Note [L.S.]: This month, the National Bureau of Economic Research released two separate papers (here and here) on the effects of e-cigarette flavor bans on vaping and smoking among teens and young adults. Both analyses find that these bans modestly reduce youth vaping but increase youth cigarette smoking. Given the relative risk profiles of electronic and combustible cigarettes, the net public health effect of these policies is likely negative. These latest results reinforce an argument I made last fall. That post is reprinted below.

Enumerating the failures of the U.S. public health establishment over the last few years would be a considerable undertaking. Although lapses in the COVID-19 pandemic response are perhaps the most salient, regulatory blunders around e-cigarettes may — in the long-run — prove even more damaging.

Over the last few years, hundreds of localities and a handful of states have implemented restrictions — including, in some cases, outright bans — on the sale of flavored vaping products. Policymakers, such as New York State’s Commissioner of Health, have claimed the measures “will protect our children” and prevent “Big Tobacco [from] target[ing] young New Yorkers for a lifetime of nicotine addiction.”

An op-ed in the New York Times, co-written by the president of the Campaign for Tobacco-Free Kids, warned that vaping was “turning millions of young people into addicted customers” and insisted that “banning flavored e-cigarettes is the most important thing we can do to reduce use among young people.”

Deterring youth from using tobacco products is a laudable goal. But evidence is mounting that targeting flavored e-cigarettes is a disastrously misguided strategy. While the policy does achieve its first-order goal of reducing e-cigarette sales, its backers too often ignore its downstream effects. As any student of ECON 101 could have predicted, restricting access to flavored e-cigarettes causes consumers to search for alternative sources of nicotine, and many end up smoking more combustible cigarettes.

A recent study led by researchers at the Yale School of Public Health finds that for every one less 0.7 mL e-liquid pod sold due to flavor restrictions, an additional 15 cigarettes are purchased, including from brands disproportionately used by underage youth. To put that in perspective, the average user vapes about 2-3 mL of e-liquid per day. That implies that reducing e-cigarette use by the average daily consumption volume for one person comes at the cost of increasing smoking by three packs of cigarettes. Hardly an attractive tradeoff, especially in light of the fact that e-cigarettes — though far from risk-free — carry fewer toxins and are less lethal than combustible tobacco products.

This latest research adds to a growing literature documenting the unintended consequences of cracking down on e-cigarettes. A paper published earlier this year found that taxing vaping products led to large increases in cigarette smoking among youth that “may considerably undercut or even outweigh any public health gains.” Another study estimated that taxing e-cigarettes at the same rate as combustible cigarettes would deter more than 2.75 million smokers from quitting in the U.S. over a 10-year period.

Cigarette smoking remains a public health scourge in the U.S., causing 480,000 deaths (more than the number of COVID-19 victims in 2020) and more than $240 billion in preventable health care spending annually.

Over the last decade, e-cigarettes have contributed to declining rates of smoking among both adults and adolescents. Yet draconian restrictions on e-cigarettes threaten to reverse these gains by giving more lethal combustible products a competitive advantage over less dangerous alternatives.

Medicare Part B pays for outpatient physician services, among other things. It reimburses clinicians on a fee-for-service basis through the Medicare Physician Fee Schedule. This means that the more services are provided, the more money Medicare spends — no matter whether the services are worth the expense and actually doing good to the patients.

Moreover, the Centers for Medicare and Medicaid Services (CMS) sets the prices at the federal level for the entire country (with adjustments reflecting regional differences in the cost of inputs). To do so, they rely on recommendations from a small group of American Medical Association-appointed doctors, most of whom are from specialty societies. The short of it is, specialists tell CMS how much they should be paid. What could go wrong?

Lots. As my colleagues and I detail in a new study:

• Prices of complex services keep rising at the expense of primary care services. Because Part B is subject to a budget-neutrality rule, clinicians achieve higher reimbursement rates for specialty services by recommending cuts to primary care services, generally speaking.

• The rate differential between primary and specialty services drives up the provision of complex services and drives down the provision of low-reimbursement services.

• Providers don’t compete. In the real world, businesses that sell goods and services compete amongst each other by offering the best possible value, i.e., high quality at competitive prices. The MPFS rates aren’t set in a competitive fashion: Instead, when the input costs go up, CMS systematically responds by increasing the reimbursement rates. As a result, prices keep going up and up.

• The rising prices of the MPFS lead to cost increases throughout the healthcare system. Most insurance companies use the MPFS as the basis for what they reimburse. When MPFS rates go up, insurance prices go up even more. If you’re trying to understand a core reason for our healthcare cost crisis, look no further.

In sum, Part B drives up prices and spending by design. It can be tweaked, but the bad incentives won’t go away. The best existing alternative to Part B is Part C (Medicare Advantage), because it pays private insurers a flat fee per beneficiary per year that the insurer can use to compensate providers. This encourages competition and efficiency, as insurers have an incentive to make sure that whatever money they spend on a beneficiary’s care is worth it. It’s not a perfect system, and the brief linked below proposes fixes to improve it. But it is the best alternative available right now, in my view.

To go further, please check out the following pieces:

• The Hill: Boosting Medicare Advantage can improve healthcare quality and costs (John O’Shea and Kofi Ampaabeng)

• Mercatus policy brief: The Medicare Physician Fee Schedule: Overview, Influence on Healthcare Spending, and Policy Options to Fix the Current Payment System (John O’Shea, Elise Amez-Droz and Kofi Ampaabeng)

The U.S. health system’s complexity is staggering. Those of us who earn a living studying it are forced to hyper-specialize to have any hope of mastering our subject. Still, it’s important to have some understanding of how the different pieces of the health system fit together (or, to refine the metaphor, have to be shoved together with a loud grating sound).

But short of cracking open an undergraduate textbook, it can be challenging to get a basic, comprehensive grasp of the many programs, institutions, and laws that define our health system. Most of us pick up bits and pieces of information over time, with no systematic way to avoid large gaps of knowledge.

Happily, filling those gaps just got a little easier.

The Kaiser Family Foundation (KFF) recently put out a concise overview of the U.S. health system, broken up into about a dozen chapters. KFF is not sure what to call it — they variously refer to it as “Health Policy 101,” a “primer,” and a “mini ‘textbook’.” You get the idea.

There are separate chapters for Medicare, Medicaid, the ACA, how private health insurance is regulated, the federal health policy process, etc. Each chapter is compiled by KFF scholars specializing in that area. Much of the content isn’t new — they’ve just pulled together work previously published in KFF reports, policy briefs, and testimony into a cohesive whole. The text is peppered with helpful graphs and tables, and hyperlinks help the curious reader access more in-depth information.

The entire document is short enough to skim in an afternoon, and is well worth the time.

Although its progressive slant will not come as a shock to anyone familiar with KFF’s other work, the material emphasizes basic explanations over ideological tirades. Most controversial issues are presented in a reasonably balanced way. In fact, you won’t find policy recommendations, or even much discussion of health policy controversies. Nor will you find much in the way of historical context through which to understand and interpret contemporary issues. Hopefully KFF will consider including both in any future revisions.

Check it out!

Medicaid is one of the black holes that comprise the unsustainable fiscal outlook in the U.S. While most people want to reduce government waste, improper payments are an area that’s arguably not receiving sufficient attention. In 2022, Medicaid programs made around $80 billion in improper payments, an incredible figure—almost twice Medicaid's spending on prescription drugs ($44 billion), and thus potentially an area where reform could save taxpayers without harming those in need.

Subscribe now

Understanding Improper Payments: Case Studies of Fraud and Administrative Errors

Improper payments are payments that do not meet CMS program requirements. These include fraudulent billing, administrative errors, insufficient documentation, and other similar sources. The Government Accountability Office has repeatedly identified Medicaid as a high-risk area due to its susceptibility to improper payments, emphasizing the need for systemic improvements to ensure fiscal responsibility​. Let’s look briefly at some examples:

The U.S. Attorney's Office, Western District of Virginia, recently fined Health Connect America $4.6 million for billing Virginia Medicaid for services not provided. In another case, a doctor who owned medical practices in Maryland and Delaware was fined $3 million and had to surrender his medical license because he improperly billed Medicaid programs. However, these are just a few instances. According to a Reuters investigation, "more than one in five of the thousands of doctors and other healthcare providers in the U.S. prohibited from billing Medicare are still able to bill state Medicaid programs." Additionally, the District of Columbia paid $79 million to 269 of the 1,800 providers after their terminations elsewhere.

Such cases naturally paint a grim picture, but the data shows that while fraud receives a lot of attention, the majority of improper payments are simply payments where there’s insufficient information to determine whether the recipient was eligible.

Source: Supplemental data, 2022 Payment Error Rate Measurement (PERM) Report

Balancing Privacy and Efficiency

One policy solution is for CMS and state Medicaid programs to make better use of (often) sensitive data to determine program eligibility. By using our rapidly growing capabilities in AI and analytics, technology could be used to verify eligibility and service provision, potentially even in real-time, and thus prevent improper payments before they occur.

So, the question becomes, are we willing to trade off some privacy in order to improve eligibility determinations and thus reduce the rate of improper payments? We are curious to see what you, our loyal readers of OpenHealthPolicy, think about this question. Should CMS and state Medicaid programs be equipped to better use (often) sensitive individual data to reduce improper payments? Let us know what you think by answering the poll and in the comment section if you are so inclined.

For more posts related to this topic, see our previous posts:

The Sexy and Not-So-Sexy Future of AI in Healthcare - Markus Bjoerkheim, Matt Mittelstadt, and Sam Alburger

Precision Policy: How Big Data is Reshaping U.S. Healthcare - Ali Melad

Red ink: Congress can’t control spending without reforming how we pay for healthcare - Markus Bjoerkheim

The Misallocation of Federal Medicaid Dollars - Liam Sigaud

Medicare Spending: Physician Services Are a Much Bigger Problem than Drug Prices - Elise Amez-Droz

Over the last 13 months, more than 20 million people have been disenrolled from Medicaid in what has become known as the “Great Unwinding,” one of the largest contractions of a social assistance program in U.S. history.

