red.

Green zone: usually firm footing

Green zone does not mean no one will disagree with you. It means the post is more likely to be defensible because it is respectful, contextualised and consistent with professional standards.

A GP posting about the impact of delayed cancer referrals, unsafe workload, poverty, housing, vaccination uptake, NHS funding, workforce morale or access to care is usually participating in legitimate public debate, for instance. The same may apply to careful comment on political or moral issues, provided the language remains lawful, proportionate and non-abusive.

You are generally on safer ground when you are:

commenting on health policy, NHS pressures, workforce issues, patient safety or public health
expressing a genuinely held personal view respectfully
making clear what is opinion rather than fact
avoiding patient details, personal attacks and inflammatory language
willing to explain the post calmly if challenged

Amber zone: pause and assess

Amber zone does not necessarily mean “do not post”. It means slow down.

Ask yourself: can I say this more clearly, more fairly, more proportionately or with better context? Have I separated fact from opinion? Could this be interpreted as hostility towards a group rather than criticism of an idea, policy, government or institution? Could a patient reasonably wonder whether I would treat them fairly?

Take extra care where the post involves:

highly charged political, moral or religious issues
war, conflict or international events
humour, sarcasm or shorthand that could be misread
comments about colleagues, employers, patients or organisations
patient-adjacent stories, even if anonymised
closed groups, WhatsApp messages or private forums
claims linked to your medical status or expertise

Red zone: high risk

Red zone posts can quickly move beyond “controversial opinion” and into professional conduct territory. They may raise concerns about patient safety, confidentiality, discrimination, public confidence or the doctor’s judgement.

Avoid posting, or seek advice first, where the content could involve:

identifiable patient information
abuse, bullying, harassment or discriminatory language
personalised clinical advice to an individual online
misleading or unevidenced health claims
threats, support for violence or dehumanising language
posts suggesting you may not treat certain patients fairly
anything you would struggle to justify to the GMC, your practice, a patient or a responsible officer
Before you post, ask five questions
Is it lawful, honest and proportionate?
Have I protected patient confidentiality completely?
Am I clear what is fact, opinion or allegation?
Could a reasonable patient question my fairness, judgement or ability to treat them without discrimination?
Would I be comfortable explaining this post if it were screen captured and sent to the GMC, my practice or a patient?
If the answer to any of these questions gives you pause, that does not always mean you should stay silent. It may mean you should reword the post, add context, remove unnecessary heat, wait before posting, or seek advice.

24/08/2026

What doctors can – and can’t – say on social media
20 July 2026

Society loses out if doctors feel they cannot exercise their freedom of speech online, but there are important limits to know about.

Doctors are citizens. They are entitled to hold views, express opinions, campaign, criticise policy, participate in public debate and speak out about matters they consider important. That includes political, moral, religious, scientific and social issues.

Doctors often get to see up close the impact of public policy, inequality, health service pressures, misinformation, social division and conflict. It is therefore unsurprising that many feel a professional, moral or personal duty to speak.

The important point is this: being a doctor does not remove your right to freedom of expression, but it does change the context in which your expression may be assessed.

A comment made by a doctor can carry more weight with the public. It may be interpreted as a professional view, even where it was intended as a personal one. And if a concern is raised with the General Medical Council (GMC), the question is not simply whether the doctor was entitled to speak. It is whether their conduct was compatible with professional standards, patient trust, public confidence and the rights of others.

That is where the grey area often lies.

The starting point: freedom of speech
Article 10 of the Human Rights Act 1998 protects the right to freedom of expression. This includes the right to hold opinions and to receive and impart information and ideas. It applies to expressions online as well as offline. It also protects expressions that may be unpopular, controversial, uncomfortable or offensive to some.

However, Article 10 is a qualified right. This means it can be restricted in certain circumstances, including where restrictions are lawful, necessary and proportionate for aims such as protecting the rights of others, public safety, health, or the reputation and rights of others.

For doctors, this means two things can be true at the same time. First, you are entitled to express personal views. Second, your regulator, employer or another body may scrutinise those views if they are said to cross a professional, legal or ethical line.

