Luke Kellaway

NHS Federated Data Platform, by the numbers: → 13 capabilities promised, 3-4 partially delivered → ~200 trusts signed up, only ~50% live → Only 1 in 4 report benefits → £182m for the first 3 years The frontline is voting with its feet. #NHS #DigitalHealth

Travelling team doctors and Heads of Medical carry extraordinary pressure on game day. Elite teams don't want a supplier. They need partners who just get it. What's the most challenging situation you've had sourcing scans abroad? #SportsMedicine

In elite sport diagnostics, the scan is rarely the hard bit. It's everything around it: → Non-GMC referral routes → Bloods back in 4 hours → Pre-briefing the rad on team protocol → Images home in a format the team can actually open #SportsMedicine

One diagnostic centre. Two key players. 48 hours to fixture. Both game plans depended on them not seeing each other in the building. Nobody asked us to keep the camps apart. We just did. That's the part of elite sport diagnostics nobody talks about. #SportsMedicine

"Fit for the Future" describes 2026/27 as a "developmental year." Moving to population-based budgets with outcome contracts across primary, community, mental health & acute is the most ambitious NHS restructure in a generation. Let's be honest about that. #NHS

The NHS is moving to population-based budgets and outcome-based contracts across every layer of care. The paper says it's not a costly or disruptive re-organisation. Is that honest about what's required, or are we understating it again?

"Fit for the Future" introduces 3 new NHS contract types: → Single Neighbourhood Providers → Multi-Neighbourhood Providers → Integrated Health Organisations IHOs will hold whole-population budgets across primary, community, mental health & acute. #NHS

NHS England just restructured how every pound of care is commissioned in England. Almost nobody noticed. Population-based budgets. Outcome-based contracts. Three new provider types. One of the most ambitious restructures in a generation. #NHS

The clinical encounter is no longer the starting point. It's the last stop after forums, AI summaries, TikTok explainers and unregulated online pharmacies. Nobody designed that pathway. Nobody is governing it. Who owns this problem?

Two years ago, 10% of people reported seeing misleading health info online. Today it's 41%. Nearly 95% of websites selling prescription medicines operate illegally. This isn't a social media problem. It's a patient safety problem. #NHS #DigitalHealth

66% of people use AI for health questions. Only 3% trust it. That gap is where clinical risk lives.

Heidi: Australian. Lyrebird: Australian. Microsoft Dragon: American. Of the AVT vendors winning at NHS scale, TORTUS is the only UK-founded one — and it just lost the flagship deployment built on its own evidence. #NHS #HealthTech

The largest AI scribe deal in NHS history just went to an Australian start-up — built on a UK company's evidence. A thread on what NHS procurement is actually rewarding.

Honest question for NHS digital strategy: If a UK start-up generates the national evidence base for an entire AI category, and still loses the commercial deployment that follows… what is the value proposition for a UK founder choosing to build in healthcare?

TORTUS's NHS AI scribe trial: → 9 sites → 17,000+ patients → 23.5% more patient interaction time → 13.4% more A&E patients per shift → £834m national value case The contract went to an Australian competitor. #NHS #HealthTech

A UK start-up generated the £834m NHS business case for AI scribing. Then lost the contract to a Melbourne competitor. If building the national evidence base doesn't win the deployment, what's the value proposition for UK healthtech founders?

Widening participation in medicine is not a communications problem. It is a pipeline, financing and retention problem that begins at age eleven and ends at registration. Access courses are the easy bit. #NHS

1/3 of English state schools have never had a pupil apply to medical school. Half have never had one accepted. Over 1,000 schools where medicine simply isn't a career. 2,000 access places over three years won't crack that. #NHS

The state holds equity in the Sovereign AI cohort. Align selection to a prioritised list of NHS operational problems and the returns recycle into the next cohort instead of leaving the system. That's a sovereign flywheel. #NHS

Start with clinical AI and you hit the regulatory ceiling. Start with scheduling, coding, workflow and interoperability and you find billions in quiet NHS spend. Unglamorous plumbing. Measurable wins. Exactly where overseas vendors have captured most value.