Beginning in April 2023, states restarted Medicaid eligibility checks that had been paused during the pandemic. Over the previous three years, Medicaid’s rolls had swelled from about 65 million to 86 million enrollees, causing federal and state spending to soar. A program originally intended to address the health needs of a few narrowly defined groups was covering more than 1 in 4 Americans.

Since the unwinding began, many enrollees have lost their coverage because their income is too high or their family structure has changed (e.g., minor children have become adults). Others have been dropped from the program for failing to provide adequate documentation of eligibility.

Last year, some observers predicted a public health catastrophe. The Medicaid unwinding, we were told, would cause the number of Americans without health insurance to surge, with disastrous consequences for vulnerable populations. Their logic was straightforward: Before the unwinding began, about 25 million Americans were uninsured. Adding another 20 million people implies a massive 80 percent increase in the uninsured rate.

But that calculation is far too simplistic.

Instead of spiking, the uninsured rate hasn’t budged. From March 2023 to April 2024 (the latest data available), the proportion of U.S. adults without coverage has held steady at 7.8 percent. In fact, some groups — including Blacks, Hispanics, and households with annual incomes below $25,000 — have reported gains in health coverage over the last year.

As I argued in an op-ed last summer, this shouldn’t be surprising. The vast majority of those leaving Medicaid had access to alternative sources of health insurance, usually from an employer. Decades of research shows that expanding public health insurance programs encourages people to drop their private coverage and enroll in the government plan. Economists call it “crowd-out.” What has unfolded over the last year is the same phenomenon in reverse: Ending Medicaid’s COVID largesse has shifted millions of people back onto private coverage.

Today, Open Health Policy is launching our inaugural podcast featuring a conversation on the future of AI in healthcare. Join me and my colleague Matthew Mittelsteadt as we explore how AI’s potential extends from revolutionary advancements you wouldn’t think possible to less sexy, but just as beneficial, everyday tasks, hosted by Sam Alburger. Enjoy!

Transcript

Note: While transcripts are lightly edited, they are not rigorously proofed for accuracy. If you notice an error, feel free to reach out to mbjoerkheim@mercatus.gmu.edu

Sam: Welcome in. This is Sam Alburger, and I'll be your host today for a conversation on the present and future of AI and healthcare. I am joined today by two esteemed scholars who should be able to sort out some of my pressing questions. First, we have Markus Bjoerkheim, a research fellow specializing in healthcare at the Mercatus Center. Also joining us from the Mercatus Center is Matthew Mittelsteadt, a research fellow focused on AI. Gentlemen, welcome.

Markus: Glad to be here.

Matthew: Thank you.

Sam: Let's jump into this. Markus, tell us a little bit about your work as a research fellow working in the realm of health care.

Markus: My background is that I did my PhD in economics at George Mason University, where I studied nursing homes. Essentially how regulations and policies can prevent or cause the worst kinds of health outcomes, the accidents and adverse events for patients in nursing homes. Then I came here to Mercatus, started with the Open Health Project, and I've continued studying those kinds of questions in the realm of Medicaid.

Sam: Wonderful. Matt, AI is the most rapidly evolving field in the world right now. Tell us a little bit about your role as a research fellow in the space.

Matthew: As you said, this stuff is rapidly evolving. In terms of my role, in terms of what I'm trying to do, a lot of it's keeping up, because this stuff is changing constantly. A lot of it's also trying to figure out how is this technology transforming the wide, wide variety of industries like health care that this general purpose set of technologies is touching. Now, in terms of policy questions, that makes it hard to do this work.

Overall, I would say that my big goal in terms of policy choices is try to push policymakers to write good policies that specifically emphasize across the board AI diffusion. We have this technology, I don't want to see it sitting on shelves collecting dusts in the form of a research paper. I want to see it actually being applied to real-world use cases, so it can actually start transforming people's lives as we've been talking about for decades now. That should be the emphasis of policy making sure that this tech is, of course, used safely, but gets into people's hands so it can start transforming the world. That's largely what I'm trying to emphasize in my research.

Sam: Absolutely. Gentlemen, we're truly at ground zero when it comes to AI advancements. Let's talk about the early stages of AI. That's the first thing we'll talk about today, the early stages of AI when it comes to health care. Currently, the craze with AI is in the form of LLMs, such as ChatGPT, and how they can comb over mass amounts of data. Markus, can you tell me about the project you're currently working on that takes advantage of this?

Markus: I can. This is a project where researchers like me, we usually use a natural experiment in text to try to find the cause and effect of some regulation or policy on people out in the real world. The great thing, as you mentioned with LLMs, is they can go over huge amounts of text and find patterns in that text. I'm working on a project where we use LLMs to identify those patterns that would allow researchers like me to study a policy, find out, "Hey, is this a policy that is helping people, or is it a costly policy that is preventing healthcare providers from providing good care and that we should look into getting rid of."

Sam: Absolutely. Matt, there's currently the first AI drug being developed, and it's in clinical trials. Can you tell us a little bit about this drug and why it's a major breakthrough for drug development?

Matthew: Just to give some context here, the same sorts of technologies that can be used to comb through policies and figure out whether or not they're working can, of course, be used in health care with things like drugs to treat diseases. In recent years, we've seen some massive advancements in what AI can do in terms of figuring out how drugs work and how they impact the human body.

As a result of this, this is pretty amazing, we're starting to see AI systems both develop and target, which is a critical piece of the puzzle, drugs at the human body. There's this company Insilico Medicine, they're a Chinese biotech firm. They actually developed the first AI-generated drug. The AI actually invented the molecule that will go into your body and AI-targeted drug, meaning the AI figured out, "Okay, we have this molecule, how does it interact with the human body? How does it play with your proteins? How might it impact this specific disease?"

They developed this drug for a disease called IPF. This is a terminal disease, and it impacts roughly five million people globally at any given time. They invented this drug using artificial intelligence. Not only that, they invented it in three years time, which, based off what I understand, traditionally, it takes about 6 to 10 years, that's correct, to develop a drug. Three years is really fast. You can imagine the cost savings. Not only that, you can imagine how quickly can we get drugs out to people to cure these diseases? There are so many things without cures today, and we now have this drug.

What’s amazing about this is that this drug has now entered FDA phase 2 clinical trials. It's being tested actively on humans. Really, what this should illustrate to you is not only the promise of artificial intelligence in domains like drug development, but the fact that this is a reality. This is in human trials right now, going into people's arms. We don't know if this is going to be approved or not. This is a test case of this whole potential technology, but it really should show you that everything that we've been saying is promise for so long is now reality, and we need to start focusing on policies that can harness that reality.

Sam: Absolutely. Markus, healthcare outcomes, they're super important to you. We're going to, what can anybody do with ChatGPT? This is what companies are doing to develop drugs, but what can I do as an individual? Healthcare outcomes are super important to you and your work. AI can be used to expedite and synthesize information for everyday people such as building a workout plan or creating a diet. Can you tell me about these ways and others that AI can improve healthcare outcomes?

Markus: Great question, Sam. I think AI can help us a lot with healthcare outcomes both in the healthcare system and as you say, our health behaviors, how we take care of ourselves on a regular basis. Most of my research is in the healthcare sector. I'll start there. One thing that I'm very excited about in terms of AI is that AI can flag those times when doctors or nurses make potentially deadly mistakes.

In the healthcare sector, we have about 50- to 100,000 deaths occur every year from mistakes that doctors and nurses make. We obviously go to the doctor, to the hospital to get better. Just like any other profession, doctors, nurses make mistakes. Sometimes they're deadly. If we can integrate AI into the workflow of our healthcare providers, they can flag that one time when the mistake is potentially deadly.

I like the analogy to airplanes today. They have what's called the ground proximity warning system, which basically tells the pilot, “You have lost track of where you are. In 60 seconds, you're about to fly into a mountain.” Most pilots will never hear that, but the one time they do, they know exactly how they should respond, pull the plane up. They will get an alarm in their air that will call out, “Pull up, pull up.”

AI, if we integrate it into the workflow of healthcare providers, it can flag those same instances where, “Oh, the mistake is potentially deadly. You need to look into this right away. The patient's test result is back. We can save the patient, but the surgery needs to happen as soon as possible.” When the doctor gets that result and then something happens, the power goes out, someone walks in with an incident, their son or daughter calls. These things happen to everyone. The one time that happens, AI can flag it and say, “You need to remember this thing.” That's I think one area where I am very optimistic about AI in healthcare.

Sam: That's wonderful. You touch on the fact that healthcare providers aren't right 100% of the time. It also brings into a tricky question where AI currently as it stands is not right 100% of the time. There's a tricky situation going on here in regards to the ethics of AI, especially when it comes to healthcare. Either of you, could you tell me a little bit about the future of liability, accountability? Markus, I know we've talked about this previously, patient empathy and how that applies to AI.

Matthew: The liability question is a tricky one. I think overall, and this is going to be a trial and error thing, we have to figure out how to integrate these systems into our workflows and how doctors are going to be using them. We need to figure out how to gear the system at incentivizing people to reduce errors as much as possible. At the end of the day, I think what's key here is that AI systems, while they can flag cancer in radiological images, while they can make diagnoses, while they can do X, Y, and Z, these are tools. The person who is applying this tool to the patient and using that tool needs to know full well how that tool works, what types of patterns in diagnoses it has, any blind spots it has because the patient is not going to be informed about that. The person, the doctor using the tool will have much more access to the information about these tools. Therefore, the doctor should be reading up on when and when not to use these technologies and be taking it upon themselves in terms of liability to control those potential failures.