Doctors are entitled to express personal views. But regulators or employers can scrutinise those if they cross a legal, professional or ethical line

The GMC’s social media guidance recognises this balance. It states that medical professionals, like everyone else, have rights to freedom of belief, privacy and expression, but that exercising those rights as a medical professional must be balanced against the possible impact on other people’s rights and interests.

That recognition is important. The GMC is not there to prevent doctors from participating in public life. Nor should doctors assume that any complaint about a strongly held view means they have done something wrong. A complaint may be distressing and disruptive, but it is not the same as a finding of misconduct.

Professional standards still apply online
The GMC’s position is clear: the standards expected of doctors do not disappear because the communication takes place on social media rather than face to face, in print or in a professional setting.

That does not mean every post by a doctor is treated as clinical practice. Context matters. A personal account, a political comment, a clinical education thread, a professional LinkedIn post and a message in a private WhatsApp group are not all the same. But the boundaries between personal and professional life can blur quickly online, particularly where a doctor is identifiable, discusses healthcare, refers to their work, comments on colleagues or engages with patients.

The British Medical Association (BMA) makes a similar point in its social media guidance. Social media can blur the line between the professional and the personal, creating ethical tensions around confidentiality, professional boundaries, reputation and public trust.

For GPs in particular, this is especially relevant. GPs are often visible in local communities. Patients may recognise them online. A comment intended for a small audience can be screen captured, shared, stripped of context and presented to an employer, regulator or journalist. Privacy settings reduce risk, but they do not remove it.

When doctors should feel on firm footing
Doctors should generally feel more confident speaking where they are expressing a genuinely held view in a lawful, proportionate and non-abusive way, particularly where they are contributing to public debate, patient safety, healthcare policy, professional standards or matters of legitimate public concern.

There is a strong public interest in doctors being able to speak about NHS pressures, unsafe systems, discrimination, public health, workforce concerns, clinical standards and patient care. Doctors also have the right to participate in wider political and moral debate as private citizens.


Doctors should ‘routinely ask’ young patients about social media use
Medical leaders have warned the dangers of social media to the health of children is comparable to smoking and not wearing seatbelts.

A doctor is usually on firmer ground where the post is clearly framed as opinion, avoids personal attacks, does not disclose patient information, does not target individuals or protected groups, does not present misinformation as fact, and does not exploit the trust attached to medical status.

It may also help to distinguish between criticism of governments, institutions, policies or ideologies, and hostility towards individuals or groups based on protected characteristics. The former may fall squarely within legitimate public debate. The latter may raise serious concerns about discrimination, harassment, patient trust and professional conduct.

When to be cautious
The risk increases where the post touches on clinical advice, patient stories, colleagues, identifiable organisations, conflict, protected characteristics, public health claims or emotionally charged political issues.

Doctors should be particularly cautious before posting anything that could be read as abusive, discriminatory, bullying, harassing or demeaning. The GMC expressly warns doctors not to use social media to abuse, discriminate against, bully, harass or deliberately target any individual or group.

The lesson is that context, language, intent, content, evidence and proportionality matter

Confidentiality is another high-risk area. A patient does not need to be named for confidentiality to be breached. A combination of details may be enough for the patient, family, staff member or local community to recognise the case. This is especially important in general practice, where communities can be small and clinical situations distinctive.

Doctors should also avoid giving personalised clinical advice in comments, direct messages or public threads. General information and signposting are usually safer. Advice to a named individual with specific symptoms, medications or circumstances can create risk, even where the person is not formally your patient.

Another area requiring care is health misinformation. The case law shows that professional regulators may be justified in acting where a doctor uses medical credentials to promote claims that undermine public health, contradict widely accepted medical opinion without proper basis, or risk damaging public confidence.

The grey area: controversial views and public confidence
The difficult cases are rarely about professional educational posts or obviously unacceptable abuse. They tend to involve strongly held beliefs, public controversy and contested language.

