Women’s Health Wins, Insurers Feel the Heat

With Annalisa Merelli and Meredith Nolan

We officially have our third women’s health unicorn. Midi Health is now valued at more than $1 billion, following in the footsteps of Maven Clinic and Pomelo Health. All of these founders deserve a ton of kudos for being trailblazers in the women’s health category. The market has taken far too long to recognize that women are driving healthcare decisions in most households globally, and therefore are not a “niche” market. Funds may no longer be talking much about DEI. But opportunities are missed when only 11% of decision-makers in venture capital are women.

Whether it’s menopause, weight management, or sexual wellness, women are flocking to virtual care. Last week, Wheel, a venture-backed company that works with virtual care clinics to help them scale, released a report showing that women’s health drove approximately half of all virtual visits last year alone. That data was based on 1.3m patient visits collected by Wheel over that period. 

Women’s health is still a challenging sector, and there are headwinds to come in the next few years pertaining to policy, reimbursement, reproductive rights and more. So stay tuned for an interview with Maven CEO Kate Ryder in the upcoming Second Opinion “Lifers” podcast. We talked about everything from the power of storytelling, the growth of the consumer business and the importance of finding allies in the fundraising process. Subscribe to Lifers wherever you get your podcasts (iTunes, Spotify, YouTube), so you don’t miss an episode. 

Our latest episode features Niyum Gandhi, CFO at Mass General Brigham. With my co-host Mike Desjadon from Anomaly Health, we discuss what’s existential for health systems in this economy, how he thinks about vendors, and what feels urgent versus a “nice to have.”

And now for our roundup of the news, featuring Annalisa Merelli and Meredith Nolan (our HBS intern), plus an interview on AI trends in healthcare communications with entrepreneur and technologist Guillaume de Zwirek, CEO of Artera. 

The news: For the first time since 1989, UnitedHealth has announced a drop in revenue following the government’s announcement of lower Medicare Advantage reimbursement. Shares took a tumble afterward. 

The numbers: Not only is the company anticipating a loss, but with shares down nearly 20%, it risks losing $60 billion in market value if there isn’t a rebound. 

They are not alone: No insurer offering Medicare Advantage plans is happy about the changes, which curb overcharging. Humana’s stock lost 20%, CVS Health and Elevance Health 13%, and Centene more than 10%.

POV: Tune into our editor-in-chief, Christina Farr’s LinkedIn on Wednesday from 10 am - 10.30 ET with Thatch CEO Chris Ellis and Anomaly CEO Mike Desjadon. We’ll be unpacking the impact of all the shifts in the health insurance market, and ultimately what it means for consumers. You can register for the event here. 

The news: The Center for Medicare and Medicaid Services announced a list of 15 drugs subject to negotiation. It includes Eli Lilly’s Trulicity, a drug for Type 2 diabetes, and AbbVie’s Botox, including Botox Cosmetics. 

There’s more: The agency also unveiled the final prices for the drugs it has already negotiated. These will be in effect starting 2027. The estimated savings for the drugs already negotiated could be up to $12 billion, with more coming from the next round. 

AbbVie is fine with it: Or so it has to be. The company has reached a voluntary agreement with the Trump administration, pledging to invest $100 billion in US-based research and agreeing to other price cuts in order to avoid tariffs. 

 POV: The U.S. fundamentally doesn’t like the idea of the government negotiating drug prices – the norm in most other developed countries. Increasingly, now, it’s happening. But only in pockets. Perhaps it’s time to pick a strategy here that will bring costs down, versus letting drugs get priced at what the market will bear. 

The news: At-home lab tests and screenings are becoming more common, but they come with the transfer of sensitive health data that may not be secure, and end up in the wrong hands (say: a lender). 

The danger: “As more people rush to direct-to-consumer health tests driven by a desire to catch cancer before symptoms emerge or to find out if they are at risk for Alzheimer’s, experts say it’s conceivable that banks and insurers could use any health data they can to mitigate their own risks,” writes Mohana Ravindranath in STAT.

What experts say: “For someone who is healthy, who does not have cancer or does not have this increased, heightened risk, it might be worth taking that into consideration prior to doing a genetic test,” said Tyler Stokes, a genetic counselor at the University of Maryland Medical Center’s Greenebaum Comprehensive Cancer Center, about services such as 23andMe. 

POV: This has long been a problem, and it’s not just related to mortgages. There have also been reports that this kind of health information can be shared with insurers to help them underwrite life insurance and long-term care. That said, there are actionable steps individuals can take to protect their health if they do have the information. 

Mendra launches with $82 million: The company closed an oversubscribed Series A co-led by OrbiMed, 8VC, and 5 AM Ventures. It will acquire, develop, and commercialize solutions for rare diseases by leveraging AI. 

Orbem raised €55.5 million ($66 million): The company, which developed an AI-enabled non-invasive MRI platform, closed a Series B round led by Innovation Industries (Amsterdam).

BrightInsight raised $13 million: The company, a digital health platform developer, raised venture funding from Mayo Clinic, New Leaf Venture, and Eclipse Ventures. 

