Biotech and Pharma

In this case study:

• Brief History of the Biotechnology and Pharmaceutical Industry
• Nuclear and Pharmaceuticals: An Industry Comparison
• Role of the State
• Importance of Intellectual Property
• Regulator Comparison
• Lessons Learned for Nuclear

The modern pharmaceutical and biotech industries look very similar to traditional nuclear reactor development in the early research and development phases. Both industries rely on early research and development in the public sector—at research universities and national laboratories, in the case of nuclear. To bring a new product to market, both industries spend significant amounts on development and making their way through stringent licensing processes.

However, the market structure and business models for the two industries are quite different. For pharmaceuticals, while the upfront development costs are large, manufacturing costs are almost trivial, leading to significant and immediate profits once a drug is approved. For nuclear, the development costs are similar, but having a design approved is only the beginning. The real proof of concept comes in the construction and operation of reactors, a process that can take decades. Intellectual property also plays a much larger role in the pharmaceutical industry, with larger firms frequently buying out start-ups for their patents. Smaller firms, as a result, can focus on proving the science of their product without worrying about longer-term business models.

The regulators of both industries—the Food and Drug Administration (FDA) and the Nuclear Regulatory Commission (NRC)—each increased their stringency in response to major failures in their respective sectors: the thalidomide crisis for the FDA, and Three Mile Island for the NRC. Increased regulation at the FDA following the thalidomide crisis caused significant consolidation across pharmaceutical firms, as only the largest firms could afford the newly required staged clinical trials. There is, however, a significant philosophical difference between the two regulators. While pharmaceuticals must be proven safe and effective, the FDA also recognizes the large benefits to public health that new drugs bring, and it plays a secondary promotional role for the industry as a result. For nuclear reactors, the technology is regulated purely with the aim of mitigating harm, and nuclear power is treated as a commodity with no recognizable benefits to public health. If there were a single agency regulating the public health impact of coal, gas, and nuclear, this outcome might be different.

The major lessons that the pharmaceutical industry has to offer nuclear apply at the intermediate stage of development, after basic research, when small start-ups are developing their products and undergoing the first stages of licensing. For the nuclear industry, there should be more support for taking new technologies from the university lab to start-up companies. Nuclear also likely needs a staged licensing process, or at least more transparency and finite timelines for decisions. Smaller firms might need to focus more on intellectual property as well, to make them more attractive for acquisitions by large incumbents with the capital to move designs through development and licensing. But most importantly, the agencies overseeing nuclear development—the Department of Energy (DOE) and the NRC—need to more explicitly recognize and promote the benefits of nuclear power compared with other energy sources—namely, clean air and low-carbon, reliable, affordable power.

Brief History of the Biotechnology and Pharmaceutical Industry

The pharmaceutical industry, not unlike the nuclear industry, emerged as a byproduct of World War II. Most of America’s large pharmaceutical companies today originally grew out of the postwar boom, particularly as a result of the state’s demand for penicillin.¹ After experiencing rapid growth in the 1950s, the industry was upended by the thalidomide crisis in 1962. The drug was developed in Germany and marketed as a cure for a wide range of conditions, but was particularly useful for treating morning sickness and sleep problems in pregnant women. Prescribed in 46 countries, the drug was consumed more commonly than aspirin in some places. Unfortunately, it took several years to discover that thalidomide was the cause of severe birth defects in over 10,000 children, more than 40% of whom died before their first birthday.² The United States was one of the only developed countries that did not approve the drug, thanks to a pharmacologist at the FDA who was skeptical of the drug’s safety and repeatedly asked the manufacturer for better evidence and more studies.³

Although the drug was never approved in the United States, the existing regulatory review process was incredibly lax, with no clear methodology for evaluating supporting evidence. To remedy this situation, Congress passed the Kefauver-Harris Amendment in 1962, which included such common-sense measures as a “proof of efficacy” requirement that is still the basis of the drug approval process today.

The immediate effect of this regulatory change was industry consolidation. With the massive investments and long lead times required to bring new drugs to market, only larger firms such as Pfizer, Merck, and Johnson & Johnson had the necessary capabilities to compete.

