NCPA Responds to Drug Channels

My little ol’ blog caught the attention of the National Community Pharmacists Association (NCPA).

Yesterday, NCPA posted a response to Shhhh! Owning a Pharmacy is Very Profitable entitled Independent Community Pharmacy Today.

I’m flattered that NCPA has validated Drug Channels as a source of industry information worthy of rebuttal. Muchas gracias.

In keeping with my blogging philosophy (“Everyone is entitled to his own opinion, but not his own facts.”), I want to show you clearly why my computations are not “exaggerated,” as NCPA claims. Please note that NCPA’s response neither disputes my figures nor provides any alternative data. U.S. Census Bureau data on pharmacy margins further supports my conclusions.

I also want to clarify one point about the availability of the Digest’s financial data—although my original critique still holds true.

Finally, I want to remind you that I have no conflicts of interest—although my interest in facts can create conflicts.

OWNER’S DISCRETIONARY PROFIT

On Tuesday, I wrote: “[I]ndependent pharmacy owners are doing financially quite well, averaging almost $300K in personal income in 2008.”

Here’s my math.

• The NCPA Digest reports a figure called Owner’s Discretionary Profit, defined to be the sum of Net Profit Plus Owner Compensation before Tax. (These are 2 of the 3 ways a pharmacy's gross profit gets spent per Tuesday's post.)
  
  
• Median Owner’s Discretionary Profit Percentage As a Percentage of Total Sales in 2008 = 7.5% (per page 9 of the 2009 NCPA Digest)
  
  
• Average Annual Sales Per Pharmacy Location in 2008 = $3.881 million
  
  
• 0.075 x 3,881,000 = $291,075 i.e., “almost $300,000”

Q.E.D.

Are you worried that I multiplied a median by an average? Don’t be. See my note for math geeks at the bottom of this post.

CENSUS DATA SHOW EVEN HIGHER PHARMACY PROFIT MARGINS

In my post, I point out: “Despite what you may have heard, gross margins for independent pharmacies have remained in a relatively stable range from 22% to 24% since at least 1999.”

U.S. Census Bureau data reports that Gross Margin as a Percent of Sales for NAICS 44611 (Pharmacies and Drug Stores) was 24.4% in 2007, which is actually higher than the NCPA data. The 95% confidence interval for the Census' estimate is 25.1% to 25.9%.

The 1999-2007 average was 25.5%—also higher than the NCPA average.

You can view the Census data here: http://www2.census.gov/retail/releases/current/arts/gmper.xls. The link will open an Excel spreadsheet. See Row 19 for NAICS code 44611. Here is the link for the standard error spreadsheet: http://www2.census.gov/retail/releases/current/arts/salescvs.xls. (See row 37.)

Granted, the NCPA data come from a self-selected sample. NCPA does not release information on samples size or provide confidence intervals around the point estimates in the Digest. Given the potential public scrutiny of the NCPA data, one could imagine that NCPA respondents might even try to make their profit data look worse than reality.

THE NUMBER OF INDEPENDENTS IS NOT SHRINKING (MAYBE)

The NCPA’s post critiques my profit computations in part by noting: "[T]he 2009 Digest data documented a net reduction in the overall number of independent community pharmacies – 590 fewer than the previous year.”

Curiously, the NCPA issued the following statement on October 19, 2009:

“The overall numbers of community pharmacies were relatively steady at 22,728. In fact, between December 2007 and December 2008, 1,225 new community pharmacies opened their doors, following the dream of being one's own boss.”(Source: Resilient Community Pharmacies Rely on Patient Services to Weather Economic Downturn, Competition, According to 2009 NCPA Digest)

The 2009-10 NACDS Chain Pharmacy Profile also reports a slight increase in the number of independent pharmacies, although “independent” is defined differently.

Whatever…there’s a standard error around all of these estimates, so I’m not sure how much we can statistically conclude one way or the other.

