Weekly Health Tech Reads | 3/30/25

PHTI on AI scribe adoption, robotic surgery and AI, 23andMe files for bankruptcy, and more

Author

Kevin O'Leary
March 30, 2025

👋 If you enjoy the free weekly newsletter, consider joining the Health Tech Nerds Community, a members-only Slack community designed for networking and knowledge sharing!

Sponsored by: Venturous Work

Struggling to find the right executive talent for your healthcare startup?

Venturous connects you with vetted, experienced fractional executives who are ready to hit the ground running—whether it’s for a product launch, scaling, or filling critical gaps.

They handle 80% of the vetting, saving you time and ensuring alignment with your needs, including cultural fit. Maximize impact while keeping costs in check.

Access (or join!) their network of healthcare leaders today.

If you're interested in sponsoring the newsletter, let us know! June availability for Q2, booking Q3.

AI SCRIBES

PHTI on health system adoption of AI Scribes

The Peterson Health Technology Institute (PHTI) published a new report on the state of the AI scribe market, focusing on how health systems are adopting the solutions and the impacts they are seeing. The summary visual is below, which articulates the key metrics health systems are evaluating AI scribes on and the early results thus far. I don’t think it should come as much of a surprise that the early adopters have seen strong positive impacts in a handful of areas like clinician burnout, with mixed results elsewhere.

✍️ Going Deeper

If you’re interested how health systems are implementing AI scribes, I’d highly recommend checking out this report. A couple of the key takeaways as I read it, beyond the headline above:

  • There doesn’t seem to be a dominant way that health systems are choosing to roll out AI scribes with their clinicians. Instead the document highlights three different approaches at three different health systems. Oschner Health is taking a very targeted approach, rolling out department by department. It spends 4-6 weeks deploying with each department, and estimates this process will take ~2 years. Providence and Mass General Brigham (MGB) took broader approaches, offering it to all physicians who wanted to use it. Providence is seeing ~15% uptake among its providers. MultiCare purchased a limited set of licenses, but apparently any provider that wanted to try it could.

  • There is wide variability in adoption across these differing approaches. Apparently when scribes are widely available, adoption is 20% - 50% of clinicians. With a more targeted approach, a system is seeing 70% - 80% adoption. MGB shared that 90% of its PCPs have requested access to an AI scribe. And among providers who have adopted an AI scribe, that doesn’t necessarily mean they’re using it 100% of the time. One health system reported that providers who are using AI scribes only use them on 30% - 40% of visits on average.

  • The ROI question looms large over AI scribes, and the general sense I get from the document is that we’re exiting a honeymoon phase and health system are starting to look for ROI beyond just the fact that clinicians love the tools. The report notes at least one system involved in the process is not planning to scale AI scribes systemwide because of a lack of ROI.

  • Related to the ROI point, it seems we’re going to quickly see AI scribe companies expanding use cases. Revenue cycle management (RCM) will inevitably be at the top of the list given the clear ROI impact that it can have. The report notes at one point that Commure is offering its AI scribe solution for free to health systems using its RCM products.

As I shared in Slack this week, this report had me interested in applying an innovation diffusion curve to AI scribes to see what a model of adoption of these tools might look like over the next several years. I tried starting with ChatGPT to generate some of the parameters required for a diffusion model. It was marginally helpful, but I made some adjustments to get to a chart that seemed to better fit for the adoption over the past couple of years. The result is a chart that looks like this:

Source: ChatGPT and my estimates of AI scribe market adoption

Given the rapid rate of adoption we’ve seen in the past twenty four months, it would seem that the adoption of AI scribes will likely peak in the next year or two, before slowing significantly. It implies that five years from now, the market will be saturated at around 800,000 providers in the US using AI scribes (ChatGPT estimated this as the market size, with the remaining few hundred thousand providers in the US choosing not to use AI scribes for various reasons).

If the details of the model are at least directionally accurate, it seems like it will place a heavy emphasis on companies extending beyond scribing capabilities very quickly, which the PHTI report indicates is happening as we speak.

ROBOTIC SURGERY & AI ADOPTION

The coming wave of surgical robot innovation and Bill Gates’ prediction that AI will make doctors obsolete

Continuing on the general theme of new technology adoption in healthcare, there were two separate interviews that caught my attention this week as it relates to the adoption of new technologies in healthcare. The first was Bill Gates joining the group of folks suggesting that AI could make doctors (and other professions) no longer necessary for most things within the next ten years.

The second was an interview in Healthcare Dive with HCA’s medical director of robotics, discussing how HCA is “all in” on robotic surgery and has been leading the way in the field. HCA claims to have done the most robotic surgeries of any health system, having now done over a million surgical procedures with robot assistance A quote from the article highlighting that there are now 100 robots that are currently FDA-approved or soon-to-be approved indicates how much the market for surgical robots is growing at the moment. In 2023, Bain also published an report on this coming wave of surgical robotics innovation and how the market is primed to grow significantly over the coming decade.

