Thursday, March 08, 2012

AARP Distorts Drug Price Data (Again)

For years, the AARP has been releasing periodic drug price reports, authored by Stephen Schondelmeyer of the University of Minnesota and Leigh Purvis of AARP. The latest study claims that “retail prices for a combined set of widely used prescription drugs finds the cumulative change in prices from 2005 through 2009 was almost double the rate of inflation.” See the AARP press release.

While the new study makes some mild concessions to prior methodological critiques, Dr. Schondelmeyer and Ms. Purvis again draw a completely specious comparison between drug prices and overall inflation. This unscrupulous misrepresentation is discouraging, especially when the media dutifully covers the "news." To be fair, even the New York Times has been shamed into presenting both sides of the story.

Read on for the depressing details and let me know if you agree.

BETTER, BUT STILL FLAWED

Many people, including yours truly, have criticized the obvious methodological flaws in the AARP price reports. For example, previous reports have included brand-name drug prices even after generic alternatives became available. A Different Perspective: The AARP Sponsored Schondelmeyer-Purvis Studies, a 2010 joint IMS-MIT study, does an excellent job of explaining why the AARP studies do not meet scientific standards.

In this latest iteration, Dr. Schondelmeyer and Ms. Purvis have addressed some of the critiques. For instance, the study finally acknowledges that the generic dispensing rate is above 0.0%. (Shocking, I know.)

The difference is dramatic. Here’s the chart from the August 2010 study, which included brand-name drugs only:

Here’s the chart from the March 2012 study, which included brand and generic drugs:

Some big problems remain. For example, the market basket is based on drug use in the Medicare Part D program, weighted by revenue…in 2006! (Yes, the last year that simvastatin was still sold as Zocor.) In other words, the new index still does not account for increases in the generic dispensing rate over the past 5 years. This creates an upward bias to the numbers.

TORTURING THE DATA

Despite the bias, the March 2012 analysis shows drug prices roughly tracking the Consumer Price Index (CPI) ... until September 2008.

So, in the finest tradition of Orwellian Newspeak, Dr. Schondelmeyer and Ms. Purvis simply reframe the discussion to “cumulative” price changes over the entire time period from 2006 to 2009. Presto chango, drug prices “rise steeply.” QED.

But wait. Did anything happen (besides the Presidential election) that could explain September 2008’s CPI drop, thereby rendering the results meaningless?

Well, as a matter of fact, the housing bubble popped and the economy tanked. (You may have heard something about it.) Way back in September 2008, I even provided Drug Channels readers with a cartoon guide to the subprime mortgage crisis in The Subprime Primer: A Brief Diversion. (Trust me, this cartoon is definitely worth your time.)

For those who don’t know, about 30% of the CPI is comprised of "Shelter," which is measured using two CPI indexes: (1) Owners’ equivalent rent of primary residence (OER), and (2) Rent of primary residence (Rent). See this BLS article for more details. Another 15% of the CPI is “Transportation,” including motor vehicles and fuel.

In 2008, the U.S. housing market tanked, oil prices dropped, and auto sales plummeted. Thus, almost half of the CPI components declined sharply for reasons that had nothing to do with the pharmaceutical industry or the election. A “cumulative CPI” that incorporates 2008-9 is an especially poor benchmark--unless you are trying to make a political point rather than a fact-based economic argument.

As my stats professor used to say: If you torture the data long enough, nature will confess.

Shameful, really.

P.S. There’s a lot more to the AARP than meets the eye. Check out Behind the Veil: The AARP America Doesn’t Know, a report by the U.S. House Ways and Means Subcommittee on Health. Scary stuff.

7 comments:

  1. I agree, but would love to see more.  What is a better measure of the brand and generic inflation rates?  Could you compile some stats for the PBM drug trend report or other studies to give your readers a more realistic picture of the price trend?

    ReplyDelete
  2. It's a different type of information. AARP is purporting to present a price index -- the change in the prices of a basket of goods. The government-collected CPI for retail drugs does a better (but still not perfect) job. Here's how the government does it: Measuring Price Change for Medical Care in the CPI

    The PBM drug trend data present the change in the total expenditures of a payer, which include changes in both prices and utilization. The reported drug trend data are also measured in a non-transparent way. See Is Caremark really better at managing drug trend?.

    ReplyDelete
  3. It's a different type of information. AARP is purporting to
    present a price index -- the change in the prices of a basket of goods. The
    government-collected CPI for retail drugs does a better (but still not perfect)
    job. Here's how the government does it: Measuring Price Change for
    Medical Care in the CPI



     


    The PBM drug trend data present the change in the total
    expenditures of a payer, which include changes in both prices and utilization.
    The reported drug trend data are also measured in a non-transparent way. See
    Is
    Caremark really better at managing drug trend?
    .

    ReplyDelete
  4. This comment has been removed by a blog administrator.

    ReplyDelete
  5. I thought the PBM data looked at how each component: price, utilization and drug mix contributed to the overall trend.  I was wondering if the price component could be used to get to more accurate brand inflation rate.

    ReplyDelete
  6. Yes. I like the straightforward breakout in the Medco drug trend report: "unit cost" vs. "utilization." See the summary chart in Insights from the 2011 Medco Drug Trend Report.

    Most investment bank publish data on brand inflation rates, but usually not as a formal index. 

    Keep in mind that AARP is trying to represent price changes in "widely utilized" drugs, so looking only at brands is misleading.

    ReplyDelete
  7. Msoumya InfoMarch 10, 2012

    This comment has been removed by a blog administrator.

    ReplyDelete