Yujan Shrestha, MD

Sunday hot take: We won't see AI impact clinical care until we have more physicians as builders. The debate on technical vs non-technical or how technical shouldn't build without clinical is so tired. Because right now the entire focus should be on how to make non-technical CLINICAL technical. And no, not technical in the sense of retraining or understanding how to code. But technical in the sense that they pair program with AI directly within clinical applications and care delivery environments. EVERYTHING meant for clinical should be built this way, from scientists in labs, to doctors in rural health care settings. Builder first, not just user or buyers. ------- Hi, I'm Dr. Mario Amaro, Founder and CEO at Cline and we're working on something new that will finally help empower doctors to transition from being "just buyers" to becoming BUILDERS. Follow us to learn more. #HealthcareOnLinkedIn

52

13 Comments

🧠 AI Can’t Handle How Doctors Actually Write Radiologists don’t speak in absolutes and that’s by design. In clinical documentation, especially radiology, hedge language is everywhere: 👉 “Cannot rule out…” 👉 “May represent…” 👉 “Recommend follow-up for correlation…” These aren’t sloppy habits. Rather, they’re legally informed survival strategies. Defensive documentation has been shaped by decades of malpractice risk and diagnostic uncertainty. But when our team trained NLP models to identify whether a report actually recommended a follow-up scan… We hit a wall. 🔍 Nearly every report suggested a follow-up, even when clinically unnecessary. Why? Because radiologists have internalized the risk: it’s safer to over-recommend than to under-caution. For AI, this is a labeling nightmare. If your model is learning from unchecked hedge language, it’s going to overpredict follow-ups, degrade specificity, and become clinically useless. 🛑 That’s why AI needs more than just raw text. It needs intent-aware labels curated by experts who understand the nuance behind the language. If you’re building AI in healthcare, ask yourself: Are you training your models on what was said or what was meant? And I'd love to know: How are you accounting for hedge language in your AI models — ignore it, label around it, or something else entirely? #HealthcareAI #AIinHealthcare #HealthTech #ClinicalNLP #MedicalAI #DataLabeling #DigitalHealth #RadiologyAI #MachineLearning #HealthInnovation

35

12 Comments

The explosion of "per square centimeter" reimbursement codes is reshaping healthcare economics before our eyes. This visualization tracks HCPCS codes that include "per sq cm" in their description. The trajectory is striking: from barely registering in 2015-2017 to nearly 60 distinct codes by 2024. What do you see? Procedures that get paid by areas substantially increased in recent years. Look at the 2025 volume - it's just the first quarter amount, already matching full-year 2023 levels. Yes, these contain the famous skin grafts, if you know what I'm talking about. Why this matters: Area-based reimbursement creates fundamentally different economic incentives than time-based or dose-based approaches. When payment scales with treatment area, product selection and application techniques become critical financial drivers. For manufacturers, securing unique HCPCS codes remains the golden ticket - with each new product driving code proliferation. For providers and payers, this creates an increasingly complex reimbursement landscape to navigate. The skin substitute market exemplifies how coding structures directly shape market dynamics. As this trend accelerates, we should expect continued fragmentation before inevitable consolidation efforts emerge. If you are on mimilabs, this is a query you can quickly run to see for yourself: SELECT * FROM mimi_ws_1.cmscoding.hcpcs WHERE short_description ILIKE '%per sq cm%'; #HealthcareEconomics #Reimbursement #SkinSubstitutes #MedicalCoding

39

2 Comments

It’s getting REAL in the PT EMR world 🌎 Raintree Systems acquired AIScribe Yoomi making it clear they’re building out serious AI infrastructure. From voice-driven documentation to smarter patient engagement — they’re laying down the foundation for automation and scale. This isn’t just about speeding up notes. It’s about redefining workflows across the entire patient journey. Then you’ve got Net Health locking in Limber Health to level up how we capture and act on functional outcome measures. That’s a strong move. This is a big boost to their FOTO Outcomes and engagement are going to matter more than ever as payers start to expect proof of value, not just volume. And then… enter Prompt or should I say Prompt Health now. They didn’t just rebrand. They’ve been busy acquiring PT Wired, Inc. and PredictionHealth to go from an EMR to a true end-to-end platform. They’re building something that brings content, automation, and AI under one roof — and if you know Adam Baliatico and Mike Dwyer you know they’re playing the long game. This isn’t a tech stack. It’s a strategy. — They are turning friends into family! So where does that leave everyone else? I’m looking at Jane App, SPRY, WebPT and a few other big names. The field’s evolving fast, and private practices are watching closely. I have a feeling PtEverywhere is going to be coming out with some amazing update 🔜. Just a feeling 😉. Let’s not forget StrataPT 🦈 and Athelas they are also making moves. Both of these companies started with the billing. Will their UX design and front end get a crazy update and new look in 2026? Here’s my take: Whoever builds the best outcomes engine — one that helps practices track and report clinical impact in a way that matters to payers — is going to win big. Because value-based care isn’t a buzzword anymore. It’s a direction. And if we can’t measure our value, we won’t get paid for it. Honorable mention to Comprehend Health. I’m sure you are going to make moves too. SaRA Health - Stress Free RTM dropping some crazy updates soon. Indie Health is about to come out with some crazy technology that will help your small clinic feel like you have a team of AI agents. Just watch 👀 you might not need that new office hire to grow your team. This is the future of rehab. It’s coming faster than we think. #PhysicalTherapy #PrivatePractice #PromptHealth #RaintreeSystems #NetHealth #AIScribe #LimberHealth #ValueBasedCare #DigitalHealth #RehabTech #OutcomesMatter #PTEMR

292

36 Comments

𝗠𝗼𝘀𝘁 𝗽𝗲𝗼𝗽𝗹𝗲 𝘁𝗵𝗶𝗻𝗸 𝗔𝗜’𝘀 𝗯𝗶𝗴𝗴𝗲𝘀𝘁 𝘄𝗶𝗻𝘀 𝗶𝗻 𝗵𝗲𝗮𝗹𝘁𝗵𝗰𝗮𝗿𝗲 𝗮𝗿𝗲 𝗱𝗶𝗮𝗴𝗻𝗼𝘀𝘁𝗶𝗰. They’re not. The quiet revolution is happening somewhere else…   Acquisition and access.   Look at FDA-cleared products like Caption Guidance and UltraSight. Both were designed to help non-experts capture diagnostic-quality echocardiograms.   Same story in cardiology. Apple Watch ECG and AliveCor pushed AFib detection out of clinics and into living rooms.   In short: The 2020-2025 gains aren’t coming from black-box diagnosis replacing clinicians. They’re coming from AI that extends skill, simplifies testing, and expands access.   👇 𝗪𝗲 𝗺𝗮𝗽𝗽𝗲𝗱 𝘁𝗵𝗲 𝗧𝗼𝗽 𝟱𝟬 𝗔𝗜 𝗔𝗽𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻𝘀 𝗶𝗻 𝗣𝗮𝘁𝗶𝗲𝗻𝘁 𝗖𝗮𝗿𝗲 Most of the high-impact ones aren’t reading scans. They’re 𝘤𝘳𝘦𝘢𝘵𝘪𝘯𝘨 𝘤𝘢𝘱𝘢𝘤𝘪𝘵𝘺.   What’s your take, is AI’s real value shifting from diagnosis to access?   #AIinHealthcare #DigitalHealth #MedicalAI #PatientExperience

29

2 Comments