Darren Devitt

The Single Patient Record is finally moving from concept to evidence, and NHS England has kicked off the Test and Learn phase. Three technical proofs of concept are being tested in the Test and Learn phase: 1. A Hub & Spoke model to link regional records via shared APIs 2. A Central Integration Model for real-time national access 3. A Virtual Data Layer that joins data on demand, not by duplication Each approach asks the same question 👉 Can we join up care safely, securely, and simply - without starting from scratch? For years, we’ve talked about interoperability like it’s a technology problem. It isn't. It’s a design and delivery challenge - and this phase is a chance to test what "good" really looks like at system scale. The interesting part is the approach is open, collaborative, and grounded in reality - discovery with frontline users, not just design in meeting rooms. If it works, it could redefine how health data flows across the NHS. Not a single system. A single understanding of the patient. Finally.

217

43 Comments

Epic announced 160+ AI projects at its recent UGM conference, while rival Oracle Cerner just launched an AI-backed EHR. But can Epic or Cerner really “run care operations”? A hospital executive I spoke with (whose hospital runs Epic) was clear: “No, Epic cannot run operations.”   I wrote down a few thoughts:   1️⃣ 𝐋𝐞𝐠𝐚𝐜𝐲 𝐄𝐇𝐑𝐬 𝐰𝐞𝐫𝐞 𝐧𝐞𝐯𝐞𝐫 𝐛𝐮𝐢𝐥𝐭 𝐭𝐨 “𝐫𝐮𝐧 𝐨𝐩𝐞𝐫𝐚𝐭𝐢𝐨𝐧𝐬.”   Hospital EHRs began as systems of record: tools for admin tasks and billing.   Running operations is a different challenge: it’s about process optimization and automation.   EHRs were the accountants; now they’re willing to become COOs.   2️⃣ 𝐍𝐞𝐰 𝐂𝐚𝐫𝐞𝐎𝐩𝐬 𝐩𝐥𝐚𝐭𝐟𝐨𝐫𝐦𝐬 𝐚𝐫𝐞 𝐞𝐦𝐞𝐫𝐠𝐢𝐧𝐠.   Provider to provider: Sword Health, a leading digital & AI rehab scale-up from Portugal, launched Sword Intelligence, sharing the tools it built to optimize its own operations (see visual)   Oil & gas to #healthcare: DNV Imatis, a CareOps player from Norway, has its roots in the oil and gas industry. The company originally optimized processes to transport oil from offshore platforms in the North Sea to continental Europe. A physician on its board pointed out that similar process optimization was urgently needed in hospitals, leading DNV-Imatis to expand into healthcare more than 10 years ago.   Airports to hospitals: Singapore’s ST Engineering, originally managing airport flows, was brought into hospitals during COVID to improve patient flow.   3️⃣ 𝐓𝐡𝐞 𝐂𝐚𝐫𝐞𝐎𝐩𝐬 𝐛𝐮𝐬𝐢𝐧𝐞𝐬𝐬 𝐜𝐚𝐬𝐞 𝐢𝐬 𝐜𝐨𝐦𝐩𝐞𝐥𝐥𝐢𝐧𝐠.   ST Engineering: A 600-bed hospital typically spends ~$6M/year on 60 senior nurses to manage operations. With ST Engineering’s 300+ standardized processes and command center model, the same work can be done by 3 non-clinical staff for ~$150k/year (and much better results)   DNV-Imatis: Clients report 20% productivity gains through automation and better resource use. Sword Health, the digital rehab scale-up, delivers 3.7× ROI per member per year in MSK care with its AI Agents. Key drivers: ◾14× clinician capacity ◾74% lower session cost ◾71% less clinical workload Now spinning off its #CareOps tech into Sword Intelligence, powered by AI Care Manager Agents: ◾Missed calls cut from 40% → 0% (24/7 coverage) ◾84% of patients scheduled instantly (eligibility auto-verified) ◾$100k new monthly revenue in 4 weeks 4️⃣ 𝐓𝐡𝐞 𝐜𝐫𝐢𝐭𝐢𝐜𝐚𝐥 𝐞𝐧𝐚𝐛𝐥𝐞𝐫: 𝐨𝐩𝐞𝐧, 𝐦𝐨𝐝𝐮𝐥𝐚𝐫 𝐭𝐞𝐜𝐡 𝐬𝐭𝐚𝐜𝐤𝐬.   Legacy EHRs aim to vertically integrate everything: from databases to CareOps to care pathways.   But advances in interoperability and AI now allow hospitals to build open, modular stacks.   Unfortunately, most hospital IT teams are underfunded and tasked with maintenance, not transformation.   To shift, budgets would need to rise from ~2% to 10%+ of hospital spend, and hiring would need to shift from traditional IT to AI and tech talent. --- Follow HGM Advisory for daily reposts of the most relevant #HealthTech news

