Earlier this week, Bloomberg reporters Robert Langreth, David Ingold, and Jackie Gu published their results of a fascinating deep dive into Medicaid generic drug prices in their article, “The Secret Drug Pricing System Middlemen Use to Rake in Millions.”
The piece, which has been making the rounds on social media over the last few days, did an excellent job explaining the ins and outs of the hidden pricing spreads that exist on generic drugs, and it featured some intuitive visualizations that helped educate readers who may not have been familiar with these little-known drug price tactics.
If you’ve been following our work since our launch last month, you know that we are no strangers to these types of practices. Bloomberg’s methodology was nearly identical to what we used to create our Medicaid Drug Pricing Heat Map (with the main difference being 46brooklyn’s decision to measure pricing on a per unit basis rather than a per prescription basis, as Bloomberg chose to do), so not surprisingly, the results were very similar as well.
MARKUP = PRICE PAID BY MEDICAID - PRICE PAID BY PHARMACY
Whether you are looking at the green shaded area (Bloomberg) or the gap between the orange and blue lines (46brooklyn), what you are seeing is “markup.” Markup is the difference between the cost to a state Medicaid program and the invoice cost of the drug to the pharmacy.
While Bloomberg highlighted some very high generic drug markups in Medicaid (notably, in Indiana), not all states show such egregious markups. Actually, some states (notably, Washington state) appear to have unsustainably low markups. Sometimes you will even see a negative markup, meaning that a state Medicaid program paid less than the average pharmacy invoice cost for a drug.
So, how do we know which states are paying a fair rate for prescription drugs?
It’s complicated. Every state builds their Medicaid program in their own unique way. Some states pay pharmacy claims directly in a fee-for-service model, and other states pay the pharmacy claims through Medicaid managed care organizations (MCOs) and their PBMs. Or states can do a combination of both models, putting a percentage of patients in either program. While the fee-for-service programs may be more streamlined, many states rely on managed care to attempt to contain costs in different ways. States that use MCOs each have a different collection or mix of MCOs, and those MCOs have a different collection of contract structures with a different collection PBMs. Put this all together and it shouldn’t be shocking that different states are paying different amounts for the same drugs. It’s just an unintended consequence of how Medicaid is designed.
Providing a means to analyze these pricing differentials across the country was one of the main motivating factors for the development of our 46brooklyn Medicaid Drug Pricing Heat Map visualization. While fee-for-service programs had minimal variability in prices from state to state, we were amazed to see the wild variations in price within the managed care programs. If the actual drug prices are relatively stable from state to state, how could there be such discrepancies in the rates that state Medicaid programs are being charged?
To get a sense of what we’re talking about, tinker around with our Medicaid Drug Pricing Heat Map that we’ve embedded below. After you check out a few individual drugs, we’ve built in a feature that allows you to check the weighted average price of every tablet and capsule in each state. To do this, first select “All” generic drugs, and then hover over any state and look at the gap between the two lines that show National Average Drug Acquisition Cost (NADAC) and State Utilization Data. That gap is the markup.
Be careful reading into large markups when there isn’t a lot of data available. This is the reason we provide metrics at the top of each chart, so you can get a good feel for which charts are based on a sizable amount of data versus which ones are not.
WHERE IS THE MARKUP GOING?
It is very important to note that markup is NOT spread. In other words, markup is not what the PBM receives and not what the pharmacy receives. It is essentially the margin that the PBM and the pharmacy “share” on a claim. That of course leads to the logical question of how markup is divvied up.
To answer this question we dug into the data coming out of Ohio, as it has become a national focal point for the issue of spread pricing. This is due in large part to vocal pharmacists and the relentless investigative journalism of the Columbus Dispatch. It’s been reported extensively that Ohio Auditor Dave Yost found $224.8 million in PBM spread pricing between Q2 2017 and Q1 2018 within the Ohio Medicaid managed care program, $208.4 million ($6.15 per prescription) which came from generic drugs.
So we know how much the PBM received, and many folks have already voiced their opinion on whether or not this is a fair number or not. But the question is, if PBMs pocketed $208.4 million on generic drugs in the Ohio Medicaid program, how much did the pharmacy receive for actually dispensing those drugs? Applying a bit of arithmetic to the table of page 2 of the Auditor’s report, we calculate that pharmacies received, on average, $13.40 per generic prescription (($662.7M - $208.4M) / 33.9M prescriptions).
To be clear, that is the total revenue the pharmacy received per prescription, not the pharmacy’s gross margin per prescription. To estimate what the gross margin per prescription is for the pharmacy, we need to know the ingredient cost.
