Eli Lilly: the Evista Project

Eli Lilly: The Evista Project®

In early 1998 Dr. August “Gus” Watanabe, executive vice president of science and technology for Eli Lilly and president of Lilly Research Laboratories (see Exhibits 1 and 2), looked out his office window toward downtown Indianapolis. He was contemplating the future commercialization path for Lilly’s new, potential blockbuster drug, Evista®, which had received FDA approval on December 9, 1997, for the prevention of post-menopausal osteoporosis. Evista®, generically known as raloxifene hydrochloride, would be entering the estrogen replacement market, a market that had worldwide sales in excess of $1 billion in 1997.

Of even wider significance was the fact that in initial trials, Evista® appeared to lower the incidence of breast cancer and reduced total LDL in post-menopausal women without the negative side effect profiles of currently available estrogen replacement therapies. The potential of this new therapeutic and its impact on Lilly could be enormous. Some analysts predicted that Evista® might become a $1B drug for the company. With this in mind, Watanabe knew that the decision on how best to commercialize Evista® would have a profound effect on Lilly’s well-being. Should Lilly follow its traditional approach to commercialization? Or should Lilly follow a course more in line with the development approach adopted for Evista® in early 1995, which would require the organization to transform its heavyweight product development team into a focused product (commercialization) team? In a senior management meeting later in the day, Watanabe would have to make a recommendation.

Company Background

Eli Lilly and Company was incorporated in 1901 to pursue the drug manufacturing business founded in Indianapolis, Indiana, in 1876 by Colonel Eli Lilly. A pharmaceutical chemist and a U.S. Civil War veteran, Colonel Lilly founded the company to improve the quality of medicinal products by introducing scientific methods into the product development process.

In the 1920s, Lilly launched its first major blockbuster drug—insulin. Developed with the research team of Frederick Banting and Charles Best of the University of Toronto, the new therapy revolutionized the treatment of diabetes. The insulin product family and the treatment of endocrine diseases remained a focal point of Lilly’s research throughout the twentieth century. In the 1980s, Lilly introduced Humulin®, the world’s first human-health-care product created by using recombinant DNA technology. In 1997, insulin products accounted for nearly 13% of Lilly’s worldwide revenues, with sales of more than $1 billion.

In addition, Lilly had developed a number of important new anti-infectives in the 1950s, including long-acting, orally administered penicillin products and erythromycin, which was the first of its kind in a major new class of antibiotics called macrolides. It also developed, during that decade, the first agents of the cephalosporin antibiotic class and, subsequently, Ceclor®, the world’s top-selling antibiotic. As with therapeutics for endocrine disease, anti-infectives remained an important focal point in 1997, representing 15% of Lilly’s worldwide revenues, with sales of almost $1.3 billion.

With the launch of Prozac® in 1986, Lilly established itself as a leader in the development of drugs for central nervous system and related diseases. In 1996, Lilly launched the blockbuster drug Zyprexa®, generically known as olanzapine. Zyprexa® was approved by the Food and Drug Administration (FDA) on October 2, 1996, as a new anti-psychotic agent for the treatment of schizophrenia. In its first 12 months on the market, Zyprexa® accumulated sales of $550 million, making it the most successful first-year drug product, ever.

In addition to development of anti-infectives and therapeutics for central nervous system and endocrine diseases, Lilly, since its restructuring in 1994(discussed below), had focused the remainder of its drug development efforts on fighting cancer and cardiovascular diseases. In the early 1990s, Lilly received FDA approval for ReoPro®, a cardiovascular, anti-platelet agent, and Gemzar®, an oncology product. In 1997 sales of ReoPro® and Gemzar® were $254 million and $175 million, respectively.

The Drug Development and Approval Process

The creation, approval, and commercialization of a new therapeutic drug in the United States comprised four stages—discovery, development, registration, and launch (commercialization stage), a process best described as long, complex, and risky (see Exhibit 3).

In the discovery stage, large numbers of molecules were screened for therapeutic efficacy. For just one approved drug, pharmaceutical companies often had to screen upwards of 10,000 molecules. After a molecule passed this initial screening, it faced extensive toxicological and animal tests. If the molecule passed these tests, the pharmaceutical manufacturer applied to test the compound in humans, and development was started.

In the development stage, the pharmaceutical manufacturer’s first step was to file an Investigational New Drug (IND) application with the FDA. The IND was necessary to begin testing in humans. A typical IND might be more than 2,000 pages long and include information about pre-clinical testing and a description of proposed clinical trials. Unless the FDA ordered a hold, clinical trials were permitted to start 30 days after the application was filed upon final approval by a review board created at the institution or institutions where the trials were being conducted. Of the drugs entering clinical trials between 1980 and 1984, only 23.5% were expected to become marketed drugs.