Concurrently to industry consolidation, the seeds of the biotech industry were sowed. Biotechnology, at least in pharmaceuticals, is distinguished by its reliance on the use of genetic engineering to produce new compounds. A series of breakthroughs throughout the 1970s culminated in the first genetically engineered human insulin in 1982, the fruit of collaboration between Caltech and the first biotech firm, Genentech. This paved the way for the emergence of other small start-up biotech firms in the 1980s, typically only composed of a few successful scientists.⁴

While the number of pharmaceutical companies remained constant during the 1970s, the 1980s was a period of rapid growth driven by the emerging biotech field.⁵ Though the industry itself was evolving, little changed in terms of output, not only because drug development takes a long time, but also because biotech research was still a relatively small part of total pharmaceutical R&D. Roughly 20 new molecular entities (NMEs) are approved each year (though year-to-year variation is large),⁶ and by 1988, only 5 had actually come out of biotech research.⁷ By the end of the 1990s, however, the FDA had approved more than 125 biotech drugs.⁸

As biotech grew, the dominance of the largest pharmaceutical firms fell. From 1950 until the early 1980s, the 15 largest pharmaceutical companies were responsible for roughly 75% of NMEs per year; today, their share has stabilized around 35%.⁹ Generally, large pharma has been good at generating successful follow-on approaches (70% of follow-on approaches come from large pharmaceutical firms) but much less successful in novel treatment options. Biotech has taken up much of this slack. From 1998 to 2008, biotech companies have been responsible for nearly half of new drugs with novel mechanisms (a subset of “especially innovative” NMEs) and 70% of orphan drugs in the pharmaceutical industry.¹⁰ Biotech had also increased its share of blockbusters—drugs whose annual sales exceed $1 billion—from 8% to 22% by 2007.

The biotech-pharma networked model suggests that smaller firms can play a key role in industry innovation.

For the last few years, anywhere between 5 and 10 biotech drugs have been approved,¹¹ and perhaps more significantly, biotech has been growing at roughly twice the speed of the pharmaceutical industry as a whole (roughly 10% per year since 2009 versus 6%).¹²

The emergence of biotech also led to a steady increase in inter-firm collaboration, whether in the form of joint research projects, strategic alliances (where one firm does one part of the process, and the other another), or mergers and acquisitions. By 2000, roughly 25% of corporate-financed pharmaceutical R&D came out of joint ventures, 3 times as much as in 1990 and 20 times as much as in 1980.¹³ From 1990 to 2010, mergers and acquisitions deals increased fivefold. As a subset of the industry, biotech is especially reliant on external forms of collaboration, with nearly half of biotech R&D funding coming in the form of partnerships (since 2009 at least).¹⁴

Since most biotech firms are small and rarely have the means to take their drugs all the way from preclinical trials to commercialization, collaboration is a necessity for them. For larger pharmaceutical companies, the growing interest in external collaboration can be traced back to a paradigm shift in research methodology.

With the tremendous advancements in genetic engineering, chemistry, and computational biology, the 1980s opened up the possibility of rationalizing the drug development process.¹⁵ Prior to that period, companies principally relied on “random screening” to identify promising compounds for drug development.¹⁶ This process involved testing a large number of chemicals to determine how they interacted with the targeted disease. By the 1980s, this process had run into diminishing returns, and the industry switched to a “guided search” model,¹⁷ thanks to the advances in computation and genetics. As a consequence of this shift, large research labs and capital were no longer as important to drug research, and the value of genetic expertise increased, bolstering the comparative advantage of highly specialized biotech firms.

According to Gottinger and Umali (2011), this paradigm shift wasn’t in itself sufficient to push large pharma toward more collaboration. Initially at least, pharmaceutical companies were convinced they could do much of this guided research themselves.¹⁸ But the early and remarkable success of Genentech changed this perception. Its first two blockbuster drugs, Humulin and hGH, were developed and commercialized in close partnership with larger pharmaceutical companies. Genentech started working with Eli Lilly in 1978 and jointly created Humulin, approved in 1982 and a blockbuster a few years later.¹⁹ Genentech also worked with the giant Japanese firm Kabi starting in 1977 on hGH, which was approved in 1985. Genentech’s success helped shake up the pharmaceutical industry and encouraged other big players to seek out partnerships with biotech firms, which they did in much larger numbers from the early 1990s onwards.