124

7 Comments

We’ve solved harder technical problems than adoption. In healthcare, the engineering behind AI is rarely the bottleneck. We can fine-tune models, train on multimodal data, even deploy agentic AI that adapts and learns in real time. 🤔 The real hurdle? Scaling belief. I’ve sat in boardrooms where AI projects had the right architecture, governance, and infrastructure — yet stalled because the people using them didn’t see themselves in the story. A physician might hear “automated utilization review” and think “more bureaucracy.” A patient might see predictive analytics as surveillance, not protection. The most advanced AI still depends on a human operating system. And that OS is powered by narrative: • For clinicians: stories that show how AI reduces cognitive load and restores time with patients. • For executives: stories that connect algorithmic precision to measurable outcomes and sustainable margins. • For patients: stories where AI is a partner with personally-felt impact, not just intervention. We don’t build trust in healthcare AI with technical documentation alone. We build it by showing how the technology behaves in the real world. How it works alongside people to make care more timely, equitable, and effective. If you want AI to scale across a complex health enterprise, architect the communication layer as deliberately as you architect the tech stack.

83

7 Comments

OpenAI's #OpenSource #GPTOSS was released with strong claims of performance on #healthcare tasks. However, our benchmarks show that it disappoints on benchmarks that weren't built by OpenAI... from medical question answering and reasoning to information extraction and summarization. JSL-MedReasoning-32B, John Snow Labs reasoning model, achieved an average accuracy of 84.99% across public medical benchmarks, surpassing OpenAI’s gpt-oss-120B which scored 77.18%. Both models are optimized for inference on a single 80GB GPU, proving that performance gains don’t require added compute costs. Even JSL-MedS-14B, not a reasoning model, outperformed gpt-oss-20B, and almost matched gpt-oss-120B despite being one-eighth its size. #ai #generativeai #llm #llms #healthcareai #healthai #datascience

58

4 Comments

Three types of healthcare AI companies are emerging—and which one you bet on could determine if you succeed, get acquired, or vanish in the next five years. Here's what you should know: 1. The Platform Players: These are the ambitious companies trying to become the operating system of digital health. They unify patient records, analytics, communication, and care pathways under one umbrella. The dream? To lock in customers (hospitals, clinics, providers) for a decade. The risk? Big, slow, and at risk of bad rollouts. Epic and Innovaccer unify records, analytics, comms, and care into true digital health “operating systems”—locking in hospitals for years. 2. Agile Point Solution Specialists: These are nimble companies going deep on a single problem: automating documentation, accelerating prior authorizations, catching hidden diseases, or improving coding accuracy. They get pilots quickly and prove ROI. The upside? Easier entry and fast wins. The catch? They might get copied, acquired, or sunsetted when platforms roll out “good enough” features. Tempus (AI oncology) and Aidoc (radiology triage) dominate single high-value workflows and quickly prove ROI before expanding. 3. Infrastructure Enablers: Quietly powering the rest, these companies focus on secure interoperability, privacy, and compliance for everyone else. They offer the tools that allow others to scale or launch new features. They’re essential but low-key, often acting as the backbone for the industry. AKASA (revenue cycle automation) and Cedar (billing/payment AI) quietly power system-wide efficiency and compliance for everyone. Why does this matter? Because healthcare is about to consolidate again—platforms will be acquisitive, point solutions will either scale or sell, and infrastructure is the glue holding it all together. If you’re picking a partner or building your own offering, ask: Are you a platform, a point solution, or enabling critical infrastructure for the rest? And who’s best positioned for the next wave? What healthcare AI play are you watching—and have you seen a category-busting startup others have missed? Share below! Check out quick.md, @ryanlissack on X for more insights.