The challenge is some people think that liability should fall onto companies, and the problem is that companies don't know exactly how these technologies are going to be used in the real world. They could design it for a very specific use case, but on the ground, a doctor could still end up using it for a different use case. That's not something the company can control, but that is something the doctor is actively controlling, and therefore that person who chooses when to apply these systems should take on liability.

Sam: Absolutely, and Markus.

Markus: What I would add is that I completely agree that if AI is going to actually make clinical decisions without any oversight by an actual doctor, there's a liability question that we need to think about. What seems to be a key point to me is that there's an awful lot of benefits on the path from where we are today until we get to a stage where AI is making decisions autonomously .

Self-driving cars, people are worried about what are they going to do when they have to crash? Are they going to crash this way or that way into the two adults or the four children? Yes, that's a liability question, but there's an awful lot of benefits on the way there. Right now, the cars are braking if you're about to hit the car in front of you. It's the same with AI, they can assist in a lot of ways with our oversight. I think a lot of the benefits are going to be on the path to that state where AI is making decisions themselves. A lot of these liability questions, they come in down the road. There's a lot we can do with AI before we need to solve some of these difficult liability questions.

Matthew: I just want to add on that in terms of the benefits here, we need to make sure that liability, any laws or whatever that we pass regarding this, do keep in mind a certain amount of error should be acceptable, because we want these systems in part to work as a double-check.

The doctor makes a diagnosis and then the AI reviews that diagnosis and offers its opinion as just a second check to make sure the doctor has thought about X, Y, and Z. I think in that specific circumstance, the AI can be wrong and we should expect that it will be wrong, and we don't want to deter people because of liability questions from using it and perhaps getting a wrong result. A doctor should be able to use this thing as a double check so we can get overall better outcomes and feel confident in the ability for certain mistakes along the way.

Markus: The relevant comparison is how often are the doctors wrong? We shouldn't expect the AI to be right 100% of the time, doctors are not right 100% of the time.

Matthew: Absolutely. I was actually just on that note reading a story. This woman went in for a mammogram, and she had two doctors tell her she did not have cancer. Then they ran the screens through an AI system, and the AI saw some little gray patch or something like that, that turned out to be stage two cancer. Two doctors got this wrong. Actually, two doctors they're doing due diligence there to have a second opinion on this very critical screening still got it wrong. The AI then filled in and got it right. That's the type of thing we're trying to prevent here, doctors get stuff wrong all the time and we just need this double-check.

Sam: Absolutely. Markus, you had a fun tidbit on the empathy piece of AI. We think of AI as this robotic future. For a lot of people AI is very scary, but you've found some research that that may point to the contrary, actually.

Markus: Yes. Initially, I think a lot of people thought, “Well, these tools might be helpful, but they're never going to be able to reproduce the human connection a doctor has with their patient.” What turns out to be the case is that doctors are often not all that personable. They have a lot of doctors speak, they talk in a way that's not always that empathetic. One study found that actually, a well-designed chatbot, patients find it to be more empathetic than the doctors. That doesn't mean that we should replace the doctor entirely, but it's clear that it can already assist the doctors in delivering messages to patients or talk through questions patients may think of after they've seen the doctor and received the diagnosis. “Oh, you have a follow-up question? Maybe a chatbot is the first place you should ask.”

Sam: It's a wonderful nugget. It's a great place especially-- as we talked about previously, everybody has a MyChart now, and so they can receive the results and you can actually put those PDFs into something like ChatGPT and have them read your results. You just had an experience with this recently.

Markus: Yes, exactly. My beloved dog Uno, he had trouble-- he got himself into a bottle of melatonin and ate the whole bottle. Melatonin wasn't dangerous for him, but it was strawberry-flavored and had xylitol in it, which is dangerous for dogs, so we had to take Uno to the emergency vet. Me and my background, I was obviously worried for Uno as anyone would be, but I also know how these healthcare systems work. I immediately knew that it's worth getting a second opinion here. Are we treating appropriately here?

I called our regular vet. She didn't call us back. It could be because calling people and giving a second opinion isn't something you can bill for. I did ask the place for a PDF of just all the clinical notes, test results, everything they had, and I gave it to Chat. Chat was perfectly able to guide me through all the questions that I had along the way. Initially, it was like, "Look, are we giving the appropriate treatment here? These are the results." It was comforting, actually, to know that, "Yes, these results are serious. It should be treated this way, and it is being treated that way."

The other thing that I'll just mention is we know that, there's this asymmetric information between doctors and patients. Doctors know what the results are, what's best. Patients don't know that. One great thing about the LLMs is that they are reducing this asymmetric information problem because you can say, "Hey, Chat, you are a clinical doctor specializing in this. Is this appropriate?" The next questions that I then started asking Chat was, "Look, are we overtreating, Uno? He received three plasma transfusions on the same day." I still don't know how much they cost, but I know there was a lot of zeros. It was also comforting to know that those values had gone up a lot and it needed to be taken care of. It needed a lot of treatment.

That's one place where we might reduce some wasteful healthcare spending if patients have this ability to double-check in both directions, are we doing too little? Are we doing too much? I think one area where that aspect patients just go through their patient portal, downloading everything that the doctors have, notes, test results, and feeding them to AI, can help in the areas where regulations might prevent AI from being integrated, like VA hospitals, government-run healthcare facilities. My guess is they're going to be the last place.

Matthew: Probably, given current regulations, yes.

Markus: Probably, the last place where AI will be integrated. Patients can be empowered by AI here. I think that's also going to have a lot of benefits.

Sam: Absolutely. Is Uno home safe?

Markus: Thank God. He is back. It's like he's 10 years younger. He is powering throughout, patrolling the yard, chasing squirrels. Uno is back.

Sam: Wonderful. Matt, one final question for you, and this may be an ambitious question. I'm going to be asking you to be a fortune teller here. LLMs currently are all the rage. We're finding incredible uses for them, and they're getting more and more advanced by the day. Beyond LLMs, where does the future of AI lie? Because AI and LLMs are not exactly the same thing, even though the general public equates them to be very often.

Matthew: Especially within health care, there's so many areas. Obviously, LLMs can help with things as we just discussed, patient charts, interpreting those to people, providing second opinions, providing just a 24-hour doctor, if you will, that could potentially analyze your test results and whatnot. There's so many different automation cases that this stuff can be used for. Triage, for example, in hospitals, whether it be a dog or a human. Triage is one area in which stressed out humans are probably not the best decision makers. ER docs and support staff are being flooded with patients all the time, patients who have different linguistic needs, who are yelling and screaming and doing all these crazy things. That's not a great place for good decision making about who should receive care start.

AI systems, perhaps, and there's good evidence to suggest this, perhaps could do a much better job at this. If you get triage right, if you automate that well, you can save a lot of lives. Just to give you an example, there's a system called eStroke that was recently trialed in the National Health Service in Britain. What they found was that by improving the automation of stroke triage, specifically, they're able to cut down the time to care by about an hour. By doing that, the result was that people who achieve functional independence after their hospital visit for the stroke increased from 16%, which is exceedingly low, to 48%, about half. That's crazy.

This isn't like some fantastical technology. This isn't an LLM or any of this cutting edge stuff. It's more traditional technologies that focus on automation, but by integrating those types of things, we can really change people's lives. We should be thinking about that type of thing, too, when we're talking about this conversation. It's not just ChatGPT. It's analyzing studies to improve policies. It's automating triage. It's developing new drugs. It's detecting cancer in diagnostic images. It's all of these things. We really need to be thinking about AI in this holistic manner because it's so diverse. If we take advantage of that full diversity, that's how we get this transformation.

Markus: One thing that I would add to that is I think some areas where we'll see a lot of benefits might not sound that sexy when we talk about them here at a podcast. One example that I'm excited about is that doctors have to write these pre-authorization requests to get the insurance company to pay for a lot of services. On average, doctors write 37 of these per week, and it takes a couple of minutes, five minutes each of them. AI can both write those and the doctor can just have a quick look at it. "Yes, this is what it should say," and send those to the insurance company. If AI can save doctors one minute for each of those, that's more time that they can spend actually doing what they're good at and have more time with patients.

Similarly, if AI write these pre-authorization forms, sends them to the insurance company, the insurance company can have AI that reads them and reviews them, and in a second, it can be approved. 80% of them are approved anyway. Instead of having to wait three to four weeks for it to get to an actual human's desk before it gets approved, this can be quicker so patients can go and get that service, that treatment, faster.

I think there's a lot of use cases, and some of them might not sound that sexy, but will produce a lot of benefits by allowing doctors, the most highly skilled people, just giving them more time to do what their skills are needed for, not menial tasks that universally anyone could do.

Matthew: To really emphasize this, I've talked about things like drug development, but it is the boring stuff that really will matter, especially with doctors, as was insinuated. We actually have a pretty large labor shortage right now. If we can unlock 50% more of a doctor's time, that's a doubling of labor. I mention to people frequently, it's like when you think of what a doctor does, you think of someone at the bed injecting things and whatnot. What a doctor actually does is paperwork, and we need to change that. AI could potentially go a long way towards fixing that problem.

Sam: Absolutely. Thank you both so much for enlightening us on the present and future of AI and healthcare. I look forward to seeing more great work from both of you in the future.

Markus: Thank you, Sam.

Matthew: Thank you.

Note [AB]: Last week, our team was delighted to see that the paper described in this post was trending as the #2 paper in economics, according to OOIR. It’s encouraging to see an increased interest in policy solutions that can provide greater healthcare accessibility to people in rural communities. We invite you to dive into the full study (linked below) and share it with your friends and colleagues.

Our research — recently published at the Southern Economic Journal — indicates that repealing CON laws increases hospital access, including in rural areas.

Rural Americans fare worse than their urban counterparts on many indicators of health, including mortality, chronic disease and behavioral risk factors. A long-standing challenge is that acute shortages of healthcare providers, in addition to already long driving times to healthcare facilities, make it difficult for many residents of sparsely populated areas to receive needed care.