Clinical trials (part of the development stage) were conducted in three phases. In Phase I, safety studies were conducted on 20 to 100 healthy volunteers. Potential side effects were identified, and a dosage range was determined. Phase II trials determined the drug’s effectiveness. Approximately 100 to 300 volunteers who had the targeted disease participated. Phase III trials, used to monitor the development of adverse reactions to long-term use, typically involved 1,000 to 3,000 patients (and sometimes thousands more) in clinics and hospitals. Patients were closely monitored during this stage to assess the drug’s efficacy and safety. Following the completion of Phase III trials was registration.

In the registration stage, a company compiled all information from these trials and, if the data successfully demonstrated safety and efficacy, they were submitted as a New Drug Application (NDA) to the FDA. Containing all the scientific information the company had gathered, NDAs typically were 100,000 pages long and required large-scale and increasingly lengthy and complex efforts to prepare for submission (see Exhibits 4 and 5).

The commercialization stage officially began with the FDA’s approval to market the drug, although pharmaceutical companies started preparing for launch and commercialization several months before the expected approval date. The most important aspect of the FDA’s approval was what “indication” would be permitted on the label, i.e., which disease(s) the drug could be used to treat and whether the drug had proven efficacy in the prevention of the disease(s) or in its treatment. Simply put, the indication on the label, to a large extent, determined market potential.

R&D in the Pharmaceutical Industry

Research-based pharmaceutical companies spent 19.4% of sales on R&D in 1995 (see Exhibit 6), a figure suggesting the importance the pharmaceutical industry placed on discovering and commercializing new therapeutics. And while R&D had always been critical, the advent of formularies in both the public and private sectors magnified the importance of productivity in R&D.

Formularies were instituted in both the private and public sector as a way of managing health care costs. A formulary was a list of prescription drugs that health care insurers would reimburse a patient for. Formularies added to the downward price pressure put on pharmaceutical manufacturers in four ways.

First, wherever possible, only prescriptions for generic substitutions of branded drugs were reimbursed. Second, health care insurers negotiated a bulk discount rate from the pharmaceutical manufacturer for any drug listed on their formulary. Third, even if more than one drug might have the same therapeutic effect, only one would be listed on the formulary for reimbursement. For example, there were at least three drugs a physician could prescribe for the treatment of allergies, any one of which had essentially the same therapeutic effect. Therefore, the health care insurer had significant leverage to lower price by pitting one pharmaceutical manufacturer against another in a price war and approving only one of the three drugs for reimbursement. Fourth, health care insurers as well as independent pharmacies sought to maintain the fewest number of drugs as possible on their formularies in order to lower their own inventory overhead costs. “Me-too” drugs (branded non-generic equivalents) were removed from formulary lists and inventories because they were seen as unnecessary.

The result was even more pressure on R&D organizations to develop radically new, differentiated, and more efficacious therapeutics. Otherwise, a newly developed drug would not be placed on a formulary after being approved by the FDA. At the same time, R&D expenditures were increasing (see Exhibit 7). The cost of bringing a new therapeutic to market had risen rapidly over the past two decades (see Exhibit 8); major drivers of development costs, including the number of required clinical trials and patients in each trial, had doubled.[1] During the 1990s, the average time required to develop a drug increased to 15 years (see Exhibit 9), thereby increasing the cost of capital needed for R&D. By 1997, the cost of developing a new therapeutic had risen to more than $500M.

Patent Protection

The combination of high development costs and long development times made intellectual property protection essential. Patent protection for the pharmaceutical industry covered the therapeutic molecule itself as well as, in many cases, the process by which the molecule was manufactured. Patents offered research-based pharmaceutical companies the protection they needed to recoup their R&D investments and generate returns.

Patents were particularly effective in the pharmaceutical industry vis-à-vis other industries because they were granted for the molecule itself, with the benefit of the molecule being specific to its structure and atomic composition. Thus, changing one bond in the chemical structure of an anti-cancer therapeutic might change it from being a powerful tool used to fight disease into a useless placebo or, worse, a lethal toxin.

While patents offered pharmaceutical companies the ability to protect their intellectual property, a finite patent life also implied a tremendous pressure to develop and commercialize new therapeutic drugs as quickly as possible. Longer development times directly consumed the company’s profits by reducing its drug’s exclusivity in the marketplace.

During the 1980s, the profitable lifetime for drugs in the United States—the time companies had to recoup development costs and generate returns for stockholders and funding for R&D—significantly shortened. This was due to the passage of the Hatch-Waxman Act of 1984, which allowed for the quick approval of generic copies of branded therapeutics, once a patent had expired.

Hatch-Waxman allowed generic manufacturers to circumvent the safety and efficacy trials conducted by the original patent holder simply by referencing the safety and efficacy data in the originator’s patent. If the manufacturer could demonstrate that the generic had the same active ingredients as in the original drug, it could avoid retesting, thereby saving five to eight years in time to market and considerable cost. For research-based pharmaceutical companies, the increased pressure to recoup R&D investments, due to a shorter time of exclusivity—12 years as opposed to 14 to 17 years pre-Hatch-Waxman—was substantial.