While there is no set model for how these partnerships work, some broad trends can be observed. One common pattern is for biotech to focus on the early stages of new drug development, especially preclinical trials. If the early signs are promising, they will enter a licensing agreement with a larger firm. Another option is for the larger firm to simply acquire the start-up, a pattern that is very common. Just last year, large pharmaceutical firms spent a few billion dollars buying the rights to biotech drugs or the companies themselves.²⁰ At this stage, it is rare for any single company to have invented, tested, and commercialized an NME solely internally.

Though again the setup varies from one case to the next, many biotech firms preserve much of their independence even once they have been bought. They maintain their separate offices, research direction, and internal structure, taking on the role of “centers of excellence.”²¹ Generally, biotech firms boast human capital better equipped to harness the cutting edge of research, as the majority of their employees have PhDs and strong ties to leading research universities.²²

The nuclear industry needs more support in turning new technologies into start-up companies.

From the perspective of biotech firms, there is little doubt that this collaboration has been beneficial. Apart from Genentech and a few others, very few biotech firms have actually succeeded in bringing their inventions to market. Reviewing the evidence from the 1990s, Owen-Smith and Powell (2004) find that network ties are a significant predictor of performance in biotech.²³ More recently, Munos (2009) finds that acquisitions lead to a 120% increase in NME output for small companies.²⁴ Interestingly, increased collaboration or consolidation between large firms isn’t as strongly associated with increases in NME output.

Since the approval of NMEs is actually relatively rare, most partnerships and external collaborations don’t actually lead to a new drug approval. In most cases, they are formed in the hope of achieving the next milestone on the long road toward drug approval. Indeed, the average R&D alliance in biotech lasts less than 4 years, whereas the drug development process takes closer to 10–12 years.²⁵ The fact that large pharma has been willing to bet substantial sums on biotech ventures still far away from commercialization has played a key role in driving the growth of the latter, and has allowed smaller firms to secure funding (often from VCs) despite having a product that is still many years from commercialization.²⁶

In sum, there has been a very clear shift toward a more networked innovation model over the last couple of decades in the pharmaceutical industry. This shift has been driven by three factors: the relative fall in the dominance of large pharma, a shift in the research paradigm, and the emergence of small research-focused biotech firms. However, this three-part explanation is incomplete, as it gives insufficient credit to one of the main drivers of this transition: the state.

Nuclear and Pharmaceuticals: An Industry Comparison

The parallels to the nuclear industry are obvious. In both sectors, development of new products stretches over multiple years (or even decades) and requires significant upfront investment. Both commercialization processes also require significant investment. However, pharmaceuticals are a much larger industry than nuclear in the United States, comprising 23% of all private R&D in 2013. Pharmaceuticals add over $1 trillion to the US economy every year.²⁷ But where the pharmaceutical industry spends over $40 billion on R&D annually,²⁸ the US nuclear industry spends under $500 million. The pharmaceutical industry releases new blockbuster drugs every year, while the nuclear industry is struggling to deploy its first new designs in 30 years.

Role of the State

The pharmaceutical industry has always been closely interwoven with the state. The state’s wartime demand for penicillin created the industry, the Kefauver-Harris Amendment drove consolidation, and the DNA and genetic research conducted in university labs across the world in the 1970s kick-started the biotech revolution. If these contributions are widely recognized, the state’s role post 1980 is perhaps less well known, but no less critical, especially for the development of biotech.

