[Pharm-policy] R*D Cost, by component

[James Love]

The following is a stylized presentation of a typical "cost of
drug development" study.  I'm working at home without access to
some papers.  The data are loosely based upon the DiMasi et al,
1991 study and the 1993 OTA report, but are not exact.  For
example, I left out the animal studies, and I don't think all the
other numbers are quite right, because I'm doing some of this
from memory.  Joseph DiMasi is on this list, can correct the
numbers if we wants to.  This is mostly presented to show people
how the costs are allocated, so it isn't such a black box.
                              
The OTA asked Joseph DiMasi to recalculate his 1991 study
numbers, with the assumption that capital costs were 14 percent
for the pre-clinical stage, and from 14 to 10 percent for other
expenditures (as a linear function).  Joe had used the following
data for average time from start of a development phase to the
NDA, and the phase length, and from that I have calculated the
cost of capital figures:


               Start to  Phase     mid-point Cost of
               NDA       Length    to NDA    Capital

Preclinical    11.8      3.6       10.00     14.00%         
Phase I        8.2       1.3       7.55      13.02%    
Phase II       6.9       2         5.90      12.36%
Phase III      5         3         3.50      11.40%
NDA Review     2.5       2.5       1.25      10.50%         
               


Here I just took the 1991 data on clinical trials and increased
them by 12 percent (OTA used 1990 numbers and Joe used 1987
numbers, and I didn't bother to check the CPI), and I assumed the
out of pocket pre-clinical expenses were roughly 1 percent of the
risk adjusted pre-clinical expenses (Joe didn't break them out
separately, and I'm open to suggestion as to how this should have
been done.)  I have presented four columns.  The "out of pocket"
column is for the outlays on a particular drug.   Most of this is
for the clinical trials.  The column labeled "risk" is the "cost"
of unsuccessful trials, before capital costs.  The column
"capital" is the cost of capital for both the successful and the
unsuccessful trials, using the variable discount rates given
above.  

I have two tables, one that has dollars, and one that has
percentages.  Table 2 is probably the more useful.  In this
example, the out of pocket costs are about 6 percent of the
total.  The risk adjustment, when you throw in the big pre-
clinical numbers, are about 27 percent of the total.  And the
cost of capital is 67 percent of the total -- twice as much as
actual outlays, even after adjustments for risk.  By stage of
development, pre-clinical expenditures are about 78 percent of
the total, after risk and capital costs are included.  

There are several things one can say about this type of analysis.
First, if you were to use a discount rate of something close to 9
percent, the cost estimate would fall by nearly one third, on
this adjustement alone.

Second, if a drug company acquires a product when it is ready for
clinical trials (such as d4T, for example), 78 percent of the
costs are already gone.  

Third, very few policy makers understand how to interpret capital costs.

How close are these numbers to reality?  Well, there are lots of
problems with the data.  As we know from the Orphan Drug Tax
Credit data, the cost of clinical trials can be very small, for
some drugs.  The assumed 12 years from preclinical to NDA is
pretty far off for some classes of drugs, such as HIV/AIDS drugs,
for example, where the average time from patent application to
NDA approval is only 4.4 years (Particularly considering that in
the example above, 67 percent of total costs are capital costs). 
And we don't have any real data on the pre-clinical stage (which
is nearly 80 percent of the total), and this is particularly hard
to measure in areas like cancer, HIV/AIDS and other illnesses
where the government support for R&D is so large.

If people find it useful, I'll redo this using the right numbers
from the 1993 OTA study, and run it by Joe to make sure we can
agree on the numbers.  


    Jamie



                             Table 1
                      Allocation by dollars
                              
               out of                        
               pocket    risk      capital        total     
                                        
Pre-Clinical   0.700     70.000    191.401        262.101

Phase I        2.390     8.002     15.791         26.182    
Phase II       4.428     5.413     9.732          19.573    
Phase III      14.337    6.144     9.404          29.886    

               21.856    89.559    226.327        337.741   
     


                            Table 2               
                   Allocation by Percentages 
                                        
                                        
               out of
               pocket    risk      capital        total     
                                        
Pre-Clinical   0.21%     20.73%    56.67%         77.60%    
                                        
Phase I        0.71%     2.37%     4.68%          7.75%     
Phase II       1.31%     1.60%     2.88%          5.80%     
Phase III      4.25%     1.82%     2.78%          8.85%     
          
               6.47%     26.52%    67.01%         100%

James Love, Consumer Project on Technology    
P.O. Box 19367        | http://www.cptech.org 
Washington, DC 20036  | love@cptech.org       
Voice 202/387-8030    | Fax 202/234-5176