Some policymakers claim that by controlling where additional health services can be offered, they can direct providers to invest more in rural and other underserved areas. That has been one of the primary justifications for Certificate-of-Need (CON) laws to remain on the books since 1986, when the federal government repealed its CON mandate on states after just 12 years. These laws require that healthcare providers receive approval from a state board before opening a new facility, offering additional services or making major capital investments in a given community. Proponents of CON often claim that these regulations ensure care will be provided to residents in rural, geographically underserved or economically depressed communities.

For example, the West Virginia Health Care Authority, the agency tasked with administering the state's CON program, argues: “In West Virginia, the CON program offers some protection for small, often financially fragile, rural hospitals and the underinsured population they serve by promoting the availability and accessibility of services." The Kentucky Hospital Association has claimed that repealing CON “would be another nail in the coffin for rural communities."

Yet, this may be just an excuse to restrict competition. In many states, the CON application process is lengthy and expensive, imposing a disproportionate burden on providers that are smaller, newer or from disadvantaged communities. Larger and more politically influential healthcare entities are often better positioned to navigate the bureaucratic process and exert both public and behind-the-scenes pressure to secure favorable outcomes. Indeed, there is evidence that political contributions affect the likelihood of CON application approval.

Maureen Ohlhausen, former commissioner of the U.S. Federal Trade Commission, asserts that CON laws “effectively serve primarily, if not solely, to assist incumbents in fending off competition from new entrants." Courts have repeatedly struck down CON restrictions for being blatantly anti-competitive, such as when “a state program was found to have denied a proprietary hospital's application because of a hidden preference for existing facilities."

So which story is true: Are CON laws used to direct providers to rural areas, or is that just an excuse used by incumbents to limit competition? In our recent research, we and our co-authors show strong evidence in support of the latter. We conduct a careful causal estimation of the effects on hospitals and find that repealing CON causes a substantial increase in hospital facilities per capita in both rural and urban areas. Moreover, we find that the repeal encourages smaller hospitals to enter the market, suggesting that CON laws are allowing existing large hospitals to suppress competition.

Given the careful analysis and clear results in our forthcoming paper, our conclusion is that repealing CON laws would go a long way toward increasing access and competition in healthcare, including in rural areas.

Many papers in empirical health economics report findings that aren’t all that surprising. Mandating more benefits makes health insurance more expensive; offering free health insurance to millions of people reduces the uninsured rate; raising tobacco taxes reduces smoking.

But occasionally a result comes along that’s tantalizingly counter-intuitive. Over the last 25 years, solid evidence has emerged showing that economic downturns in the U.S. tend to be accompanied by declines in mortality — such that, according to the leading scholar on the topic, Christopher Ruhm of the University of Virginia, the U.S. would have experienced 40,000 more deaths during the first year of the COVID-19 pandemic if the unemployment rate hadn’t gone up.

Another recent analysis concluded that “the Great Recession provided one in twenty 55-year-olds with an extra year of life.”

Let’s get something out of the way. People thrown out of work during recessions do not suddenly get healthier. In general, losing one's job has a very negative impact on health. One study estimated that “the health-damaging effect associated with being jobless is similar to the effect of about 10 extra years of age.”

So if recessions cause a lot of people to become unemployed, why don’t we see a substantial increase in mortality?

The quick answer is that the declines in health experienced by those who become unemployed during recessions appear to be more than offset by health improvements in the rest of the population.

One explanation is that a contracting economy usually means better air quality due to fewer commuters on the road, fewer vacationers traveling, and fewer factories billowing smoke. Poor air quality, according to some estimates, is responsible for up to 200,000 U.S. deaths annually, so declines in pollution can have a meaningful effect on mortality. Reductions in travel also translate to fewer motor vehicle fatalities, the leading cause of death in the U.S. for those under age 55.

Other possible explanations are more speculative, with different studies offering inconsistent results. It’s plausible that by reducing incomes, recessions might dampen consumption of unhealthy substances like cigarettes and alcohol. On the other hand, economic uncertainty might make people more anxious, which could increase smoking, drinking, and drug use. Some studies suggest that recessions increase heavy drinking, especially among men.

A third possibility focuses on time costs. When employment opportunities dwindle, the opportunity cost of leisure declines, which might encourage people to engage in behaviors that promote health but are time consuming, such as exercising regularly, preparing home-cooked meals, and maintaining an active social life. Here again, the empirical evidence defies simple characterizations. For example, while some studies find that exercise increases during recessions, total physical exertion tends to decline because some people lose physically-demanding jobs and don’t increase their exercise routines enough to make up the difference.

So while the pro-cyclicality of mortality in the U.S. is well-established, we haven’t yet nailed down the mechanisms responsible for it.

__

Covid-19 exposed a lack of medical personnel in the United States to meet a national emergency, but a shortage of doctors has been a problem in the US for years. The number of doctors is a function of how many new doctors are admitted to the profession and how many leave the practice of medicine each year. The supply of new entrants is constrained primarily by the Centers for Medicare & Medicaid Services (CMS), a federal agency that provides the bulk of the funding for hospital residencies.

Hospital residencies are positions for recent medical school graduates to work in a clinical setting – usually at a hospital or doctor’s office – treating patients and continuing their training in a particular subfield. Residency durations vary by specialty. Even after medical school, an MD cannot obtain a license to practice medicine without at least one year of residency. The number of residency slots directly determines the number of licensed doctors entering practice.

CMS residency funding was capped beginning in 1997 at 1996 levels, and has only been raised once since then in Section 126 of the Consolidated Appropriations Act of 2021. From 1987 to 1997 the number of residents grew by 20.6%, while from 1997 to 2007 the growth in residencies was only 8%. The 2021 law attempts to address several issues by slowly increasing residencies in specific underserved situations such as rural areas. This officially just began with 200 new residencies nationwide this year, climbing to 1,000 additional residencies per year in five years. It is only a drop in the bucket compared to the 140,000 resident doctors in the US in 2020, the majority of whom were federally funded. Using a rough estimate from Census population data for 2000 and 2020, there are at least 50 million, or 18%, more people in the nation today than when the law was passed, making the current change a rather small effort.

This growth in population without a corresponding growth in the doctors we train each year leads to higher salaries for doctors and higher costs for patients. More people are demanding the time of a similar number of doctors, driving up the price of doctor’s time, even before accounting for the US’ aging population.

The law of supply and demand predicts that a larger supply of doctors would drive down the cost of doctors’ services as well as how much they are paid. The 1997 restriction on supply of residencies was originally lobbied for by the American Medical Association (AMA), the main professional association and lobbying group for doctors. The AMA now recognizes the shortages this created and is encouraging Congress to remove the limit. US healthcare costs are rising at a pace that the profession is actively asking for a change that will increase competition for its members. Increased supply of anything, even doctors, means decreased prices, all else being equal.

General (or family) medicine takes three years of residency while specialties can require up to seven years. Compared to general medicine, these extra years provide a doctor with an increased salary netting $1 million to $3 million dollars extra of lifetime earnings. While having an adequate supply of doctors is beneficial to patients and hospitals, practicing medicine is massively lucrative compared to almost any other occupation. Income for doctors is a large incentive to pay for education. The lowest average annual salary for physicians in a state is $149,557 per year in Georgia, according to ZipRecruiter, while some specialties have average annual incomes of a quarter to over a half a million dollars a year.

Solution 1: Just raise the cap above 100,000 residents

Removing the CMS caps will certainly increase the number of residents, but it requires more federal spending when the federal government is already running annual deficits. Half of hospitals receive more than $150,000 per resident, and the quarter of hospitals receiving the lowest payments can range up to more than $100,000. Even if politically achievable, it runs the risk of further increasing the national debt, without some form of new dedicated funding.

There remains the problem of trying to calculate from the top down the right number of doctors needed today, let alone anticipating needs in in five or 10 years. Dictating from above the number of medical professionals needed is arbitrary and static, not taking into account changes in the healthcare market.

Central planning disrupts the ability of the price system to communicate how many doctors are needed though the costs of training and salary. A static and rigid system will constantly under-provide new doctors if it is tailored to past needs, which is likely given how slowly it can be changed. Case in point, the 1997 levels were held constant almost for 25 years.

Removing caps would likely encourage hospitals to apply for funding for more residencies. Currently not all applications by hospitals are granted, indicating that a removal of caps would increase spending by CMS for residency programs. Medical residents are relatively cheap labor compared to nurse practitioners and physician assistants, incentivizing hospitals to try to get more funding for more residencies.

Solution 2: Require hospitals to pay for more of residents’ costs

It is clear that residents are an economic boon for hospitals. The salary of a resident doctor is lower than for nurse practitioners and physician assistants but the resident doctors work more hours. There is also evidence that resident doctors improve patient outcomes, and that teaching hospitals – with more residents – have better outcomes than non-teaching hospitals. Residency programs allow for 24-hour physician coverage and multiple assessments of a patient.

One study looked specifically at the ways in which residents are both an input to hospital productivity and an output for learning purposes and found residents to be an overall input to hospital production, with productivity around 37% of senior physicians. As an overall input we should expect that hospitals would take this burden on so long as they could pay the residents accordingly.

Given the financial incentives to hire residents, hospitals could take on a larger proportion of funding their residencies. One way to achieve this would be to cap the total amount given for each residency. This would allow for a larger number of residencies to be supported by the same money, while shifting more of the responsibility for training residents onto hospitals to reflect the benefits hospitals receive from residents.

The downside of this idea is that increasing hospital costs will need to be paid for by someone, and this may be reflected in hospital pricing. Increasing the overhead costs of hospitals will increase the cost of running the hospital and thus healthcare prices potentially as well. This may be mitigated to a degree by the fact that residency programs already have a net positive effect on hospital finances.

