Show Notes
In this episode, we hear how Generative AI is making it into the consultation room - but not through NHS endorsed routes - surveys suggest that ⅔ of doctors are using AI, for backoffice tasks - but also increasingly for information and diagnosis.
David Navarro, a research fellow in generative AI at the Australian Institute of Health Innovation, Charlotte Blease, associate professor at the Participatory eHealth and Health Data Research Group at Uppsala University, and Marcus Lewis, GP in London, reflect on what we know about the real way in gen AI is being used - and what “triadic care” (doctor, patient and AI) will mean for the future of the therapeutic relationship.
We also hear from Teppo Järvinen, professor of orthopaedic surgery at Helsinki University, about surgical subacromial decompression - a 10 year follow up of a double blinded placebo controlled trial, confirms that surgery is no more effective than standard care. Yet surgical interventions continue - we hear why.
Finally, we go to a Cholera clinic in Nigeria, where Médecins Sans Frontières are running cholera treatment centres, which you can help by donating to our Christmas appeal.
Links Generative AI and the clinical encounter
The BMJ appeal 2025-26: Inside MSF’s response to cholera in Nigeria: a day in the life of an emergency doctor
Arthroscopic subacromial decompression versus placebo surgery for subacromial pain syndrome
More from Medicine and Science from The BMJ
In this episode of podcast we explore two clinical dilemmas - unrecognised drug side effects, and the outcomes from hip and knee replacements. When an unrecognised adverse drug effect is mistaken for a new symptom, the resulting prescribing cascade can end in polypharmacy and potentially preventable patient harm. New research on nearly 2.3 million adults in Ontario, has found 24 high-priority, potentially inappropriate prescribing cascades—including laxatives after iron, pain relievers following statins, and sedatives after cholinesterase inhibitors. Paula Rochon, Jerry Gurwitz, and Antonio Cherubini, three of the study's authors join us to discuss. Over 100,000 hip and, 100,000 knee replacements are performed annually in the NHS, and health systems are investing hundreds of millions to rapidly scale robot-assisted surgery, but is this expensive new tech delivering better patient outcomes? Two target trial emulation papers, using the UK's National Joint registry aim to find out if robots are better at join replacement. Hassan Mohammad, one of the papers' authors shares their findings, and Teppo Järvinen joins the conversation to discuss challenging orthopaedic surgery standard practice. Guest List Prescribing Cascades Discussion
Referrals for ADHD and autism assessments have reached unprecedented levels, leaving NHS waiting lists stretched into years and families struggling to access support. But is the rise driven by a genuine increase in neurodevelopmental conditions, increased awareness, or systemic overdiagnosis? This prompted the government to commission the Independent Review into Mental Health Conditions, ADHD, and Autism; and we're joined by vice Chair of that review, Professor Sir Simon Wessely, Regis chair of psychiatry at King’s College London. In this episode, we discuss: Community prevalence vs. administrative prevalence, the roles of underdiagnosis, misdiagnosis and overdiagnosis. Why the current focus on diagnosis is a mistake, and why environmental factors are really key to the changes we're seeing. We also cover thecore recommendations of the report.
Last year, in July 2025, the independent review of the physician associate and anaesthesia associate roles, carried out by professor Gillian Leng, was published. The review followed the contentious role out of PAs in the NHS - where their scope of practice differed across trusts, and GP practices, and concerns were raised about clinical safety. Some of the key recommendations were around seeing undifferentiated patients, training needs, and supervision - and we spoke to Gill at the time, and the government accepted her proposals. But questions remained about exactly how some of those recommendations might work. Today, with the publication of a draft clinical protocol - and we clarify some of those questions.
The VAERS system in the US monitors adverse events following vaccines, but a new investigation on BMJ.com has found problems with the data mining systems used to spot patterns - and more worryingly, that the agencies involved were alerted about the issues, and didn’t take steps to fix the problem. David Willman, Pulitzer Prize winning investigative journalist, joins us to explain what he found.
Something different for the podcast today: a live event. It’s the inaugural BMJ lecture, in partnership with the Academy of Medical Sciences and hosted at the historic Royal Academy of Arts. We want to talk about the big issues facing healthcare professionals around the world - and we’re starting with one of the biggest: how do we make decisions in the face of real risk and genuine uncertainty? Well, who better to talk through those big questions than the man who had to try to navigate those risks and uncertainties during the pandemic? Chris Whitty is England’s Chief Medical Officer and Chief Medical Adviser to the UK Government, advising on all aspects of healthcare. He is also a professor of public health at the London School of Hygiene and Tropical Medicine, where he was director of their Malaria Centre. In his lecture, Professor Whitty argues that medical training must better prepare doctors to manage risk, highlighting that "higher risk decisions, and accountability for these, have moved to later in training", causing a "substantial and sudden step up" in risk for new consultants. He acknowledges AI’s potential to speed up clinical diagnostics, but warns against public-facing AI tools that present diagnoses with false certainty while ignoring medical risk and uncertainty. Read the related Opinion article: "We must be more systematic about risk holding throughout doctors’ careers" This recording is available to view on YouTube.