While summer isn’t officially over, it sure feels that way. That’s because the kiddos are back in school, which means less interruptions as we stare at charts and numbers and peck away at our keyboards.
We took advantage of this newly-acquired, uninterrupted time to update and study our Brand Drug Price Change Box Score visualization, with a specific focus on midyear price increases, which had gotten some negative press back in early July.
It turns out there initially wasn’t much to study. We updated the box score viz and found that in July and August, price increases were only announced on 18 brand-name drugs, while two drugs decreased in price.
So only 16 net price increases. That’s down from a net:
53 price increases in Jul/Aug 2018
209 price increases in Jul/Aug 2017
353 price increases in Jul/Aug 2016
548 price increases in Jul/Aug 2015
If you are more of a visual person here’s what that looks like:
Of course, when we saw this our first reaction was 😁.
But that quickly was replaced by 🤨. We had a feeling we were missing something.
We were missing something
Back in January, we went into copious detail on how we constructed our Brand Drug Price Change Box Score dashboard. If you are looking for a refresher, click here to read our January post on how to use the viz.
Long story short, we figured out that anytime a manufacturer published a change to its listed Wholesale Acquisition Cost (WAC), Myers and Stauffer (on behalf of CMS) picked up this change and published it in its publicly-available National Average Drug Acquisition Cost (NADAC) database, which tracks surveyed pharmacy invoice acquisition costs on prescription drugs.
If you have read any of our work, you know we love NADAC. We don’t have very high standards when it comes to drug pricing databases. We’ll fall in love with any database that:
Is available to the public (either free or at a reasonable cost)
Contains prices that (unlike WAC, AWP, or SWP) are actually based on the principles of supply and demand
Despite these criteria being way too broad and undiscriminating, there is sadly only one database out there that meets both of them, and that’s NADAC. So, we are grateful it exists.
Having said that, NADAC has three considerable limitations that we have also discussed in our work. First, NADAC doesn’t include rebates or any off-invoice discounts that pharmacies can receive from wholesalers on most generic drugs. Second, NADAC doesn’t include rebates that drug manufacturers pass to health plans, Medicaid programs, and PBMs. Third, it is a voluntary survey of invoice-based acquisition costs from the retail community pharmacy “class of trade” (i.e. no specialty, mail-order, long-term care, or hospital pharmacies are included).
This third limitation is what we believe is the source of our current problem of what seemed to be missing data from the NADAC prices that live behind our Brand Drug Price Change Box Score, so let’s drill into it a bit deeper.
A Voluntary Survey of Retail Pharmacies
When a pharmacy receives a NADAC survey, it can choose if it wants to complete it. It’s “voluntary.”
Some pharmacies choose to complete the survey. This choice helps federal and state government officials better understand what drugs really cost, and helps government set more accurate benchmark pricing for programs like Medicaid fee-for-service.
Other pharmacies choose to ignore the survey. This choice hides the pharmacy’s acquisition cost from the government, which may worsen the quality of the NADAC survey benchmark that federal and state governments rely upon to approximate drug costs and to set pharmacy reimbursement.
According to CMS and Myers and Stauffer, of the ~2,500 pharmacies that are randomly surveyed each month for the NADAC calculation, about 450-600 will respond. While that’s a meaningful sample size, it sadly tells us that the majority of retail pharmacies that are surveyed are choosing Option #2: Ignore.
It shouldn’t come as a surprise that the response rate is so low. Put yourself in the shoes of an executive at a very large pharmacy chain (think CVS, Walgreens, Rite Aid) who likely has exceedingly low acquisition costs spread across many retail outlets. By submitting your costs, you risk releasing competitive, proprietary rates – big companies avoid that like the plague. But additionally, by submitting your costs, you risk lowering NADAC, which would lower your Medicaid fee-for-service reimbursements or any other reimbursements that are linked to NADAC. That sounds like a surefire way to lose your job.
However, by not responding to the survey, you ensure that NADAC is based on the higher submitted acquisition costs of your smaller pharmacy peers, thereby securing higher Medicaid reimbursements relative to your costs. Congratulations, you have a corner office in your future!
While a bit dated, the data appears to prove this out. In December 2012, CMS and Myers and Stauffer reported that 64% of total survey responses came from independent pharmacies while 36% came from chains. However, independents accounted for 35% of all retail community pharmacies, while chains accounted for 65% – the complete inverse of the response rate. Anyway, do the math on these percentages and you will find that independents are 3.3 times more likely to respond to the survey than are chains.
Let’s move to the second key word – “retail.” Currently, the NADAC survey is only sent to retail community pharmacies. It does not include specialty pharmacies.
This is a growing problem for NADAC as time passes. Since PBMs are successfully driving many specialty prescriptions to their own specialty pharmacies (and away from traditional retail pharmacies), that means NADAC will be missing out on obtaining data on many specialty drugs (and their hefty prices).