Solution 3: Doctors repay some of the cost of their residency once it is completed

What is preventing private funding in the form of loans or some other mechanism from forming to arbitrage this high salary situation? Currently the private sources of funding for residencies are through the higher rates private insurers pay to teaching hospitals that supplement these programs. However, it is difficult to calculate the amount. Despite private gains in the form of higher wages, we don’t expect residents to take out loans to pay for any portion of this part of their education, despite the higher wages a medical school graduate can command after completing a residency. If the problem with hiring more residents is the cost of them, then borrowing the money to pay the hospital for training arbitrages the potential higher earnings.

Personal loans can respond to market forces to balance the supply and demand for doctors with incoming entrants via the price mechanism, by balancing the salaries of doctors against the loans required to earn them. Currently the salaries of all physician fields are enough to cover the cost of these loans, despite the high costs of residencies. Shifting to a system where residents are expected to bear some of the costs of their post-med school training would allow CMS financial support for residencies to support a larger supply of new doctors without increased federal spending..

Alternatively to loans, doctors whose residencies are supported by CMS could be required to repay a portion of the cost to CMS over a period of several years after the completion of their residency. This method would reduce the risk that a resident would have large loans despite not completing their residency by requiring only those who incur the benefits of higher earnings to repay CMS.

The entirety of the cost does not need to be placed on doctors. This can easily be combined with the suggestion above, as hospitals benefit from the resident doctors as well. Doctors typically graduate medical school with about $200,000 of debt. However, this is also what many doctors make in a single year. While it is a lot of money to be paid back, it is relatively small compared to the increase in earnings that debt allows.

While government funding is not going away, it could be better used to supplement a price incentivized system instead of acting as top-down central planning. Given the high salaries of doctors, taking on a portion of the cost is unlikely to deter enough medical students that expanded residency slots would go unfilled. Federal support for residencies could reflect the provision of public goods, as well as encourage doctors to locate in rural and other underserved communities. Rural doctors, for example, often earn less and could be disincentivized compared to higher paid city doctors and CMS could waive part of their repayment to encourage doctors to practice in those communities.

Adding an additional cost for medical training would be a disincentive for people choosing to become a physician. However, the barrier to more students is currently due to the reliance on a top-down system that limits the number of residents. Rebalancing the way residencies are paid for would allow for more doctors in the long run and could incorporate some private payment for the private financial benefits of residencies. A system of loans to pay for some portion of medical residencies would be imperfect but far more responsive and effective at meeting demand for care than having one agency guess at the medical needs of more than 330 million people.

Medicaid is usually regarded as a health insurance program for low-income people. That view isn’t wrong, of course, but it is incomplete. Another, less often noted, objective of Medicaid is to provide macroeconomic stability by automatically injecting money into the economy when aggregate demand slackens. The trouble is, it’s pretty bad at it.

Before getting into the details of how Medicaid reacts to economic downturns, it’s important to appreciate the sheer size of the program. Medicaid spending accounts for more than 3 percent of U.S. GDP and represents — by a very wide margin — the single largest intergovernmental subsidy in the country. Each year, the program transfers about $500 billion from the U.S. Treasury to state coffers. In 2019, Medicaid accounted for 65 percent of all funding flowing from the federal government to the states. (As a point of comparison, only 8 percent of federal grants that year were targeted to transportation projects.)

Consequently, even small shifts in federal Medicaid funding can have a large impact on state budgets and the macroeconomy. In 2016, Medicaid accounted for more than one-quarter of all state spending. In most states, the only budget category more costly than Medicaid is K-12 public education.

Medicaid enrollment and spending are counter-cyclical, increasing during times of economic distress when employment-sponsored health coverage shrinks as jobs are lost and the number of low-income Americans grows. On top of that, economic downturns usually cause state revenues to stagnate or decline, reducing states’ fiscal capacities precisely when more people are reliant on the social safety net. And since nearly all states face balanced budget requirements, barring them from running deficits to accommodate higher Medicaid spending during recessions, the infusion of federal Medicaid cash helps states stay afloat.

On the surface, Medicaid’s funding rules seem reasonably responsive to changing economic conditions. Each state’s share of Medicaid spending is primarily determined by the Federal Medical Assistance Percentage (FMAP), derived from a formula that shifts a larger proportion of Medicaid expenditures to the federal government in states with lower per capita personal income relative to the national average. Conversely, states with higher per capita personal income are responsible for a larger share of their Medicaid expenditures. So if a state enters a recession, its personal income per capita should fall and automatically trigger larger federal Medicaid payments, right?

In practice, it’s not so simple.

First, the FMAP formula is based on the ratio of state and national per capita personal incomes. When a single state or region experiences a dip in personal income, the FMAP formula ensures that these states receive more federal aid. But if the country enters a recession that affects all states similarly – by reducing personal income fairly evenly across the country – each state’s FMAP doesn’t change much, since the ratio of each state’s per capita personal income to national per capita personal income stays about the same.

Second, there’s a substantial time lag in the collection and calculation of the personal income data used to determine FMAP rates. It takes years for statisticians to generate state-level estimates of personal income, and the problem is compounded by the fact that the figures used in the FMAP formula are based on a rolling three-year average. As a result, FMAP rates in effect at any given time are based on data from three to six years earlier. In setting FMAP rates for fiscal year 2021, for example, the federal government used personal income data from 2016, 2017, and 2018. In other words, while the country was still recovering from a deep recession, federal Medicaid payments to states were still being calculated based on pre-COVID data.

The FMAP formula’s sluggish response to recessions has led federal lawmakers to repeatedly intervene, passing legislation to temporarily increase the FMAP rate during recent economic crises (including the recessions of 2001, 2007-09, and 2020). These post hoc revisions to the FMAP rate have several limitations. They are prone to political squabbles, are often enacted months after crises hit, and create significant uncertainty for states.

For example, the Families First Coronavirus Response Act — the first major pandemic-related legislation, signed into law on March 18, 2020 — provided states with a temporary 6.2 percentage point increase in their regular FMAP rate. There’s no reason to think that the 6.2 percentage point figure was the result of any careful analysis. In fact, we know it wasn’t: Congress just recycled the exact same figure that it had approved in the American Recovery and Reinvestment Act of 2009. Despite starkly different circumstances — it was obvious that the COVID-19 pandemic would impact Medicaid differently than the Great Recession had — Congress didn’t make any effort to tailor its response.

Instead of relying on a chaotic federal policymaking process to pass FMAP rate increases during recessions, a better solution would be to modify the FMAP formula to automatically deliver additional federal Medicaid funds to a state if the health of its economy falls below some threshold.

Concrete proposals along these lines have been put forward by the Government Accountability Office (GAO) and the Brookings Institution. The details differ — the GAO favors using the employment-to-population ratio to gauge the economy’s health, while the Brookings plan focuses on the unemployment rate — but the main idea is the same: Create an automatic trigger mechanism based on high-frequency data that sidesteps the political dysfunction of Washington and turns Medicaid into a macroeconomic stabilizer worthy of the name.

There's no doubt that residents in nursing homes with more staff experience better outcomes—the literature on this topic is vast. The CMS-commissioned ‘Staffing Study’ includes a review of over 30 studies from 2019 to 2022 alone. However, most studies on staff and outcomes are descriptive, illustrating that more staff correlates with better outcomes, but stops short of asserting a causal relationship. 

But if the goal is better policy, the quality of the evidence should be the primary consideration. The sheer number of papers, whether 100 or 200, shouldn’t matter much, if what they predominantly show that healthier, financially better off residents are able to get care in nursing homes with more staff. 

Studies attempting to estimate the causal effects of increasing staff through minimum staff policies present mixed findings: one study (Lin 2014) shows a positive relationship for nurses, but no effect for nurse aides; another finds no effect from adding more social workers. Evidence suggests that while total staff hours increase, an unintended consequence is a reduction in registered nurses. Not surprisingly, several studies also conclude that the added costs negatively impacts the financial performance of nursing homes, potentially leading providers to exit and/or deter entry. 

An important question, therefore, is how policymakers respond to the quality versus the quantity of evidence. An interesting policy experiment is currently underway:

The proposed Minimum Staffing Rule was announced September 1, 2023 with a comment period of 60 days. In that time, the federal government received over 46,000 comments (!), which CMS now has to sift through and, theoretically, use to inform their next steps. The sheer amount of comments has already delayed the expected implementation date until 2025 (I call this as a point for team-quantity.)

Research shows that policymakers do make changes based on comments, so it will be interesting to see how the final rule differs from the proposed version. While comments on proposed rules is not where you'd expect high-quality research evidence to have the most influence on policymakers, the mere four mentions of "Lin 2014" among 46,000 comments is nevertheless  interesting.

Given the high-quality evidence, it's clear that strict enforcement of the rule by CMS would necessitate lowering the requirements, or risk triggering massive bankruptcies throughout the industry. The research also shows far more careful consideration is necessary regarding the impact on staff composition. Is CMS sure patients will benefit from more nurses and nurse aides, given the very predictable reductions that will occur to licensed practical nurses, a category not covered under the proposed rule and the obvious place nursing homes will cut costs to stay afloat? So what do you think, will CMS consider the insights from high-quality studies, or will the sheer volume of public feedback steer the course of policy?

While Medicare’s rosier-than-expected spending trajectory deserves to be celebrated, less attention has been paid to a similar pattern unfolding in Medicaid — only in the opposite direction.

In the early 2010s, after the ACA’s passage but before the implementation of its core provisions, the Centers for Medicare & Medicaid Services (CMS) projected that the adults becoming newly eligible for Medicaid under the ACA would cost less — a lot less — to insure, on average, than other non-elderly adults already on the program. That prediction turned out to be completely wrong.

