Reducing The Real Cost of Medical Innovation

One thing is certain. Many people will be deconstructing Matthew Herper’s two articles in Forbes on the high cost of drug development. Herper estimates that the cost of developing a new medicine from scratch is, on average, about $5 billion and can be as high as $13 billion. Either way, that’s a lot of money and a lot higher than previous estimates, which have risen over time.

Just how accurate are these estimates? There are many reliable measures and sources of drug development costs, such as the study from the Office of Health Economics, that show it takes about $1-2 billion to bring a new drug to market. That’s a lot of money too. Matthew was apparently dramatizing the fact that drug development costs are staggering and that lower costs would benefit all. He suggests that non-profit advocacy groups can help pre-qualify new drugs before they enter clinical trials and that companies could do a lot better in killing new projects before they go too far into development. While it’s not the first time these points have been raised, I happen to agree with them and will elaborate on these ideas in a later post.

Viewing new drug costs in context

The chart below compares the total cost of bringing a new drug to market with an approximation of the relative value of that cost each year since 1990. Taken into consideration are adjustments in the value of that investment plus the average inflation rate. While these figures are estimations, they illuminate how drug development costs have increased and why.

For most years since 1990, it would appear that the cost of a new drug equaled an average increase in relative value plus inflation. Then, in 2005 the cost of developing a new drug increased much faster than the relative value plus inflation.

This is NOT an estimate of what a new drug is worth. It’s not even a real estimate. Instead, it’s a way to highlight the fact that drug development costs have increased dramatically since 2005. And it’s a way of suggesting that the increase has something to do with changes in the drug development process.

2010 estimates, as do all previous estimates, take into account the opportunity and direct costs of prior years of investment. And direct costs not only include what is spent on existing projects but on future projects as well.

Meanwhile, the opportunity cost of drug discovery, relative to previous generations, has declined. Genomics and high-powered computing has allowed companies to pinpoint treatments that might work in specific patient groups with greater efficiency. So why have costs shot up?

Drug development: breaking bad, getting worse

While companies are becoming better at increasing the number of drugs in early development, the percentage that makes it across the finish line has been dropping. That is, companies have not gotten better in breaking off the bad bets from the good ones. In fact, they might have gotten worse.

Here’s what former Eli Lilly Senior VP of Research, Steve Paul notes: “Encouragingly, recent benchmark data on Phase I WIP (work in process) across the industry indicate that most companies have begun to substantially increase investments in the earlier stages of drug discovery; this is reflected by the number of candidates entering Phase I trials, which has increased significantly.^{9, 17, 18} However, based on the benchmark data, for most companies, the number of NMEs entering clinical development and progressing to Phase II and III are still insufficient to achieve 2–5 launches per year⁹; this reflects many years of operating at WIP levels below what would be required in the earlier stages of drug discovery and development. Thus, inevitable pipeline gaps will arise (as they have) and given the CT (cycle time) of the process (see chart below), such gaps cannot be filled quickly through traditional means.”¹

The model defines the distinct phases of drug discovery and development from the initial stage of target-to-hit to the final stage, launch. The model is based on a set of industry-appropriate R&D assumptions (industry benchmarks and data from Eli Lilly and Company) defining the performance of the R&D process at each stage of development. See details

Lower costs call for a break from tradition

The difference between the anticipated rise in drug development costs and what we see now could possibly be a result of two factors:

1. The existing approach to drug development — within self-contained large corporations — is becoming more expensive to maintain at a time when networks of scientists, entrepreneurs, and individuals are emerging as the source of potential new products.
2. The emergence of parallel tracks in drug development — one is based on earlier technologies and another is based on biomarkers and other tools that allow companies to screen out medicines unlikely to work in specific groups. By way of analogy, drug companies are using both mainframes and tablets to do computing for the same project.

Let’s eliminate resource redundancies

Pharmaceutical companies are trying to still do it all when a more networked or crowd-sourced approach might make more sense. In effect, each company is reluctant to parcel out aspects of the development process to more agile enterprises, and instead, they create their own individual versions.

Dr. Paul concurs: “The R&D enterprise…is predominantly 'owned', operated and fully controlled by a given company (Fully Integrated Pharmaceutical Company or FIPCo)…” instead of one “that is highly networked, partnered and leveraged (Fully Integrated Pharmaceutical Network or FIPNet). Traditionally, large pharmaceutical companies have pursued the discovery, development, manufacture and commercialization of their medicines largely by owning and controlling each component. In part, past reliance on the FIPCo model was as much a necessity as a choice. Today, however, the opportunity to partner virtually all elements of R&D through a coordinated and global network or FIPNet could (if effectively managed) substantially improve R&D productivity by affordably enhancing the pipeline from early discovery through to launch. A FIPNet will theoretically allow greater access to intellectual property, molecules, capabilities, capital, knowledge and, of course, talent.”²

This is part of the solution and one to which Herper alludes. But there’s one more reason for increased costs that Herper does not address — the rising cost of complying with FDA regulations and the additional cost and uncertainty of developing information to convince health plans and health systems to cover new treatments after they receive FDA approval. See the costs of delayed access.

These are the biggest obstacles to innovation, and they will be discussed in my next post. Start your own discussion. Share this article.

• 1. How to improve R&D productivity: the pharmaceutical industry's grand challenge Steven M. Paul, Daniel S. Mytelka, Christopher T. Dunwiddie, Charles C. Persinger, Bernard H. Munos, Stacy R. Lindborg & Aaron L. Schacht
• 2. Lindgardt, Z., Reeves, M. & Wallenstein, J. Waking the giant: business model innovation in the drug industry. In Vivo 26, 1–6 (2008).
• 9. Mathieu, M. P., ed. Parexel's Bio/Pharmaceutical R&D Statistical Sourcebook 2008/2009. (Parexel International Corporation, Waltham, 2008).
• 17. Booth, B. & Zemmil, R. Prospects for productivity. Nature Rev. Drug Discov. 3, 451–456 (2004).
  Booth and Zemmil framed the issue of productivity in the pharmaceutical industry and provide a qualitative discussion of some of the drivers and potential solutions.
  Article
• 18. Hu, M., Schultz, K., Sheu, J. & Tschopp, D. The innovation gap in pharmaceutical drug discovery & new models for R&D success. [online], (2007).