A recent Medical Defense Society case study illustrates this. A GP was investigated by the GMC following a complaint about personal social media posts concerning the Gaza war. The complaint alleged that the posts were antisemitic, offensive and potentially criminal. The GP denied wrongdoing and maintained that she was expressing legitimate political and moral concerns. Medical Defense Society assisted with a response arguing that the posts were defensible, that the GMC could not determine alleged criminal offences, and that the doctor’s Human Rights Act rights, including Article 10, were relevant. The GMC investigation was closed with no action.


Why doctors are expected to solve everyone’s problems
Our jobs are literally to fix things for people – but if we take that approach in all aspects of life, it is a recipe for disillusionment and burnout

A regulator must balance competing considerations. It must consider public protection, public confidence and professional standards. But it must also respect lawful freedom of expression. A controversial opinion is not automatically misconduct. Causing disagreement is not the same as causing harm. A complaint is not proof that the line has been crossed.

At the same time, doctors should not assume that attaching a political or moral label to a post makes it immune from scrutiny. Posts that glorify violence, use racist or antisemitic tropes, target colleagues, support unlawful acts, spread dangerous health misinformation or suggest that a doctor may not treat patients fairly can become much harder to defend.

Social media risk often comes from what might be called context collapse. A post written for one audience can be seen by another, separated from the tone, timing and conversation that shaped it. A GP may believe they are speaking as a private citizen, while a patient, employer, journalist or regulator may read the same post through the lens of professional status. That is why the safest approach is not to avoid difficult subjects altogether, but to think carefully about how the post may travel.

Doctors should not be frightened out of public life. Society benefits when GPs contribute thoughtful, informed and principled views. But the safest and most effective professional voice is usually one that is clear, evidenced, proportionate and respectful.

A doctors’ guide to social media posts
A useful way to think about social media risk is to place a post into one of three zones: green, amber or red.

Green zone: usually firm footing

Green zone does not mean no one will disagree with you. It means the post is more likely to be defensible because it is respectful, contextualised and consistent with professional standards.

A GP posting about the impact of delayed cancer referrals, unsafe workload, poverty, housing, vaccination uptake, NHS funding, workforce morale or access to care is usually participating in legitimate public debate, for instance. The same may apply to careful comment on political or moral issues, provided the language remains lawful, proportionate and non-abusive.

You are generally on safer ground when you are:

commenting on health policy, NHS pressures, workforce issues, patient safety or public health
expressing a genuinely held personal view respectfully
making clear what is opinion rather than fact
avoiding patient details, personal attacks and inflammatory language
willing to explain the post calmly if challenged
Amber zone: pause and assess

Amber zone does not necessarily mean “do not post”. It means slow down.

Ask yourself: can I say this more clearly, more fairly, more proportionately or with better context? Have I separated fact from opinion? Could this be interpreted as hostility towards a group rather than criticism of an idea, policy, government or institution? Could a patient reasonably wonder whether I would treat them fairly?

Take extra care where the post involves:

highly charged political, moral or religious issues
war, conflict or international events
humour, sarcasm or shorthand that could be misread
comments about colleagues, employers, patients or organisations
patient-adjacent stories, even if anonymised
closed groups, WhatsApp messages or private forums
claims linked to your medical status or expertise
Red zone: high risk

Red zone posts can quickly move beyond “controversial opinion” and into professional conduct territory. They may raise concerns about patient safety, confidentiality, discrimination, public confidence or the doctor’s judgement.

Avoid posting, or seek advice first, where the content could involve:

identifiable patient information
abuse, bullying, harassment or discriminatory language
personalised clinical advice to an individual online
misleading or unevidenced health claims
threats, support for violence or dehumanising language
posts suggesting you may not treat certain patients fairly
anything you would struggle to justify to the GMC, your practice, a patient or a responsible officer
Before you post, ask five questions
Is it lawful, honest and proportionate?
Have I protected patient confidentiality completely?
Am I clear what is fact, opinion or allegation?
Could a reasonable patient question my fairness, judgement or ability to treat them without discrimination?
Would I be comfortable explaining this post if it were screen captured and sent to the GMC, my practice or a patient?
If the answer to any of these questions gives you pause, that does not always mean you should stay silent. It may mean you should reword the post, add context, remove unnecessary heat, wait before posting, or seek advice.

