Nicole Durden Sobolewski

Using GenAI models for administrative tasks is only the beginning. We all see their potential and capability for clinical help. They can go beyond producing notes to ensure that the insights within the notes are comprehensive, correct according to the patient-specific clinical context, adhere to organizational protocols, and are backed by the correct clinical reasoning behind their decisions. I enjoyed speaking with Adam Carewe MD about the path we still need to pave in order to make GenAI live up to its clinical potential and how we as an industry can join forces in bringing reliable clinical solutions to market. Kahun Medical #digitalhealth #healthcareAI #clinicaloversight

31

1 Comment

What is one of the country's leading safety net health systems focused on right now? Gabriel M. Cohen MD, CMIO for Population Health at NYC Health + Hospitals shared his perspective with us in this week's 𝘌𝘹𝘦𝘤𝘶𝘵𝘪𝘷𝘦 𝘐𝘯𝘴𝘪𝘨𝘩𝘵𝘴: 𝗪𝗵𝗮𝘁 𝗮𝗿𝗲 𝘆𝗼𝘂𝗿 𝘀𝘁𝗿𝗮𝘁𝗲𝗴𝗶𝗰 𝗽𝗿𝗶𝗼𝗿𝗶𝘁𝗶𝗲𝘀 𝗳𝗼𝗿 𝟮𝟬𝟮𝟰? 1️⃣ Patient engagement – leveraging omnichannel communication to efficiently close care gaps 2️⃣ Interoperability – putting usable data in front of our clinicians at the point of care to support care gap closure 3️⃣ Care coordination – building the tech infrastructure for our social workers, community health workers, and care managers 4️⃣ Care infrastructure for asylum-seeking and incarcerated patient populations – multiple initiatives, with a particular focus on providing housing for homeless populations 𝗪𝗵𝗮𝘁’𝘀 𝗮 𝗿𝗲𝗰𝗲𝗻𝘁 𝘃𝗲𝗻𝗱𝗼𝗿 𝘀𝘂𝗰𝗰𝗲𝘀𝘀 𝘀𝘁𝗼𝗿𝘆? Our partnership with Findhelp. Addressing health-related social needs is a critical priority for NYC H+H Our organization already has a high penetration of social needs screening, so our focus today is connecting patients with the right resources to address the identified disparities. Findhelp helps us with this by: • Ingesting the health-related social needs screening data • Identifying those patients who have a need • Locating the best local resource(s) for that patient • Connecting the patient to the resource(s) • Facilitating end-to-end communication, including informing the physician and care team whether the patient accessed the resource(s) Key to this initiative's success has been its tight integration with our current workflows in Epic, to the point that it should largely feel seamless for our clinicians. 𝗪𝗵𝗮𝘁 𝗮𝗿𝗲 𝘆𝗼𝘂 𝗳𝗼𝗰𝘂𝘀𝗲𝗱 𝗼𝗻 𝘄𝗶𝘁𝗵 𝗿𝗲𝗴𝗮𝗿𝗱 𝘁𝗼 𝗔𝗜? As a safety net health system, NYC H+H’s priority is to ensure that the AI tools and models we employ can sufficiently manage potential bias. We’re currently prioritizing governance-related workstreams (e.g. governance councils, frameworks, and processes) that will drive a safety net health system playbook for appropriately evaluating and managing AI. The specific AI use cases we're currently prioritizing include: • Chart summarization • Inbasket management Other future areas of exploration: billing automation and radiology automation — For more insight from leading health IT execs, as well as Elion's latest research and can’t miss industry news, subscribe to our newsletter 📬 👉 https://bit.ly/49z6AUW

46

4 Comments

Big congrats to our outstanding rising clinical research stars and global health practitioners at IGHS, UCSF! After months of dedicated hard work, we've all completed our qualifying oral exams. I believe these projects hold the potential to lead to the development of new healthcare technologies, services, or interventions that address pressing global health issues and aid underserved populations. The breadth of our research is evident in the diverse range of global health issues we explored, including but not limited to: 1. Women's empowerment and reintegration programs. 2. Telehealth systems for Native American children with complex medical needs. 3. Performance of COVID-19 exposure notification systems. 4. Mental health interventions during natural disasters in low- and middle-income countries. 5. Cancer screening and treatment access in underserved communities. 6. HIV care and support services. 7. Predictors of disease severity in pediatric febrile illnesses. 8. Cost-effectiveness analyses of various interventions (malaria detection, CAR T-cell therapy, cervical cancer screening). 9. Social media analysis of tobacco product marketing strategies. 10. Pediatric emergency triage tools for resource-limited settings. 11. The impact of food insecurity on mental health. 12. Acculturation, diet quality, and their connection to childhood obesity. 13. Nursing perspectives on COVID-19 pandemic interventions. This wide range of topics underscores the multifaceted nature of global health challenges and the critical need for innovative, interdisciplinary solutions. Furthermore, cost-effectiveness analyses, cost-of-illness studies, and implementation studies can inform strategies for sustainable and scalable solutions.

