Digital Health Tools

In healthcare, progress only matters when it reaches patients. Digital pathology is helping make that happen. For decades, pathology relied on tissue samples on fragile glass slides and manual workflows. Getting a second opinion could take time and, in some cases, physical samples had to be shipped, with the risk of delay or damage. Digital pathology is changing that. High-resolution imaging and AI can help pathologists see more, faster and with greater precision. They can also help identify patterns beyond what the human eye can detect alone. And once slides are digitised, cases can be shared instantly, anywhere. What inspires me most? How this innovation breaks down barriers. Digital pathology enables collaboration across institutions and borders, helping specialists connect and bringing answers to patients faster. In 2021, Roche launched the Digital Pathology Open Environment, which encourages collaboration among software developers to improve patient outcomes and expand personalised healthcare through innovative image analysis. Today, together with partners such as Dr Dolores Lozano Escario and the team at the Clínica Universidad de Navarra in Pamplona, Spain, we are helping redefine how we study and detect disease. When technology supports human expertise, the impact is real: better insights, greater confidence and ultimately better outcomes for patients.

Google unveils AI-powered healthcare innovations spanning drug discovery, enhanced search, and integrated medical records: 💊In drug discovery, new open AI models (TxGemma) are designed to understand both text and molecular structures to help predict the safety and efficacy of potential therapies 💊An AI co-scientist tool built on Gemini 2.0 assists biomedical researchers by parsing scientific literature, generating novel hypotheses, and proposing experimental approaches 💊These tools will be available through the Health AI Developer Foundations program, aiming to streamline the early stages of drug development 🔎 In search, expanded health knowledge panels now cover thousands more topics and use AI to provide quick, credible answers to health-related queries 🔎 The "What People Suggest" feature aggregates user discussions from online platforms to offer personalized insights based on shared experiences with specific health conditions 🔎 These enhancements support multiple languages, including Spanish, Portuguese, and Japanese, and are initially rolling out on mobile devices in the U.S. 💿The global launch of Medical Records APIs for the Health Connect platform on Android enables apps to read and write standardized medical data, such as allergies, medications, immunizations, and lab results 💿The APIs support over 50 data types, integrating everyday health tracking with official medical records from healthcare providers 👇Links to source articles in comments #DigitalHealth #AI #Google

Telehealth only leads to over-utilization of care. It does not actually replace in-person visits, it only costs us more money. 🤨 Sound right? That concern has shaped the policy conversation for years. It began due to a widely cited study demonstrating DTC telehealth increased total utilization. It became a benchmark for both caution and debate, especially in Medicare policy debates (can read here: https://lnkd.in/erZ4qYXV) Btw, that study was from 2017. Yet we still had that story ingrained and was hard to shake despite some research showing different results. Well in a new study, we have new evidence to may get rid of this belief once and for all. This study analyzed 100% of Medicare Fee-For-Service (FFS) claims from 2019 to 2024 to assess how telehealth has affected outpatient visit volume. It focused on evaluation and management (E&M) visits across three specialties with different levels of telehealth use: 🔹 Low: Orthopedic surgery 🔹 🔹 Medium: Primary care 🔹 🔹 🔹 High: Behavioral health Here’s what stood out: 1️⃣ Telehealth stabilized: After its initial spike, telehealth found its place. In 2024, it made up 38.4% of behavioral health visits, 6.3% in primary care, and just 1.2% in orthopedics. 2️⃣ More telehealth didn’t mean more visits. Total E&M visits were actually lower in specialties that used telehealth more: 📉 Behavioral health: 4.1% relative decline 📉 Primary care: 7.2% relative decline (Compared to orthopedics as a baseline) 3️⃣ Telehealth was substitutive, *not* additive: This is a key difference. Virtual care mostly replaced in-person visits rather than creating new demand. It met patients where they were without overwhelming the system. 4️⃣ Overall utilization stayed steady: Despite new care models, visit rates held consistent. Telehealth expanded flexibility, but capacity constraints and clinical workflows still shaped how care was delivered. These findings challenge long-held assumptions (I can't believe that it has been 8 years). We now have strong, early data suggesting that broad telehealth adoption doesn’t drive overutilization in Medicare. That’s a meaningful shift. It is time to move beyond outdated fears and into more thoughtful, evidence-based policy. As someone who’s worked in both emergency medicine and telehealth, I’ve seen how virtual care can meet real needs without excess. I hope we continue building systems that reflect that. Not just in theory, but in how we support access, quality, and sustainability in practice. 🧠 Curious to hear your thoughts especially if you're working in policy, digital health, or any corner of the system where these questions come up daily. 🔗 Read the full study here: https://lnkd.in/eyam4jHF Note: this is a preprint so might be more to add post peer review #digitalhealth #telemedicine #telehealth

