Inside the wild universe of prescription drug markups

Over the past couple weeks, we’ve been reflecting more on how neat Bloomberg’s Ohio Medicaid markup bubble chart was in their recent article, “The Secret Drug Pricing System Middlemen use to Rake in Millions” (Figure 1).  We’re kicking ourselves now that we neglected to delve deeper into this in our report, “Bloomberg Puts Drug Pricing ‘Markups’ on the Map”. But like fine wine, good visualizations only get better with time, so it took us a few weeks to fully realize the possibilities that such data analysis could open up.

Figure 1
Source: Bloomberg

Bloomberg’s excellent visualization does, however, leave a few open questions. What do other state Medicaid managed care programs look like? How do they compare to state fee-for-service programs?  What do all states look like? To answer these questions, we set out to build a new visualization dashboard to compare drug markups between state Medicaid programs. The finished product is embedded below – we call it the “Medicaid Markup Universe” (because it looks very celestial).  You can also access it through the 46brooklyn Visualizations page, along with instructions on how to the use the new dashboard.

One very important difference between our visualization (a.k.a. “viz”) and Bloomberg’s is that we chose to swap the fields that are represented by the size and the color of the bubbles. In other words, in our viz, the size of the bubbles represent the markup per prescription, while the color represents the number of prescriptions. In Bloomberg’s viz, it’s the exact opposite.  

The great part about what we do is that there is no right or wrong way to create a visualization.  Each viz tells a story, and small changes to the design of a viz, which may on the surface seem inconsequential, are often necessary to unearth completely different learnings from the data.  This is exactly what happened here, and it resulted in a new way for us to explore the staggering variability in markup between drugs.

There is a disturbingly large difference in markup across generic drugs in managed care programs

The first observation that immediately jumps out at us when looking at the bubble chart for Medicaid managed care is the disparity between the size of the bubbles (where each bubble represents the average markup on one generic NDC Description - e.g. Aripiprazole 5 MG Tab). Tableau makes this easy for us to see, concentrating the largest bubbles/markup in the center of the “universe” with the bubbles/markup progressively getting smaller as they approach the perimeter of the universe. Simply put: the bigger the bubble, the bigger the markup.

As shown in Figure 2, in the center of the “universe,” the national average Medicaid managed care markup can reach as high as $3,200 per prescription for generic Gleevec (Imatinib Mesylate 400 MG Tab). As “markup” is essentially the amount that a PBM and pharmacy split, in the case of Imatinib, that $3,200/prescription is an estimate of the combined “take” for the PBM and pharmacy supply chain components.

Figure 2
Source: 46brooklyn Research

In stark contrast to Imatinib, Figure 3 shows that the outer edge of the visualization is chock full of mature generics such as Amoxicillin 500 MG Capsule, which carries a markup of just over $2 per prescription.

Figure 3
Source: 46brooklyn Research

There is a valid argument that the supply chain should be compensated different sums for the delivery of these two drugs (although 1,600 times more is a bit hard to defend in our view). One is an expensive blockbuster oral chemotherapy generic, while the other is an inexpensive antibiotic that has been around since 1979. But what about statins? Let’s look at generic Crestor (Rosuvastatin) versus generic Lipitor (Atorvastatin).  These two therapeutic equivalent drugs should have similar markups, right? Wrong.

As shown in Figures 4 and 5, the markup for the most commonly dispensed strength of Rosuvastatin (20 MG) was $25.18 per tablet in Q1 2018, while the markup for the most commonly dispensed strength of Atorvastatin (40 MG) was $5.67 per tablet in Q1 2018. So the supply chain is receiving nearly 4.5 times more margin to dispense one statin over another. Or viewed through a different lens, the state is getting charged 4.5 times more for one statin over another.

Figure 4
Source: 46brooklyn Research

Figure 5
Source: 46brooklyn Research

Spread-based contracts allow PBMs to set drug prices

The logical next question is, how and why is this happening? If the service that the supply chain renders is essentially the same, how can one product be so much more profitable than a comparable alternative? The answer is that when a managed care organization (or for that matter, any payer) enters into a traditional “spread” contract with its pharmacy benefit manager (PBM) the PBM graduates from being a claims adjudicator for generic drugs to actually setting the price for individual generic claims.

This is a substantial promotion!  It is akin to Visa being granted the ability to set prices of individual products that you are buying at Target, and then charging 4.5 times more for a bottle of Pepsi than a bottle of Coke.  This may sound like an outlandish example because we are all used to dealing with a highly efficient retail marketplace, but it’s unfortunately not an exaggeration of how the prescription drug supply chain works.

The ability for PBMs to set prices is interesting considering the industry’s rhetoric against drug manufacturers. Pot, meet kettle:

Of course, once a payer is in a spread contract, pricing lists are set by the PBM and are typically confidential. So, the rationale on why certain drugs are set at premium margins versus other drugs is hidden to others.

