Back to Insights
TrustWorks On Call Newsletter Header

Super Intelligent Spelling Mistakes

October 1, 2026

Welcome to TrustWorks On Call, here with your healthcare business and strategy 411 for the week. If you enjoy our work, please consider forwarding it along to a friend and encouraging them to subscribe. 

This week, we go Beyond the Whiteboard to show the latest data on declining physician independence, before Dialing In on whether to monetize patient portal messages. But first the news, which only has one AI story this week, in an exercise of restraint. However, that means we’ve saved some tokens to note that the recent White House accord aligning the top AI—sorry, we mean SI—executives on safety controls contains a not-so-intelligent spelling error. Perhaps a frontier model could have caught this.

Behind the Headlines

Unpacking the forces driving healthcare's biggest stories.

1. Essentia, HealthPartners announce intent to merge.

  • Two Minnesota health systems—Duluth-based Essentia Health and Bloomington-based HealthPartners—have announced their intent to merge, forming a $13B nonprofit system with 22 hospitals operating in Minnesota, Wisconsin, and North Dakota; HealthPartners also operates a 1.4M-member health plan in Minnesota and Wisconsin.
  • Effective January 1, 2027, pending regulatory approval, the systems will combine under the HealthPartners name, although branding on Essentia facilities will not immediately change.
TrustWorks Take: Hospital merger activity has rebounded across the country, but Minnesota has experienced a particular flurry. This merger announcement comes after South Dakota-based Sanford Health completed its acquisition of North Memorial Health in September to enter the Twin Cities market, and Minneapolis-based Allina Health agreed to merge with California-based Sutter Health last May. The systems’ CEOs cited the usual benefits of scale, such as investment capacity and supply-chain pricing, as motivating factors, and claimed that patient care and labor contracts will not be negatively impacted; however, increased negotiating power with payers usually results in better rates and higher prices—which are the only proven outcome of hospital mergers, according to academics. Nonprofit systems need to do a better job of turning the theoretical benefits of merging into measurable improvements felt by patients in their community. 
 
Essentia-HealthPartners and Allina-Sutter both hope to avoid the fate of the scrapped 2023 merger between Minneapolis-based Fairview Health Services and Sanford, which Minnesota Attorney General Keith Ellison effectively killed via an extended regulatory review process. (Minnesota’s Attorney General does not approve or deny transactions but can sue to block them.) Ellison, who has already called for community input on both pending combinations, is faced with an interesting quandary: while any of these deals alone might not alarm regulators, they could have a big impact on Minnesotans when taken together. The regulatory system is not set up to manage this.
 

2. MA insurers restrict plan offerings again in 2027.

  • The Centers for Medicare and Medicaid Services (CMS) shared that Medicare Advantage (MA) plan offerings will be “relatively stable” this year, with the total number of plans available nationally dropping from 5,553 to 5,532 and weighted average premiums falling 16 percent; however, this obscures a pullback by major insurers.
  • The six largest MA payers scaled back the number of unique individual MA plans they offer, ranging from 3,000 fewer for Centene to 150 fewer for Elevance, and five of the six reduced the number of counties they serve.
  • The total number of available plans is relatively stable because major insurers are offering more special needs plans (SNPs) and some smaller insurers are expanding their offerings.
TrustWorks Take: CMS, which wants to project a stable MA market, is in a narrative battle over these preliminary figures with Wall Street analysts, who are commending payers for cutting less profitable plan offerings. In response to payers projecting total MA enrollment will decline by about 2M, CMS stated it expects actual enrollment to be “be more robust than current plan projections indicate,” because most MA enrollees shed by one plan will hop to another, whether that payer wants them or not. And in contrast to CMS’s report that weighted average premiums will be down next year, industry analysis expects flat premiums and a 10 percent increase in out-of-pocket limits, as MA payers increase cost-sharing.
 
Policy changes to risk adjustments and star ratings, as well as rising utilization from a sicker MA population, have eaten into MA profits, prompting this retrenchment strategy from payers, which is now into its second year. Beneficiaries are starting to notice, as MA plan customer satisfaction has also declined for two consecutive years, from 652 to 611 out of 1000. (For reference, commercial health plans scored 562 this year.) MA beneficiaries who become dissatisfied with their choice of plans may have a tough time switching back to traditional Medicare, as Medigap insurers are allowed to deny coverage or hike rates based on the results of a health screening after the initial Medicare enrollment window.
 