36

13 Comments

Severe mental illness remains one of the most underserved areas in mental health, possibly because it is difficult to address with the type of virtual, easy-to-scale models that VCs love to back. We're so excited that Sonia Priscilla García and Stas Sokolin at Amae Health have been able to break the mold, bringing better in-person solutions for patients and families dealing with SMI, and attracting strong venture partners to do so cc: Lisa Jacobs Blau, Alison Ryu, Claire Akkan Quiet Capital, Healthier Capital, Jan Ellison Baszucki, Karin Kissane, Mike Volpi

48

2 Comments

On May 2nd, we’re unveiling a workshop called 𝐒𝐚𝐥𝐞𝐬-𝐅𝐨𝐜𝐮𝐬𝐞𝐝 𝐂𝐥𝐢𝐧𝐢𝐜𝐚𝐥 𝐒𝐭𝐮𝐝𝐢𝐞𝐬, which will be taught by digital health luminary, 𝐁𝐞𝐧 𝐑𝐨𝐬𝐧𝐞𝐫, 𝐌𝐃, 𝐏𝐡𝐃.  ​ In addition to being the Manos Health Lead Medical Advisor, Dr. Rosner is also an Associate professor at UCSF, a veteran digital health researcher, and a practicing hospitalist physician. ​ Dr. Rosner has the perfect background for this topic, because he has expertise in running complex clinical studies at the large AMC level and has practically navigated generating evidence as part of previous digital health launch teams. ​ Basically, he knows the science and the business. ​ He will be walking you through the practical steps you need to take to run a validation study that will support your sales. This is not an academic class on how to run a randomized control trial solely for investigative academic journal publishing. Our process focuses on designing and executing studies that prove your solution works so that you can attract your dream customers. ​ This is the same method we have used with our multi-billion-dollar clients to build, prove, and launch profitable digital health products: ​ It is perfect for those who are investing heavily in building digital health products (or healthcare adjacent, even if it’s not tech) but aren't seeing the expected returns. Or if you are currently building out your product and want to ensure you are “sales-ready” down the road. ​ Why is this workshop a game-changer for you? Sales-focused clinical trials are the key to success in healthcare: learn how to change a cost driver into revenue generation. Knowing when you are ready is the first step: understand how to know and prepare for a study and figure out what type is best for you (and most cost effective). Executing a clinical trial doesn’t have to be overwhelming: walk step-by-step with us as we take you through 10 basic steps, from choosing where to start, to marketing your results. Converting our study to sales and marketing will get you to the next level: discover tactics for converting a scientific study to marketing materials to attract your dream buyers. ​ This workshop is not just about collecting evidence; it's about transforming that evidence into a strategic advantage and a clear path to sales. ​ Who Should Attend: Founders, Health-Tech Executives, Heads of Commercialization, Scientific Chairs and Product Leaders. ​ Are there any bonuses? You bet. We’re going to give away our: ​ 🧬 Study Type Diagnostic: A list of questions that will help you target what types of study might be right for you. 📈 Marketing Dissemination Templates: Documents that will help you promote your clinical study in marketing materials and on social media. ✅ Getting Started Checklist: Use this to make sure you are ready to engage in a study and avoid re-work and costly missteps down the road. ​ Get your spot here 👇

180

21 Comments

Hey - interesting and slightly frightening read in The San Francisco Chronicle on seemingly backlash from essential workers in the nursing space on the deployment of AI. I wonder how effective an introductory course in GenAI on what it is, how it works, etc. would solve some of these concerns. It would appear that rapid deployment of new technologies does not always sit well with staff if there is a lack of explanation or insight on the value added component. Lots of assumptions in that statement but something prompted this response and nurses are CRITICAL to everyone. There are great training providers out there that offer these programs for the front line worker demographic and could likely mutually benefit all parties if deployed as part of an Education as a Benefit (EaaB) strategy for both incumbent staff and new recruits -IMO. #EducationasaBenefit #EaaB #Educationasasolution #Healthcare #GenerativeAI #Pangea