First-Mover Advantage

The creation of a new therapeutic class meant monopoly rents for the manufacturer. In 1986, Lilly received FDA approval to launch Prozac®, its first blockbuster drug for central nervous system and related diseases. Prozac® was the first of its kind in a new and important therapeutic class of drugs called Selective Serotonin Re-uptake Inhibitors (SSRI), used for the treatment of depression. Prozac® revolutionized the treatment of depression by its improved efficacy and a greatly reduced side effect profile. The drug was an instant success, reaching sales of $96 million in 1988, $313 million in 1989, and $645 million in 1990.

Not until 1992 did the FDA approve other SSRIs. That year, Pfizer launched Zoloft® and SmithKline introduced Paxil®. By this time, however, Lilly had established its hold on the marketplace, with Prozac® sales topping $1B. Lilly maintained the lion’s share of this market. In 1997, Prozac® sales, of almost $2.5B, accounted for over 30% of Lilly’s worldwide revenues, and Prozac® represented 44% of worldwide sales in this therapeutic area; Zoloft® and Paxil® sales stood at $1.4 billion each.

Drug Development at Lilly

When Gus Watanabe assumed leadership of Lilly Research Laboratories (LRL) in 1994, one of his primary goals was to improve the product development process. After serving as chairman of the department of medicine at Indiana University, Watanabe had joined Lilly in 1990 as vice president in charge of cardiovascular research. During the next several years he assumed increasing responsibility and in 1994 was promoted to executive vice president of science and technology, responsible for all of Lilly’s research and development efforts.

Watanabe and other LRL senior managers were acutely aware that product development deadlines were not being met. Indeed, timelines for product development milestones were moved back monthly, if not quarterly. “Timelines didn’t mean very much,” affirmed one researcher. “It was almost expected that you wouldn’t make your deadlines.” Indeed, from 1985 to 1995, Lilly had launched only two major new therapeutics: Humulin® in 1986 and Prozac® in 1987. In an industry where R&D productivity was critical, this poor performance was a major issue.

Not surprisingly, this performance had taken its toll, particularly in income growth from 1991 to 1995 (see Exhibit10). In addition, from early 1992 to mid-1993, Lilly lost $18 billion in market capitalization.[2] Combined with this dry spell in R&D productivity was the onslaught of discontinuous change in the competitive landscape. Many governments – including those of Germany, Spain, Italy, and the United Kingdom—imposed policies affecting the pricing and usage of medicines. Concurrently, the United States began grappling with health care reform.

In 1992 and 1993, Lilly restructured to better align itself with the changing marketplace. In mid-1993, Randall Tobias was appointed chairman and chief executive officer; in January 1994, Lilly announced that it was separating its medical devices and diagnostics businesses from its core pharmaceutical business to “better focus its resources on global pharmaceutical operations and further maximize shareholder value.” As a result, in June, Guidant Corporation was formed, consisting of five of Lilly’s nine medical device and diagnostics businesses; in December, 20% of Guidant’s stock was offered to the public, with Lilly retaining an 80% equity share. Also in 1994, Lilly purchased PCS Health Systems, a managed-pharmaceutical-care company.

Lilly was determined to focus its efforts on its core pharmaceutical businesses, expand its reach in this sector, and enter an elite group of top-tier pharmaceutical companies. To that end, on June 5, 1995, Lilly announced that it planned to sell its remaining equity in Guidant.

Function-Based Product Development

As LRL’s senior management searched for causes of Lilly’s product development problems, they encountered a number of disturbing facts—not the least of which was that it was almost impossible to diagnose specific reasons for delay in detail because product development resources were shared across all of LRL, with little accountability on a project-by-project basis. Furthermore, there were no systems to track resources and measure progress except at the macro level.

Staffing, as well, was a serious issue. Project managers were generally assigned four to five projects to manage, with functional team members also assigned to four or five projects. This might seem an efficient use of resources theoretically, but in practice a team member could rarely, if ever, focus on a particular project long enough to be effective. Thus, someone would be called to fight a fire on one project while simultaneously being responsible for important work on another. Once again, assigning accountability, tracking individual performance, and relating that to project progress were all extremely difficult.

Also contributing to significant delays was the fact that teams often met only monthly to decide important project issues. Compounding this problem was the absence of detailed project plans, which would have forced team members to plan for future uncertainties, determine internal and external interdependencies, and forecast future potential road blocks.

The Development of Evista®

On January 13, 1995, Watanabe announced the formation of two heavyweight product development teams “to expedite the successful, timely introductions of selected drug candidates”—a turning point in Lilly’s drug development history. Each team (one each for Evista® and Zyprexa®) was to “focus exclusively on a single compound during Phase III clinical trials.” These dedicated, co-located, and cross-functional groups represented a significant departure from Lilly’s traditional approach to drug development, whose management was characterized by the loose coordination of contributing functions. Effective product development required the integration of specialized capabilities. Integration was difficult in most circumstances, but was particularly challenging at Lilly, with its strong functional groups, extensive specialization, large numbers of people, and multiple, ongoing operating pressures.[3]