A CLARIFICATION ON TRANSPARENCY

On the day before my post, John Norton at NCPA told me: “Non-NCPA members do not have access to the financial data collected for the Digest, but they can purchase a ‘Print’ version of the Digest which contains demographic data, but no financial data.”

I erroneously concluded that non-NCPA members had no access to the data, but this is not strictly accurate. Apparently, the NCPA Digest with financial details is available in the library of your local pharmacy school. Good news if you are a pharmacy student or faculty member.

However, my original assertion still stands: NCPA used to make these data readily available to non-members for purchase but now puts them behind the NCPA firewall. Tough luck if you are a health plan, Medicaid agency, PBM, wholesaler, analyst, or anyone else without an NCPA membership. Most associations simply charge more to non-members, which is what NCPA did in past years.

I joined NCPA as an associate member and can get access to the financial data. But as I type this post, I'm starting to wonder if my membership will be revoked later this morning...

AM I CONFLICTED? DECIDE FOR YOURSELF

Rather than rebut my facts, NCPA chooses instead to impugn my reputation, writing: “Drug Channels/Pembroke Consulting does not disclose a meaningful list of paying clientele. Given the opinions expressed on Drug Channels, advising readers of possible conflicts of interest would appear to be a prudent step.”

Guess what? I don’t get paid to write this blog. I make Drug Channels freely available as part of my mission to educate, inform, and challenge people about an oft-misunderstood corner of the U.S. healthcare system. Whenever possible, I provide links to the original source material so that you read it for yourself and make up your own mind. I am congenitally opposed to unsupported accusations, overwrought hyperbole, or just plain misrepresentation. See Question 9 on my Frequently Asked Questions list for more.

Obviously, I hope that the occasional blog reader feels compelled to hand me a bag of money, but a quick glance at Drug Channels' blog traffic should convince you that those wise folks are a statistical minority of the readership.

As for disclosure, I have a small private company, not a lobbying group or non-profit 501c(6) association. My clients expect discretion and confidentiality. The analyses on Drug Channels are based on information and data that are in the public domain. I NEVER (a) disclose material, nonpublic information about a public company; (b) disclose information that I have a duty to keep confidential (e.g., by agreement, fiduciary duty, etc.); or (c) disclose information that I obtained from any person who expects me to keep it confidential. A lengthy disclaimer appears at the bottom of every page.

WHAT’S NEXT?

I have neither the time nor the resources to engage in an extended tit-for-tat with a well-staffed national trade association. The data speak for themselves. Make up your own mind.

Oh, and keep reading Drug Channels. Thanks!

---

BONUS COMMENT FOR MATH GEEKS

The non-math inclined should stop reading now.

Still here? Cool.

As I note above, I multiplied a median by an average. This is not necessarily a problem.

• If the distribution of Owner’s Discretionary Profit Margins is symmetric, then the average equals the median. My computation stands.
  
  
• If the distribution is skewed (non-symmetric), then the median does not equal the average and further analysis is required.

Since I don’t have access to the underlying data histogram, I computed a simple skew estimate called Bowley Skewness, a.k.a. the quartile skewness coefficient, using the limited information in the 2009 Digest. I assume a unimodal distribution and reasonable sample size (n>25).

My computation shows that the distribution of Owner’s Discretionary Profit Percentage in the 2009 Digest has a very slight negative Bowley skewness (-0.17), i.e., the average is less than the median. However, this figure is not statistically significantly different from skew=0, i.e., symmetry. Therefore, the average and the median are pretty close, so my computation stands.

To reassure myself, I computed the Bowley skewness for the 2008 Digest data and found that it had a slight positive Bowley skewness, i.e., the average is greater than the median. Again, the skewness was not statistically different than zero.

Don't even get me started on kurtosis...

P.S. The Flattery photo above comes from uber-snarky demotivational retailer Despair.com. They have great holiday gifts for your office breakroom.