I published this article on KevinMD on 1/9/25, but I want a record of it here as well since this is my central repository for all of my healthcare writing.
For regular followers of this blog, the principles discussed in that article are not new. But the framing for how I’m putting some of them together is new, so I hope it helps give people a more rounded understanding of the futility of many of CMS’s alternative payment model programs. Their outcomes are just what I predicted in my 2016 article.
It’s amazing how predictive a theory can be when it’s based on an accurate understanding of the principles at play.
After thinking and writing about healthcare for so long, I don’t often learn new things that turn out to be a big part of the puzzle. But this is one of them. I hope you can follow my thought flow as I explain.
First, I was thinking about something I’ve directly experienced a number of times. In talking to the upper-level managers of healthcare providers about their pricing, increasingly often they’re creating a list of cash-pay prices, which they use to determine how much to charge patients who either don’t have insurance or who, for whatever other reason, want to pay out of pocket for a service. But they’re almost never willing to have those price lists shared publicly.
Why?
Quick sidenote: The fact that providers are doing this so much more, and as pushed by some of the government transparency initiatives as well, is huge! Before, providers would use the chargemaster price when a patient wanted to self-pay. And since chargemaster prices aren’t even close to grounded in reality (for reasons that I’ve discussed before and won’t rehash here), it was a major ripoff if you didn’t purchase through insurance.
Anyway, the answer is that they worry that it will get them in trouble with insurers, presumably because if a cash price turns out to be less than the negotiated price with even one insurer, if the insurer finds out about it, the insurer will refuse to pay their previous (higher) price, so the provider will lose bigtime during the next round of contractual price updates.
Essentially, this means that providers’ attempts to create self-pay price lists (in an attempt not to screw over uninsured patients) could cost them a lot of revenue from insurers. So they keep them mostly secret.
And I suspect this is such an issue because of gag clauses in provider-insurer contracts. Gag clauses prevent either party from divulging the agreed-upon prices, which probably also indirectly prevent providers from being willing to share any prices publicly.
So that’s where my mind was at, thinking about this whole issue of price opacity and the fact that prices between private parties in healthcare are set through bargaining power-based negotiations between insurers and providers. And it led me to start thinking harder about the upsides and downsides of negotiated pricing versus standard pricing. Here is the definition I’m using for each of those terms:
Negotiated pricing is setting prices based on the relative bargaining power between the buyer and the seller.
Standard pricing is a seller setting their price wherever they want, and the buyer has a choice to buy from that seller or not.
The big realization I’m making–which also has big implications on why our healthcare system performs the way it does (as I’ll describe at the end of this post)–is that negotiated pricing kills a ton of value-sensitive decisions. Even worse, opaque negotiated pricing kills even more value-sensitive decisions.
The effect of something preventing a large chunk of value-sensitive decisions is that it weakens the link in that market between (1) delivering high-value products and services and (2) earning more profit. So the value delivered by that market no longer rises over time like it otherwise would because competitors are too busy competing over the new things (rather than value) that determine how much profit they will win. New things like market share, which empowers them to negotiate up prices. And we wonder why the healthcare market has rolled up (i.e., become consolidated into a small number of very large companies) like it has. How else can you earn more profits when prices are set mostly based on bargaining power-based negotiations?
Another side note: For those interested in business strategy, I’ve never heard anyone talk about this as a factor that pushes an industry to consolidate, but it’s probably a major factor in many industry roll-ups. Another factor, more well know, is if there are significant economies of scale to be had, that will also push an industry to consolidate.
So how does negotiated pricing kill value-sensitive decisions?
I’ve tried several times to explain this part, and I’m still failing to some degree, which means I haven’t pieced everything together well enough yet. So here’s my best shot, which might still be confusing.
When all the prices between insurers and providers are set through negotiated pricing, they aren’t as closely connected to what is truly valued in the market. This relates to the importance of prices in a free market and how they are the most important thing that conveys the value of something. Hayek wrote about this. So the prices don’t connect to what the true market price would be, which makes for an inefficient market when people are trying to determine and choose the highest-value options. There would still be a link between value and profit, it would just be a lot weaker. But, make it even worse by adding in the propensity for negotiated prices to be opaque, and now you’ve taken away the ability for individuals to know relative prices at all (assuming the prices are opaque to the decision maker), so you’ve completely killed value-sensitive decisions.