I think the evolution of the robotic surgery market provides for a really interesting analog for the adoption and impact of AI in healthcare. Both technologies have narratives around the potential to displace clinicians and move the marginal cost of care delivery towards zero. And while there are differences between the two technologies that go further than just hardware vs software, I think it’s instructive to look at the growth of the robotic surgery market when thinking about AI adoption.

Intuitive Surgical (ISRG) has long been the only meaningful player in the robotics space as the market has grown over the past two decades. ISRG’s stock performance over that time (up 24,000%) has reflected that, but it will have a lot more competition moving forward. That said, just take a look at how ISRG is trading for a perspective on the market opportunity here — it currently has an enterprise value of around $172 billion and 2025 projected revenue around $9.5 billion, meaning it is trading at ~18x forward revenue, even with this wave of competition coming to market. Keep in mind the health tech sector is trading closer to 2x revenue.

The doctor who started ISRG clearly has understood that the technology can replace surgeons and make them obsolete. Yet ISRG in part has been so successful over the last two decades because it has taken the exact opposite approach — driving adoption by supporting surgeons and driving a positive ROI for hospital administrators.

To drive this home, check out this fascinating expert interview with the former VP of Clinical Sales at Intuitive Surgical (ISRG) discussing in detail how ISRG successfully drove sales of its robots at hospitals. It does a great job highlighting the different segments of hospitals, how they think about the financial impact of new technologies like this, the various stakeholders involved in decision making, and how ISRG approached each. It was particularly interesting to see how ISRG intentionally sought to create a divide between clinicians and administrators and coached clinicians on how to convince their hospitals to buy the robots by telling the administrators that they’d take their volume to a competing location in town that had a robot. If you’re looking at driving adoption of a new technology in hospitals I’m not sure you’re going to find a more thoughtful interview on how to do exactly that.

Notice that no part of the narrative is about replacing surgeons or radically reducing the cost of healthcare. Instead it is all about convincing surgeons of the utility of the robot, convincing hospital administrators of the brand and positive financial ROI from having a robot, and helping grow the careers of folks responsible for these programs. After twenty five years of growth in the market, surgical robots now play a meaningful role in the ecosystem but also aren’t threatening to replace surgeons anytime soon. Meanwhile, the theoretical conversation about whether they may ever do so continues on the periphery. I’m pretty sure the investors behind ISRG are pretty happy with the outcome here regardless.

Now, tying this back to Gates’ prediction about AI replacing physicians in the next decade while moving us toward zero-cost healthcare. I’m guessing Gates is probably right that the technological capability will exist; it goes without saying that he is much better equipped to opine on that than I. At the same time, I think the statement underestimates the practical challenges that have nothing to do with the tech that will emerge when trying to drive change like this. Namely, that AI healthcare startups backed by for profit investors will need to generate positive financial returns, and the clearest way to do so is to tap into existing profit pools of incumbent health systems and payers. As long as that holds true, the visions of AI transforming healthcare in this country seem rather far-fetched to me.

Other Top Headlines

  • HHS announced a massive reorganization this week, as you’ve likely already heard. It will reduce its workforce by about 10,000 full-time employees, and hopes to save taxpayers $1.8 billion a year as a result of the move. In all, HHS has now trimmed its workforce by about 25%, prompting concerns about the potential implications of these moves.

  • 23andMe filed for bankruptcy and Anne Wojcicki resigned from her role as CEO in order to pursue a bid for the assets of the business. As part of the filing, 23andMe received a $35 million loan to support ongoing operations. The bankruptcy filing triggered individuals deleting their genetic data from 23andMe en masse this week, causing a huge spike in traffic to 23andMe’s website, which apparently crashed a number of times.

  • This is a good read in Forbes on Gather Health, a newish primary care model that launched in Boston a few years ago for dual eligibles. Gather launched in 2022 and currently has 2,500 patients across four clinics in Boston. It expects to reach 5,000 patients by the end of 2025, reaching $44 million of revenue for the year. If I’m doing my math right, that implies they’re doing ~$12k of revenue per member per year ($44 million / 3,500 average patients). The article notes that Gather has raised $17 million on a $47 million valuation. Gather appears to be taking an interesting approach here, with a heavy emphasis on being a community hub and integrating EMTs into the care team.

  • The WSJ dug into Medicaid payments this week, finding that insurers collected $4.3 billion in double payments from the government because some enrollees were signed up for Medicaid in two states at once between 2019 - 2021. As discussed in the HTN Slack during the week, while the headline seems troubling here, it seems the analysis here is more complicated than the headline might suggest.

