Different River

”You can never step in the same river twice.” –Heraclitus

February 24, 2005

How much does it cost to invent a new drug?

Filed under: — Different River @ 7:57 pm

In a historic 1991 paper, Joseph DiMasi and colleagues calculated an estimate of the total cost to a drug company of developing a new drug. The innovative part of the paper was that they took into account the fact that most drugs fail to reach the market, and the drugs that do reach the market take many years, maybe even decades to develop. So, instead of just taking a couple of case studies of drugs that made it, they calculated total research costs, including for drugs that didn’t make it. This is the right way to do it, since a drug company doesn’t know in advance which of the (say) 5,000 compounds it tests will turn out to be the one that becomes the new drug. In any case, they calculated that the average drug costs $231 million to produce, counting only costs of research and development, and of getting the drug through the FDA approval process (i.e., not including actually producing the pills, or marketing). The $231 million is in 1987 dollars, which if you adjust for inflation, is equivalent to $350 million 2000 dollars.

In 2003, DiMasi and his colleagues published another paper available here, recalculating the cost based on more up-to-date data. They found that the cost had risen to $802 million (2000 dollars) — in other words, the cost of developing a new drug more than doubled in 13 years.

Needless to say, the issue is much more controversial now than it was in 1991, and anti-drug-company lobbyists came up with some rather silly criticisms of the methodology. Alex Tabarrok of George Mason University debunks one of the big criticisms here.

Now, Alex points out a paper by two other economists, Christopher Adams and Van Brantner, have tried to validate their data in a new paper entitled, “Estimating the Costs of New Drug Development: Is it really $802m?” Their answer is … no, it costs more. As in, between $839 million and $868 million (in 2000 dollars).

In other news, the new survey by the Pharmaceutical Research and Manufacturers of America (PhRMA) finds that drug company research spending in 2004 is up 12.6% over 2003, to $38.8 billion.

That’s a lot of money.

References:

  • DiMasi, Joseph A., Hansen, Ronald W., Grabowski, Henry G., Lasagna, Louis, 1991. “Cost of innovation in the pharmaceutical industry,” Journal of Health Economics 10:107–142.
  • DiMasi, Joseph A., Hansen, Ronald W., Grabowski, Henry G., 2003. “The price of innovation: new estimates of drug development costs,” Journal of Health Economics 22:151-185.
  • Adams, Christopher and Brantner, Van V., “Estimating the Costs of New Drug Development: Is it really $802m?” (December 2004). http://ssrn.com/abstract=640563

5 Responses to “How much does it cost to invent a new drug?”

  1. Dave Schuler Says:

    Sounds like the effects of Gammon’s Law to me.

    Any idea of how the total cost breaks down between actual research and company-borne approval costs?

    And, of course, the percentage increase in research spending would be even better news if it were rising as a percentage of sales (it isn’t). It’s rising as a percentage of expenditures.

  2. Different River Says:

    Dave Schuler writes:
    And, of course, the percentage increase in research spending would be even better news if it were rising as a percentage of sales (it isn’t). It’s rising as a percentage of expenditures.

    Why would it be better news if research spending were rising as a percentage of sales? That would imply that a dollar of research spending was producing less sales, that is, that research is becoming less productive or less efficient. Why would that be good?

    As it happens, over the long run research spending is increasing as a percentage of sales — in 1970 it was 12.3% of domestic sales, and it increase fairly steadily up to 18.8% in 2004. There was a “bulge” to the 21% range in 1993-1998, but there have been bulges like that before, so it’s not clear the trend is now reversed. For the figures, see page 50 of the PDF file here.

    The factors affecting drug company research spending are complex. For some recent research on the issue, see this paper.

  3. Dave Schuler Says:

    The numbers that I’m looking at “R & D as a percentage of sales” (from the link you supplied on the PhRMA site) are

    1997 21.6
    1998 21.1
    1999 18.2
    2000 18.4
    2001 18.3
    2002 18.4
    2003 17.7

    That looks like a downwards trend to me. But you’re right it is complex.

    Why would it be better news if research spending were rising as a percentage of sales?

    Maybe I’m thinking shallowly but profits and executive salaries seem to be going up pretty quickly in the pharmaceutical industry and I’d like to think that they were plowing more of those profits back into the core business.

  4. Dave Schuler Says:

    Oh, and thank you very much for the link to the paper. It’ll take me a while to digest it.

  5. Different River Says:

    On the “R & D as a percentage of sales”:

    Yes, it’s down from the 1993-1998 bulge, but there is a general upward trend over the long term. And, the 2004 number is a substantial increase from the 2003 number.

    Do you have any figures on profits and executive salaries? You said earlier that R&D was increasing as a percentage of expenditures. Executive salaries are part of expenditures. The only way R&D could be increasing as a percentage of expenditures by decreasing as a percent of sales would be if non-R&D expenditures were dropping (or increasing slower than R&D), and sales were increasing faster than R&D. That would imply huge increases in the efficiency of non-R&D operations, like manufacturing and marketing (i.e., more sales per dollar spend on marketing and manufacturing). That would mean that (for example) we should be seeing fewer drug ads, but they should be more effective. Do we?

Leave a Reply

Powered by WordPress