Tuesday, April 13, 2010

Latest IMS Data Shrinks Pharma by $26 Billion

You surely saw news stories based on this press release: IMS Health Reports U.S. Prescription Sales Grew 5.1 Percent in 2009, to $300.3 Billion.

What you may not know is that IMS Health also restated sales data from 2005 through 2008, resulting in a $25.7 billion, four-year drop in U.S. market size versus previous reports. No, this is not a belated April Fool's joke. IMS' explanation for the updated data is included below.

Chain stores and Non-Federal Hospitals had the biggest downward adjustments over the period, while Independents and Mail Pharmacies had small adjustments. I’ll comment more on share movement by distribution channel in an upcoming post. In the meantime: Caveat marketor!

IMS Health kindly makes certain industry data freely available on its Top-line Industry Data web page.

The 2009 Channel Distribution by U.S. Sales report (published in April 2010) presents data from IMS Health National Sales Perspective™ (NSP) based on sales *into* the various distribution channels tracked by IMS. As I understand the data, NSP represents product purchases from wholesalers or manufacturers, not revenues of the buyer. IOW, you won’t see data on pharmacy revenues. See below for more details on the NSP data.

What got my attention were the changes from the 2008 Channel Distribution by U.S. Sales document (published in April 2009). The older report shows market size that is higher every year from 2005 through 2008. In other words, the market has apparently shrunk—a lot—as measured by IMS.

Here are the changes to the top-line market size data:

Observations and comments
  • The NSP sales into Chain Stores (pharmacies) were adjusted downward by -$13.5 billion (-3.5%), accounting for 53% of the total restatement.

  • The source of the discrepancy changes over time. Chain stores account for 71% of the difference in 2005, but only 36% in 2008.

  • The 2009 report on Number of U.S. Dispensed Prescriptions did not make any revisions to previously published data of total market size. Hmmm...

  • Will these revisions affect the government’s already shaky predictions of future drug spending? See Drug Forecasts: Oops!...They Missed It Again for background.

  • Ironically, I called on IMS to restate its data in 2008, although I was focused on reclassification among channels, not total market size. See IMS Recounts, but Independents Still Growing.

So, what happened?

The good folks at IMS told me that there were "material enhancements" to NSP data. These included items such as returns, shipping adjustments and historical corrections not previously reflected. IMS restated 72 months of history, leading to the market-level declines in sales volumes across all channels. IMS clients heard about the changes in January.

Okey-dokey. But I wonder if these type of mega-revisions will encourage manufacturers to rely more on EDI 852 and 867 data from wholesalers, or perhaps depend on channel data integrators such as Drug Channels sponsor IntegriChain?

A note about NSP data: NSP represents sales at invoice pricing, not sales at a list price such as Wholesale Acquisition Cost (WAC) or Average Wholesale Price (AWP). Contract pricing, such as a discounts processed via a wholesaler chargeback transaction, are reportedly reflected in the IMS NSP measures of price and sales. Off-invoice discounts and rebates, such as rebates paid directly by the manufacturer to a third-party, are not reflected in NSP data.

I'd appreciate hearing your thoughts on what's behind the latest numbers.


  1. Nick BastaApril 14, 2010

    1. First, let's give a tip of the hat for IMS' publication of these data--however squishy they are; it's not something they're obliged to do, and you, i and a lot of other people make use of them.
    2. Apart from the channel-by-channel variations, how significant is a 2.3% variation in the data from one year's report to the next? Admittedly, we'd all like the finest possible precision, but given all the variables in prices, discounts and the like, channel by channel and month by month, how much precision is reasonable to expect? (It's noteworthy that the differences in the reported data are conssistantly a shade above 2%, which points to a systematic change, or error if you will).
    3. I continually wonder at the difference between the IMS data and that reported by the feds, in National Health Expenditures. For 2008, it reported 234.1 billion in "retail" sales, which doesn't agree with the IMS figure (291.5 billion) even when you back out hospital, LTC and federal facility sales (46.8 billion, per IMS). Even the year-over-year percent changes are different (3.2% change in fed figure for 2008, vs 1.85% per IMS).
    4. IOW?

  2. How significant is a 2.3% variation in the data from one year's report to the next? Maybe not much to consultants, but in the real business world it means making or missing a quarter. Currently, we are looking at our actual sales (pharma co.) vs our projections. Our projections were done a year ago off historical IMS data and refined as time goes forward. If our projections were off miss reported data, we have goals that are artificially high. 2% on a million or billion dollars becomes a lot of money to make up. It is easy to say it is not very big, until you have to make up the difference when it was never there! Nick, try finding $20 million dollars a year out of the air or $5 million a quarter. If you have a billion dollar drug, this is what you have to do, because your projections were based on bad numbers. Real bosses do not care, that it was IMS fault. The number is the number and has been communicated to the street last year. Pharma pays IMA a lot of money to get it right for this reason! Most people realize that everyone has a different number, and that is not big deal. The same company (like IMS) can not has numbers that change or you get big problems. LITTLE PERCENTAGES CAN STILL EQUAL HUGH DOLLARS.

  3. Nick,

    In addition to the points made in the comment above, I wrote the post to highlight an interesting fact that no other publication had bothered to mention.


  4. Adam--
    Not sure where you're going by saing "no other publication had bothered to mention" since, as you know, I reported last week that the historical data had changed (see, but obviously i didn't make as big a deal of it as you did.

    To "Anonymous" (whom i will forever think of as "Hugh Dollars" in return for implying that i'm a consultant!) I would repeat that i value accuracy as much as the next person, but my question still stands: how much precision should be expected? If not 2.3%, then 1.0%? 0.5%? 0.01%? Or (now i'll REALLY start an argument) should US pharma report their exact, actual sales figures per product every Dec 31st for the preceding year, and then we could all dispense with IMS Health? When you add in the complexity of per-channel sales, i think the answer is that, in fact, the pharma industry depends on IMS (and its competitors) to tell it how much revenue it should be seeing--which in turn points to the frustrating opaqueness of today's pharma distrbution channels.

  5. I stand corrected since I didn't read your article closely enough. I hereby amend my statement to read: " interesting fact that only one publication..." OK?

    I don't expect unobtainable precision. I'm just puzzled by large dollar revisions from four years ago given only minor revisions in past years.


  6. If pharma and others really cared about inaccurate forecasting, counting, etc. they would pay IMS less for their products and services or pay competitors to do it better. IMS would be concerned about the reduction in brand equity and do something to fix the situation.

    None of these seem to be occurring, which to me suggest that no one cares too much about this data, or "close enough" is all buyers are willing to pay for.

  7. AnonymousMay 17, 2010

    Trust me, people care a lot about the data being accurrate! There are multiple issues with the way the data are collected, processed and then reported upon. IMS processes more data than the IRS so this isnt a simple accounting method. Buyers of the data are very aware of the issues but need the very detailed level of info that they can provide - not just at the macro level which you are looking at. Can competitors do it better? If they can, IMS has a habit of buying them out!
    They charge premium prices because they can...there arent many good alternatives