In its 2013 report on Medicaid’s financial outlook, the last analysis released before Medicaid expansion began, CMS projected per enrollee Medicaid spending for different groups of beneficiaries. That data is reproduced in Figure 1 for newly eligible adults and other non-elderly adults on Medicaid. In the first year of expansion (2014), expansion adults (the violet line) and other non-elderly adults (the blue line) were expected to cost virtually the same. But by 2016, expansion adults, on average, were projected to only cost between two-thirds and three-quarters of what other non-elderly adults cost. Moreover, the gap was expected to gradually increase over time, with per enrollee spending on other non-elderly adults rising slightly faster than per enrollee spending on newly eligible enrollees. In 2022, the last year in the projection window, CMS predicted that expansion adults would cost $4,875 per enrollee, while other non-elderly adults were expected to cost $7,195 per enrollee — nearly 50 percent more.

CMS thought that newly eligible enrollees would be relatively inexpensive to insure for three reasons: 1) they would be in fairly good health — otherwise they would qualify for Medicaid under pre-ACA rules (e.g., by receiving Supplemental Security Income); 2) large numbers of people with low health spending, especially young adults, would join Medicaid under the ACA, dragging down the average spending per enrollee; and 3) per enrollee spending on newly eligible adults would moderate after an initial surge in costs in the first year or two as individuals who had not previously had their medical needs adequately met began accessing services (i.e., “pent-up” demand).

Though its assumptions appear reasonable, CMS’ actuaries blundered, recent data shows. Table 1 contrasts the predicted levels of spending for each eligibility group with the actual levels of spending in 2019. (More recent data is available, but may be distorted by the COVID pandemic. CMS can hardly be blamed for failing to foresee a once-in-a-century health crisis.)

Not only did actual spending on newly eligible adults far exceed projections, but actual spending on other non-elderly adults came in well below expectations. What’s going here?

I don’t think anyone has a fully satisfactory answer. But with hundreds of billions of dollars at stake, more scholars should be tackling this question. 

One potential explanation is the opioid epidemic. The severity of the opioid crisis over the last decade could not have been accurately predicted in 2013. Medicaid expansion has increased treatment for opioid addiction at a substantial fiscal cost, and it seems plausible that expansion adults may be more likely to use drugs and avail themselves of addiction treatment than other non-elderly adults.

My Mercatus colleague Charles Blahous has argued that institutional incentives may have played a role as well. The enhanced federal funding for expansion adults may have made states less vigilant about containing costs for this population.

Some of my ongoing research touches on this theme. The data suggests that states have reclassified some Medicaid enrollees from the “non-expansion” group (i.e., individuals who are eligible under pre-ACA rules) to the newly eligible group. Since per enrollee costs in the non-expansion group are much higher than in the newly eligible group (driven mostly by high spending on the aged and those with disabilities), these reclassifications serve to artificially inflate per enrollee spending in the newly eligible group. Audits carried out by the federal government in California, New York, and Colorado (all of which expanded Medicaid) suggest that about 10-25% of enrollees classified as newly eligible under the ACA are, in fact, either ineligible for coverage or misclassified.

Forecasting is, of course, a highly imprecise exercise. I don’t fault CMS for making mistakes. But the evidence from Medicare and Medicaid over the last decade should make us skeptical of anyone’s ability to accurately foresee all the consequences — intended and unintended — of policy reforms.

COVID-19 tore through nursing homes with staggering lethality. As of March 2022, two years into the pandemic, more than 150,000 US nursing home residents had died of COVID-19 – roughly 10% of the total U.S. nursing home population. Sadly, well-intentioned lockdowns made things worse.

Prior to the development of vaccines, the primary strategy for protecting nursing home residents from COVID-19 was physical distancing and isolation. Residents were often confined to their rooms, unable to socialize with family and friends or even interact with fellow residents.

From the outset, it was clear that these draconian measures would have devastating effects on nursing home residents’ quality of life, cognitive functioning and mental health. In mid-2020, only 5 percent of nursing home residents surveyed reported having visitors three or more times per week, compared to 56 percent before the pandemic.

Many nursing home residents expressed anguish at having their social lives suspended. “I have become more anxious and depressed due to the separation from my loved ones. I have little appetite and am losing weight,” shared one resident. “If the virus doesn’t kill me, the loneliness will,” said another.

This suffering was justified, some argued, because isolation was effective at blunting the spread of COVID-19. It wasn’t until September 2020 that the Centers for Medicare & Medicaid Services (CMS), the agency overseeing the two federal health insurance programs, eased the visitation restrictions it had imposed in March 2020 and began permitting nursing home visits under limited circumstances.

In recent research, I show that the overall effect of COVID-19 nursing home lockdowns on residents’ health was sharply negative. Using cell phone tracking data, I create an index of isolation measures at more than 10,000 nursing homes in the U.S.

I find that while stricter isolation measures were associated with fewer COVID-19 cases and deaths, they were also predictive of substantially higher rates of non-COVID and total deaths in the second year of the pandemic. Overall, I estimate that isolation measures were associated with substantially more total deaths: A one standard deviation decrease in isolation in all nursing homes is predictive of 7,305 fewer overall deaths.

The effects are largest in facilities with higher proportions of residents with dementia. Because regular social contact is critical to slowing the progression of diseases like Alzheimer’s, this finding supports the hypothesis that isolation measures harmed residents’ health.

These results join other post-mortem analyses of the pandemic response, exposing egregious — and, in cases like this one, deadly — miscalculations by public health officials. Learning from these mistakes is essential. COVID-19 will not be the last pandemic.

My findings also have immediate relevance. Even in the absence of a pandemic, airborne pathogens are common, and CMS guidelines urge nursing homes to implement physical distancing and isolation to limit transmission. Yet there are very limited data on the effectiveness of isolation procedures in reducing infections. Nor have the broader impacts of these policies on mental health outcomes been sufficiently scrutinized.

There’s no escaping it: A trade-off exists between shielding nursing home residents from infectious disease risk and inflicting harm on them through denial of social interaction. Let’s do better at striking the right balance.

The Swiss Cheese Model, developed by James Reason, is a simple model that’s widely applied in aviation. The model illustrates how multiple layers of safety systems and procedures, none of which are individually foolproof, when stacked make aviation incredibly safe. In fact, with enough carefully designed layers, you prevent all accidents except for those rare occasions when the holes align perfectly, often due to a unique combination of previously unforeseen circumstances.

This model is also a good dynamic description of accident prevention in aviation. Accidents and even minor incidents are investigated by the airlines and aviation authorities across the world, leading to reports with recommendations to add more, or tweak existing layers, in order to prevent the previously unforeseen alignment.

To give you an example, the Airbus A320 that Captain Chesley "Sully" Sullenberger famously landed on the Hudson River had fly-by-wire flight controls powered by the aircraft's engines. In the rare event of an engine failure (ChatGPT says a rate well below 1 in 100,000 flight hours), the pilots would still be able to control the plane using power from the other engine. In the event of a dual-engine failure (an event so rare ChatGPT won’t give me a rate, but instead lists just two examples), the plane actually has two additional systems! The Ram Air Turbine, an almost cartoonish windmill device that will automatically pop out from under the plane and generate enough electricity to power the most essential systems, and an Auxiliary Power Unit (APU), a gas-powered turbine capable of maintaining all flight-control systems, including the more advanced safety features (ChatGPT tells me there’s are also emergency batteries that can power the most critical systems and instruments, but I think we all get the idea at this point).

That’s great! It would be crazy if we didn’t try to learn from our failures, and in the Swiss Cheese model, it is these kinds of redundancies that make the whole enterprise safe. But while it’s a great starting place, much like economists will start to think about virtually anything through the lens of supply and demand, let this dynamic play out long enough, and perhaps a different model would be better.

Even if the simplest version of this model will always get more safety by adding layers, you’ll notice that as you keep adding layers, each subsequent layer is almost entirely redundant. This should make you wonder what the model is not showing?

The obvious answer is that the model fails to convey the cumulative costs that adding more and more mostly redundant slices can have in the other direction. Gordon Tullock likened this kind of regulatory accumulation to throwing rocks into a river: while no single rock noticeably alters the river's course, the collective impact over time directs the stream in unintended and unpredictable ways.

Captain Sully’s experience with US Airways Flight 1549 , which hit a flock of geeze shortly after takeoff, actually highlights the crux of the issue. Standard procedure with a dual-engine failure involves the pilots executing check-lists of specific steps (not unlike the MDS forms nursing homes are required to complete on a regular basis, for example). The check-list Captain Sully and his first officer were meant to execute can be seen here, but was unfortunately designed for dual-engine failures at cruising altitude, calling for multiple rounds of specific configurations of buttons and switches, each round followed by 30 seconds of wait time, before attempting to reignite the engines.

Keeping his wits about him, Sully’s decades of experience allowed him to realize immediately that complying with the checklist, designed for a scenario that differed from the one he was facing on just one critical margin, would doom his flight of 150 passengers. Sully instead chose to deviate from protocol by activating the plane’s Auxiliary Power Unit just two seconds after hitting the birds. This “Sully” moment, an almost instantaneous decision, returned safety features to the flight control systems and was pivotal to saving the lives of all 150 passengers.

Readers of this blog will not be surprised to learn this incident led to a series of requirements and recommendations, including — you guessed it — a checklist for dual-engine failures at low altitude. Now that list may be a good idea, even if dual-engine failures at low altitude are extremely rare, and I certainly won’t take a stance on whether there are too many safety layers in aviation after one flying lesson.

But what I will point to, is that while Sully is rightly portrayed as a hero both in the movie starring Tom Hanks and in the NTSB’s official report, that same report actually describes another Sully moment that we can learn just as much from. As Sully was landing on the Hudson, the report describes that he experienced "task saturation" due in part to the extensive and lengthy checklists and protocols that were poorly designed for the scenario he was in. While intending to realign a previous hole in the layers of the swiss cheese, the inclusion of procedures for scenarios both with and without fuel remaining, along with several ditching procedures, actually contributed to the complexity of Sully’s situation, and led him to approach the landing on the Hudson at a lower than intended speed. Interestingly, the safety features Sully brought back by activating the APU were critical to the landing being relatively smooth despite the low speed, and ultimately to the survival of everyone on the plane.

