Is marketing a bad thing? How much money does Coca-Cola spend on Research & Development of its premier soft drink? Nothing. When something works, you go with it. New Coke taught them that. But they market it like crazy.

Yet whether pharmaceutical companies are primarily interested in research and development or marketing is central to the cultural debate about medicine.

Marc-André Gagnon and Joel Lexchin, writing in PLoS Medicine, state that information on promotional expenditures from IMS, the most widely quoted authority that surveys pharmaceutical firms, isn't reliable. Well, they say "may not be reliable" but that's as close to definitive as you can get - and all it will take to make headlines in the mainstream press.

They accessed data from CAM, a market research company that surveys physicians to audit promotional and marketing activities undertaken by the pharmaceutical industry, and compared its estimates to IMS and arrived at a new figure for total spending.

Using their method, they say pharmaceutical promotion (2004 numbers) is as high as $57.5 billion, more than double the estimate given by IMS.

And much higher than R&D, rather than lower. With U.S. domestic sales of $235.4 billion, that higher estimate would place 24.4% sales revenue geared toward marketing versus 13.4% for R&D.

There are two issues; first, is the number accurate and second, so what?

In 1977 there was one of those bad uses of statistic that caught the public's attention, namely that there were half as many divorces that year as marriages. To journalists that meant 50 percent of marriages were ending in divorce. Of course, it meant nothing of the sort but that statistic became part of the culture and divorce was suddenly considered almost typical. Divorces rose until the actual divorce rate is almost 50 percent.

IMS estimates a $27.7 billion expenditure for marketing. The comparison by Gagnon and Lexchin says it is twice that. If the lower number is accurate their marketing is just under 12% of sales and below their 13.4% R&D expense.

Lexchin and Gagnon say the IMS, using published annual reports, is wrong because it surveys companies and uses annual reports. Yes, those CEOs who will go to jail for virtually any infraction under Sarbanes-Oxley are somehow underestimating marketing to look better in the eyes of a public that has no clue what their marketing costs are and don't much care.

The IMS lacks accuracy because they are not monitored by the government, the writers also contend. Politics aside, that's a difficult argument to make. The same government that can't even manage damage control and logistics for a hurricane that hits land, and where the damage is really obvious, is unlikely to be able to ferret out false data when clever CEOs that are already going to jail for publishing it are apparently still able to hide it.

These authors instead use CAM which, like IMS, uses surveys, but of doctors, and CAM has a 'validation committee' that vouches for its accuracy and that committee consists of industry representatives for companies like Merck, Pfizer, Bristol-Myers Squibb, etc.

I am not sure why a validation committee made up of industry representatives is better but perhaps it is. I certainly doubt they would sign off on any report that says their annual report is so inaccurate their CEOs could go to jail for it but let's put logic aside for the moment.

Does a higher number from CAM mean IMS is inaccurate? Not at all. The difference is not in tangible things, like direct-to-consumer advertising (US $4 billion according to both) or journal advertising (US $0.5 billion for both) but rather in intangibles, detailing and samples and undefined unmonitored promotional expenditures.

IMS estimated the amount spent on detailing at US$7.3 billion [4] versus US$20.4 billion for CAM [10], and while IMS gave a retail value of US$15.9 billion for samples [14], CAM estimated a wholesale value of US$6.3 billion [10].

So how did they resolve the numbers? They didn't. They simply took the higher number that IMS provided for samples and the higher number CAM provided for detailing, added

other marketing expenses after correcting for the 30% estimate of unaccounted promotion

and came up with a whopping difference. A difference that is not actually borne out by objective analysis but takes the worst of all possible scenarios and a bit of speculation and draws a conclusion.

Now, if the pharmaceutical industry is $235 billion, they certainly don't need any help from me. I have never gotten any money from them and if any pharmaceutical company has ever so much as placed an ad here through one of the companies that sells our stuff, I am not aware of it. But this is a science site and we have to defend ourselves from being drawn into class warfare by misuse of statistics.

So the numbers may or may not be accurate. It all depends on whether or not you want to believe them. But if they are accurate, is that a bad thing? Like the marriage example from the 1970s, a higher marketing expense, per year or in one year, versus research and development, in one year or per year, doesn't tell much of a story about what's really going on.

If a drug company spends seven years on research and development, they're going to recoup that expense and they're going to want to win against competitors to do it. That necessitates a marketing budget and it's the price society pays for having the best pharmaceuticals in the world.

Like any industry, we want to assume the best and monitor for the worst and companies like IMS and CAM help us do that. Combining the worst numbers from each service and padding the total by 30% to make a stronger case against an industry is cultural politics, not science.

We could all use less of that.

Article: The Cost of Pushing Pills: A New Estimate of Pharmaceutical Promotion Expenditures in the United States. PLoS Med 5(1): e1 doi:10.1371/journal.pmed.0050001