AcademyHealth, a health services and policy research member-based organization, recently released a blueprint for modernizing federal health evidence and the agencies that produce it. AcademyHealth President and CEO Aaron Carroll, M.D., M.S., spoke with Healthcare Innovation about the recommendations for the Agency for Healthcare Research and Quality (AHRQ) and other agencies. 

Healthcare Innovation: Before we dive into the blueprint document, could you talk about the recent developments with AHRQ? I interviewed you in April, when AcademyHealth was warning that despite congressional funding, AHRQ was not issuing grants. Now the GAO has ruled on this. Could you talk about what it said and the potential impact?

Carroll: Basically they ruled that the impoundments that the government had been seeking were illegal. A lot of that comes out of lawsuits from 2025, so I imagine we're going to now go through the same thing in 2026. HHS is claiming that that the money that was appropriated in 2026 has been spent, but as far as I know appropriators still haven't been shown where it's gone. What we do know is that tons of grants got canceled, which means that money which we thought was going to get spent wasn't, and really no new grants have been issued. So that's a lot of money that no one really knows where it went yet.

HCI: Is Congress showing any interest in investigating this?

Carroll: I’ve heard appropriators are talking to HHS and demanding answers. I don't think they have any yet.

HCI: Maybe it'll have to wait until the new Congress.

Carroll: It’s possible, but it doesn't matter who's in power; there are appropriators on both sides of the aisle, and usually they want to know that the budget that they all voted for and passed is how the money is being spent.

HCI: Does what's happened with AHRQ prompt some of the recommendations in the AcademyHealth blueprint as far as protecting the independence of federal health research agencies?

Carroll: Yes. I mean a big part of it is that some of these laws exist, but they don't really have teeth. The impoundments have been illegal since  a law passed in 1974, I think. We're learning part of the problem is that once you find that the impoundments are illegal, what do you do about it? What are the next actions? So it is about shoring up and fixing the entire enterprise infrastructure, and part of that is accountability, and figuring out what you do when money isn't spent like it's supposed to be. One of the things we have is a recommendation on how new laws could be passed to give that teeth so that people could follow up on it.

HCI: Another is that the heads of the agencies AHRQ and the National Center for Health Statistics (NCHS) get a fixed term of office, statutory qualifications for the job and removal only for cause….

Carroll: We had five main goals for this. The first was to protect the independence of the research infrastructure. The second was to modernize it and bring it up to date, especially with respect to data. Third is to fund it in ways that allow it to be durable. Fourth was about building the workforce so that we could make sure that we staff it appropriately, and finally, to work to embed evidence into decisions and to make sure that all the money that we're spending on research and building evidence actually gets thought about and used when policies and care decisions are being made that affect people. 

HCI: Another is to solidify AHRQ's role, and part of that is bringing in NCHS and a couple of other groups under the AHRQ umbrella, right?

Carroll: We need a home for all of this infrastructure. Especially when we talk about connecting evidence to policy, AHRQ is a natural fit. It's the institutional home for health services research. It has a bunch of core data functions, but it needs to be reauthorized in a way that gives it the same level of protections that other agencies have. It also needs a better funding floor, statutory director qualifications, and removal protections that are drafted to survive recent Supreme Court decisions. We need to differentiate between the evidence generation part of the infrastructure and the policy.

Policymakers can look at the same set of evidence and come up with different ideas because they have to make decisions that reflect their constituencies and willingness to spend and prioritization, and they're the ones that are held accountable. That's how it should be, but that's not how the evidence generation and production should be, right?

We're trying to empower AHRQ within the evidence infrastructure to be that nonpartisan, nonpolitical, curator and manager of the evidence enterprise that political appointees and policymakers can use to help make better policy.

HCI: I want to zero in on some of the data issues because that's where our readership is focused. A lot of them are CIOs, CMIOs, and chief data officers. One recommendation is to give the HHS chief data officer binding data governance authority for a core set of functions. What what would that accomplish? 

Carroll: Part of the the issue is that a lot of the different things that might need to happen are spread among all different agencies, so people don't have true authority and ability to make things happen. We need to do better. We can't just make some things voluntary. We're not going to get consistent data. Part of what we are trying to do is move things so that there would be fewer redundancies, less ability to skirt regulations and to give people who need to make these kinds of decisions the real authority they need to do what they're responsible for.

HCI: A few other recommendations focus on public health data modernization and things like health IT certification for public health data. It seems like every few years there's a renewed focus or a bolus of funding for public health data infrastructure, and then the attention seems to drift away from it. Are there still a lot of ways that we're falling short there that more attention and more resources would help?

Carroll: Lots of different ways. The ways that we collect data at a federal level are often really unorganized and unconnected. Often they're unfunded mandates because the federal government isn't necessarily collecting the data; often it's the states that are collecting it. But it needs to be standardized and it needs to be funded. It needs to be monitored. The ways that we track births, deaths, and disease outbreaks are really not great. We need national standards to say how fast the data needs to get there. And we need to make sure that that data is open and available by default, so that people can use it in ways that they need. And it needs a governance body with teeth. We need someone who has the authority to set those milestones, coordinate the budgets, and who is also protected by law against being suspended or ignored.

HCI: There are a couple of recommendations around the Center for Medicare and Medicaid Innovation. Is there an overall critique of the way that CMMI has run its models, or communicated, or used data that these recommendations grow out of?

Carroll: Well, some of it is trying to bring the ways that CMMI does its evaluations up to the same standards that we use for other types of research. Some of that involves registering the evaluations up front in the same way we register trials up front. It is also making sure that the people who design and run a model are not the same people who evaluate it, because obviously there are conflicts of interest. Also, if they find something doesn't work, it should turn off. If they find something does work, it should be expanded, and that should happen sort of by default, not at someone's whim. 

Some of the recommendations we make towards CMS are also about standardizing access to Medicaid data, fixing it so that Medicare Advantage data is just as good as traditional Medicare data, especially as more and more Americans choose to get their health care through Medicare Advantage.

HCI: I think another important one that you've already touched on a bit is making the relationship between CMS and AHRQ more operational, so that the evidence on treatments and procedures more clearly determines what Medicare pays for or doesn’t. 

Carroll: Exactly, and some of these things exist; they just don't necessarily get used, and they should be used. We want AHRQ assessing treatments that Medicare is deciding whether to cover. That makes sense. It's what everybody should want. We need to make sure that AHRQ is doing that in a non-biased way where everyone agrees on the metrics up front. That's the kind of work we need to make sure that Medicare is covering the best treatments at the best price and getting people the most bang for the buck.

HCI: Another example might be all this money they're spending on the Rural Health Transformation Program. How are they going to evaluate whether that's really having the impact that they want it to have?

Carroll: Absolutely. It’s an excellent example. I mean, you're going to spend billions and billions of dollars, and states are going to do different things. So why don't we actually see which of those things really works and which of them don't? Why don't we find out which are providing the most bang for the buck? If we do those kinds of evaluations, everyone could learn from it and it could lead to better policy in the future.

David Raths is a Contributing Senior Editor for Healthcare Innovation, focusing on clinical informatics, learning health systems and value-based care transformation. He has been interviewing health system CIOs and CMIOs since 2006.

 Follow him on Twitter @DavidRaths