We’re excited to share research from our health team at Microsoft AI: a proof-of-concept showing that AI can master medicine’s most intricate diagnostic challenges by following the same step-by-step reasoning expert physicians use. There's more detail in our pre print paper & blog Paper-> https://lnkd.in/egDiNsqR Blog-> https://lnkd.in/esGFhSeB Sharing what I'm most excited about from this work. 1. Benchmarks Traditional medical benchmarks like the USMLEs condense clinical cases into neat multiple-choice questions—far from the real clinical workflow. We’ve approached things in a different way: Sequential Diagnosis Benchmark (SDBench) deconstructs 304 of the most diagnostically complex and demanding cases in medicine published in the New England Journal of Medicine. SD Bench requires models—and physician—to begin with an initial presentation, ask follow-up questions, order tests, and converge on the confirmed diagnosis—just as in routine clinical practice. You can see how this works in a video with Xiao Liu on our blog. 2. Performance With this new benchmark we tested a suite of the best known generative AI models against the 304 NEJM cases with impressive out-the-box performance. Beyond this we developed the Microsoft AI Diagnostic Orchestrator (MAI-DxO). By emulating a virtual panel of physicians with diverse thinking styles, MAI-DxO boosts raw model accuracy and solves a remarkable 85.5% of NEJM cases. For comparison we evaluated 21 practicing UK/US physicians and on the same tasks, these experts achieved a mean accuracy of 20%. 3. Costs One of our concerns was that AI would default to ordering every investigation to arrive at the correct diagnosis. So we set the system up such that each requested investigation also incurred a cost. This allowed us to evaluate performance against both diagnostic accuracy and resource expenditure. As MAI-DxO is configurable it is seen to operate along a Pareto frontier of accuracy versus resource use. What’s next: While for now exciting research, we believe this kind of superhuman clinical reasoning will in future reshape medicine. A particular focus for our group is on consumer health. Today, Bing and Copilot answer over 50 million health queries daily—from a first-time knee-pain search to finding a late-night pharmacy. We’re committed to bringing rigorous and reliable AI support into these journeys, backed by clinical evidence and robust commitments to quality, safety and trust. A huge shout-out to everyone on our new team who contributed, our partners across Microsoft, and particularly to Mustafa Suleyman for his vision. He saw the opportunity for AI to improve healthcare more than a decade ago and it now feels like this is the right time to deliver on the opportunity. Harsha Nori Mayank Daswani Christopher Kelly Scott Lundberg Marco Túlio Ribeiro Marc Wilson Xiao Liu Viknesh Sounderajah Bay Gross Peter Hames Eric Horvitz Charlotte Cooper Simpson, PhD

As a physician-and-patient, I see the medical world from both sides of the white coat. As we celebrate the end of the year, so many healthcare innovations stand out to me-- including 3D printing, AI/machine learning, gene editing, and ... of course telemedicine. With telemedicine, there’s less hassle for our patients’ basic clinic visits – no more travel, parking fees, or staring at walls in waiting rooms. Quality care comes to each person in their living room. But it's not just about convenience. Telemedicine is bridging geographical gaps & bringing essential specialist care to underserved areas. For example, Equum Medical recently partnered with a rural hospital to provide remote nephrology services, making dialysis accessible locally. This means patients can receive their treatments closer to home-- surrounded by their own support network-- without getting transferred out. 🚑 Tele has the potential to ease the burden on patients and their families. It can also help alleviate the strain on hospitals and healthcare workers, esp in areas facing staff shortages. I'm inspired by what's ahead in medicine, telemedicine, and beyond. It means a lot to me, as someone who is on “both sides." What is everyone else looking forward to in 2025? #medtech #womeninmedicine #telemedicine

Everyone in this room is staring at a floating human brain. 🧠 Not through a VR headset. Not on a 2D monitor. But as a life-sized, interactive 3D hologram reconstructed from an MRI scan. For decades, doctors have interpreted hundreds of flat MRI or CT slices, mentally reconstructing anatomy in their minds. Today, AI, spatial computing, stereoscopic displays, and real-time rendering are changing that. Imagine what this means: 🏥 Surgeons can visualize complex anatomy before making the first incision. 🧠 Medical students can literally walk around a human organ. 🤖 AI can automatically segment tumors, blood vessels, nerves, and organs in seconds. 📊 Multiple specialists can collaborate around the same 3D model instead of scrolling through thousands of images. ⚡ Faster decisions. Better planning. Potentially safer procedures. This isn’t just about making images look “cool.” It’s about reducing cognitive load. Our brains evolved to understand the world in 3D—not as thousands of grayscale image slices. By transforming medical scans into spatial, interactive objects, technology lets clinicians focus on diagnosis and treatment instead of mentally reconstructing anatomy. And this is only the beginning. As AI continues to advance, we’re moving toward a future where every MRI, CT scan, ultrasound, or even live surgical feed becomes an intelligent, interactive digital twin of the patient. The convergence of: • AI • Spatial Computing • High-performance computing • Advanced GPUs • Real-time visualization will redefine medicine over the next decade. The hospitals of the future won’t just display medical data. They’ll let doctors step inside it. The question isn’t whether AI will transform healthcare. The question is how quickly hospitals can adopt the computing infrastructure needed to make it reality? #AI #Healthcare via @royrodenhaeuser #MedicalImaging #SpatialComputing #DigitalTwin #Holograms #Innovation #FutureOfHealthcare #MachineLearning #HighPerformanceComputing #GPU #Technology