Related:
What I’m saying is that negotiated prices have caused two problems in our healthcare system: (1) The person making the decision on where they will receive care (i.e., the patient) no longer has any idea of the price of things, and they usually don’t care anyway because they usually pay the same amount regardless (through a flat co-pay, for example). (2) Sellers are unwilling to make any other prices public–including self-pay prices–for fear of it interfering with their gag clause or costing them money in the future when others demand that same price if it ends up being lower than a negotiated price.
The overall effect of shifting our healthcare system from its original standard pricing to negotiated pricing is that it eliminated a ton of value-sensitive decisions. And, once you hit a critical mass of non-value-sensitive decisions, the sometimes harsh incentive of the market for competitors to improve value (or else go out of business) gets significantly weaker.
So how did we come to have negotiated pricing in our healthcare system?
We had some amount of it even from the 1930s with Blue Cross and Blue Shield plans, although the details of these arrangements suggest that they didn’t get translated into taking away the ability of patients to identify the price of their different options and choose accordingly.
But what really caused the shift was the Health Maintenance Organization Act of 1973.
Here’s some context for that Act: Healthcare innovation was increasing how much providers could do for patients, and some of those new things were very expensive. Consequently, the price of insurance was rising fast enough that people were trying to come up with new ways to lower healthcare spending. The big idea that took hold was to try to reduce the number of unnecessary services provided by providers by having insurers review proposed services and approve or deny them. It makes sense–fee-for-service providers get paid more when they do more, so they have an incentive to do more things for patients, including things that could be of dubious value. The solution worked, to an extent–the insurance companies were able to prevent a lot of questionable-value services from being performed.
This caused problems, too. The usual big complaints about HMOs is that (1) they would inappropriately deny things and delay services and (2) they caused a ton more administrative costs. All this is true. But possibly the biggest drawback is that it started us down this road of shifting to negotiated pricing.
When insurance companies starting forming HMO plans, they needed to have a limited provider network so that they could restrict their beneficiaries from receiving care outside of an HMO contract. Limited provider networks based on contracts between the insurer and the providers opened up another possibility for insurers: If a big insurer has a lot of market power by virtue of insuring a large percentage of the people in the region, they could start requiring providers to accept lower payments from them as a condition of being included in the HMO network. The providers were often willing to agree to that because of the additional volume it would bring them. But the insurer would also require the provider not to tell anyone else how much less they were getting paid by that insurer (a gag clause), which helped the insurer in a lot of ways, such as by preventing its competitors from finding out just how little they were paying, thus preventing competitor insurance companies from shifting their negotiation anchor regarding what kind of prices they could expect to have to pay.
So, overall, what I’m saying here is that an unanticipated side effect of HMOs is that they led to our healthcare system shifting from standard pricing to negotiated pricing, which eliminated a large chunk of the value-sensitive decisions in the market. And negotiated pricing through limited-network insurer-provider contracts remains even after the most aggressive versions of HMOs have gone away.
Now, what’s the big implication from this that I referenced above?
What would you expect to see if an entire industry of a country loses a ton of value-sensitive decisions in a relatively short period of time? Especially if this happens through a big shift toward opaque prices for the people who are choosing between different competitors, prices would no longer be constrained by the market, so they will start rising rapidly. There would also be an impact on quality–it would stop rising as fast as it was rising before.
If you wanted to take a more formal analytical approach to this, you could use a difference-in-differences approach by looking at how that country’s results change over time (before and after the change that reduced value-sensitive decisions) and compare that to how other countries’ results change over that same time period. Specifically, you would want to use as comparators countries whose markets didn’t go through that same change.
I’m not going to do a formal analysis like that right now. But, as you take a look at this graph, keep in mind that even though the HMO law was signed in 1973, HMOs didn’t really gain a ton of uptake until the 1980s:
Here’s another one showing the same thing but with more countries. It just doesn’t go back as far in time:
I couldn’t find a longer-term graph to show, but I’ve seen them many times that show that the U.S. was middle of the pack for decades before 1980, and then something changed in the 1980s, and I could never figure out why.
And for those of you who think that it’s because the U.S. started becoming way more wealthy than others around that time, take a look at this:
It looks like our rise in GDP per capita probably has some impact on why our spending increased as much as it has, but we don’t see Norway and the others becoming outliers like we have, so that implies the GDP issue is not the main factor here.
And on the topic of quality, it’s harder to compare country to country, especially when we don’t have universal coverage, so I won’t try to get into that in detail. Even looking at life expectancy is unreliable because a country’s healthcare system has such a small impact on that compared to the country’s culture surrounding diet and exercise.
