[Ip-health] The Office of Health Economics study

Jamie Love james.love at keionline.org
Mon Nov 17 03:22:38 PST 2014

On Sun, Nov 16, 2014 at 12:40 AM, Dimasi, Joseph A. <Joseph.Dimasi at tufts.edu
> wrote:

>  What exactly was confusing to you about our paper?  I resent your slur
> about intent.  You have no basis for it, and it certainly isn't helpful.

To answer, "What exactly was confusing" about the 2003 paper, I would say
that the following would have made the paper both more accessible, and less
prone to misinterpretation by policy makers and journalists.

1.  To have made the paper more accessible, the assumptions regarding the
adjustments for risk of failures could have been presented in a manner
similar to the BIO paper published in nature, or some of your own published
work.  I thought that the assumptions you made about risks for clinical
trials were a useful and solid part of the paper, but not everyone was able
to go from the probability of entering a Phase (Table 1, page 162) and the
"final clinical success rate" of .215 percent, to get a risk of failure for
investments at that stage, and this could have been presented differently,
like it was in the Nature paper, IMO.  This is true particularly since you
are inclined to jump all over people if there is even a minor quibble in
sorting out the math to make it more accessible.

2.  The costs presented in the paper are really a set of cost elements,
which are often confused by people, and this is particularly true for the
term "out of pocket costs." The NIH abstract on your paper says "The
estimated average out-of-pocket cost per new drug is 403 million US dollars
(2000 dollars)."   http://www.ncbi.nlm.nih.gov/pubmed/12606142.  Lots of
people think this means that companies spent $403 on average in cash
outlays on the development of a new drug, and this is before adjustments
for the risks of failures.  But this is a completely false reading of the
paper, which uses the term "out of pocket" 54 times, and both to describe
outlays before risks are taken into account (Table 1), and outlays adjusted
for risk adjusted (for clinical at bottom of page 165, for pre-clinical on
page 166).    Lots of people read the $403 million in "out of pocket" costs
cited by the NIH abstract to mean that a company risked $403 on projects
with long odds of success, when the adjustments that give you $403 means,
with 100 percent odds of success.   I'm sure your consulting clients and
you both recognize that this misunderstanding leads to a significant
over estimation of the costs, in the minds of many influential readers.  I
look forward to the new paper to see how this is addressed in the

3.   Many people thought the paper was estimating a meaningful average for
all drugs.  Your paper does say that this is not the case, as it excludes
products developed by smaller firms and licensed-in products, but even
then, the sample is a selection of drugs, for an industry where the
variance in costs are huge, even for big companies.   Here it would be
useful to provide more details of the underlying data that gives you an
average of 5303 patients in trials, showing, for example, the distribution
of numbers for patients in trials.  This also could have been compared to
other data, such as the number of patients in medical reviews for NMEs, or
subsets of drugs, such as orphans or cancer drugs, where the pricing issues
are particularly concerning.  If your sample is irrelevant to the drugs
where the disputes over pricing is most acute, well, this is something that
could be made more clear to the non-technical readers.

4.  There could have been a better and more clear explanation to what was a
historically high estimate of per patient costs in trials, including how
the issue of overhead, when compared to direct costs, was done.

To answer " I resent your slur about intent.  You have no basis for it, and
it certainly isn't helpful," I will respond as follows.

1.  I assume that the big drug companies that are among your consulting
clients have interests in exaggerating or confusing people about drug
development costs, particularly when a company like Novartis triples the
price on Gleevec and gets more than $35 billion for a drug they spent less
than $100 million (risk adjusted) to register, to mention only one example
highlighted recently by 60 minutes, and an example highlighted also by
Brian Druker, the NIH funded researcher who won the Lasker prize for his
work on the drug.   I don't think it is controversial to describe the
intentions and interests of the drug company clients.  If your work
advances their cause, and never undermines it, it is reasonable to ask if
your presentations lack the type of balance and fairness that lead to
deeper public understanding, or deeper misunderstanding.

I do consider you a very knowledgeable expert, and one who tells the truth,
assuming the person answering the question knows what to ask, and this is
why we have frequently sought your advice on the work that we do -- we
expect you to identify flaws in our work, and to deepen our understanding,
and your counsel has been useful, as I have told many people.

But if your work is consistently used to justify prices for drugs where the
facts you present are at odds with the realities for those drugs, or when
your work is wildly misquoted (how many times do people quote the $802
number as a figure out-of-pocket costs before adjustments for risk or
capital costs?) and this is never cleared up, then people, including me,
will assume their is an element of advocacy, on behalf of your clients.


James Love.  Knowledge Ecology International
KEI DC tel: +1.202.332.2670, US Mobile: +1.202.361.3040, Geneva Mobile:
+41.76.413.6584, twitter.com/jamie_love

More information about the Ip-health mailing list