Tuesday, December 13, 2011

Generic Drug Prices are Rising, according to latest AMP data

In case you haven’t been paying attention, the Center for Medicare and Medicaid Services (CMS) continues to pump out monthly Average Manufacturer Price (AMP) data. The September 2011 data was just released.

The data confirm that generic drugs remain very inexpensive. More than half of generic drugs are sold by the manufacturer for less than 15 cents per unit. Remember, these are actual transactional prices, not list prices.

For fun, I created an index of generic prices. Based on the first three months of AMP data, this Index of Generic Weighted Average AMPs (IGWAAMPS, pronounced “igg-whamps”) rose by 4.9% from July to September.

Read on for the surprising details.


CMS has a new website, so some old links don’t work anymore. Bizarrely, the links are embedded within text paragraphs of this page—kind of like playing Where’s Waldo! If you don’t want to search, here are direct links to the zip files of data:

For more background on these data, see my two previous posts: Hello, Transparency: CMS Publishes its First AMP Data and The Pharmacy Reimbursement Hit from AMP-Based FULs.

OBSERVATION #1: Most generic drugs sell for pennies per pill

The table below shows the distribution of WAAMPs for the 826 product groups included in the September 2011 data. Two-thirds of the products groups sell for less than 25 cents per unit (pill, tablet, capsule, etc.). More than half are less than 15 cents per unit.

Note that these data reflect prices paid to the manufacturer by: (1) wholesalers for drugs distributed to retail community pharmacies, and (2) retail community pharmacies that purchase drugs directly from the manufacturer. Thus, the average acquisition cost paid by a pharmacy (to a wholesaler or a chain warehouse) will be higher.

OBSERVATION #2: Generic drug prices are rising

CMS is still working the kinks out the system, judging by the fluctuating number of products groups with a weighted average AMP:

  • July 2011: 719 product groups
  • August 2011: 887 product groups
  • September 2011: 826 product groups
However, 593 product groups showed up in all three months. To assess whether overall prices rose or fell, I computed an equal-weight index of these 593 product groups using three consecutive WAAMP values. I set July 2011 to equal 100.

As the chart below shows, this hypothetical basket of generic drugs rose by 4.9% from July to September 2011.

Here’s what happened in July vs. September:
  • WAAMP increased in 317 product groups (median increase: $0.028; +17%)
  • WAAMP decreased in 276 product groups (median decrease: -$0.016; -13%)
Yes, I realize that an index with 3 data points is not too exciting, but hopefully you can see what the WAAMP data will provide—a real-world look at generic drug prices.

Is the trend an artifact of draft data? A quirk in how CMS is averaging? Hard to tell, but I’ll keep an eye on this index as a way to judge the generic drug prices over time.


  1. Mike WinkelmanDecember 13, 2011

    Adam:  As usual, you're digging into an important topic. It is possible that what you're describing is due to the rather strange calculus mandated by CMS, by which generic manufacturers must calculate their AMPs.  You might find it informative to dig into this methodology.  

  2. Believe it or not, CMS has still not issued final regulations regarding AMP computations! 

    Anyway, I only included product groups with three months of data. I am assuming that a similar set of manufacturers provided data over this period, so any errors would be stable and not affect the month-to-month change in index value. 

  3. Adam, I believe that the 4.9% increase in this market basket of generic products is more related to the methodology of AMP/FUL pricing determination than actual generic pricing increases.  A 4.9% increase in 3 months is not what you would expect for generics. AMP/FUL prices will fluctuate with no actual price point changes, due to the weighting variances in the methodology from month to month.  This is another problem with accepting these price points as benchmarks.  CMS has talked about a "smoothing" process, but I am not shure that will help much, and will prolong the pricing corrections for real significant market price changes.

  4. If you go into the detailed view in the spreadsheet you are going to realize that these AMPs include the Reference Listed Drugs (RLD), i.e. the innovator drugs as well:
    1. That means the AMP that is shown is not the true generic price but a mix of RLD and generic pricing. Depending on how much market share the RLD retains in any particular month it can have a fairly large impact particularly in the first months after the products become generic when the RLD share is still higher
    2. That means part of the increase you are observing may be due to RLD price increases which are far more common than increasing generic prices.

  5. Mark,

    Agreed. There are many methodological mysteries within these data. Note that 47% of the products groups saw declining prices.


  6. True. This is the same dynamic as ASP. Brand share is pretty low for mature generics, so I presume that the weighted average for any particular product group primarily reflects the AMPs of multi-source products. 

  7. Speaking from minimal experience - those months in your chart probably aren't the best to judge generic price increases on. When you enter fall and children go back to school - supply and demand and/or perceived demand always causes prices to spike in antibiotics, allergy, cold and steroid meds - which account for a large majority of Fall generic sales volume.  My old boss would always base our upcoming year's bonuses on the revenue numbers from October to December of the present year (we were idiots, we went for it) and act as if we were on a bell curve of monumental prop'.s. The trend would die in January and pick back up in February only to die again - quickly - in April.  Same sales pattern every year. No bonus.

    Check out the AWP's for Z-packs in Fall and Spring over the last ten years.  If the drug were publicly traded we'd be posting on WSJ right now.In the Summer you can flat out steal a 3-pk 250mg Z-pak for $6.  In late August, when you really start needing it ... It's $7.46.Not discrediting the fact that you're most likely right due to this economy, but, it was a reality I grew painfully accustomed to ...Volume from July to Sept on those 593 product groups would seal the deal for your theory I would think.  If the volume is consistent - you're right. If it's lower or higher, who knows?

  8. These are "draft" data with many unknown methodological issues. I wouldn't take them too seriously yet. Over time, the patterns will become clearer and more meaningful.