7

Thank you Arik Hill for the shout out and thank you to Shirley Wang and The Wall Street Journal for including our work at Lyssn, as well as my colleague Tim Althoff in her thoughtful piece on the use of and impacts of chatbots in mental health.  (Gift link in comments.) Properly validated and tested AI does indeed have a role in improving both access and quality in mental health, human services, and crisis intervention. And, as the article notes, there are both significant risks and growing opportunities to do both in a way that enhances patient outcomes while supporting clinicians on the front lines. At Lyssn, we use a research-led model of product development to ensure that our platform supports the use of evidence-based practices and minimizes risk, including bias.  Even in high-risk situations like crisis calls and 988, our NIMH-funded work with @protocall is demonstrating the platform’s safety and utility. We've been working on AI and mental health for over 15 years. Now, change is coming more rapidly. We believe that clinically-validated tools, including chatbots, will make a meaningful difference in the lives of people suffering from mental health challenges.

44

1 Comment

To my amazing network - I am thrilled to announce the launch of Pave Health Ventures, an early-stage venture fund dedicated to broadening access to healthcare and improving the overall well-being of our communities. I’ve been very fortunate in my healthcare career. As the Founder & CEO of MedPilot, we capped off a seven-year journey with a successful exit when we were acquired by Vytalize Health. For the last three years, serving as Chief Innovation Officer, I’ve had the honor to contribute to the remarkable growth at Vytalize, impacting the lives of hundreds of thousands of patients by enhancing the quality of care they receive. I am immensely grateful to Matt Buder Shapiro, not only for being the perfect partner and co-founder but also for being an incredible friend. Additionally, I extend my heartfelt appreciation to Faris Ghawi, Amer Alnajar, and Jordan Brown for their outstanding leadership and for broadening my perspective on the significance of value-based care. This journey has not only provided me with industry expertise but has allowed me to cultivate invaluable relationships and an incredible network. As I embark on this new chapter, I am profoundly grateful to be supported by top executives and thought leaders in the investment, wellness, and healthcare industries. Within the healthcare ecosystem, we are represented by key stakeholders from health systems, insurers, governmental organizations, retail health businesses, provider groups, and value-based care organizations. At Pave Health Ventures, we believe the timing couldn't be more opportune to help support health-focused startups. With the growing prevalence of illness and chronic health conditions impacting our communities, consumers and key Healthcare Stakeholders are increasingly prioritizing proactive and preventative care. This shift in focus is fueling demand for technology, products, and services aimed at improving health outcomes, unlocking trillions of dollars in growth potential within the healthcare landscape. Pave Health Ventures is all-in. We are actively seeking solutions that tackle the root causes of illness, manage chronic conditions, and optimize care and care teams. Please support us in our mission by opening your networks and introducing us to amazing founders and entrepreneurs. Together, let's help pave the way to a healthier tomorrow. Jacob Myers

140

34 Comments

Someone broke down the estimated costs of a $15 Sweetgreen salad. We need these cost graphics for digital health to help future founders. Especially for tech enabled cos, the greatest cost is labor. Much of the "value created" is in curation and matching, which means the gross margins are reliant on labor arbitrage. Pay someone $X, but get paid $X+Y% per visit. The challenge with healthcare is that you often have to pick one of two business models: 1. specialized, low volume, high cost, low gross margin % 2. generalized, high volume, low cost, high gross margin % I'd argue founders should focus on gross dollars over the lifetime of the customer (like Jeff Bezos said) instead of gross margin %'s, but we'll leave that for another time. The challenge is that many tech enabled digital health cos have high cost of goods sold (labor) and moderately high product, design, and engineering budgets. We must build differentiated solutions for patients, clinicians, and clients to be out incumbents, however, we're realizing that companies are ultimately valued on the discounted value of future cash flows. And the high SG&A costs at most organizations might be inbalanced to the unit economics of the business. I'm not sure the math pencils out for everyone unfortunately. But I'm just a guy on the internet, would love to hear the thoughts from those smarter than I (you!). Comment below. --- p.s. I have no idea the accuracy of the graphic and not an investor in Sweetgreen. Cheers. Credit: David Crowther