The price a payer is charged for any given generic claim may be close to the actual drug cost. Or it may not.  The amount the PBM pays the provider/pharmacy may cover the drug cost and/or pharmacy operating costs. Or it may not. The PBM has complete flexibility to set all of these variables for each claim, so long that it satisfies its annual financial commitments to the payer, which as we have written, are most likely tied to an outdated and stale generic benchmark called average wholesale price (AWP) or set arbitrarily via subjective PBM-generated maximum allowable cost (MAC) prices.

The difference in drug markups creates warped supply chain incentives

These pricing tactics are exactly how Ohio Medicaid managed care PBMs now famously raked in $208 million in spread off of generic drug claims between Q2 2017 and Q1 2018. But this problem runs deeper than just spread – arbitrary generic pricing creates warped incentives through the supply chain to dispense certain drugs over others, and as a result, serve some patients over others.

This is deeply troubling to us.  Let’s return to our Amoxicillin and Imatinib examples, and pretend that the pharmacy is actually receiving the entirety of the markup (which would only be the case in a non-traditional full pass-through PBM contract).  Based on the price set by the PBM, the pharmacy would have to fill 1,600 Amoxicillin prescriptions to generate the same gross margin of just one Imatinib prescription. The pharmacy is a “price taker” – it has no control over this math.  It is just left to manage its business to maximize its profit under the pricing terms set by the PBM. And this math does not favor dispensing Amoxicillin.

That’s also a problem if you are a patient that is on Lisinopril, Omeprazole, Ranitidine, Hydrochlorothiazide, Azithromycin, or any of the other 891 generic drugs whose Q1 2018 nationwide average markup was set below $5 per prescription.  The pricing structure that has been put in place by the PBM is valuing the treatment of you and your condition at less than half the margin that your pharmacy needs to cover its typical operating costs. Meanwhile, the PBM may value the treatment of other people with different ailments at multiple times the actual cost to dispense their medications.  This is patently wrong, in our view. Patients are patients, and they deserve a system that encourages the same level of service from their pharmacy, PBM, insurer, or anyone else, regardless of their ailment. Warped incentives that undermine this goal have no place in our supply chain.

Fee-for-service programs have done a much better job eliminating warped incentives

Thanks to recent CMS policy changes on state Medicaid fee-for-service programs, most states are now using objective drug pricing benchmarks that represent actual acquisition costs (AAC) and objectively-defined dispensing fees. Ohio is one of the many states that have adopted a full pass-through NADAC-plus-dispensing-fee pharmacy reimbursement model in its fee for service program, but have continued to allow the PBMs to set markups in managed care.  As shown in Figure 6, the contrast between the two models is stark. Medicaid managed care shows large discrepancies between the size of bubbles from the inside of the “universe” to the outer edge. Medicaid fee-for-service shows almost no discrepancy between the markup across the generic universe.

Figure 6
Source: 46brooklyn Research

In other words, the fee for service model has largely removed the economic incentive to dispense one drug over another, leaving only the incentive to serve more patients.  And how do you serve more patients? You take market share from your competitor by providing better service than your competitor. This is a good first step: Create supply chain incentives that push the marketplace to provide better patient service rather than incentives that increase the risk of limiting service for a large contingent of patients.

Heads, I win. Tails, you lose.

The comparison between Ohio managed care and fee-for-service markups is even more staggering when viewed using a histogram (Figure 7).  It also provides us with a hint on how PBMs are “saving money” for states in managed care.

Figure 7
Source: CMS State Utilization Database, CMS NADAC Database

It turns out that all of those teeny, tiny bubbles (markup of $0 to $3 per prescription) around the outer edge of the managed care universe in Figure 6 comprise over 60% of generic prescriptions dispensed in Q1 2018.  In other words, the state got a fantastic deal on these prescriptions by paying Ohio pharmacies at most 30% of what the state itself determined to be their cost to dispense (determined by Ohio to be $10.49 per prescription, on average).

Meanwhile, the state got a terrible deal on 11% of the prescriptions that were charged above even a small pharmacy’s cost to dispense (determined by Ohio to be ~$13 per prescription, on average). Of course, in a spread model the PBM is free to keep the proceeds from the drugs that were overcharged while passing through the savings on the drugs that were priced well below operating cost.  For the PBM, it’s heads, I win; tails, you lose.

This is why Ohio Medicaid is claiming that managed care still saves money over fee for service, despite the fact that the PBMs collectively took $224 million in spread alone in just one year through the Ohio Medicaid managed care program. In examining the data, it’s easy to see why these savings claims may be true.

The PBM can save money for the state by under-reimbursing the pharmacy on the preponderance of inexpensive generic claims and passing through these savings, and then pay itself through excessive spread on the remaining generic claims.  This isn’t improved efficiency – it’s simply a margin shift from the pharmacy to the PBM, and one that comes along with the unfortunate set of side effects that we have argued can be so detrimental to patient service.