Sword buys Kaia: Sword Health, which provides physical therapy programs for pelvic health and pain management, acquired its competitor Kaia Health, which develops digital therapy for musculoskeletal health and COPD. The deal was valued at $285 million, and Sword is additionally looking to raise $500 million in the first quarter.

Spring Health buys Alma: The digital mental health provider is acquiring the maker of tools and software for mental health providers. The acquisition should close in the next quarter, and the details of the transaction are not available to the public. 

NOCD acquires Rebound Health: The virtual care provider for OCD acquired the trauma disorders provider. The financial terms of the deal haven’t been disclosed, but both products will now be under the newly launched parent brand Noto. 

Wisp acquires TBD Health: The merger between the virtual women’s health provider and the sexual health provider marks Wisp’s first acquisition and foray into hybrid care. Financial details of the transaction were not disclosed. 

VB Spine to acquire SpineHawk: The largest privately held spine company plans to acquire the Robotron Surgical Technologies product with the aim of expanding its capacity for software-enabled spine care. The financial terms of the deal are not public. 

Maven launched a Clinical Health Institute: The digital clinic focused on women and family health has launched an institute to study the impact of digital health. It currently features more than 40 studies, abstracts, and academic collaborations. 

WellTheory launches postpartum program: The platform offering virtual care for autoimmune patients will offer dedicated support to postpartum patients, with a focus on how hormonal shifts can trigger symptoms. 

1) You've written a lot about how health systems are using AI. Do you see there being increasing evidence in this market of a product-led growth motion in healthcare? In other words, providers using these tools on their own, and in some cases for free, and then that groundswell of adoption leads to a larger enterprise sale? Do you think PLG can work in healthcare?

The more advanced uses of agentic AI don’t support a PLG motion. Interoperability is too much of a barrier -  even with friendly EHRs, API fees become a second-order inhibitor (see the rate table from Brandon Keeler’s recent piece on Epic vs TX). This technological and financial overhead has encouraged many vendors to seek out “special” relationships with EHRs (a la Epic’s now-defunct Workshop). These deals go way beyond the economics of traditional marketplace partnerships we see in SaaS (e.g., 30% take rate with Salesforce, Apple, Android Play stores) and have started looking a lot like a fast lane to getting yourself an EHR native competitor/clone. I don’t think this trend is great for healthcare.

2) When it comes to physician workflow, what's your view on integrating with the EHR? And how can startups best work with Epic?

You have to meet the user where they are. Too much investment has gone into the EHR to be able to circumvent it, but at the same time, too much market power has consolidated among the big players to tie your destiny too closely to them. In terms of striking the right balance… we believe in leveraging open/free standards wherever possible (i.e., HL7, direct database) even though they don’t represent the most modern technical access points. This strategy provides protection around the uncontrollable predatory price increases by the EHRs. At the same time, it is impossible to deliver a great product without going proprietary (FHIR, API, Interconnect). We design our products to try to limit API calls and seek out win/win business arrangements with the EHRs where possible. I’d love for all APIs to be free, and we’ve put our money where our mouth is with our own marketplace, which has 50+ partners and processes hundreds of millions of requests for free each year.

3) You're passionate about agentic AI. How are you personally leveraging AI agents at Artera to do your own job more effectively? I'd love one example. 

Our product development lifecycle has been completely rewritten. Claude reads our PRDs, writes the code, documents changes, learns from each release, and pings the team when it ships to production. Every bug fix and code review becomes a permanent lesson—what one engineer called "compounding engineering."

My executive team hosts a weekly strategic meeting where teammates present memos on meaty topics for debate. I now assume work output is AI-assisted, and we have an agent that listens in, then summarizes the discussion and decisions in the exact format we've perfected over years of manual iteration.

More recently, I've been deep in recruiting. We have an AI agent that listens to our interviews, provides coaching after the fact (I was told I talk too much, and reprimanded for asking about the origin of a candidate's name), and delivers a first-draft interview summary with ratings. I even had the AI craft a rejection letter based on the full log of my conversations with a candidate. It required light editing, but it captured nuance I would have struggled to articulate myself - the kind of thoughtful closure candidates deserve but rarely get because leaders don't have the time.

In healthcare, where compliance and sensitivity matter, these agents don't replace judgment; they give me more time to exercise it.

4) If you're a health system considering agentic AI for applications like scheduling or prescription refill, how do you do that successfully within the context of the EHR? What are some best practices?

The technical capabilities are table stakes; the operational readiness is not. Physician templates need to be digitized and reimagined to allow patients to self book vs routing to a human for review. This work takes time, physicians are particular, and many don’t even realize the friction they impose on patients (or that their staff have found a way to work around). A well-designed AI will follow instructions shockingly well, and those potential template dead ends ruin the utility of agents and worsen the patient experience.

Our favorite hack is to start with known patient <> appointment pairs. Verifying, rescheduling, and canceling existing appointments can make up more than 30% of call volume, and requires very little template cleanup. These workflows can often be enabled from existing integrations. This is the low effort / high reward entry point to a much more robust agentic workforce.