153

9 Comments

really good post from Ayokunle Omojola on actually deploying agentic workflows when it comes to healthcare billing - but I thought this point here in particular was poignant There are lots of interfaces where information is presented – machine-readable interfaces, human readable interfaces, and human-human phone calls. Machine-readable interfaces in legacy tech parts of the world seem to degrade more and are constrained by things like legacy data standards or terrible APIs. Human readable computer interfaces are richer in information + people complain louder when they’re bad, but they’re also constrained by what’s presented in the user interface. Human-human (aka calling someone) has no restrictions and is the most flexible at passing information, but currently the least scalable as long as one of the two people talking is a human being There are lots of experiments now being run to access information in each of these interfaces when it’s possible, and move to the next interface when it’s not. Ayo has a really interesting workflow that does this with API calls + browser agents in the post

59

2 Comments

Plenty of companies talk about AI in healthcare. Fewer talk about bias in AI - and even fewer do something about it. At Medow Health, we’re building AI that doesn't just “work” - it works fairly, safely, and transparently. We’ve seen how easily generative models can hallucinate or reinforce assumptions. And in healthcare, that’s not just a glitch - it’s a risk to patient care. Here’s what we do differently: – No inference: our models respond to real clinical audio, not imagined facts. – Real-time clinician control: edit, accept, reject — doctors stay in charge. – Fine-tuning with real clinicians: every update is validated with real anonymised data and expert eyes. – Specialty-specific models: language and structure tailored for actual subspecialty needs. The goal isn’t just “automate the note.” It’s: do it right, do it responsibly, and never let AI drift into shortcuts that put patients at risk. If you’re working on clinical AI - or just care about equity and accuracy in healthcare - I’d love to swap notes.

45

1 Comment

OpenEvidence went from a $1B to $3.5B valuation in 5 months. Last week, Doximity bought their competitor Pathways for $63M. OpenEvidence is positioning itself as the single source of truth for medical decisions, basically ChatGPT meets UpToDate. They've locked down partnerships with 13 major scientific journals (including JAMA) and claim 40% of US doctors already use their platform. Doximity—once "LinkedIn for doctors," now much more—just acquired Pathways for $63M ($26M cash, $37M equity). Pathways is essentially an OpenEvidence clone, and according to Kevin from Health Tech Nerds, they're generating maybe $1-2M ARR. That's a 30x revenue multiple. Doximity already has "Doximity GPT" for summarizing charts and reports. So Pathways fits their strategy. But what's amusing is that, as Brendan Keeler reported, OpenEvidence sued both Pathways AND Doximity for allegedly "stealing" their system prompt. So what happened is that Doximity realized they can't beat OpenEvidence, so they bought the company that also can't beat OpenEvidence (see visual), betting that their distribution moat can overcome OpenEvidence's first-mover advantage in medical decision support. Now we have two distribution powerhouses on a collision course. OpenEvidence owns 40% of physicians through product/marketing superiority. Doximity owns massive reach through platform dominance. Both want to capture the critical moment when doctors need answers. In Belgium we say "a cat in trouble makes weird jumps," but maybe this cat knows exactly where it's landing. PS: The above was co-written with Blake Madden from Hospitalogy, go subscribe as it’s my fav newsletter ❤️

193

64 Comments