So, that’s where I’ll leave it. I know there’s some uncertainty in this, but I am fairly confident that I finally have the answer to why the U.S. diverged in its healthcare spending in the 1980s–because we shifted from standard pricing to negotiated pricing as a result of HMOs.
And I have new evidence that is very consistent with my theory about how value-sensitive decisions are central to determining how much the value of goods and services provided by a market increase over time.
Update: I haven’t yet looked into how much the Price Transparency Act is likely to help with all of this. I’ll have to weigh in after I’ve analyzed it more thoroughly.
The background for this new bundled payments initiative is that CMS has been trying out bundled payments in a voluntary capacity for a while, and they have lowered costs only slightly but paid out a lot more than that in bonuses.
If I keep talking about how important bundled payments are, does that mean this is evidence against what I’ve been saying?
That’s an important question. We tend to subconsciously ignore or justify or minimize or forget information that is dissonant with our worldview, so it’s something I have to be deliberate about avoiding.
The only way to sustainably improve the value delivered by our healthcare system is by increasing value-sensitive decisions, which will enable higher-value providers and insurers to win more market power, which translates into more profit.
And bundled payments don’t achieve that by themselves.
For a value-sensitive decision to be made, you need patients to have (1) multiple options, (2) both the price up front and also some information about quality, and (3) they have to pay a little more if they choose a more-expensive provider or less if they choose a less-expansive provider.
So, no, I wouldn’t expect bundled payments to have a big impact immediately and in isolation. But they’re a super important piece of the solution to start integrating permanently into our healthcare system, so I hope CMS keeps up these efforts regardless of the short-term costs and benefits of these programs!
And one last point on the mandatory nature of this new bundled payment program, which is called the Transforming Episode Accountability Model (TEAM). (Man that’s a name that clearly started with the acronym.) I am not totally convinced that mandating hospitals to join a program is a good thing, but it may offer a track to solving one of the other challenges providers always face when they join alternative payment models: non-uniform incentives.
When a provider is rewarded differently from each payer, they don’t have the uniform incentives necessary to really optimize toward any single set of incentives. So even better than mandatory participation in a single bundled payment model would be to get all the insurers together in a region and have them all offer the same model to the providers, and then they have the choice whether to join or not.
And if you could also get the providers to report useful quality metrics and get the insurers to implement some differential cost-sharing requirements for those same care episodes, you’d be well on your way to seeing a region’s value start to dramatically shift upward for those specific services.
Basically, I’m describing the pilot program I designed for Utah as an intern at the Department of Health, but our grant proposal didn’t get funded. That was a heartbreaker. I do think that Departments of Health have a role here as a convener of these various parties to help them solve some collective action problems (in this case, getting uniform incentives for providers when implementing alternative payment models).
These sort of changes are so implementable if only the people running CMS knew about them.
Oh, and for those who have been following this blog for a while, here’s a quick update: I’m still working on revising my Theory of Money series. You’ll know it’s done when I release the summary post that gives an overview of each part in the series.
I write regularly about the need for patients to be able to make “value-sensitive decisions.” It means they make decisions about where they receive care (and also which insurance plans they choose) while considering both price and quality. One important step to helping patients do this is enabling them to know beforehand what the total price of the care episode will be.
That’s where bundled payments comes in. Their most important function is to give patients an apples-to-apples comparison of the total price of the care episode because they set a single price that includes all the things that might be needed in the care episode.
But when you start trying to turn bundled payments into pay for performance programs, that’s when they should be called bungled payments.
Which brings me to Medicare’s bundled payment models. There was a great article in JAMA recently by Joynt Maddox, Shashikumar, and Ryan, entitled Medicare’s Payment Models–Progress and Pitfalls.
Here’s an issue they wrote about regarding benchmarks: “Continuing to base benchmarks on historical costs will lead to a feedback loop disadvantaging participants with low spending. Those who reduce spending will have lower historical spending in future years and will receive lower benchmarks. They may find it more difficult to meet progressively lower benchmarks and will thus eventually pay penalties, despite being efficient.”
This sounds like what I wrote a few weeks ago: “Sure, this incentive has gotten us better quality for more money, and yeah eventually we’ll probably have higher quality overall, but it’s going to be at the cost of a lot of consternation of providers as we repeatedly take away their quality bonuses when we raise standards. Overall, this quality bonuses idea is just a frustrating and generally ineffective way to improve value.”
They suggest defining an “efficiency floor” that exempts anyone who meets that threshold from penalties. That’s reasonable for sure, but I suggest getting rid of the bonuses/penalties thing altogether.