While the immediacy of the consequences differ, my claim is that a regulatory approach of requiring very specific procedures for almost any aspect of healthcare delivery leads doctors and nurses to be task-saturated every single day.

Take something seemingly simple like administering medications. Nurses are required to double-check medications against the patient's chart, confirm the patient's identity, and document the administration details in a medication administration record (MAR). Now do this for 20 patients, and you see where I’m going. While each of these steps could prevent the unholy alignment of holes in the layers of swiss cheese, there’s no safeguard in this model that ensures we prevent that alignment in the lowest cost way.

Of course, following a slightly imperfect procedure won't harm a patient in a nursing home, the way Sully's impossible check-list would doom his passengers — it will just waste some time. But that's exactly the point. The accumulation of slightly imperfect procedures followed for every single patient, every single day, will add up to dramatically less time for patient care. But with each imperfection being minor, never would a doctor or nurse recognize the higher imperative, and deviate from their checklist, the way Sully ignored his checklist to activate the Auxiliary Power Unit.

We need a collective Sully moment in healthcare regulation, but the point I’m trying to make is not that we should ignore rules and procedures. A "Sully" moment in healthcare regulation would be for us all to recognize the cumulative costs of ever more protocols. We should recognize that no one set of procedures can be optimal for the vast variety of circumstances that occur across 15,000 nursing homes on any given day, and that it is impossible to devise optimal procedures for all or even most circumstances, without simultaneously creating an extremely large and cumbersome set of procedures that wastes crucial time and resources.

No amount of tweaks can prevent anything bad from ever happening. What we can prevent, however, is that the accumulation of layers and protocols becomes a burden that directs well-meaning people away from serving their patients to remain compliant. The cumulative effect of compliance is not just delays, but ultimately reduced time for patient care, so we should reevaluate our approach, to ensure that the safeguards we have in place enhance rather than impede the ability of providers to deliver care, and that we start getting rid of the layers that don’t.

Over the last few years, a number of high-quality papers have presented convincing causal evidence that giving poor children Medicaid coverage produces large and durable improvements in their health, reduces criminal behavior and incarceration, and enhances their performance in the labor market as adults. There’s even evidence of positive intergenerational effects.

A recent working paper from CBO estimates that Medicaid enrollment during childhood increases later-life earnings (thereby boosting tax revenues and reducing spending on anti-poverty programs) enough to plausibly offset about half of Medicaid’s initial outlays.

In short, the funds devoted to Medicaid coverage for children (about $3,600 per child per year) might be some of the best-spent money in the entire U.S. social safety net.

It seems that an important mechanism behind these effects — unsurprisingly — is that children on Medicaid receive more healthcare services than they otherwise would have, setting them up for a lifetime of better health.

Much of this research, however, leverages changes to Medicaid that were made in the 1980s and 1990s, when eligibility rules were tighter, total enrollment was much lower, and children made up a larger proportion of Medicaid enrollees. A lot about the federal-state program has changed since then, most notably the expansion of Medicaid under the Affordable Care Act (ACA), which added millions of adults to the program and created fiscal incentives for states to prioritize these newly eligible adults over other eligibility groups, including children.

On the one hand, expanding Medicaid eligibility for adults tends to increase Medicaid enrollment among children, too (the “woodwork effect”). That’s what the Oregon Health Insurance Experiment showed, and evidence from the ACA finds fairly large effects as well. So Medicaid expansion may be indirectly responsible for more children benefiting from Medicaid.

But there are other factors to consider. The influx of new enrollees under Medicaid expansion has strained some aspects of the health system. As I’ve noted before, several studies have linked Medicaid expansion to longer wait times for appointments, slower ambulance response times, and greater delays in the emergency room. These obstacles may make it more difficult for children to obtain care, eroding the historical benefits of Medicaid coverage for this population.

In a study with my Mercatus Center colleague Charles Blahous, I find that per capita Medicaid spending on children in states that expanded Medicaid in 2014 was only 5.9 percent higher in 2019 than it was in 2013 — an annual increase of less than 1 percent per year. Among non-expansion states, per capita Medicaid spending on children grew 22.7 percent over the same period, implying that Medicaid expansion may have redirected resources away from children.

This phenomenon has likely weakened the beneficial effects of Medicaid coverage for children, and ought to be considered when assessing the overall merits of Medicaid expansion — especially because any direct health benefits to adults from gaining Medicaid coverage are hard to detect.

Our research — recently accepted at the Southern Economic Journal — indicates that repealing CON laws increases hospital access, including in rural areas.

Rural Americans fare worse than their urban counterparts on many indicators of health, including mortality, chronic disease and behavioral risk factors. A long-standing challenge is that acute shortages of healthcare providers, in addition to already long driving times to healthcare facilities, make it difficult for many residents of sparsely populated areas to receive needed care.

Some policymakers claim that by controlling where additional health services can be offered, they can direct providers to invest more in rural and other underserved areas. That has been one of the primary justifications for Certificate-of-Need (CON) laws to remain on the books since 1986, when the federal government repealed its CON mandate on states after just 12 years. These laws require that healthcare providers receive approval from a state board before opening a new facility, offering additional services or making major capital investments in a given community. Proponents of CON often claim that these regulations ensure care will be provided to residents in rural, geographically underserved or economically depressed communities.

For example, the West Virginia Health Care Authority, the agency tasked with administering the state's CON program, argues: “In West Virginia, the CON program offers some protection for small, often financially fragile, rural hospitals and the underinsured population they serve by promoting the availability and accessibility of services." The Kentucky Hospital Association has claimed that repealing CON “would be another nail in the coffin for rural communities."

Yet, this may be just an excuse to restrict competition. In many states, the CON application process is lengthy and expensive, imposing a disproportionate burden on providers that are smaller, newer or from disadvantaged communities. Larger and more politically influential healthcare entities are often better positioned to navigate the bureaucratic process and exert both public and behind-the-scenes pressure to secure favorable outcomes. Indeed, there is evidence that political contributions affect the likelihood of CON application approval.

Maureen Ohlhausen, former commissioner of the U.S. Federal Trade Commission, asserts that CON laws “effectively serve primarily, if not solely, to assist incumbents in fending off competition from new entrants." Courts have repeatedly struck down CON restrictions for being blatantly anti-competitive, such as when “a state program was found to have denied a proprietary hospital's application because of a hidden preference for existing facilities."

So which story is true: Are CON laws used to direct providers to rural areas, or is that just an excuse used by incumbents to limit competition? In our recent research, we and our co-authors show strong evidence in support of the latter. We conduct a careful causal estimation of the effects on hospitals and find that repealing CON causes a substantial increase in hospital facilities per capita in both rural and urban areas. Moreover, we find that the repeal encourages smaller hospitals to enter the market, suggesting that CON laws are allowing existing large hospitals to suppress competition.

Given the careful analysis and clear results in our forthcoming paper, our conclusion is that repealing CON laws would go a long way toward increasing access and competition in healthcare, including in rural areas.

The opioid overdose epidemic in the U.S. continues unabated. In 2021, more than 80,000 deaths were recorded involving opioids — nearly four times more than in 2010. More recent provisional data suggests that things have only gotten worse over the last few years.

While the causes of the crisis are still debated, the basic story we’ve all heard goes something like this: Unscrupulous pharmaceutical companies downplayed the addictive risks of new opioid products like Oxycodone. Doctors over-prescribed these drugs without fully understanding their side effects. At the same time, illicit opioids (especially fentanyl) proliferated on American streets as the increasingly porous Southern border gave cartels easy access to the U.S. market, and declining economic prospects for many working-class Americans created robust demand for substances to blunt physical and emotional pain.

Other possible factors have been less widely discussed. For example, how does the largest expansion of health insurance coverage since the 1960s — driven by changes to Medicaid eligibility under the Affordable Care Act — fit into this picture?

There are plausible arguments on both sides. With more Americans visiting the doctor, Medicaid expansion may have helped fuel the opioid epidemic by making the use (and misuse) of prescription opioids more accessible. Figure 1, which shows total U.S. opioid deaths by year, is certainly suggestive. In 2014 and 2015, just as most states were adding millions of people to the Medicaid rolls, opioid deaths began to increase sharply. Moreover, expansion states tend to have higher drug overdose deaths per capita than non-expansion states.

But each of these points isn’t quite as compelling as it first appears. The vast majority of drug overdoses involve illicitly manufactured substances, which the Medicaid program obviously does not supply. (Of course, it’s possible that some people were first prescribed opioids through Medicaid, developed an addiction, and later succumbed to illicit formulations.)

As for the coincidental timing of Medicaid expansion and the uptick in opioid deaths around the country, it’s not clear which way causality runs. It could be that state policymakers adopted Medicaid expansion in an attempt to address the deepening opioid crises in their states.

Some have argued that the opioid crisis would be even worse if Medicaid expansion hadn’t expanded access to substance abuse treatment. This view is supported by research showing that Medicaid expansion increased admissions to mental health facilities and Medicaid-reimbursed prescriptions for medications used to treat mental illness. There is also evidence that Medicaid expansion reduced self-reported rates of psychological distress among those who gained coverage, at least in the short-run.

Several studies have attempted to measure Medicaid expansion’s effect on drug overdose deaths by comparing trends between expansion and non-expansion states. Some find that expansion was associated with slight reductions in overdose mortality, particularly in counties that saw the largest gains in insurance coverage due to expansion. Other research fails to detect any significant relationship between Medicaid expansion and opioid deaths.

Overall, although there are reasons to worry that Medicaid expansion has had negative unintended consequences, there isn’t much evidence that worsening the opioid crisis is one of them.