Further Reading
Human Rights Act 1998, Schedule 1, Article 10 This protects freedom of expression while recognising that the exercise of that right may be subject to lawful, necessary and proportionate restrictions.
Using social media as a medical professional. This GMC guidance covers freedom of belief, privacy and expression, public trust, confidentiality, boundaries, discrimination, bullying and harassment.
Good medical practice This GMC framework includes trust, professionalism, confidentiality, honesty, respectful behaviour and public confidence.
Conflict in the Middle East: how we are responding to queries and concerns (PDF) This BMA guidance, also linked to in the body of the article, highlights that social media can blur the professional and personal and create ethical tensions around confidentiality and professional boundaries.
Social media, ethics and professionalism This protects freedom of expression while recognising that the exercise of that right may be subject to lawful, necessary and proportionate restrictions.
Fitness to Practise & Freedom of Speech This article published by the Medical Defense Society and also linked to in the body of this piece includes the Dr A case study and the arguments around Article 10, legitimate political expression and GMC fitness to practise proceedings.
Social Media and Dentistry – How to Post Without Getting Sued: 6 Top Tips This article by the Dental Defence Society includes practical points on consent, confidentiality, professional tone and online risk.
Adil v General Medical Council [2023] EWHC 797 (PDF) This case and later appeal material, illustrates that Article 10 is engaged in regulatory cases, but may be restricted where necessary and proportionate, particularly where public health misinformation and public confidence are in issue

Rohan Simon is CEO of the Medical Defense Society

iStock has millions of exclusive stock photos, videos, vectors, and illustrations that are researched and shot by a diverse global community. © 2023 iStockphoto LP.

24/08/2026

Defensive medicine: why doing less can be the hardest thing

By Michael Marshall

Doctors sometimes make clinical decisions not because they are necessary, but to avoid potential complaints. But when does this count as "defensive medicine" and how big a problem is it really?
Many doctors have done it – most, even, according to recent surveys.

It can take different forms: a patient comes in with a complaint that nobody can really do anything about, so to keep them happy you refer them to a specialist even though you know it won’t do any good, for instance. Or the patient has an infection that’s probably caused by a virus, but they demand antibiotics and you feel sure they’ll complain if they don’t get them, so you give in and prescribe them anyway.

Defensive medicine is fear dressed up as thoroughness or completeness

The definition of “defensive medicine” is a bit loose, but generally speaking, it refers to medical decisions made not because they will be clinically effective or worthwhile, but rather to avoid an argument, formal complaint or potential lawsuit.

Research suggests it is common – in the UK, and in other countries around the world – at great cost to healthcare systems due to wasted resources and time that could be far better spent.

But the full scale of the problem remains unclear. Partly that’s because there is a vast grey area between true defensive medicine, or interventions that the clinician is confident will not be of use, and simply dotting your Is and crossing your Ts – and in that expanse lie the very many scenarios in which the probability of a useful outcome may be low, but the implications of missing something important by skipping that test or intervention are potentially devastating. Add to this that most of the time these decisions are being made when the patient’s true condition is uncertain and time is short.

As concerns about defensive medicine grow, though, so too do calls for more proactive steps to counter it – in part by clarifying what it actually refers to, but also potentially by creating guidelines that don’t just lay out when a doctor should do something, but also when they are justified in not taking action.

Unnecessary care
In a recent survey of GPs conducted by Pulse, 78% said they had deviated from standard medical practice to prevent backlash and avoid potentially being referred to a regulator.

This figure is consistent with previous studies. A 2013 survey of 204 hospital doctors in the UK also found that four out of five reported practicing defensive medicine – mostly unnecessary tests or pointless referrals to specialists.

