Injecting molecules from a sea slug that received tail shocks into one that didn’t made the recipient animal behave more cautiously.
Peer review varies in quality and thoroughness. Making it publicly available could improve it.
Costs fo Drugs: Discovery-ApprovalDiscovery• Target identification• Target validation• Lead identification• Lead validationPre-Clinical Studies• In vitro validation• In vivo validationPre-Clinical Studies• In vitro validation• In vivo validationClinical Trials• Phase I trials• Phase II trials• Phase III trials• Regulatory affairsPercentages represent the total cost associated with that activityThe process of discovering new
October 25, 2004|
• Target identification
• Phase I trials
The process of discovering new drugs and getting regulatory approval is becoming ever more costly. At present, the total turnover of the world's pharmaceutical industry is roughly $350–$400 billion, while the cost of R&D is estimated at about $45 billion per year. In comparison, R&D investment was $5 billion in 1986 and $30 billion in 2001. In spite of the significantly larger spending on R&D, the total number of new chemical entities (NCEs) approved by the US Food and Drug Administration has not shown a commensurate increase. The number of NCEs licensed in a given year has remained relatively constant over the years, varying from 20 in 1986 to between 25 and 30 during 1999–2001. Simple arithmetic therefore suggests that the cost of discovering a new drug, defined as the ratio of the total R&D expenditure per year divided by the number of NCEs licensed in that year, has climbed from about $250 million in 1986 to nearly $1 billion today. Of this amount, roughly half, or $400–$500 million, represents the actual expenditure, while the remainder represents the opportunity cost; that is, the cost of locking up one's capital over a long period.
The pharmaceutical industry is among the most profitable in the world, with a collective profit margin of about 20%. In contrast, the "sunrise" information technology (IT) industry shows a profit margin of about 6%–8%. One could argue that the pharmaceutical industry does not yet have the same incentive to reduce costs as do other industries, such as banking, insurance, and even software. In most Western countries, the revenues of pharmaceutical companies come primarily through reimbursements from various insurance programs. Due to changing demographics, these insurance schemes are becoming more and more difficult to sustain. Thus, despite their impressive profit margins, pharmaceutical industries are finally realizing that they need to reduce costs.
One way to reduce costs occurs during the drug discovery phase. While "drug discovery" is used both to represent the broad activity ending with regulatory approval, as well as the narrow activity ending with a validated lead, in this article the phrase is used in the narrow sense, to span the following four activities: target identification, target validation, lead identification, and lead optimization. These activities are estimated to contribute about 15% of the total cost of discovering an NCE.
Courtesy of Mathukumalli Vidyasagar
Since the annual R&D expenditure of the pharmaceutical industry is estimated at $45 billion, it would seem that the total market for those offering discovery services is 15% of this amount, or roughly $6 billion per year. However, since roughly half of the R&D expenditure consists of the opportunity cost, the actual market for those wishing to enter the discovery arena is about $2–3 billion per year.
Once a drug has received regulatory approval, a pharmaceutical company has significant postregulatory activity, including manufacturing, sales and marketing, distribution, and postapproval regulatory affairs. If this postregulatory activity is excluded, the range of activities of a pharmaceutical company consists of three broad categories (see Figure).
Increasingly, pharmaceutical companies are restricting themselves to just the clinical trials category, and vacating the first two activities. The discovery phase and the preclinical phase are increasingly being filled by small, agile companies, which develop promising molecules, validate them through in vitro or animal testing, and then license them to pharmaceutical companies for a combination of up-front payments and future payments against milestones. This trend has serious implications for the discovery process, since small companies are often short of cash, and need to come up with optimized (or at least identified) lead compounds as quickly and as cheaply as possible.
At present, lead identification consists of testing every chemical compound or molecule in one's library (often numbering in the millions) against the therapeutic target, which is typically a gene or a protein. This is a relatively expensive process known as high-throughput screening (HTS). Using modern computational methods, it is possible to adopt a "structure-based" approach to lead identification, consisting of: determining the three-dimensional shape of the therapeutic target, using informatics methods to weed out all but a few hundred possible leads, synthesizing minute quantities of only the chosen compounds and then testing them against the therapeutic target, and determining the 3-D structure of the interaction between the lead compound and the therapeutic target.