113

27 Comments

Exciting news from Monica Chopra, Jeffrey De Flavio, MD, Samantha Adelberg, and the rest of the Auxa Health team as they officially announce their $5.2M Seed round. The team's deep industry expertise and drive for change have led them to solve critical infrastructural friction to providing quality healthcare and adoption of value-based care. Auxa's AI benefit navigation platform is simplifying the complexities of healthcare by streamlining all of the fragmented and siloed data and information to make it actionable in real-time for healthcare organizations. For the 70% of Americans who aren't getting access to the health and social care benefits they deserve, Auxa's solution will be pivotal in enabling equitable care. We're thrilled to partner with co-investors Zeal Capital Partners, AlleyCorp, Laconia, and Chaac Ventures on this journey and look to share many more milestones ahead. Cc K50 Ventures, Ryan Bloomer, Daniel Vásquez, Nick Talwar, Stefanie Ng https://lnkd.in/dPDY7-Db

44

4 Comments

Two companies in the Kidneyverse orbit are joining forces to leverage the power of human touch to unlock the "last mile" of social impact and cultural connection. SameSky Health is merging with GroundGame.Health, which closed a $17 million funding round led by 7wireVentures after the companies joined. The combined company will accelerate a shared mission to improve health and advance health equity by closing the loop around health-related social needs and gaps in care at scale. Together, their coverage spans all 50 states. [1] SameSky is a DaVita Venture Group portfolio company. We also shared an article in last week's Across the Kidneyverse written by an investor from 7WireVentures that (naturally) referenced GroundGame as a noteworthy examples of a digital health company embracing local communities and on-site services. Other examples mentioned included Live Chair and Waymark. [2, 3] Speaking of social determinants, you might have seen the Fierce Healthcare article sharing how Papa is looking to drive deeper support for insurance plans through new offerings focused on Star ratings and SDOH. The company is pitching Papa as a way for health plans to improve their Star ratings—by enlisting pals to drive members to annual wellness visits and preventive disease screenings. [4] Is it just me, or can we find some way to adapt and scale these programs for people, households and communities impacted by kidney disease? We have a home dialysis bill in the mix, a new FDA VR/AR Home as a Health Hub project, and a push for more last mile, in-home support. [5] How long before these trends come together and make a dent in US health/care? Let us know what you think! 👇🏼 Sources: [1] https://lnkd.in/gqKjfWSF [2] https://lnkd.in/gNFmJ8Pr [3] https://lnkd.in/gz_-PSaM [4] https://lnkd.in/g-Rj998A [5] https://lnkd.in/gh9gt8Cd

34

1 Comment

As I look at startups and incumbents using LLMs / Generative AI in healthcare for provider workflows, I am trying to separate fact from hope, and current reality vs. the likely progress in the next 1 - 2 years. Wondering if there is anything in the "it's real now" bucket or in the "it seems very likely in the near future" one. 3 basic questions I am hoping to be educated on - if you have a strong POV, I'd love to learn from you: 1. Ambient listening: Given Nuance knows a thing or two about dictation and speech, and Msft about LLMs, what is the real angle for new players? Is it fewer hallucinations? More accuracy? More actionable guidance based on proprietary medical knowledge? I am assuming everything needs to go back into the EHR so why is this not simply Epic + Msft wins? I get the stuff about elephants not dancing and such but still...? 2. Clinical Diagnoses: Is this for real? Will we really see a day that my historical and real time medical record, (not just the stuff I see on Apple Health) can be fed into a fine-tuned model with all the RAG, RLHF bells and whistles, and out will pop helpful guidance for medication or other meaningful interventions I should take? And for all the talk of "provider in the loop / AI is an assistant not the agent" talk, is there a real example of this capability saving time / $$ / creating access? (Outside of radiology - I get how images are a well solved problem) 3. Workflow integration: Related to 2. above, has anyone seen a real LLM use case integrated into nursing / clinician / other provider workflow in a useful way that doesn't make them use multiple screens, create new and often artificial alerts and triggers, and add burden as opposed to take it away? Above all, are hospitals and staff buying into / trusting any specific use cases? I'd love to see a case study or two of real ROI delivered with ecstatic users who would hate to lose the capability. This is revealing my ignorance, so please educate me on the reality and near term (1 - 2 year) potential, I get all the stuff about the long term promise. Thanks in advance!