First, the state enacted a suite of legislation to boost the industry. The widely known Bayh-Dole Act of 1980 allowed research sponsored by the National Institutes of Health (NIH) to be patented. The slightly less-well-known Stevenson-Wydler Act of the same year required publicly funded research institutions to form technology transfer offices and to do more to make their research available to businesses. In 1983, the Orphan Drug Act was passed; its aim was to encourage the development of drugs for relatively rare diseases, through generous tax credits, funded research, extended IP, and FDA fast track. The Orphan Drug Act played an especially important role in supporting the nascent biotech industry. By the early 2000s, 90% of the revenue of the four biggest biotech firms—Genentech, Biogen, Idex, and Serono—came from drugs that benefited from the Act.²⁹

Through the NIH, the state also provided direct research funding for breakthrough drugs. NIH funding increased significantly between the mid-1980s and the mid-2000s, increasing at an average of 2.9% per year. Average spending in the 1980s was $10 billion, compared to $35 billion today.³⁰ By 2000, NIH funding accounted for over half of nondefense public R&D, up from 30% in the mid-1980s.³¹ This research proved remarkably effective for the biotech industry—Vallas et al. (2011) estimate that 13 out of the 15 blockbuster biotech drugs that were on the market in 2007 benefited from NIH funding in the early stages. Equally influential was the Small Business Innovation Research (SBIR) program; set up in 1982, it was tasked with funneling federal dollars to R&D in small businesses, many of which were in the biotech industry.³²

The combination of direct funding, subsidies, and supportive regulation was clearly a major factor in explaining the growth and success of the US biotech industry. In fact, the broad range of policy support gave the US biotechnology industry a unique advantage over its international competitors. While other countries also provided direct funding for genetic research, they lacked complementary institutions like the Orphan Drug Act or SBIR, which inhibited the growth of their own biotech industries.³³

Drug Development Costs

Estimates of the cost of new drug development aren’t easy to pin down. Certain drugs can cost little more than a hundred million dollars to develop, whereas others can exceed the billion-dollar mark. What’s more, there is no universally agreed method to estimate drug development cost.³⁴ One widely cited study out of Tufts University estimated the average cost of drug development to be $1.4 billion in 2016, a number that includes the cost of failures as well as success. Tufts estimates the opportunity cost of investing in drugs to be $1.2 billion, which is an estimate of forgone return during the 10–12-year investment period. Tufts’s estimate has been the subject of considerable controversy, especially since they haven’t released the raw data from which these numbers are derived. Still, a review of these debates strongly suggests that most drugs end up costing in the hundreds of millions to develop. As a comparison, NuScale will spend about $45 million to have the NRC review their license application, and the process will take 3 years.³⁵ However, Nuscale has spent closer to $450 million and taken 10 years to get to the point of license application with the NRC. ³⁶

The Tufts study is perhaps more useful to get a sense of how the cost of drug development has evolved over time. The center has been estimating the cost of drug development since 1987, and estimates have been on a consistently upward trajectory, even after adjusting for inflation. This increase in costs can be explained by two main factors: the increasing number of failures, and the increasing cost of post-approval R&D (phase IV), from complying with foreign standards or testing for previously unobserved side effects.³⁷

Drug Development Timeline

In terms of timeline, the range is slightly smaller than it is for cost, with most NMEs taking between 8 and 12 years to develop. Tufts estimates it takes roughly 128 months to get a drug from synthesis to approval but “only” 96 months from the beginning of clinical trials to FDA approval (provided the drug gets that far). Each phase of the clinical trials typically takes one to two years (and gets a little longer with each phase). The majority of drugs go through three pre-approval stages: stage 1 involves testing the drug for any adverse effects on humans (50–100 participants), stage 2 tests the drug for actual effectiveness (100–300 participants), and stage 3 tests its effectiveness on a much larger pool of patients (1,000–3,000). If a drug successfully passes each of these stages (and the probability that it will increases with each), the firm will submit a New Drug Application (NDA) to the FDA. If the drug is classified as a priority (like an orphan drug), the roughly 100,000-page application will be reviewed in 6 months. If it isn’t, the process takes between 10 and 12 months.

However, the FDA isn’t solely involved in the final approval of the drug; it also works with drug companies to create a schedule for the trial phases, as well as the set of criteria used to evaluate the drugs following each trial stage. To even begin a stage-1 trial, firms have to submit an Investigational New Drug Application (IND) to the FDA. This IND must contain detailed information on animal pharmacology (a large part of preclinical trials involves testing the drug on nonhuman species), the process of drug manufacture, and proposals for the clinical protocols to be followed. This initial FDA review typically engages a range of external experts (often in universities) to help review and refine the trial protocols. This period is also when a firm will receive its patent (and the 20-year exclusivity period begins then).