The US, perhaps because its insurance-based healthcare system drives high rates of litigation, is the country where defensive medicine has been most intensively studied. A 2005 survey of more than 800 US doctors estimated that 93% of all US physicians did it – although that survey focused on specialties at high risk of litigation, such as emergency medicine, surgery and obstetrics. As well as unnecessary interventions, 42% of the US doctors reported avoiding procedures that were prone to complications and refusing to treat patients with complex medical problems. These avoidance behaviours were much less common in the UK survey, where only 9-21% reported them – perhaps because declining or avoiding care is simply less possible in a publicly-funded system.


That’s not imposter syndrome you’re feeling
The catch-all label has become a way to convince ourselves that major problems we face in medicine are rooted in personal crises of confidence. That’s seldom true

Globally, around 76% of physicians report practicing defensive medicine, according to a 2023 systematic review and meta-analysis of 64 studies across 23 countries. Of all surveyed specialties, anaesthetists were the most likely to resort to defensive medicine: 92% claimed to have done so.

“We need a culture shift from defensive practice to clinical courage,” consultant in emergency medicine Dr Iain Beardsell at the University Hospital Southampton NHS Foundation Trust wrote in the BMJ in September 2025. Beardsell argued that the problem is that doctors are afraid, constantly asking themselves, “What if the thing that looks like a viral illness turns out to be something I’ve only seen once before and missed this time, and I get sued and referred to the GMC and fired?” Defensive medicine, he wrote, is “fear dressed up as thoroughness or completeness”.

“If you look at the UK, I don’t think there’s a huge amount of evidence that we do have a culture of that kind of litigation,” says Professor Carolyn Tarrant, a health services researcher at the University of Leicester. “I think it’s more of a hypothetical fear of negative consequences from the action that you’ve taken.”

Why it happens
While there has been a fair bit of research quantifying the scale of the problem, there has been relatively little investigation into physicians’ underlying motivations for defensive medicine.

That is starting to change, though. Tarrant has studied the overuse of antibiotics in emergency departments, variously interviewing doctors and observing them at work. “That’s a particular patient group and a particular pathway in a particular dynamic,” she says. Patients in the ER are often acutely unwell, with symptoms suggestive of infection or sepsis, but with little clinical information. “Doctors are often having to manage them under conditions of uncertainty.”

The UK has focused on sepsis in recent years, following some high-profile cases where it was not picked up as quickly as it might have been, leading to avoidable deaths. Successful treatment of sepsis requires rapid action, says Tarrant. “There’s an evidence base that certain things need to happen in quite a quick time to improve their chances of survival, one of those being administration of antibiotics in an hour,” she says.

Overprescribing is a rational response to uncertainty and risk, so how then do we make people feel safer not to take action?

Under those circumstances, “the safest route is to prescribe a broad-spectrum antibiotic,” says Tarrant. While the doctor waits to see if this is effective, they should order blood culture tests to find out if the patient really does have an infection or sepsis and refine the treatment accordingly. But this doesn’t necessarily happen. “What we see in our research is that, because the microbiology results take 24 to 72 hours to come back, and the patient has moved on, then that review and focusing of treatment doesn’t always happen reliably.”

The result may be unnecessary antibiotic use, but these decisions are not irrational, says Tarrant. Time pressure and lack of staffing only exacerbate the pressure in emergency rooms. “Overprescribing is a rational response to uncertainty and risk, so how then do we make people feel safer not to take action?” she asks.

Other areas of medicine such as general practice do not usually operate under such conditions of urgency, but similar dynamics apply. Claire Hastings is a paramedic who is now doing a PhD at the University of Leicester, supervised by Tarrant, looking at how and why GPs choose not to act. She has interviewed 32 healthcare practitioners, including both GPs and advanced clinical practitioners.

“Every single person I spoke to all very candidly said that they at times will potentially order a blood test… or maybe write a prescription when they probably don’t need to clinically,” says Hastings. But when she asked why, “it gets a bit complicated".

One factor is time. A 10-minute GP appointment often doesn’t offer enough time to explain to a patient why the doctor is declining to perform a test. There are good reasons to refrain, says Hastings. “If you do have a test that you don’t need, you may have to take time off work to go and get that test,” she says. Worse, the test might randomly give an anomalous result, leading to a host of unnecessary further interventions. But explaining this takes time, which stressed GPs don’t have.