  • AI startup Akido Labs was again in the news this week with the WSJ highlighting a new partnership in NYC for Lyft and Uber drivers. Akido will staff clinicians in offices and mobile locations around NYC who will use its AI technology to help increase access to care for drivers who might otherwise forego care. The article highlights that it is still clinicians responsible for the medical decision making, they’re just using the AI as a clinical decision support type tool to help increase access.

Quote of the Week

Virtual cardiology clinic Miga Health’s Co-Founder and CEO shared a helpful perspective in Endpoints on the business learnings that led to it winding down clinical operations last year. These couple of sentences stood out to me:

If you want to build a successful healthcare services business that reaches millions of people, taking insurance is an important part. But that’s not necessarily a high-margin business and is expensive to fund — potentially requiring hundreds of millions of dollars of investment to reach scale.

That, Aguirre realized, was going to be hard to do in a venture-created company.

“We just don’t believe that there will be companies like Miga that take insurance and get to the scale that we think is necessary,” Aguirre said.

It provides a fascinating lens into the decision to wind down clinical operations, with the quote above underscoring the learnings about the business model challenges that a business like Miga recognized it would have faced as it attempted to scale. It’s going to be interesting to watch how the healthcare services funding model adjusts for the next wave of innovation in the sector.

Startup Funding Announcements

  • Navina, an AI copilot that supports risk-bearing providers, raised $55 million. It has recently announced partnerships with agilon and Privia Health, among other orgs.

  • Solace, a company that connects Medicare members with patient advocates, is reportedly raising up to $40 million at a $300 million valuation.

  • Layer Health, an AI platform for chart reviews, raised $21 million. Flare Capital, one of the VCs in the round, shared their investment thesis in this blog post.

  • Silna Health, a platform that helps providers automate prior auths, raised $27 million.

  • Taxo, a propriety LLM designed to automate administrative tasks, raised $5 million.

  • Arlo, a level-funded health plan for small businesses, raised $4 million.

  • Marit Health, a salary data aggregator for clinicians, raised $3.2 million.

  • Scribe Health, an AI scribe, raised $1.2 million.

  • Xealth, a platform that integrates digital health vendors into clinical workflows, announced a strategic investment from Morningside Ventures.

Other Worthwhile Reads

Rewiring healthcare payers: A guide to digital and AI transformation by Florian Niedermann, Karl Kellner, Mathis Friesdorf, and Rob Levin
A McKinsey team discusses the opportunity for payors to adopt AI faster, along with the organizational transformation needed to do so. Read more

Modernizing The Social Safety Net by Emily Queen
A good look at safety net programs. Some of the quotes remind me of this 2024 Stat reporting about Homestyle Direct and how there seems to be a lot of opportunity to do better in the medically tailored meal world. Read more

Affordable Nutrition is More than Food by Ellen Brown
An interesting read on the Supplemental Nutrition Assistance Program (SNAP) soda debate, suggesting the issue is much more than whether SNAP covers soda. It argues the issue is a much deeper one facing our entire food supply system and how we make whole foods more accessible. Read more

It’s time to block-grant Medicaid by Tony LoSasso
This post argues that the Federal Medical Assistance Percentage (FMAP) has created a perverse incentive for states to spend more in order to receive more federal dollars for Medicaid, and suggests the solution is to move towards block grants for states. The post wades into some of the complexity with moving to block grants, but suggests the potential for fiscal savings outweighs the concerns. Read more

Inside New York's Thriving Digital Health Alumni Network by Susana Rojas & Martin Mignot
New York City has a thriving digital health startup ecosystem, in part driven by the alumni networks of a crop of companies started in the last decade, including Oscar, Ro, Flatiron, and Thirty Madison. It’d be fun to see someone do a fuller analysis of startup alumni networks — my anecdotal impression is something of a unicorn in terms of how many companies its alumni have gone on to start. Read more

How the American Medical Association Screws Doctors by Matt Stoller
A good perspective on the issues posed by the American Medical Association having a monopoly over CPT codes. Read more

Business Development Partner at Float Health, a full-stack platform for Specialty Pharma home infusion. Learn more.
$135k — $145k | Remote

VP, Health System Partnerships at Midi Health, a digital health company treating perimenopause and menopause. Learn more.
Remote

VP of Product at Spark, a platform for independent Medicare brokers. Learn more.
$250k — $290k | Remote

Program Manager, Strategic Initiatives at One Medical, Amazon’s primary care business. Learn more.
Up to $142k

Healthcare Investment Associate at Upfront Ventures, an early stage VC firm based in Southern California. Learn more.
On-site (Santa Monica)

Contact us to feature roles in our newsletter.

Want to share feedback with us?

Pick the option that fits best - we read all the feedback!

Login or Subscribe to participate in polls.

Show your support by sharing our newsletter and earn rewards for your referrals!

Reply

or to participate.