[Ip-health] Novartis outlays on Gleevec R&D, and the 2003 DiMasi estimate of R&D costs

Jamie Love james.love at keionline.org
Fri Apr 5 07:09:07 PDT 2013

Given the number of queries I have had about the Novartis R&D outlays
on Gleevec, I thought it might be useful to highlight a few of the
comparisons between what Novartis did for Gleevec for its May 2001 FDA
approval, and what DiMasi's 2003 paper (the $802 million drug
development cost paper) data suggest companies do to register a drug.

1.  Pre-Clinical Research

DiMasi assumed the company financed the pre-clinical research
themselves.  In DiMasi's work, this is a big number in his final cost
estimates. DiMasi estimates that the (risk adjusted) out-of pocket
costs of pre-clinical research were $121 million per approved new
chemical entity, and $335 million after taking into account the
opportunity cost of capital.  Note that this is 42 percent of DiMasi's
$802 million number.

In the case of Gleevec, most (but not all) of the pre-clinical
research was funded by the NIH, the Leukemia Society and other
non-profit donors.

Bottom line:  Novartis doesn't get to claim it paid for all of the
pre-clinical research, because they didn't.

2.  Phase I tests

DiMasi assumed the company financed the Phase I tests.

In the case of Gleevec, the NIH financed the first Phase I test, and
co-funded subsequent Phase I tests.

Bottom line:  Novartis doesn't get to claim it paid for all of the
Phase I tests, because they didn't.

3.  Phase II tests

DiMasi assumed the company financed the Phase II tests.

Novartis did in fact finance the Phase II tests.  There were three,
involving 1028 patients.

Bottom line:  Novartis paid for Phase II testing involving 1028
patients.  Now do the math.   Depending upon per patient cost
assumptions, this can be estimated at $10 to $24 million, with the
later assuming DiMasi's estimate (which was more than twice the BMS
oncology group estimated per patient cost).   If anyone thinks
Novartis spent more than $24 million on 1028 patents in Phase II
tests, explain what the money went for.  Buying the patients rollex
watches and diamond rings?

4.  Phase III tests.

DiMasi and others note that is is the most expensive phase of clinical
testing, and they figure big in his estimate of there being an average
of 5303 patients in clinical trials for an NME FDA registration.

Novartis was not required to provide any Phase III tests to the FDA to
register Gleevec in 2001.

Bottom line: For the 2001 registration of Gleevec, the cost of Phase
III tests for Novartis were ZERO.

4. Risks of Failures

I used the DiMasi risks of failure to adjust the out-of-pocket costs.
The Phase II out-of-pocket outlays were multiplied by 3.3.

5.  Opportunity cost of capital

According to DiMasi, the opportunity cost of capital was HALF of the
cost of drug development:

      5.6. Total capitalized cost per approved drug

      Our full cost estimate is the sum of our preclinical and clinical
      period cost estimates. Our base case out-of-pocket cost
      per approved new drug is US$ 403 million, while our fully
      capitalized total cost estimate is US$ 802 million. Time costs,
      thus, account for  50% of total cost. (page 166).

DiMasi assumed an 11 percent real return for the opportunity cost of
capital.  He also assumed that clinical testing began 90.3 months
before product registration.

I used the DiMasi assumption of an 11 percent real return cost of
capital.  But the overall opportunity costs of capital were estimated
at 11 to 20 percent, rather than 50 percent.  Why?

DiMasi assumes the company financed pre-clinical research, when most
of this was funded by the NIH and the Leukemia society.

For clinical testing, the time from the start of testing to to FDA
approval was just 35 months from the first Phase I trial, which the
NIH funded, and the Phase II trials that Novartis funded probably took
place in 2000, for a product registered in May 2001. (Novartis is
welcome to tweak these dates if they have better ones).

Concluding comments

The financial press has become lazy in reporting drug development
costs for particular drugs, frequently repeating vague industry
assertions, or quoting a number from couple of industry funded
consultants reports that may or may not have any relevance the facts
or context for the specific case they are discussing.   Within the PR
framed narrative, all drugs are above average as regards drug
development costs, regardless of any facts related to what actually

Most people reading these estimate compound the errors by mistaking
figures that are adjusted for risks of failures and the opportunity
cost of capital for estimates of outlays that account for neither.

Ironically, the estimates that are most based on facts and actual
events are the least believable to the financial press (and many
academics), because they seem so much different from the narrative
they they repeat, but never take the time to understand to examine as
regards its veracity or relevance.

James Love.  Knowledge Ecology International
http://www.keionline.org, +1.202.332.2670, US Mobile: +1.202.361.3040,
Geneva Mobile: +41.76.413.6584, efax: +1.888.245.3140.

More information about the Ip-health mailing list