In contrast, the NRC’s review of license applications can take several years—3 years in the case of NuScale. What then follows is a 5–10-year construction process before the plant starts generating revenue.

Importance of Intellectual Property

If a drug successfully clears each of these hurdles, firms can finally commercialize it, and use their patent protection to charge far above the marginal cost of production. While new drugs can be priced at a little more than a few hundred dollars, others can easily cost thousands if not hundreds of thousands of dollars per prescription. Since the marginal cost of drug production is low,³⁸ the firms stand to make a substantial profit and recoup much of their investment.

Clearly, this profit is almost entirely contingent on regulation. The majority of drugs are easy to copy, and were it not for patent protection, other firms would quickly offer cheaper alternatives. The last few years has proven to be a very stark illustration of this fact, as many drugs lost their exclusivity. For instance, Pfizer’s Lipitor, an anti-cholesterol drug, was the world’s top-selling drug for eight years but lost market protection at the end of 2011. In 2012, its sales ranking dropped to 14th, a nearly 61% decline in revenue in one year—from $12.9 billion to $5.1 billion. Bristol-Myers and Sanofi’s Plavix, a blood thinner that was number 2 in sales in 2012 at $9.5 billion, dropped to number 12 the next year, to $5.2 billion. Patents on Plavix expired in several European countries in 2011 and in the United States in 2012. Reviewing a range of drugs in the early 2000s, Conti and Berndt (2016) estimate that prices fall between 30% and 50% in their first year following the loss of exclusivity and even more after that.³⁹ Though sales of the protected drug will fall, total sales of the drug (i.e., adding up branded and generic) tend to increase, as does revenue, consistent with a typical supply-demand story.

If IP is the main profit driver, pharmaceutical companies do also spend substantial sums on marketing. In fact, frequent media reports suggest they spend more on advertising than on R&D. Those figures, however, should be taken with a grain of salt since advertising spending is often bundled up with management and operational expenses.

A meta-study from Johns Hopkins found that total marketing expenditure for the industry was $31 billion in 2010.⁴⁰ Total marketing expenditure actually peaked in 2004 but has stayed fairly consistent at 8–10% of sales throughout 2000–2010. Most of that marketing money is directed toward physicians and doctors, but there has been an increase in direct-to-consumer spending in the last few years, around $4–5 billion per year. To put these numbers in context, FierceBiotech estimated that industry spent $67 billion on R&D in 2010, or twice as much. And indeed, Statista’s estimate of US pharmaceuticals R&D as a percentage of revenue between 1990 and 2010 hovered around 16–20% in 2015.⁴¹

Regulator Comparison

The FDA’s 2014 budget was $4.7 billion, $2.6 billion of which came from the public purse, and the remaining $2.1 billion came from user fees. This 50-50 split has been fairly consistent ever since the Prescription Drug User Fee Act of 1992, which allowed the FDA to charge fees to drug manufacturers.⁴² By contrast, the NRC has a budget of $1.1 billion, 90% of which is funded by user fees (the change was part of a wider effort by then-President George H. W. Bush to cut the government deficit).⁴³

The FDA charges a fixed fee for regulatory review: an IND cost $449,500 in 2013, and an NDA for an NME cost $5 million (if a new drug is a “me too” drug rather than a new molecular entity the fee is reduced to $1.5 million). As mentioned above, the FDA’s review process rarely exceeds a year and is often only six months. By contrast, the NRC has no fixed fee—it charges between $170 and $270 per hour—or timeline by which to deal with new applications. According to their estimates, a new reactor design certification takes five years (and the early site permit review takes three), but its most recent approvals suggest it takes a lot longer than that for innovative designs.

Two designs recently approved by the NRC are the AP1000 and ESBWR. The AP1000 was submitted in 2002 and approved in 2011; similarly, the ESBWR was submitted in 2005 and only approved in 2014. Though reliable estimates are difficult to find, the process seemed to cost in the region of $500 million in both cases. NuScale has reported that they’ve already spent $130 million in preparation for submitting their license to the NRC.

The agencies overseeing nuclear development need to more explicitly recognize and promote the benefits of nuclear power.

