[Ip-health] New estimate of R&D per approved drug - $1.7 billion

[Mike Palmedo]

http://www.bain.com/bainweb/PDFs/cms/Marketing/rebuilding_big_pharma.pdf

The link above is to an article titled "Rebuilding Big Pharma's Business
Model." The whole article is a good read, but what really caught my
attention was their estimate of $1.7 billion in R&D spending per new
drug approved. There's a sub-article on "The Rising Cost of New Drugs,"
the text of which is cut and pasted below. There are four graphics in
the above PDF that convey some of the info that seems to be missing
from the text.

To come up with this figure, the authors (Bain & Co.) used their own
data. The reader does not get to see this data, and there is no
explanation of where this data comes from, except that the time period
is 1997-2001 (more recent than the Tufts time period). They also include
costs incurred after approval. Then there are a bunch of little
adjustments that must add up, like a slightly longer development time.
The actual methodology is not laid out.

----------------------------------------------------------------------
The Rising Cost of New Drugs

Jim Gilbert, Preston Henske and Ashish Singh
Published in In Vivo Business and Medicine Report
November, 2004

Industry estimates peg the cost of bringing a chemical entity to market
at about $900 million, including post-launch studies. Based on recent
performance data, however, the true cost is nearly twice as high=97closer
to $1.7 billion per successful launch, when you also include average
launch costs of $250 million. The former estimate derives from data from
the period 1983 to 2000. Analysis of more recent data from 1997 to 2001,
taking into account both direct and indirect costs, indicates that
performance has declined substantially.

This higher total cost, combined with lower average margins and shorter
exclusivity periods, translates into single digit average returns on
investment: about 5% for an average compound. Statistical simulations
suggest that there is only a one in six chance of a new compound
achieving a return on investment of 12% or more.

One major reason for increased costs and lower ROI is a dramatic decline
in productivity. Only one compound now reaches the market for every
thirteen discovered and placed in preclinical trials, compared to one
for every eight between 1995 and 2000. (See Exhibit 2.) Attrition has
been particularly severe in Phase III development. Average development
costs per compound have increased from $131 million to $200 million,
while the chances of each compound receiving approval has fallen from
73% to 59%.

Further downstream, the commercial side has seen a similar decline in
productivity. Physician details have become almost twice as expensive,
evidenced by the drop in sales representatives=92 productivity of nearly
50% over the past seven to eight years.

Without a new model, costs will likely continue to rise. The overall
cost of manufacturing and supply operations will grow further, owing to
the increasing expense of regulatory compliance as well as the growing
complexity of the molecules manufacturers are producing. And while costs
rise, mounting payor price pressures and aggressive patent challenges
limit the total revenue potential of the average drug.

In recent years, these productivity declines for self-developed products
have made in-licensing more attractive. Companies have increased their
investment returns by licensing drugs developed elsewhere and putting
them through clinical trials. As price competition for in-licensing of
compounds has sharpened, however, the average expected returns for Phase
III in-licensing have dropped, from 12% for the period between 1995 and
2000 to about 6% today (See Exhibit 3.) Falling success rates of Phase
III trials have also played a role in driving down the expected returns
of in licensed compounds.

In the long term, productivity could improve=97and thus so could the
viability of the blockbuster model. One source of improvement is
scientific: development of more predictive preclinical toxicology
screening could increase success rates and reduce expensive failures in
the later stages of development. Similarly, the increasing adoption of
pharmacogenetic profiling could benefit clinical trial design,
recruitment and outcomes. Another source is technical: Increased
automation of clinical trials plus earlier regulatory involvement could
reduce time to market and total cost. Further still, new IT-enabled
approaches supporting physician, payor and patient sales could reduce
launch costs, increase peak sales and reduce sales and marketing costs.
But all these improvements together are unlikely to yield returns
greater than the industry=92s cost of capital. (See Exhibit 4.)