57

12 Comments

The High Cost of Healthcare: Urgent Challenges from the UCSF Rosenman Symposium (Post 1) Last week, I attended the UCSF Rosenman Institute Symposium 2024: Innovating Healthcare Access: A New Era for the Future. It was a rare and electric gathering of healthcare leaders, innovators, and experts, tackling the persistent challenges we face. Panels covered diverse topics, from rural health to mental health to AI to obesity, with passionate debates on the best paths forward. Panelists included Dr. Diana Ramos (California Surgeon General), Geoff Cook (CEO of Noom), Caroline Carney, MD, MSc, FAPA, FAMP, CPHQ, (President of Behavioral Health & Chief Medical Officer at Magellan Health), and S. Monica Soni, MD (CMO of Covered California). Attendees ranged from Susan Ehrlich (CEO of Zuckerberg San Francisco General Hospital and Trauma Center to Joanna Drake, Magnify Ventures →My primary takeaway: Peer-driven, AI-supported behavior change, like that offered by GetMotivated.ai, is the most scalable and accessible solution to address the complex crises facing healthcare today. Panel: The Balancing Act of Covering Costs and Lives Discussion Highlights: Misaligned Incentives: Sandra Clarke (EVP & COO, Blue Shield of California) highlighted how profit motives often overshadow patient well-being. → The takeaway? We desperately need to shift towards value-based care models that prioritize outcomes, not just bottom lines. Lack of Transparency: Elizabeth Mitchell (President & CEO, Purchaser Business Group on Health (PBGH) exposed the opacity of healthcare pricing, revealing how it leaves employers and individuals in the dark. → Transparency is not just a buzzword; it's a fundamental right for consumers to make informed decisions about their health. The Power of Primary Care: Beth Ratliff (COO, Premise Health) made a compelling case for investing in advanced primary care as a cornerstone of cost reduction and improved health outcomes. → Prevention and early intervention are not just good medicine; they're smart economics. The Role of Technology: S. Monica Soni, MD, (CMO, Covered California) painted a picture of technology's potential to revolutionize healthcare access and innovation. But she also issued a challenge: physicians must be actively engaged in cost management. → Technology is not a silver bullet; it's a tool that needs expert guidance to truly benefit patients. The Need for New Models: David Goldhill (CEO, Sesame) delivered a powerful call to action, urging us to break free from outdated models and embrace bold, innovative solutions. → The status quo is failing us. It's time to rethink how we deliver and pay for healthcare. →Follow me Michael Goldstrom and GetMotivated.ai to learn more from the conference and support the only *scalable* solution to our health crises: peer-driven, ai supported behavior change. ...more to come... #UCSF #RosenmanInstitute #HealthTech #HealthcareInnovation #DigitalHealth #ValueBasedCare

32

5 Comments

The best way to make progress on improving outcomes for diseases like #breastcancer is to foster collaboration between different healthcare organizations - working together to share knowledge and leverage data in a privacy preserving manner. That's why this partnership between Rhino Health and International Consortium for Health Outcomes Measurement (ICHOM)'s learning collaboratives is especially exciting.

9

On the latest Vital Signs, Nikhil Krishnan and I had a ton of fun chatting with Dr. Atul Butte. Dr. Butte is a professor at UCSF, Chief Data Scientist of the UC Health System, and a co-founder of numerous healthcare startups. In our conversation, Dr. Butte shared his takes on healthcare AI opportunities, challenges, hype, and more. A few highlights: ⚡LLM use cases, data, and growth LLMs are already helping to solve the “drudgery” aspects of medicine, such as automatically drafting SOAP notes and patient messages. Dr. Butte is excited about future applications of LLMs that can help with the “cognitive” aspects: with difficult medical problems an LLM might be able to stitch together disparate pieces of information to generate a potentially novel solution. He also highlights that prompt engineering will become the next “coding language” as so much of leveraging LLMs is knowing how to ask them the right questions. Additionally, Dr. Butte discusses how since 2016, UCSF has been working to deidentify its patient notes so that it can query them to answer clinical questions and ultimately enhance patient care. ⚡The need for patient-facing AI As patients are increasingly able to access their own personal health data in nicely structured formats (e.g., EHRs, wearables), they desperately need more tools to help them interpret and make decisions using that data. Dr. Butte understands that these patient-facing solutions are challenging to get funded but he emphasizes that patients need to receive the benefits of AI sooner. ⚡Evaluating healthcare AI solutions In order to evaluate AI solutions, we often put human specialists – such as medical professionals – on the same tasks. Understanding that specialists often disagree with each other, a significant challenge arises in figuring out how we can evaluate AI considering we can’t just “average” all the specialist opinions. Dr. Butte points out that there still remains a massive gap in AI evaluation tools. Additionally, there are so many organizations that want to have a say in how healthcare AI is regulated that the landscape is uncertain. An awesome conversation with an expert on all things AI meets healthcare. Listen to the full episode below: Spotify: https://spoti.fi/3Wr3643 Apple: https://apple.co/4acwgan