In Covid in the Nursing Homes, Alex Tabarrok and I examined the pandemic in great detail, focusing particularly on policies that failed to protect nursing home residents. I would like to briefly review our discussion of COVID testing, and how new evidence in this area is pointing toward Rapid Antigen Testing as a missed opportunity.

We document how, instead of encouraging private test suppliers to enter the market, the Centers for Disease Control and Prevention first botched the creation of their own test, then prohibited private actors from using it, and when, on top of that, the Food and Drug Administration broke long-standing practice and issued an “emergency requirement” that lab-developed tests be pre-approved, the result inevitably was very limited testing. These cascading errors meant that by the time a facility had a positive test, the virus would already have spread throughout the facility, contributing to COVID’s devastation of nursing homes, especially in the early stages of the pandemic.

What did surprise us, as we worked on the paper, was the failure of point-of-care Rapid Antigen Testing to fix this:

“Unfortunately, requirements and point-of-care tests did not turn the tide on testing, even though two-thirds of nursing homes had test capability by the middle of September [2020. ...] It took until late November–December before nursing homes were running a million weekly antigen tests and, even then, they ran more of the slower, more expensive lab tests. Why weren’t the rapid antigen tests used much more frequently?”

I’m still not sure we have an answer to that question, but it increasingly seems like we were right to call this a missed opportunity. Our analysis found testing each resident one additional time each week for a year was associated with 1.54 fewer facility COVID deaths, an 18% reduction. In a paper published yesterday in the New England Journal of Medicine, McGarry, Gandhi, and Barnett look at the same question and find a very similar answer:

“the performance of one additional test per staff member per week was associated with a 30% reduction in resident Covid-19 cases and a 26% reduction in related resident deaths.”

This is a topic I’ll likely return to in more detail, but it seems increasingly clear Rapid Antigen Tests were underutilized. At $5 per test, the authors calculate it would only cost $13,000 to save a nursing home resident’s life, a bargain compared to most other public health interventions.

With the 2024 Republican presidential campaign being dominated by indictments, wars and inflation, healthcare issues have been relegated to the back seat of policy discussions. Most candidates’ stump speeches have been practically devoid of any mention of healthcare. But a few candidates have outlined specific healthcare policy goals. And one idea keeps coming up: repealing certificate-of-need (CON) laws, which require healthcare providers to obtain approval from a government board before expanding services in an area. Currently, 35 states have these laws on the books.

Former UN Ambassador Nikki Haley cited CON law repeal as one of her healthcare priorities in a debate last year, and she did so again in last week’s Iowa debate against Florida Governor Ron DeSantis:

“And then we're going to go and eliminate Certificate of Need in this country. I did that in South Carolina, as well. That basically says if you have a hospital here, you can't have another hospital for X number of miles. They do the same thing for surgical centers, for nursing homes. We're going to put competition back in health care so that health care is fighting for the patient. That way, services go up and costs go down.”

It may be the first time a prominent, major-party presidential contender has taken such direct — and forceful — aim at CON laws. And about time, too.

Although the president has limited ability to influence state CON laws, the fact that the issue is being raised in national debates shows how important it is for states to address bad policies lingering in most of the country.

It’s now clear that CON laws’ stated purpose — to contain costs in the healthcare sector by limiting the duplication of resources — is merely a smokescreen to secure political support for regulations that allow established providers to exclude potential rivals.

Maureen Ohlhausen, a former commissioner of the US Federal Trade Commission, has stated bluntly that CON laws “serve primarily, if not solely, to assist incumbents in fending off competition from new entrants.” Predictably, less competition yields higher prices and worse quality of care.

A substantial body of research documents the harmful impacts of CON laws. Last year, the Institute for Justice reviewed more than 120 academic papers on CON laws, spanning nearly five decades. Its analysis found that 89% of results in these studies showed that CON laws have negative or neutral effects, and that negative effects were five times more common than positive effects. “Many of these studies show that CON laws are bad for patients, bad for payors [sic], bad for improving access to care (including rural care), bad for vulnerable populations, bad for mortality rates for common conditions, and bad for healthcare innovation,” the authors wrote.

Moreover, reforming CON regulations has proven politically feasible — and not just in Republican strongholds like Haley’s South Carolina, where the current governor, Henry McMaster, signed legislation last year virtually eliminating CON restrictions in his state. Since 2016, CON laws have been lifted or pared back in Connecticut, West Virginia, Montana, North Carolina, Tennessee, and New Hampshire.

Now, as the Republican primary process shifts its focus to New Hampshire and South Carolina, candidates have a chance to highlight the benefits of CON law repeal in two states that have embraced reform — and been better off for it.

Lawmakers are again scrambling to stave off a partial government shutdown by negotiating several spending bills with just a week left on the clock. 

It’s frustrating to see the disproportionate focus given to issues like tax enforcement, border/immigration restrictions, and foreign aid, when anyone familiar with our budgetary outlook and simple arithmetic knows it’s ultimately a question of healthcare and social security. 

If you don’t want to take my word for it, here’s former President Obama, who at a White House healthcare summit in 2010, highlighted that:

“Almost all of the long-term deficit and debt that we face relates to the exploding costs of Medicare and Medicaid. Almost all of it. That is the single biggest driver of our federal deficit. And if we don’t get control over that we can’t get control over our federal budget.”

A year ago, my colleagues Veronique de Rugy and Elise Amez-Droz wrote “We can’t get the debt under control without changing the way we pay for healthcare.” While there have been positive developments, notably the Great Medicaid unwinding which has shrunk our largest safety net program by over 14 million people and is on track to save around $80 billion annually, the statement remains as true today as it was then. 

I won’t indent it all, but what follows is a lightly edited reprint of the ideas proposed by my colleagues last year, tailored to address the ongoing challenges in healthcare spending. Let us know what you think!

Solutions to Get Healthcare Spending Under Control

The inclusion of ideas in this list doesn’t constitute an endorsement but is rather meant to stimulate discussion about various proposals.

Medicaid

• Establish caps on federal spending for Medicaid — Options for Reducing the Deficit, 2023 to 2032 (Congressional Budget Office). The CBO estimates that this measure could yield savings of $501 billion to $871 billion between 2023 and 2032. Liam Sigaud suggests using a funding mechanism resembling the Children’s Health Insurance Program.

• Reduce federal Medicaid matching rates — Options for Reducing the Deficit, 2023 to 2032 (CBO). The CBO estimates that this measure could yield savings of $68 billion to $667 billion through 2032.

• Fully nationalize the financing and administration of Medicaid benefits — Beyond Bailout Federalism: The Case for Nationalizing Entitlements (Chris Pope, Manhattan Institute). According to Chris Pope, this proposal has the benefit of preventing states from tapping into unlimited amounts of federal reimbursement.

• Curb waste, fraud, and abuse — Curbing Waste, Fraud, and Abuse in Medicaid (Tara O’Neill Hayes, American Action Forum). Medicaid’s improper payment rate averaged 9.8% in 2016, and the CBO estimated that cutting improper payments would reduce the deficit by more than 11%.

• Tighten Medicaid eligibility for long-term care services — Taking the Strain Off Medicaid’s Long-Term Care Program (Chris Pope, Manhattan Institute). Doing so, along with setting up tax-advantaged saving options for long-term care, as proposed by Kofi Ampaabeng, would reduce Medicaid’s role in paying for long-term care.

Medicare

• Increase the premiums paid for Medicare Part B — Options for Reducing the Deficit, 2023 to 2032 (CBO). The CBO estimates that this measure could yield savings of $57 billion to $448 billion between 2023 and 2032.

• Reduce Medicare Advantage benchmarks — Options for Reducing the Deficit, 2023 to 2032 (CBO). The CBO estimates that this measure could yield savings of $392 billion by 2032.

• Adopt competitive bidding for Medicare Advantage — Implementing Competitive Bidding in the Medicare Program: An Expressway to Solvency (Rohini Chakravarthy, Gail Wilensky, and Brian Miller).

• Adopt strict payment site neutrality — How to Reduce Health Care Costs: Understanding the Cost of Health Care in America (David Hyman, Cato Institute).

• Eliminate the requirement that Medicare Part D plans cover all approved drugs in six protected classes (immunosuppressants, antidepressants, antipsychotics, anticonvulsants, antiretrovirals, and antineoplastics) — How to Reduce Health Care Costs: Understanding the Cost of Health Care in America (David Hyman, Cato Institute).

• Move drugs from Medicare Part B to Medicare Part D — How to Reduce Health Care Costs: Understanding the Cost of Health Care in America (David Hyman, Cato Institute).

Other healthcare-related items

• Reduce tax subsidies for employment-based health insurance — Options for Reducing the Deficit, 2023 to 2032 (CBO). The CBO estimates that this measure could yield savings of $500 billion to $893 billion between 2023 and 2032.

• Limit state taxes on healthcare providers — Options for Reducing the Deficit, 2023 to 2032 (CBO). State taxes on healthcare providers are responsible for inflated federal reimbursements to states for healthcare expenditures. See this Mercatus Center paper by Brian Blase for more information. The CBO estimates that this measure could yield savings of $41 billion to $526 billion by 2032.

• Charge Original Medicare beneficiaries a copay for home healthcare services — How to Cut $343 Billion from the Federal Budget (Brian Riedl, Heritage Foundation).

• Eliminate Title X Family Planning, the Maternal and Child Health Block Grant, the Health Professions Opportunity Grants, and the Rural Health Outreach and Flexibility Grants — How to Cut $343 Billion from the Federal Budget (Brian Riedl, Heritage Foundation).

• Move drugs from prescription-​only to over-​the-counter or behind-​the-​counter — How to Reduce Health Care Costs: Understanding the Cost of Health Care in America (David Hyman, Cato Institute).

Note: The post has been updated to reflect that the deadline that is fast approaching is for a partial government shutdown, and not the debt ceiling, as was originally stated.