A selection of photos I took last night. Quite a haul, but then again, the green lady put on quite a show! @VisitTromso @visitnorway @elusive_moose #AuroraBorealis #NorthernLights

RT @but_cyclists: 99% of pedestrians deaths and 94% of pedestrian serious injuries on pavements are caused by motorists. t.co/d5Mos…

There is no "war on drivers"

One thing that always surprises me is how poorly clinical academics are treated in the UK. Yet another example of how medicine in the UK is being torn apart.

Always nice when drivers are so insecure that they have to rev as loudly as they can and announce their presence like this. Gives you time to pull over and get out the way. Even so, for a residential 20mph zone, this was a little close for comfort, and obviously far too fast!

"Someone did something undeniably good, which will make our streets safer for everyone, but it was done by a cyclist so I'm not sure how to feel about it" is such a weird take honestly

Ironically, it took several attempts to try and post this as the signal on the train is so poor. Can't even send a single tweet into the world without a strugfle. Right outside one of the UK's biggest cities!! 🙃

Trying to get anything done while on a moving train in the UK if it requires reliable connectivity/signal is so unbelievably painful. No wonder productivity is poor. Not to mention the unreliability of the rail network itself. We've really dropped a couple of big balls here!

Funny how "minor recurring errors" suddenly feel a lot more serious when they aren't happening behind a steering wheel. Would Kelvin be this forgiving if the doctor’s "lack of attention" repeatedly happened in a clinic instead of a 20mph zone?

This is exactly the tension that played out at @RoySocMed last night. 75% of doctors walked in believing AI would radically change their role — 90 minutes later, the majority reversed. The room decided: same purpose, same humanity, different tools. The thinking is being led, but it's landing on evolution, not revolution. Wrote it up here: nose.london/blog/rsm-great-d…

Attended my first drop in session with @DeepakSardiwal_ this evening, to discuss issues of road violence and speeding. Great to hear support for improving things, along with updates on what else the council are doing to support safer streets, and how residents can get involved!

*Explaining to a British guy that big cars kill more children* “Well it’s my money, why shouldn’t I be able to use it to kill more children?”

Really excited 🥳 to share two breast cancer AI papers from my time at Google, published jointly in Nature Cancer today! We set out in 2021 to answer a question that matters to millions of women: can AI safely improve breast cancer screening in the NHS? Five years, five organisations, and 125,000 women’s scans later, here's what we found. 1️⃣ Our first paper (nature.com/articles/s43018-0…) evaluated Google's mammography AI across five NHS screening services, with 39-month follow-up including interval next-round cancers: → AI achieved superior sensitivity to human readers (54% vs 44%, P<0.001) with non-inferior specificity → 25% of future interval next-round cancers detected = potential for earlier diagnosis → Reading time reduced by 32% while cancer detection increased by 18% → No systematic disparities across age, ethnicity, deprivation, or breast density → Prospective deployment at 12 sites confirmed feasibility but revealed distribution shift requiring recalibration - a critical lesson for implementation 2️⃣ Our second paper (nature.com/articles/s43018-0…) tackled what happens when AI becomes the second reader. When readers disagree today, a specialist panel "arbitrates". We studied 50,000 women's screens with 22 readers, with and without AI as the second reader: → End-to-end including arbitration, our AI-enabled arm was non-inferior to standard double reading (P<0.001) → Human reading workload reduced by 46% → AI flagged far more interval next-round cancers before arbitration, but many were overruled, even when the AI correctly localised the cancer → Future: better explainability, prior image integration, reader training, and new pathways to maximise AI success (e.g. supplemental imaging for high risk normal cases) An editorial from Allan Hackshaw and Rosalind Given-Wilson (nature.com/articles/s43018-0…) covers this work really well - thank you! Conclusion: The AI works, and it can find cancers earlier. But how we integrate it into clinical workflows will determine whether that potential translates into better outcomes for women. This collaboration between @GoogleResearch, @imperialcollege, @RoyalSurrey, @stgeorgeshospital, St George's University Hospitals NHS Foundation Trust, and Imperial College Healthcare NHS Trust was funded by the NHS AI Award. We are deeply grateful to everyone involved. Thank you to @skourti_elena at Nature Cancer. Congratulations Lucy Warren, Marc Wilson, Jenny Venton, Ken Young, Mark Halling-Brown, Megumi Morigami, Lisanne Khoo, Deborah cunningham, Richard Sidebottom, Reddy Mamatha, Hema Purushothaman, Delara Khodabakhshi, Lesley Honeyfield, Amandeep Hujan, Tsvetina Stoycheva, Andy Joiner, Reena Chopra, Aminata Sy, Dominic Ward, Lin Yang, Rory Sayres, Daniel Golden, Namrata Malhotra, Rachita Mallya, Lihong Xi, Della Ogunleye, Charlotte Purdy, Alistair Mackenzie, Jane Chang, Jonathan Dixon, Elzbieta Gruzewska, Emma Lewis, Marcin Sieniek, Shawn Xu, @DrSusanThomas, @shravyas, @fjg28_fiona, @Ara_Darzi, Hutan Ashrafian 🎉

there are no limits to what you can do with a car and still be allowed to drive again one day

This is a pretty important intervention in the superflu/NHS about to collapse due to it debate

Yet again, I have no words! 😍 I haven't posted about music (or anything, really!) in a while. But this is exceptional. Dagny is just such a phenomenal songwriter. open.spotify.com/track/2PvIz…

What happens if you ask generative AI to write an editorial for the centenary edition of @ADC_BMJ, ask 2 other AI's to review the editorial, and then ask the original AI to revise based on these reviews? A fun thought-experiment you can read here! adc.bmj.com/content/111/1/2.…

Britain’s roads in 2024: 128,272 people killed or injured = 351 people every day. Of which : 29,467 were killed or seriously injured = 81 people every day. Of which : 1,602 were dead = more than 4 people every day. And at the same time the great British public are outraged and our politicians have empathy. IT IS BEYOND TIME FOR AMBITIOUS AND BOLD ACTION TO REDUCE ROAD DEATH AND SAVE LIVES ON OUR ROADS. #TalkAboutRoadDeath #StopRoadDeath @Heidi_Labour @Keir_Starmer @transportgovuk @DHSCgovuk @wesstreeting @DavidLammy @ShabanaMahmood

Dear @David_Cameron I know you mean well, but unfortunately screening for early disease may sound logical and sensible but can also be ineffective and harmful. Recommending a bad test leads to false reassurance, and harms through unnecessary or ineffective treatment that only /

Why do ‘we’ seem to tolerate road death as “inevitable”? We have not had a national strategy since 2011 and that was pretty lacklustre, and so there has been virtually no focus on saving lives or reducing serious injuries in road traffic collisions! Many fight the “war of the motorist” (offensive to victims) and @Keir_Starmer seems to be on the side of helping drivers with his “laser like focus”. Meanwhile …. Here I share photographs of five children killed in road traffic collisions - all in West Yorkshire @WestYorkshireCA @WestYorksPolice. All recently! All probably preventable! Lillie, aged 4 💔 (Her sister) Rubie, aged 9 💔 Alice, aged 9 💔 Harrison, aged 13 💔 Christopher, aged 15 💔 There have been more … Has sufficient action been taken to prevent these deaths? What action? I would like to give you a long list of preventative action taken …. But I can’t! Why? @MayorOfWY @DeputyMayorPCWY @Heidi_Labour @LilianGreenwood @VisionZeroWY @SHinchcliffe @jeaniekerr @Alex_Ross_Shaw @JamesLewisLab @TimSwift #StopRoadDeath #VisionZero