Thanks to the Centers for Medicare & Medicaid Services (CMS) and Myers and Stauffer, we can turn to our old friend NADAC to help us out with this. We dug into the same database that lives behind our Medicaid Drug Pricing Heat Map and added up the total NADAC ingredient cost for all generic oral solids for the same time period used by the Ohio Auditor (Q2 2017 through Q1 2018) and then divided that by the total number of generic oral solid prescriptions dispensed over that time.
As shown in Figure 4, we arrived at a weighted average NADAC cost of $12.25 per prescription, meaning on average, out of the $13.40 per prescription that pharmacies received on generic drugs, pharmacies spent $12.25 per prescription in order to acquire those drugs from their distributors Based on the data in Auditor Yost’s report, this would leave an average Ohio pharmacy with a margin of only $1.15 per prescription.
This margin is in stark contrast to Ohio’s results from its own cost of dispensing survey, that aimed to capture the overall costs associated with operating a pharmacy and dispensing medications to patients. The most recent Ohio cost of dispensing survey (conducted in 2016) arrived at an average cost to dispense of $10.49 per prescription. To be clear, this is the gross margin that Ohio Medicaid determined a pharmacy needs to cover its day-to-day operating costs (e.g. pharmacists, technicians, rent, utilities, pharmacy software, licensing, etc.). So we estimate that managed care Medicaid is falling short of Ohio Medicaid’s own targeted dispensing fee by nearly 90%.
Before we move on, we should note two qualifying assumptions that we made in performing this analysis. First, we only looked at generic oral solids to arrive at an ingredient cost to avoid unit mismatches. However, this creates conservatism in our comparison (i.e. the $1.15 margin is likely too high) because oral solids are, on average, typically cheaper than creams, gels, ointments, pens, inhalers, etc. You can see this in Figure 6, which shows that Ohio fee-for-service (which is a transparent cost-plus program) paid 33% less per prescription for generic oral solids versus all other generics. So we are comparing an oral solid ingredient cost that is arguably low to the Auditor’s overall average, resulting in a higher estimated margin.
There is an offsetting factor that is unaccounted for here, and that is off-invoice rebates that pharmacies receive from their wholesalers. While these rebates are not guaranteed, not offered by all wholesalers, and not for available for every drug, these pricing discounts do matter and are not reflected in the NADAC rates. Unfortunately, as is the case with Pharma-PBM rebates, off-invoice wholesaler-pharmacy rebates are not public information, which further distorts those looking to get a handle on true pricing.
But we can at least get a feel for what rebates would have to be for Ohio pharmacies in order to turn a profit on Medicaid managed care generic prescriptions. Let’s say, as Ohio found in 2016, that the average pharmacy needs $10.49 per prescription in gross margin to cover its operating costs. If they are getting $1.15 in margin (based on NADAC invoice cost) from Medicaid, then they would need to be getting $9.34 in rebates from their wholesaler to break even ($1.15 + $9.34 = $10.49). $9.34 is 76% of the $12.25 weighted average NADAC. So, the pharmacy would need to be getting a 76% or greater rebate from its wholesaler just to cover its operating costs. We’ll leave it to Twitter to gauge the pharmacy reaction to this number, but our educated guess is that you likely have better odds retiring off of winnings from a slot machine than a typical pharmacy has of getting anywhere close to a 75%+ rebate on generic drugs from any wholesaler.
THE STEEP PRICE OF “PREDICTABILITY”
It would not be fair to take our Ohio example and extrapolate this to managed care in other states. As we mentioned earlier, each state has different MCOs with different PBMs with different contracts between the two. And as we’ve learned time and time again, the pricing practices that the supply chain uses vary from plan to plan, payer to payer, drug to drug, and region to region. Having said that, we are seeing complaints from pharmacists in other states – notably Illinois and Kentucky – that look eerily similar to the complaints coming out of Ohio a year ago.
However, if you are a state policymaker – or for that matter, any payer – and have entered into a spread contract with your PBM, you have effectively taken the price-setting keys for generic drugs away from the marketplace and handed them over to your PBM in return for the “predictability” that this structure provides. The disconnect between what you pay for any given drug and its actual cost (i.e. markup) is the price you pay for this “predictability.”
If you are now having second thoughts about the steep price you are paying for “predictability,” there is only one way to get to the facts on what is really going on, and that is to follow Ohio’s lead and do a prescription-level audit of your program to truly understand how much spread is being taken by your PBM for its services.