Another thing they wrote in the article: “Second, the current benchmark method permits payout of large reward, offsetting any savings that the models achieve for CMS. . . . While the BPCI-A, OCM, and CJR payment models are associated with reductions in clinical spending, the bonuses these programs have paid have far exceeded the spending reductions they have incented.”
This sounds, again, like what I wrote a few weeks ago in the same blog post: “If our goal is to improve value, what we’ve just done is taken the higher-value providers and increased their price, which means their value has dropped back down to everyone else’s.”
They suggest creating a “stop-gain” provision to limit potentially excessive bonus payouts. That’s reasonable for sure, but I suggest getting rid of the bonuses/penalties thing altogether.
We need, instead, to do a few things all at the same time. First, providers need the freedom to offer lower prices. Second, prospective patients need to be able to easily find those prices (and, preferably, patient-relevant quality information as well!). Third, patient’s insurance plans need to be designed such that patients end up paying less out of pocket when they choose a lower-priced provider.
If that could all happen, more people would start choosing the providers who offer better value. And what does that mean? It means the better-value providers would get more patients and more profit! And the lower-value providers would get less profit. So what we have is the same general outcome of rewarding the better providers and penalizing the worse ones, but instead of trying to do it administratively with all sorts of inefficiencies and distortions (all while not enabling more people to receive higher-value care), we have instead brought to bear on this challenge the dispersed preference information from the market. It will very effectively reward the ones who are truly worthy of being rewarded, penalize the ones who have lower value, AND more patients will get higher-value care immediately, which I would argue is our overall goal with these programs in the first place (you know, “value-based care“).
Last week, I wrote about how quality metrics are misused by healthcare reformers. They’re almost exclusively tied to bonuses or penalties from insurers. In other words, they’re used to increase or decrease the price providers get paid. This is a form of administrative pricing, which is a super economically inefficient way to set prices. And I proposed the alternative use of quality metrics–to help patients choose higher-value providers.
We give people quality metrics and they seem to generally do a good job shopping for the best value in pretty much every other industry, which drives competition over value. So why do we fail so miserably in healthcare?
The first problem is that healthcare is missing the thing that motives people to shop around for the best value: their money is on the line. I wrote about this a couple weeks ago. We need people to pay a little more if they choose a higher-priced provider. But when prices are opaque or unknowable beforehand, or when their insurance plan makes them pay the same regardless of the provider’s price (or if the insurance plan is complex enough that the patient doesn’t understand that they’ll have to pay more if they choose a higher-priced provider), people don’t perceive that their money is on the line. In that last sentence, I just listed four issues preventing people from actually caring what the price is!
And then there are the issues of having only one option (like in a rural area) and non-shoppable services (like during emergencies) and non-shopping-when-you’re-already-established-with-a-provider. Yeah, there are a lot of reasons people don’t shop for prices in healthcare! But in spite of all that, there are some good studies that show that people will actually shop for services when all the stars align.
I know that even if people have a hard time knowing prices beforehand, they theoretically could still shop just as vigorously for the highest quality.
But I think there’s something that happens when people can’t shop for price that sorta stops them from thinking about shopping for quality too. I haven’t seen any studies that prove this, but I suspect it’s a thing.
So let’s talk about the people who say, “Well if I don’t know what I’m going to pay, I might as well try to find the best quality option.” They use a variety of sources since there isn’t one single well-known and useful quality source out there. Usually they rely on recommendations from their doctor or their friends and family. If that person had a good experience, that’s a reliable indicator of quality, right?
Or maybe they decide to be brave and try Googling quality metrics. They’ll find something, certainly. But chances are they’ll find quality metrics that aren’t super relevant to what they actually care about. For example, maybe they’ll discover Medicare’s Care Compare website. What does 3 stars even mean? Even drilling down, how useful is it to know that a hospital’s safety is “below the national average” in 2 out of 8 metrics? How does that get weighed against a high recommendation of the hospital from a family member? Or, is that quality rating ignored because the hospital’s lobby is spacious and it advertises meals prepared by well-known chefs?
Compare the relative uselessness of those quality metrics to the example of Seattle’s Virginia Mason Health System when they were redesigning their low-back pain care pathway. They figured out that people care most about how soon they can get back to work (it’s expensive to live in Seattle, if you didn’t know) and, among other changes, made same-day appointments available. This was the quality metric people cared about, and their low-back-pain market share doubled.
After reading all these barriers to people shopping for the best value in healthcare, I hope you can see that (1) this problem is perfectly explainable and (2) it’s totally fixable. Can someone please tell the Medicare administrators that most of their current efforts at “value-based purchasing” are going to be close to useless? And tell them to look at getting rid of some of these barriers to patients choosing high-value providers instead.