So what is the point to this discussion? Because it’s voluntary and focused on retail pharmacy reporting, NADAC only includes drugs purchased by retail pharmacies that choose to volunteer their acquisition costs. In other words, NADAC is simply a reflection of the drugs these 450-600 Good Samaritan pharmacies purchase and dispense. And guess what these few hundred small pharmacies don’t usually purchase and dispense due to PBM restrictions – expensive, specialty drugs! So it follows that NADAC should be largely blind to the acquisition cost of many expensive, specialty drugs. And guess what we found? It is.
The missing price increases
So let’s get back to those 16 net price increases from this summer. We need to redefine what this really is. What this means is that of all of the brand-name drugs commonly dispensed at your local retail pharmacy, 18 of them increased in price this summer while two of them decreased in price.
So what about all of the drugs not dispensed at your local pharmacy? How many of those experienced price increases this summer? Answer = 281.
That’s right, we found 281 brand-name National Drug Codes (NDCs) that increased in price this summer that were NOT in the NADAC database. Several of these drugs appear to be obscure, low utilization drugs. But many others appear to be expensive specialty drugs not dispensed in the retail community pharmacy class of trade.
This second observation really pops when comparing the WAC of the drugs missing from NADAC to the WAC of the drugs captured in NADAC. Overall, the median WAC of the drugs that went up in price that are missing from the NADAC database is $113 per unit. The median WAC of the drugs that went up in price that are included in the NADAC database? Just $8 a unit.
We went back over the past few years and performed the same analysis. Figure 2 shows the number of NDCs that increased in price in July and August over the past five years. The blue series shows the NDCs that were included in the NADAC database, while the orange series shows the NDCs that were excluded from the NADAC database.
Our main takeaway from Figure 2 is that in each of the last five years, there have been more NDCs that increased in price that were missing from the NADAC database than were included in the NADAC database. In other words, the disparity we found in 2019 is not an anomaly.
It also turns out that starting in 2017, the missing NDCs have been significantly more expensive than the included NDCs. Figure 3 shows the median WAC of all brand-name drugs with price increases in July and August. In 2017, the median WAC of the NDCs not captured in NADAC shot up to over $100 per unit, and has remained there for the past three years. Meanwhile, besides a relatively small increase in 2018, the median WAC of the brand-name drugs included in NADAC hasn’t meaningfully changed.
In summary, the drugs with price increases that are missing from NADAC are considerably more expensive (i.e. more likely to be specialty drugs) than the ones included in NADAC, at least over the past three years.
A Better NADAC
If you have made it this far into this report, welcome to the drug pricing nerd’s table. We saved you a seat. We know this isn’t the sexiest of data-dives, but it’s really important. Drug pricing accountability relies on knowing the pricing, and NADAC is the best publicly-available way to access those prices and to track market trends.
We suspect that readers have come to one of two conclusions.
The first conclusion may be not to use our Brand Drug Price Change Box Score because it’s missing hundreds of (largely) specialty drugs that have experienced price increases. While we think that there is still a lot to glean from it, it’s important to keep this limitation in mind. We are disappointed that the only public data source that is available to all of us is limited due to its “voluntary” and “retail” nature, while the rest of the data that is required to paint a complete picture hides in boardrooms or behind paywalls.
But some visibility into the problem is better than none at all. To use an analogy, if you’re stuck in the middle of desert dying of thirst, you’re probably not going to thumb your nose to a bottle of Aquafina just because it’s not a bottle of Acqua di Cristallo Tributo a Modigliani. Our visualizations are no more than a necessary bottle of Aquafina for a collective public that is suffering from massive dehydration when it comes to drug pricing data. It’s not enough, but it’s the best we can do with the tools we have at our disposal.
The second conclusion is that we need a better NADAC. NADAC is quite comprehensive for generic drugs, covering 97% of all Medicaid generic claim submissions. But clearly, its coverage of specialty drugs is significantly lower than this, and it needs to be improved, especially given that government is increasingly using this benchmark to set reimbursements.
So how can CMS make NADAC better? They can support legislative efforts to change “voluntary” to “mandatory” and “retail” pharmacies to “all” pharmacies. In our view, this is the only way we will assemble a public database that has supply/demand driven pricing for as many drugs as possible. Sitting around and waiting for the supply chain to altruistically volunteer such pricing is an exercise in futility – we’ll die of thirst before that happens.
As we mentioned last month, U.S. Senators Chuck Grassley and Ron Wyden’s Prescription Drug Pricing Reduction Act (PDPRA) of 2019 aims to make several adjustments to the NADAC survey, including requiring survey participation, reporting wholesaler rebates, and examining how to obtain average acquisition costs for specialty pharmacies.
This would appear to be an excellent step towards eliminating the blind spots from the NADAC database and providing an even sharper blade of accountability for federal officials overseeing problems in the drug supply chain and for the price-setters within that supply chain that have made a habit of draining the lake.