Another key difference between the FDA and the NRC is their mission. While the NRC’s primary objective is public safety—“to ensure adequate protection of public health and safety”—the FDA has a twin mission: to protect human health by ensuring the safety of new drugs and to advance public health by speeding up the drug innovation process. In other words, part of the FDA’s mandate is to accelerate the commercialization of new technology.

In fact, some go so far as to say that the FDA is too amenable to new drug proposals and, indeed, the FDA approved roughly 90% or more of submitted drugs last year, a rate that has been steadily growing from a low of 60% in 2008. The Senate is currently debating the Republican-sponsored 21st Century Cures Act, which would both lower the bar for approval and speed up the process. Opponents of the bill are concerned it will lead to a lax regime and see a repeat of past scandals, as when, in the late 1990s, two FDA-approved drugs (Vioxx and Avendia) had to be removed following the post-approval discovery of serious side effects.

These debates notwithstanding, it is clear that the FDA is far more amenable to new technologies than the NRC has ever been. If achieving the perfect balance between sufficiently diligent yet pro-innovation regulation is a continuous process, comparing the NRC and the FDA suggests the former has the balance wrong.

Lessons Learned for Nuclear

At a superficial level, the parallels between the nuclear industry and biotech are obvious. The timelines are long and the development costs are very high. One key difference, though, is the marginal cost. Building a new nuclear plant is like building a cathedral: building the second one is only marginally cheaper than building the first, unlike a pill. The development of a blockbuster drug is actually more similar to the building of a plant, since once it’s built the O&M costs are low. Since the marginal cost to new builds is very far from zero for nuclear developers, it would suggest the appetite for innovation is weakened. If the reward for innovation is approximated as the difference between marginal cost and marginal benefit, then ceteris paribus, a higher marginal cost reduces the size of the reward and thus the motivation to pursue it.

A related difference between pharmaceuticals and nuclear is that, for nuclear, older designs’ value doesn’t collapse as soon as the IP expires. A proven design, even if it is fairly antiquated, can still be a winning strategy, especially in countries with underdeveloped technical infrastructure. Rosatom building four of its VVER-1000s in Kudankulam, India, is a case in point. For drugs, there is no significant benefit to using a brand instead of a generic, and indeed, certain drugs will only ever make it to developing countries once a generic version exists. Even if marketing can help expand the life and profitability of drugs, its effectiveness is limited. Again, the imperative for innovation is less critical to survival for nuclear developers than for pharma companies.

Keeping this in mind, the biotech-pharma networked model does suggest that smaller firms can play a key role in industry innovation, despite being incapable of fully realizing that innovation on their own. It’s also worth noting that large corporate players initially took a lot of convincing to adopt this model. It was only when a few genetically engineered compounds gained commercial viability that larger firms realized that (a) genetic engineering would be key and (b) smaller biotech firms were best equipped to handle these new advancements.

To get to that point, it’s important to recognize the critical role of policy. Whether it was in the form of direct NIH support to genetic research, the Orphan Drug Act, or favorable regulation, Genentech and other early biotech companies benefited from a tremendous amount of public support. And even today, the FDA’s systematized, transparent, and relatively efficient process plays a critical role in facilitating collaboration: every time a biotech company successfully completes a different stage of the drug development process, it boosts its chance of getting acquired.

While the market structure of building nuclear power plants will never look like the high-profit pharmaceutical industry, its early-stage RD&D could shift to be more innovative like the pharmaceutical industry’s did in the 1970s and ’80s. Universities should create more support mechanisms to spin off research into private companies, such as incubators, seed funding, and tech transfer programs. Small nuclear firms should focus more on intellectual property to make their companies more attractive for acquisition. Following the pharmaceutical sector, nuclear companies should invest more in joint ventures and collaborative R&D to solve shared challenges and demonstrate emerging technologies. The NRC may need to develop a staged or phased licensing process. But more importantly, the NRC should incorporate more transparency and offer strict timelines for application review and decisions. Finally, DOE and the NRC should explicitly acknowledge the benefits of nuclear power as compared to the alternatives. DOE should make the case for investing in commercialization of new nuclear designs from a public health perspective, the same way the FDA does for orphan drugs.