Last week I wrote about how cost sharing is misunderstood. This week I’ll continue in the same vein and talk about the same thing but related to quality measurement and reporting.
Quality measurement and reporting is becoming a pretty big thing. Just look at all the different Medicare programs (the big ones being MIPS and APMs) trying to achieve this thing they call “value-based purchasing” (which, in their estimation, seems to mean pretty much anything other than straight fee-for-service reimbursements). These programs involve lots of quality reporting requirements, and then compensation is directly tied to those quality metrics, usually through bonuses for high performers.
But this is the wrong way to use quality metrics.
Before I explain why I believe this is the wrong way, I need to clarify what my goal is with healthcare reform. I am interested in improving the value (Value = Quality / Price) our healthcare system delivers.
This is usually the part where people say, “If you want to improve value, you’ll make a lot more progress by preventing people from getting sick in the first place, so you should focus your efforts on public health initiatives!” Or, others will say, “You need to work on getting more people access to the healthcare system. Solve this issue first, then you can figure out how to improve the system’s value!”
I agree that those are very important issues. And I believe we need to work on both of them as well as this one of improving the value the system delivers at the same time. So I’ll keep writing about these things and figuring out how to fix our healthcare system in all these ways.
Anyway, let’s think about what is going on when a provider does a great job and has really high quality metrics and gets paid bonuses (say, 5% or so on top of what Medicare would otherwise have paid them) as a reward.
If our goal is to improve value, what we’ve just done is taken the higher-value providers and increased their price, which means their value has dropped back down to everyone else’s. Sure, this incentive has gotten us better quality for more money, and yeah eventually we’ll probably have higher quality overall, but it’s going to be at the cost of a lot of consternation of providers as we repeatedly take away their quality bonuses when we raise standards. Overall, this quality bonuses idea is just a frustrating and generally ineffective way to improve value. But I understand why it’s so popular–it’s an obvious way to encourage value.
Is there an alternative? Of course. We need to find some way to reward providers for providing extra quality. But how we do that, that’s the question.
What if we could find a way to get more patients to choose those higher-value providers? This would reward them with more profit, and now the providers with lower value are losing out on money because they’re losing market share. There would be no administrators at fault when a provider makes less money. No top-down program decisions to blame. PLUS, more patients would be getting higher-quality care immediately. That’s a pretty great system.
So, instead of using quality reporting to give administratively determined bonuses, we need to use them to help patients identify the best-quality providers so they can choose to receive care from them. This would involve measuring very different quality metrics–ones that patients actually care about.
Can we do it? I believe we can. There’s a lot to how we could make this happen, and I’ll talk more about that next week.
This last week, I spoke with someone who works at an insurer. When I asked if it’s relatively easy for them to identify the pharmacy benefit manager (PBM) that’s offering the best deal, they said that it is, with caveats.
Quick sidenote: I’m finding kind people to answer my probing questions about PBMs, but I’ve also faced a fair amount of hesitancy in general. I think it’s because they worry what is going to happen with the information they are giving. I’m committed to being as transparent and unbiased as possible about the information I receive, and I’m equally committed to not disclosing any of my sources. So I guess that means I cannot prove the reliability of the information I’m sharing, but it’s worth it as long as I maintain access to good sources to help me understand this stuff!
Remember that for a market to be competitive it needs (1) multiple competitors, (2) the customers need to be able to identify the value (price and quality) of each competitor, and (3) the customers need incentives to choose the highest-value option.
The PBM market fulfills all these criteria pretty well. There are plenty of competitors (three big ones, several midsize ones, and lots of smaller ones). So, when an insurer submits a request for proposal (RFP), they will get multiple offers. Identifying the value of the proposals received is doable, if a bit tricky, as discussed below. And the insurer has incentives to choose the highest-value option–getting a great comprehensive formulary with the desirable meds makes for happy members, and lowering the costs goes to their bottom line (assuming there are no annoying medical loss requirement issues).
Let’s talk about the challenges that come into play when they try to identify the highest-value option. It’s actually pretty straightforward–these are incredibly complex contracts, to the point that regular healthcare consultants are not deeply specialized enough. And PBMs leverage that by trying to define things in ways that save them money. To the extent that, if an insurer wants to identify the best PBM proposal, they will probably need a consultant that specializes in helping insurers contract with PBMs. They need the help of someone who knows all the PBMs’ tricks.
I won’t even get into all the complexities of those contracts, partly because I don’t know many of them and partly because those details don’t change the big-picture incentives I’m talking about.