3. Epic detected security flaw using frontier AI.

  • Epic Systems revealed that it allowed Anthropic’s Claude Mythos tool, a frontier AI model not available to the public, to scan its security systems for vulnerabilities, where it found “certain configurations of software might permit someone to see sensitive patient records without the intrusion being reflected in a digital audit trail.” 
  • Epic founder and CEO Judy Faulkner said that the company paused most of its tech development in order to implement a six-week plan to shore up its security weaknesses.
TrustWorks Take: The US healthcare industry is already the number-one victim of cybersecurity incidents like ransomware attacks and data breaches, due to its combination of valuable data and fragmented, outdated infrastructure. The proliferation of AI models, even those far from less powerful than Anthropic's description of Claude Mythos, are more than capable of supercharging common hacking techniques like phishing attempts that trick employees into clicking the wrong link and letting malicious actors into an IT system. As Epic’s partnership with Anthropic shows, AI is also capable of identifying weaknesses and shoring up the defenses of companies with a vested interest in state-of-the-art cybersecurity. However, most healthcare companies will not be stress-testing their security systems with Anthropic, and hackers tend to go after low-hanging fruit. 
 
Although Epic followed Faulkner’s comments with a clarification that the six-week pause will not affect the progress of their development road map, some Epic clients might enjoy a slow-down in Epic updates. Epic has been throwing so much at their clients over the past year, at an ever-quickening pace, that some systems feel like they cannot keep up. Epic's ambitions with AI are massive, but there is a risk of overrunning the userbase’s appetite for transformation.
 

Beyond the Whiteboard

Visualizing key trends from the healthcare industry

Physician Independence Drops Below 20 Percent
Since 2018, over 80K physician practices have changed ownership structure from independent (“physician-owned”) to ownership by an integrated delivery network (“hospital-owned”) or a corporate entity, like a health insurer, private equity firm, or large pharmacy chain (“corporate-owned”). The share of physicians working for health systems or corporations has risen from about 50 percent in 2018 to over 80 percent. Although corporate entities have acquired 2.5 times more physician practices than health systems have, health system employment of physicians has grown 2.7 times more than corporate physician employment, suggesting that corporate entities are buying smaller practices and reducing physician head counts at the practices they acquire. 
 
These figures can best by explained by economics and demographics. Physician-owners tend to seek health system employment for stability, as an insulation from rising costs. Corporate employment, especially by private equity firms, offers physician-owners a return on their investment and an exit opportunity. It is a natural business cycle for Baby Boomers to start a practice, grow it to maturity, and sell it to a private equity firm to fund their retirements. However, the prohibitive economics and administrative complexity of starting an independent physician practice these days makes young, entrepreneurial physicians far less likely to try. The market can still support the daily operations of independent practices, but it does not support their creation and strongly incentivizes their dissolution into parent organizations. 

Dialing In

Sharing insights from our work with clients

Monetizing MyChart Messages?
“Should we be charging for MyChart messaging? I know Cleveland Clinic does.” A physician in a primary care practice’s working group asked us this in a conversation about access channels and appointment slots. “I want to be available to my patients, but it’s hard to prioritize over stuff that earns revenue.” The conversation turned to pajama time and the—unrealized so far—promises of AI tools to eliminate it. The group consensus was that the technology is there, but their workflows had yet to catch up.
 
If a physician is spending significant time answering a patient’s portal message, it may be reasonable to charge for it, but it is more likely a sign that you are misallocating your physicians’ time. Most messages do not require a physician to reply, especially now that AI agents can automate prescription refills and basic triage. The MyChart inbox is also becoming increasingly rudimentary, as some systems and larger medical groups are setting up more public-facing agents to answer questions currently sent to individual providers, or more often than not Google. The promise of AI and the point of workflow redesign should be to maximize the time physicians spend directly with patients that requires their expertise, while automating, delegating, or redirecting lower-complexity work. Some virtual, asynchronous encounters will still require physician judgment and may appropriately generate revenue, but the question of how to monetize that work should come after the work of reducing and redistributing it.