Digital health isn’t just a trend—it’s becoming the default channel for care delivery across generations. The latest 2024 Consumer Adoption of Digital Health Survey offers a compelling snapshot of how different generations are embracing virtual care, health tracking, wearables, and data-sharing. 🔍 Key insights: - Millennials lead in virtual care use (68%) and wearable ownership (66%), showing strong engagement as Digital Devotees. - Gen Z is proactive in tracking health metrics digitally (64%) and owning connected devices (60%). - Surprisingly, older generations—especially the Silent Generation—are the most willing to share data (90%) and trust provider-shared health info (76%). This generational shift reinforces what many of us in healthcare already see: digital health is not about age, it’s about mindset and trust. As we design the future of care, the challenge isn’t just technology—it’s creating connected, trustworthy, and inclusive digital experiences for everyone. #DigitalHealth #ConnectedCare #PatientEngagement #VirtualCare #HealthTech #FutureOfHealthcare #HealthcareLeadership #WearableTech #GenerationalInsights #RockHealth #CareRedesign #HealthEquity #DataDrivenCare #VirtualFirst #TrustInHealthcare

Pt safety for AI safety... A recent conversation gave me one of those "why aren't we already doing this?" moments. We're spending enormous energy figuring out how to make #AI safe in healthcare. And we should be. The risks are real and likely to get more acute over time as the usage of #AI in Healthcare grows. What I've been wondering about is why we are treating this risk as something unusual...we already have a well-established, well-tested infrastructure for managing risk to patients sitting in every health system in the country. It's our existing #patientsafety experts and systems. When an AI model or algorithm produces an erroneous clinical result, we should treat it with the same rigor and scrutiny we'd apply to any patient-facing technology or process failure. What does that mean? → Report it through your existing event reporting systems → Execute comprehensive root cause and common cause analyses → Discuss findings in M&M conferences and risk management committees → Apply the hierarchy of controls to eliminate or mitigate the risk going forward We don't need to build something new from scratch. We need to redeploy what we've already built — the structures, the processes, the culture of safety — and extend them to cover AI-related risks. The discipline of #PatientSafety has spent decades researching and deploying best practices for how to interrogate system failures without blame and facilitating the redesign of systems and processes to prevent recurrence. That's exactly the muscle we need right now. The tools are already in your organization. Let's use them. My patient safety colleagues...what am I missing? How do we need to adapt our safety infrastructure to meet the AI moment? #HealthcareAI #QualityImprovement #PatientSafety #AI

Today, Radiology published our latest study on breast cancer. This work, led by Felipe Oviedo Perhavec from Microsoft’s AI for Good Lab and Savannah Partridge (UW/Fred Hutch) in collaboration with researchers from Fred Hutch , University of Washington, University of Kaiserslautern-Landau, and the Technical University of Berlin, explores how AI can improve the accuracy and trustworthiness of breast cancer screening. We focused on a key challenge: MRI is an incredibly sensitive screening tool, especially for high-risk women—but it generates far too many false positives, leading to anxiety, unnecessary procedures, and higher costs. Our model, FCDD, takes a different approach. Rather than trying to learn what cancer looks like, it learns what normal looks like and flags what doesn’t. In a dataset of over 9,700 breast MRI exams—including real-world screening scenarios—our model: Doubled the positive predictive value vs. traditional models Reduced false positives by 25% Matched radiologists’ annotations with 92% accuracy Generalized well across multiple institutions without retraining What’s more, the model produces visual heatmaps that help radiologists see and understand why something was flagged—supporting trust, transparency, and adoption. We’ve made the code and methodology open to the research community. You can read the full paper in Radiology https://lnkd.in/gc82kXPN AI won't replace radiologists—but it can sharpen their tools, reduce false alarms, and help save lives.

AI can quietly fix the gaps your clinicians see every day. Mental health waiting lists stretch for months. Women's health concerns get dismissed or overlooked. The system struggles to meet demand. This is where AI-driven platforms step in. Mood tracking tools monitor patterns that might take weeks to surface in traditional therapy sessions. Crisis intervention systems provide immediate support when human resources are stretched thin. Gender-specific health monitoring catches early warning signs that often slip through routine appointments. These platforms offer something your current infrastructure might struggle to provide: accessibility. A woman experiencing postpartum anxiety at 2am gets real-time support. A patient in a rural area tracks symptoms that inform their next specialist visit. Someone hesitant about traditional therapy finds a low-barrier entry point to mental health care. The technology handles what it does best: continuous monitoring, pattern recognition, data collection. Your clinicians handle what they do best: personalized care, complex decision-making, human connection. I spoke with a healthcare administrator last week. She was skeptical about AI in these sensitive areas. After exploring the applications, she realized something important. AI tools free up her clinical team to focus on the patients who need them most. The platforms handle routine monitoring and early intervention. Her specialists tackle the complex cases requiring human expertise. This approach reaches underserved populations who face the biggest barriers to care. It delivers tailored solutions at scale. It turns healthcare from reactive to proactive. The question becomes: how can you integrate these tools to amplify your existing care delivery?