But the good news is that, with the right knowledge/assistance, insurers are able to make value-sensitive decisions in the PBM market! In fact, apparently many insurers submit an RFP every few years to make sure what they’re getting from their current PBM is still competitive, otherwise they’re probably leaving money on the table.
In response to some of my recent posts on pharmacy benefit managers (PBMs), as well as my overall interest lately in understanding the drug market better, someone asked what a value-based PBM would look like. Interesting question!
When analyzing healthcare from a “value versus volume” perspective, realizing exactly what we mean by that is an important starting point.
Consider what we mean when we say that healthcare providers are “rewarded for volume.” This is usually interpreted to mean healthcare providers are paid in a fee-for-service way–they deliver a service, they get paid. Which means they make more money when they deliver more services, so the incentive is to deliver as many services as possible.
And when we say that providers are “rewarded for value,” this is usually referring to some form of capitation, which means they get paid per head (that’s where the capit part of capitation comes from). In other words, providers would, for example, get paid a monthly fee for every patient whose care they are responsible for. Which means they make more money when they deliver fewer services (which, theoretically, happens when they are doing the right things to prevent their patients from getting sick).
These two methods of reimbursement are seen as “good” and “bad.” Capitation and its variations have good incentives (to prevent illness) and fee for service has bad incentives.
But they’re not opposites, like two sides of a coin or something. They are actually two different ends of a single spectrum. That spectrum is the “breadth of products/services sold as a single unit” spectrum. (I should come up with a better name for it.)
At one end of the spectrum, you have people buying very narrowly defined things. Like if a hospital really did send you a bill for every single nursing task and bandage and bag of saline and tablet of acetaminophen you received while you were there. This is the essence of fee for service–buying narrowly defined things. A doctor visit here, a procedure there.
The other end of the spectrum is buying very broadly defined things. Like paying a healthcare organization an annual fee for covering every single healthcare need that you could possibly have during that year, all inclusive. Every surgery and cancer treatment and emergency department visit etc. would be included.
I’ve written about all this before (way back in 2013!), but the way to figure out where on the spectrum the service should sit (i.e., how broadly defined the product/service should be) is to think about it from the patient’s perspective to see what “job” they want done that it’s fulfilling for them.
The easy example is if someone needs a hip replacement, let’s say they’ve tried all the conservative measures and now their job is simply to get their hip replaced and then recover/rehab from that. So why would they pay separately for the surgeon’s time, the OR time, the anesthesiologist’s time, the medications administered, the hardware used, the physical therapy appointments, the pre-op and post-op appointments, etc., when they could just pay a single lump sum to get their job fulfilled?
When we buy a plane ticket, we don’t pay a separate bill for the airplane depreciation, the fuel, the pilot’s time, the flight attendant’s time, the snack, etc., right? No, we just pay for the single plane ticket that includes all the products and services that go into getting us from point A to point B.
Using that principle of identifying the job to be done and then defining the service as broadly as is necessary to allow the patient to pay a single price for getting that job fulfilled will allow anyone to determine where on the “breadth of products/services sold as a single unit” spectrum anything in the world should sit.
So what about the drug market?
Much of the time, we know pretty well how long we’ll be on a medication. If it’s an asthma med, such as an inhaled corticosteroid, usually the patient will be on it for years or decades, so just knowing how much it costs per month is probably the right breadth of services.
Or, if it’s not a chronic medication, such as a cure for hepatitis C, figuring out the total cost of your direct-acting antiviral regimen is pretty easy if you at least know how much each pill will cost you and how many days (weeks) your treatment course will last.
My point is that the drug market, even though you’re typically buying either a short course of pills or a monthly allotment of them, is already “value based” because the breadth of products is attuned to the job you have for the medication (“keep my asthma at bay for 1 month,” or, “cure my hepatitis C”).
How do PBMs fit into all this?
Well, they’re middlemen. As far as I can tell, even though they’re the ones making the formularies, they aren’t really doing anything to actively shift the breadth of products sold one way or another, which is good because it already seems to be sitting on the spectrum in a good place.
Is there a role for including medications in capitated arrangements so that patients’ diabetes and hypertension and heart failure meds are all included in their annual or monthly fee? I guess that’s possible–it would encourage providers to choose cheaper meds, and it would decrease financially motivated medication nonadherence. So maybe PBMs would be involved in coordinating those efforts.