In principle, a structure-based approach can save both time and money compared to the usual HTS approach. Time is saved because only a few thousand compounds are tested against the target as opposed to all compounds in one's library, while money is saved because most of the compounds are eliminated through computer simulations, or in silico computations.
Since the structure-based approach makes heavy use of informatics and computer simulations, the "entry barrier" in terms of capital investment is much lower for this approach than for HTS. Given that India has already established itself as the destination of choice for IT-related services, this also suggests an opportunity for the Indian biotechnology community to offer drug discovery services.
To explore this possibility, it is necessary to address several issues, including the skill set required for a company to offer structure-based drug discovery services, its potential customers, the cost savings and appropriate "onsite-offshore" model for a company to offer this service, and effective communication. Each of these issues is addressed in turn.
While the structure-based approach is less costly than HTS, the skill set required is actually much broader. Hence a company (Indian or otherwise) aspiring to offer this service needs an eclectic team consisting of crystallographers, medicinal chemists, computational chemists, toxicologists, and experts who specialize in protein structure prediction and analysis.
Potential customers for this kind of service would be small to medium biotech companies that have limited cash reserves, and are under severe pressure to produce lead candidates as quickly (and as cheaply) as possible. Using an Indian service would increase their chances of doing so within the available budget. Large pharmaceutical companies might not be interested in this kind of service until the pressure to cut costs increases.
The potential cost savings in contracting work to India is always a function of the onsite-offshore mix. The leading Indian IT companies aim for a mix of 40% on site and 60% offshore. With this mix, the cost savings to the overseas customer are between 40% and 50%. These are only estimates, however, and actual savings can be computed only after some experience has accumulated.
Effective communication is perhaps the trickiest point. A major reason for the success of IT outsourcing has been that an appropriate vocabulary has evolved for describing the job precisely. Once the SRS (software requirement specifications) and ATP (acceptance test plan) have been mutually agreed upon, both the customer and the vendor know exactly what is expected from each side. Thus it is possible for an India-based service provider to minimize the onsite presence (mostly for finalizing the SRS and going through the ATP), and do most of the work offshore.
In the pharmaceutical area, no such vocabulary exists. While drug discovery is inherently less straightforward than writing "pure" software, it is also true that necessity is the mother of invention, and that the evolution of a vocabulary in other domains (banking, insurance, retail, etc.) has been spurred by the need to cut costs. Until pharmaceutical companies feel similar pressure, the onsite ratio in this domain will be much higher than the 40% in conventional IT applications, and the cost savings lower than 50%.
Two major factors could prevent India from realizing its potential in this area: a hesitation on the part of potential customers to release intellectual property to India, and a shortage of skilled personnel. As for the first point, India has a reputation of respecting the treaties it signs. Once India starts adhering to the World Trade Organization patent regime as of January 2005, India will start adhering to a product patent as opposed to the current process patent regime (
As for the second point, many training programs already exist in bioinformatics, but they are of highly uneven quality. India has a greater need for broadly trained life scientists with some knowledge of bioinformatics, rather than narrowly trained diploma holders. In the short term it may be desirable for the government of India to introduce some kind of accreditation program, as it did earlier in the area of IT. The IT accreditation scheme contributed significantly to assuring a steady supply of well-trained IT professionals, and the aftermath is clear for the world to see.
As India overcomes such obstacles, one can expect in the near future that the country will become the destination of choice for drug discovery outsourcing, just as it is today the undisputed leader in IT outsourcing.
Mathukumalli Vidyasagar is an executive vice president of Tata Consultancy Services, India's largest IT firm. He received BS, MS, and PhD degrees in electrical engineering from the University of Wisconsin. The author or coauthor of nine books and more than 130 papers, he has held visiting positions at MIT, UC Berkeley, UCLA, CNRS in Toulouse, France, and Tokyo Institute of Technology.
He can be contacted at