Logic prevails in Mississippi

If you return to our visualization and start scrolling through the managed care bubble charts for each state you will see that not all states have continued to allow the PBMs to arbitrarily set generic prices in their managed care programs.  Mississippi is one of those states, as can clearly be seen in Figure 8.

Figure 8
Source: 46brooklyn Research

The managed care and fee-for-service bubble charts look almost identical for Mississippi in Q1 2018. As they should, because starting on April 1, 2017, Mississippi enacted legislation requiring all changes to its Medicaid pharmacy dispensing fee model prompted by CMS’ Covered Outpatient Drugs rule to apply to both fee for service and MississippiCAN (Mississippi’s managed care program).  What a novel idea! Figure out the answer to comply with a CMS rule and then apply the answer to all of Medicaid, rather than just part of it.  As a Florida Gator alum (Eric) it’s hard to say this, but congratulations Mississippi: you are providing an example for the rest of us when it comes to applying objectivity and properly-aligned incentives in your Medicaid program.

Another example of focusing on results over process

If you have a dumb incentive system, you get dumb outcomes
— Charlie Munger

In our initial report, “The Cancerous Design of the US Drug Pricing System,” we concluded with a discussion on how important it is to focus on getting the process right, and letting the results flow naturally from the correct process.  Outsourcing managed care generic price setting to the PBMs is yet another example of how Medicaid has been lured in by attractive “results,” but in doing so, has allowed complexity to propagate that has the potential to inflate hidden costs in our healthcare system.

PBM-driven managed care programs put the interests and whims of PBMs above the overall interests of the Medicaid program. Pharmacies that have the unfortunate displeasure of having more Amoxicillin prescriptions than Imatinib prescriptions are put at a significant market disadvantage than other pharmacies who may have a different patient base with more lucrative disease states.

PBM-driven managed care programs are already drying up access to certain types of medications as pharmacies improvise to stay afloat. In Ohio, the practices are driving many pharmacies out of the areas that arguably need them the most, and while the perception is that this has been an independent pharmacy problem, we are learning that it’s far greater than that. Target escaped a couple years ago. Rite Aid is halfway there. Kroger has started scaling back, and there’s speculation things could be far more dire for them. And just recently, Genoa was the latest shoe to drop.

And then in the sad but true realm of “dumb outcomes,” Fred’s Pharmacy recently announced that in the face of their struggling business woes, that they were going to focus more heavily on more margin-friendly products like beer, wine, and tobacco.

What does all this mean? Less access to quality healthcare and less competition on price and service. As we’ve seen in the hospital sector, overconsolidation is driving prices through the roof, and we may not be far away from that in pharmacy.

This also means that pharmacies (like any other business) will do whatever they can to turn a profit and stay open, which may not necessarily reflect the best interests of patients or our overall healthcare goals. This includes pharmacists spending their time scouring wholesalers for obscure NDCs that result in higher reimbursements, or calling doctors to switch patients from one therapeutic equivalent to another based on how much the PBM is willing to pay. These are ultimately distractions from the true role of the pharmacist, which is to mindfully and attentively serve his or her patients.

Unfortunately, this problem is not at all confined to Medicaid managed care.  In commercial plans, spread pricing contracts are the norm, not the exception, which should be clear by the fact that they are referred to in the industry as “traditional” contracts.  This means that if you are a beneficiary with a high-deductible, which incidentally the Kaiser Family Foundation just reported has increased eight times as fast as your wage over the past decade, then you are exposed to the arbitrary PBM pricing practices for generic drugs. This isn’t just an entitlement program problem; it’s a problem for all of us.

In a recent thought-provoking piece, Drug Channels’ supply chain guru Adam Fein whiteboarded the design of a new prescription supply chain if rebates were banished from the system.  In his report, drug prices at the pharmacy counter would be set as a fixed offset (the professional dispensing fee) to their ingredient cost, consistent with CMS’ recommendations in the Covered Outpatient Drugs rule. It is simple, and it aligns the incentives of the supply chain with those of the patient. It may not always be cheaper under any myopic set of calculations we can perform, but it starts to eliminate hidden costs associated the more complex system that we currently have in place.

The best part is that a transparent pass-through model will get cheaper over time.  As generic prices deflate, payers will fully realize these savings. As market share naturally flows to the pharmacies providing the best patient care and service, dispensing fees will come down (the larger the pharmacy, the lower the dispensing fee it requires to break-even).

With pharmacies less concerned about uncontrollable margin volatility, incentive structures can be introduced to shift a portion of the dispensing fee to a performance-based payment, bringing more revenue to pharmacists that are having a meaningful positive impact on payers’ total healthcare spend, and less revenue to those that are not. But the right incentives must be in place to get started.  With proper incentives, marketplaces can rapidly and efficiently drive meaningful positive change. Without them, they can be destructive, or one might say… dumb.