Ultimately, the big improvements that will change the drug market aren’t so much going to come from optimizations in the volume versus value space, but rather they will come from increasing competition and value-sensitive decisions. And maybe from limiting the degree to which PBMs distort the market? But I’m still figuring that one out.
In the “I’m still Confused by PBMs” article, I went through the step by step process of PBMs coming into existence, which they did by filling a need in the market for companies to help consumers pay the correct copay right up front when they buy a medicine.
This time, I want to think more about how they went from that simple software solution (integrating an insurer’s formulary into the pharmacy’s software system so that it can spit out the right out-of-pocket price on the cash register when a patient buys a medicine) to being seen by many as the “shady middlemen.”
The starting point is the position they found themselves in the market. They were responsible for creating formularies for several different insurers. Their goal, assuming the PBM industry is competitive, is to help insurers have the most generous formulary for the cheapest.
I bet at some point, some young upstart working for a PBM had an idea . . .
Young upstart: “If we want to offer a cheaper formulary than our competitor, why don’t we try to negotiate directly with drug manufacturers to lower the costs of the drugs?”
Old manager, feeling superior: “But how are we going to do that? You don’t understand that drug manufacturers only negotiate with pharmacy wholesalers. The wholesalers are the only party to which manufacturers sell their drugs. Nobody else buys directly from the manufacturers, so no one else can negotiate with manufacturers.” And then, with a taunting eyebrow raise, “Unless you are suggesting something radical, like that we start backward integrating to act as drug wholesalers as well?”
Young upstart, undaunted: “Not at all. How about this. Why don’t we pay a visit to a drug manufacturer that is selling a medication that has several competitors in the same drug category and make them an offer they can’t refuse. We could tell them we’ll put their medication in the lowest copay tier and all the other medications in that same category will still be in the middle copay tier. They will sell way more of their medicine and make a lot more money. But, in return, they have to pay us a “rebate” for every transaction of their medication that we process. They still make more money because they’re selling so much more of their medication, and we get some of it.”
Old manager, interest piqued: “And how is this going to allow us to offer a cheaper formulary than our competitors?”
Young upstart, gaining momentum: “We’ll simply charge less for insurers to use our formulary. Sure, they’ll have to pay a slightly larger share of the total cost of that specific medicine, but the lower price we offer them will more than make up for that. And the best part is, everyone wins! The drug manufacturer wins by increasing profit, the patient wins by paying a lower copay, the insurer wins by getting a cheaper overall formulary, and we win because we keep some of the rebate!”
Old manager, ever skeptical: “If everyone wins, then where is the money coming from?”
Young upstart: “The money comes from the other drug manufacturers, whose market share goes down. That profit that they’re losing is being divvied up among (1) us, (2) the drug manufacturer we’re contracting with, and (3) the insurers using our cheaper formularies, some of which will be passed on to patients.”
Old manager: “Ok, that makes sense, but this sounds too good to be true. You haven’t mentioned pharmacies yet–how would this impact them?”
Young upstart: “I was hoping you’d ask. This won’t impact pharmacies at all. They won’t even know about this transaction between us and the drug manufacturer. As far as they’re concerned, all they see is that they’re still getting paid the negotiated price for the medication, it’s just that patients are paying less and the insurer is paying more.”
Old manager, nodding: “So the patient pays less but the insurer pays more. Yet the insurer is saving money overall because our formulary is cheaper enough to more than compensate for that.”
Young upstart: “Exactly.”
And that is what I imagine to be the start of PBMs negotiating “kickbacks” with drug manufacturers. It was all in the name of PBMs being able to offer lower-priced formularies to insurers by orchestrating a way to help some drug manufacturers sell more drugs and get rebates/kickbacks/volume discounts in return.
This surely boosted the profitability of the PBMs that started doing it, which, when others heard about it, started doing the same thing.
Eventually, every drug manufacturer started paying some kind of rebate to PBMs, which means it became a zero sum endeavor overall for manufacturers because the net effect of having a higher market share through a specific PBM but a lower market through the others that made deals with their competitors means that they end up with essentially the same market share, the only difference being that now they are paying money to PBMs to avoid losing that market share.
Shady middlemen indeed. But I can’t blame them for doing it–this is what capitalism and competition is all about. It’s just a market failure that this specific strategy turns out to be a cost-increaser in the market.
Ever since the start of residency, the Journal of the American Medical Association (JAMA) has been delivered to my mailbox without me ever subscribing to it. They keep threatening to stop sending them if I don’t pay for a subscription, but I keep finding the journals in my mailbox. Usually I will glance at the titles of the articles and read the ones I find interesting. That is why, this week, I am writing about what I read in the viewpoint article, Reducing Low-Value Care and Improving Health Care Value, by Drs. Allison Oakes and Thomas Radomski.