Is it possible to leave work on time as a doctor?
As a GP is suspended for making up appointments to be able to leave work on time, Doctors.net.uk explores the pressures doctors face when it comes to clocking off

Relationships with patients are also a factor. Some of Hastings’ interviewees worked with groups who are reluctant to access healthcare, often members of deprived groups. Doctors reported ordering unnecessary interventions in order to maintain the doctor-patient relationship. Their argument, says Hastings, was: “if this patient feels that I’m not listening to them, I might lose them.” This would mean the doctor was unable to help them in the future.

A further factor is the sheer workload involved in handling even trivial complaints. Hastings says a common comment among her interviewees was that were not necessarily worried about having a complaint if they didn’t do something, because they knew they could justify why they’d taken that decision. But even despite that, the doctors sometimes engaged in defensive medicine because dealing with complaints is time-consuming. “It’s about the additional workload that that’s going to add, and the emotional stress that that’s going to add,” she says. “It’s about their own wellbeing and not burning out.”

Reducing defensive medicine
Defensive medicine does not happen in a vacuum. It is one of several drivers of “low-value care”: medical interventions that provide little or no benefit to patients, may cause harm and are relatively expensive given their lack of benefit.

Low-value care is a major burden on healthcare systems. Low-value imaging alone costs billions of dollars globally every year, according to a 2024 systematic review. The economic costs of low-value care as a whole have not been systematically estimated. However, unnecessary prescriptions of dependency-forming medications like antidepressants and opioids may cost NHS England more than £564 million every year.

A lot of defensive medicine comes down to the difficulty of deciding not to take action

“It [low-value care] is a really big problem within healthcare systems,” says Tarrant. However, “the question of how much defensive medicine plays into that is much more difficult to quantify.”

That’s because seemingly distinct motives often overlap, says Hastings. “Risk to patients and risk to me as a healthcare practitioner are quite closely aligned,” she says. “It’s quite hard to separate them.”

There have been attempts to reduce low-value care. Initiatives like Choosing Wisely, which launched in the US and later expanded to the UK, have had some success. A 2023 scoping review of 167 studies found that these kinds of interventions were more likely to be effective if they were multifaceted, for instance combining educational talks and training with audits and clinical decision support. Likewise, a review of 121 randomised controlled trials, also published in 2023, found that there are many potentially effective strategies for reducing low-value care – although there was little information about financial and regulatory interventions.

For Hastings and Tarrant, a lot of defensive medicine comes down to the difficulty of deciding not to take action. “People have a tendency to act,” says Tarrant. “How can we support people to feel safer not to act?”

The pair suggest that part of the answer may be to rewrite medical guidelines to include explicit information about when it is acceptable or desirable not to order a test or prescribe a medication. “Guidelines always tell us what we should do, and not what we shouldn’t do,” says Tarrant. She wants guidelines to also say “under these circumstances, it’s safe not to do this.”

Back in 2005 the National Institute of Clinical Excellence (NICE) did publish a set of “do not do” guidelines, advising doctors not, for instance, to “use blood tests routinely for the identification and diagnosis of alcohol use disorders”, or offer drug treatment for attention deficit hyperactivity disorder (ADHD) to pre-school children. However, the do-not-do guidelines were not widely used and were discontinued around 2019 (although NHS England maintains a list of items that should not be routinely prescribed in primary care). “This seems like a step in the wrong direction,” says Hastings.

Tarrant, Hastings and their colleagues also argued in the British Journal of General Practice in 2025, that doctors could benefit from improved guidelines around record keeping and documentation. “Being able to say, ‘what I did was in line with the guideline’, or ‘was in line with recommendations’, is probably the safest defence that you have got,” says Tarrant.

For Hastings, the key thing is to understand defensive medicine not as a personal choice by cautious individual doctors, but as a more insidious institutional problem. Doctors are often acting rationally given the incentives of the healthcare system they work in, Hastings says. “I think defensive medicine is part of this system.”

Michael Marshall is a science writer who specialises in covering health, life sciences and the environment. He is the author of The Genesis Quest.