Adrienne Ng is a JD candidate at Osgoode Hall Law School.
Donald Light and Rebecca Warburton (hereafter LW), in February 2011, published a paper entitled “Demythologizing the high costs of pharmaceutical research”. The paper suggests that the US$802 million price tag of pharmaceutical innovation estimated in a 2003 report is inflated. A more sound estimate would be US$43 million.
In 2003, Joseph DiMasi, Ronald Hansen and Henry Grabowski (hereafter DHG) from the Tufts Centre for the Study of Drug Development (Tufts CSDD) published a paper entitled “The price of innovation: new estimates of drug development costs.” In this paper, DHG estimate that the cost to discover and bring a new drug to market (in 2000) is US$802 million. They define a new drug as a “new molecular entity (NME),” and not a reformulation or recombination of existing drugs. To get this figure, DHG took into account drug synthesis, animal testing, clinical trials, testing for safety and effectiveness and cost of capital (applied over the time it takes for clinical trials and regulatory review). This figure, while widely cited by both government officials, drew doubts from other economists. Chief among those are Donald Light and Rebecca Warburton (hereafter LW). DHG’s 2003 study sparked the beginning of a somewhat constant academic debate with LW that has continued to this day. Since 2003, LW have published two pieces criticizing DHG’s study, and DHG have issued published replies each time.
In February 2011, LW’s past criticisms were re-stated in yet another paper entitled “Demythologizing the high costs of pharmaceutical research.” A recent Globe and Mail article gives us a “Coles Notes version” of this paper. A few of LW’s most recent criticisms are as follows:
The category of “new drugs”
The $802 million represents an estimate of the R&D cost of the costliest one-fifth of drugs, not all drugs. Most new products are a reformulation or a recombination of existing drugs and only 22% of drugs developed by pharmaceuticals are NMEs.
Cost of capital
The cost of capital depends on the length of time for a drug to get to market and a standard annual return on investment (ROI). LW claim that DHG have inflated the time estimate. According to the U.S. FDA Register, drug trials generally take 36 months (instead of 72 months) and regulatory review generally takes 12 months (instead of 18 months). As well, the ROI should be 3% and 5% in Canada and the United States respectively, and not 11% as used in the original calculation.
Generous tax treatment
LW claim that generous tax treatment for companies involved in research development (R&D) were left out of DHG’s original calculation. They state that half of corporate R&D expenses are paid for by taxpayers over the long term.
LW note that pre-clinical basic research is often performed in universities and government-sponsored labs and thus the number of months devoted to such research or it’s cost is impossible to calculate. Yet, DHG allocates 52 months to pre-clinical research, which is ultimately included as one-third of their USD$802 million estimate.
Use of mean vs. median
LW claim that DHG’s use of a mean cost instead of a median inflates their estimate.
Towards the end of the article, LW come up with their own estimate of the cost of drug development. They conclude that R&D costs companies a median of US$43 million per new drug.
Throughout the article, a theme running in the background (though not explicitly stated) is that it is perhaps in DHG’s best interest to inflate the cost of pharmaceutical development. DHG are often referred to as “company supported analysts”. This much is true. According to the Tufts CSDD website, the centre receives unrestricted grants (which represent approximately 40% of the centre’s operating expenses) from pharmaceutical and biotechnology firms. But what does that really mean?
The on-going debate between LW and DHG is perhaps an example of academic discourse at its finest. What is surprising, however, is that LW do not cite either of DHG’s past replies in their 2011 paper, even where there are particularly relevant and valid rebuttals. I encourage you to take a look at these published replies here and here. More recently Tufts CSDD has issued an official response to LW’s 2011 paper, which can be accessed on their website here. I encourage you to take a look at this too.