They start the article by talking about how there’s a lot of low-value care delivered in U.S. healthcare, even in the absence of financial incentives for delivering more care. They cite some studies from places like the VA system and the Alberta, Canada, system (my home province!), showing how they, too, deliver lots of low-value care. This is their great encapsulation of that important insight: “The provision of low-value care when financial incentives are not present suggests that there are other motivating forces that contribute to overuse. . . .”
Next, they talk about regional variation, saying that from one region to another, there are “systemic differences in care delivery.” But how could doctors act so differently from one region to another? Their answer, at least in part, is that “organizational culture influences patterns of low-value service use. Individual organizations have distinct overuse profiles.” I like that phrase: distinct overuse profiles.
And it’s true. Very true. I’ve worked in the midwest, the northwest, and the mountain west, and I see it. For example, almost no one ordered blood ammonia levels at one hospital, and at another it’s an almost expected part of any workup of confusion, even if the patient has no liver history.
Another example: Just today a colleague was telling me about how if she ever ordered a certain kind of fluids (LR), she would get multiple phone calls checking to see if she’d really meant to order that. Evidence is fairly convincing these days that LR is usually better than normal saline, but that hadn’t caught on at her old hospital. At her new hospital, there’s literally a pop-up warning for anyone who tries to order normal saline that says, “LR is better.” And it offers to switch the order to LR for you.
Another example: At my current hospital, I very frequently see a urinalysis ordered on patients who presented to the ED without any complaints that would make me suspect a UTI. And since urinalyses are commonly falsely positive, those questionable urinalysis orders frequently lead to questionable antibiotic administrations. At other hospitals, the ED physicians’ culture is to have a much higher threshold for ordering urinalyses, and even when they are ordered, the likelihood of treating “asymptomatic bacteruria” is much lower.
Another example: The likelihood of the radiologist reading a chest x-ray as having an opacity that could be pneumonia seems to be different from hospital to hospital. And even if the patient hasn’t had any other symptoms of pneumonia, if they have any sort of respiratory complaint and the chest x-ray report says possible opacity “consider pneumonia,” it seems they always end up getting admitted for pneumonia and started on antibiotics.
These are just a few examples of how the practice of clinical medicine is so different from facility to facility in 1,000 tiny ways. And I understand why it happens.
In residency, your practice patterns are being strongly shaped by what your attendings do. But they’re also shaped, to a large degree, by the personal studying you’re doing and by presentations given at noon conferences and morning reports, where the presenter has spent a lot of time reviewing the newest evidence on the topic. There is a ton of active learning, and your connection to the newest evidence is fairly strong. Although, possibly as a side-effect of this, you also seem to add rare diagnoses to your differential more often, and this probably leads to more low-yield testing.
After residency, working as a regular attending not affiliated with a residency program, the focus is very different. The overall goal of practicing medicine is the same–delivering great care for patients–but there isn’t nearly as much active talk about ways your group might be practicing low-value care. (Instead, ensuring adequate coding and documentation dominates the discussion topics.) The connection to the newest evidence is a lot weaker. And there seems to be more of a focus on avoiding malpractice, which leads to having a lower threshold for ordering tests and scans for common diseases, which also counts as low-yield testing when the diagnosis in question is unlikely in that given scenario.
Individual physicians will still learn new things through personal study, but it’s an uphill battle to justify doing something different than your colleagues. So as you take over your colleagues’ lists of patients when you come on service, and as you take over newly admitted patients from the night before, you are frequently seeing and slowly being influenced by how others in your group are practicing medicine. The practice patterns naturally homogenize. I was once asked by a colleague to send an email to my hospitalist group about the evidence backing up something that I was doing because it was different than the other hospitalists.
This is how particular practice patterns spread and homogenize throughout an organization. Combine those effects with the other active pushes for practice pattern changes that come through leadership communications and EMR integrations (like the “LR is better” pop-up warning), and distinct overuse profiles start to make a lot of sense.
All these cultural factors that lead to systemic but regionally different low-value care delivery will need more than simply top-down Medicare reimbursement changes.
I stand by my solution that, if value-sensitive decisions in healthcare became more widespread, the financial incentives for decreasing low-value care would not only change the macro incentives, but they would be strong enough to induce healthcare managers/clinical leaders to go about finding creative and effective ways to make the organizational changes necessary overcome the cultural factors I’ve talked about in this post.
Clear Thinking on Healthcare 