82

8 Comments

Just finished Ezra Klein's interview with Dario Amodei, co-founder and CEO of Anthropic, the company that created Claude. I keep being struck by how the leaders of top #AI companies are all raising the alarms on how quickly AI models are advancing and how underprepared we are as an international community to manage the impact, risks, and opportunities. Advanced AI models create new problems that we have never before faced, or even contemplated. AI also offers a means to try to combat or regulate these very challenges. This cyclical dynamic creates a very uncertain future. In #healthcare, we are used to cycles of change that feel fast from inside them but which are really quite slow. It's difficult for us to imagine that the capabilities of technology, business, the practice of medicine, and the administration of healthcare's complexities could be very different three years from now. Yet, virtually all leaders in AI are suggesting that such change likely in the near term. We need not wait for incumbent solutions to adopt such possibilities for them to reach consumers and providers. I have heard many healthcare leaders dismiss such claims as 'crying wolf,' citing past hype cycles of change in healthcare that never came to pass. But this is the very definition of disruptive change; we often fail to see it until it is upon us. Within the next several years, humans will be interacting with AI in virtually all aspects of our lives. AI models have been trained to address diagnostics, care delivery, documentation, and administration of healthcare, and their sophistication in these areas is evolving rapidly. Why would we assume all of this tech would stop at the doors of one of the biggest sectors of our economy? We are having this conversation with executive teams and boards around in the industry - interested in sharing thoughts! #AIhealthcare #disruption #innovation https://lnkd.in/eZm5nK6s

8

1 Comment

I recently sat down with Michael Batko, CEO at Startmate and Founder of Puddle Pod for Episode 11 of Insane Ambition. Startmate has invested in 230+ startups, much of it under Michael's leadership, and now has $3.5b in value across its portfolio companies. In this episode, we discussed: - Michael's journey from consulting, then living in the Amazon rainforest, to leading Australia's biggest startup accelerator - The future of Australia's startup ecosystem - The labour government's very stupid plan to make it harder for people to qualify as sophisticated investors (now thankfully shelved) - Startmate City! Startmate's plan to build Australia's first startup city. Probably the most ambitious idea I've ever heard from a startup accelerator. - The balance between visionary thinking and traditional startup advice (think Elon Musk's plan to colonise Mars, vs Y Combinator's problem/customer obsession) - "The Courage to Be Disliked", and its influence on Michael's approach to life and business. Links in the description below 👇

61

4 Comments

Be sure to tune in to this conversation with Thomas Vande Casteele of Awell and CareOps. There are many challenges for industry incumbents, digital/virtual care organizations, and primary care and specialty value-based care organizations as they plan, invest, and orchestrate clinical interventions and care processes. Thomas and the team at Awell, as well as their companion site, CareOps, provide the resources necessary for these valuable care orchestrations.

14

1 Comment

A trio of studies published just last week each took a different path to the same conclusion: GenAI isn't ready for prime time with patients. 1️⃣ The research that grabbed the most headlines came out of UCSD, finding that GenAI-drafted replies to patient messages didn’t cut down on overall messaging time. >Although GenAI reduced the time physicians spent writing replies by 6%, that was more than offset by a 22% increase in read time, while also increasing average reply lengths by 18%. Authors: Ming Tai-Seale Sally Baxter Florin Vaida 2️⃣ Another study in The Lancet showed that GPT-4 can effectively generate replies to health questions from cancer patients… as well as replies that might kill them. >Six radiation oncologists analyzed GPT-4’s responses to simulated questions from cancer patients for 100 scenarios, finding that 58% of its replies were acceptable without any editing, 7% could lead to severe harm, and one was potentially lethal. Authors: Shan Chen Frank Hoebers Hesham Elhalawani Benjamin H. Kann, M.D. 3️⃣ Another study in NEJM AI showed that four popular GenAI models have a long way to go until they’re better than humans at matching medical issues to the correct diagnostic codes. >After having GPT-3.5, GPT-4, Gemini Pro, and Llama2-70b analyze and code 27,000 unique diagnoses, GPT-4 came out on top in terms of exact matches, achieving an uninspiring accuracy of 49.8%. Authors: Ali Soroush, MD, MS Ben Glicksberg Eyal Zimlichman, MD Eyal Klang https://lnkd.in/gWyZyUPB

23

1 Comment