Weekly Health Policy Briefing 04/14/2026

The first round of ACCESS Model Accepted Participants was released this week. We’ve talked about ACCESS a lot around the HTN offices, in the HTN Slack, and to our friends and loved ones who only seem to reply with sighs and grimaces.

My take, for what it’s worth, is that:

The CMS Innovation Center describes its mission as:

I say this in the most laudatory way, but CMMI is a very American attempt to resolve the tension in the Gallup poll I screenshotted above. Should the government raise revenue or reduce costs? The American people pretty decisively say no, so we created CMMI to try to come up with a third option.

Looking through the list, I was pleasantly surprised by the number of companies that raised their hands for what seems like a sort of Project Hail Mary¹. I think we’ll still end up having to either raise taxes or reduce benefits, but I think it’s cool that we’re trying, and the expected value, despite the low probability, is still very high because we spend so much on health care.

Elsewhere in Washington:

An idea you hear sometimes is that health insurance should work more like car insurance. Usually, the person advancing this argument is mostly talking about the ACA’s essential benefits, comparing health insurance paying for annual physicals to car insurance paying for an oil change.

It’s an interesting thought experiment, but there are a couple of rejoinders you tend to hear by way of response: One flavor is about the merits of the essential health benefits themselves, either in reducing longer-term costs (an ounce of prevention is worth a pound of cure) or a political one where the voting public really doesn’t like paying premiums, having high deductibles, and not getting any sort of first-dollar coverage.

Another flavor of criticism of this metaphor is that health insurance markets and car insurance markets differ in some pretty important structural ways. The two big ones are that (1) driving without car insurance is illegal, or practically illegal, in nearly every state, and (2) car insurers are allowed to underwrite based on individual driver characteristics like age, history of accidents, and cetera.

(Interestingly, the ACA tried to adopt (1) with its individual mandate while making (2) mostly illegal, instead only allowing insurers to use “community rating” based on geography, wide age brackets, and smoking status.)

A challenge with the car insurance structure that might be nagging you if you live in a place in the United States with only limited public transport options is that some people are both a bad risk from an insurance perspective but need to drive to work, bring their kids to school, and cetera.

States devised a clever solution to this problem: assigned risk pools, which do pretty much what it sounds like. People who couldn’t otherwise can’t get car insurance apply through a state-operated program, and they are assigned to an insurance company that must accept them, but can charge them higher prices than a standard customer. This balances the insurer’s ability to profitably manage risk with people, even poor drivers, legitimate need for car insurance.

I find this to be a pretty satisfying way for a state to solve a market failure. In a pre-ACA world, states tried to do something similar for people deemed uninsurable in commercial plans, though the general consensus is that they didn’t work very well, and the ACA’s guaranteed issue provision made them irrelevant. It does seem to work better for Workers’ Comp, where states spin up an insurer of last resort company, which can end up being a pretty good business you can then spin out for a cool $300m if things get tight in the budget.

Another somewhat obscure corner of a state’s insurance obligation is providing Medicaid to foster children and young adults who have aged out of the foster care system, or at least it was obscure to me before reading this article in KFF titled “State-Run Insurance Plans for Foster Kids Leave Some of Them Without Doctors.”

Reading through the article, many of the challenges KFF identified, like not covering CAR T-Cell Therapy, limited provider networks, and inadequate mental health services, strike me as tragic, but unfortunately, rather garden-variety tragic, Medicaid problems. In short, eminently solvable problems that are unfortunately only solvable with more money.

Bracketing those problems for a moment, a state-facilitated special needs plan holds a lot of promise, with the state having a unique interest in the continuity of care, and foster-impacted children and young adults having a unique set of needs. Of course, we can’t really bracket the problems; an underfunded special needs plan doesn’t really accomplish much for those impacted by the foster system and might not be any better or even somewhat better than traditional Medicaid or an MCO. But it would be a really cool thing if a state could pull it off.

Elsewhere at the state level: