Eli lilly diabetes case study

9-696-077 R E V : A P R I L 1 5 , 2 0 0 4

________________________________________________________________________________________________________________ Professor Clayton M. Christensen prepared this case. HBS cases are developed solely as the basis for class discussion. Cases are not intended to serve as endorsements, sources of primary data, or illustrations of effective or ineffective management. Copyright © 1996 President and Fellows of Harvard College. To order copies or request permission to reproduce materials, call 1-800-545-7685, write Harvard Business School Publishing, Boston, MA 02163, or go to http://www.hbsp.harvard.edu. No part of this publication may be reproduced, stored in a retrieval system, used in a spreadsheet, or transmitted in any form or by any means—electronic, mechanical, photocopying, recording, or otherwise—without the permission of Harvard Business School.

C L A Y T O N M . C H R I S T E N S E N

Eli Lilly and Company: Innovation in Diabetes Care

“Look at these. Aren’t they beautiful?” asked Larry Ellingson, executive director of Eli Lilly and Company’s Diabetes Care Business Unit, as he showed his visitor a briefcase full of odd looking plastic devices. “They’re pens—insulin pens. All you do, “ he explained as he screwed one apart, “is put a little cartridge of insulin in here; close it up like this; turn this dial to the amount of insulin you need; poke the needle just under your skin (which he didn’t demonstrate); and squeeze this trigger. That’s all there is to it. Then you just put this cap over the needle and put it back in your briefcase, purse, or pocket until your next meal, when you take a shot again. The patients will just love it.”

If history was any guide, Ellingson was right: they would love it. Lilly’s principal competitor in the worldwide insulin business, Denmark-based Novo Nordisk, had introduced insulin pens to the European market several years earlier with great success. Pens were a more convenient way for patients to take insulin. Conventionally, patients carried a separate syringe, inserted its needle into an insulin vial, pulled its plunger out to draw slightly more than the desired amount of insulin into the syringe, flicked the syringe while holding its needle up to dislodge any air bubbles that clung to the walls of the syringe’s cylinder, and then squeezed the plunger slightly to force those bubbles— and some insulin—out of the syringe. Only then could they inject themselves with insulin. This process typically took more than a minute, whereas patients could prepare and administer a pen injection in as little as 10 seconds.

It was early in 1995, and Novo was building a new plant in the United States to produce insulin cartridges for its pens. Ellingson hoped that Lilly’s new line of pens (see Exhibit 1), the result of a multimillion dollar investment, would blunt the advantage Novo had enjoyed with convenience- conscious customers and stabilize Lilly’s share of the worldwide insulin market.

Insulin was an important product for Lilly, one of the world’s largest pharmaceutical manufacturers with sales of over $5 billion (see Exhibit 2). Insulin in fact was Lilly’s second-largest revenue producer after its widely prescribed drug for depression, Prozac.

Diabetes and Insulin

Diabetes is actually two fundamentally different diseases that share a similar set of symptoms: Type I patients produce no insulin, the hormone necessary for cells to utilize glucose, while Type II patients cannot efficiently use the insulin their bodies produce. Type I, also known as juvenile diabetes, usually begins during childhood or puberty. Type II, known as adult-onset diabetes, is manifest later in life (usually after the age of 40) and usually is associated with—and possibly caused

For the exclusive use of c. ma, 2019.

This document is authorized for use only by congyao ma in SPRING 2019 (UNGER) THE INNOVATION PROCESS-1-1 taught by BARRY UNGER, Boston University from Jan 2019 to May 2019.

696-077 Eli Lilly and Company: Innovation in Diabetes Care

2

by—obesity. Digestive processes convert most food into glucose (a simple sugar) and then pass that glucose into the blood as the body’s main source of energy. Body cells are able to burn or metabolize glucose, however, only when there is insulin present, acting as a sort of catalyst for “burning” the glucose. Because they either cannot produce insulin in the pancreas (Type I) or use the insulin they produce (Type II), those with diabetes (hereafter simply called “patients”) can have high concentrations of unmetabolizable glucose in their bloodstream.

Patients need to inject the precise amount of insulin required to metabolize the glucose produced by their digestive system. If they inject too little, the resultant high blood-sugar levels cause a slow deterioration of the body, particularly of the eyes and kidneys. Low blood-sugar levels caused by an overdose of insulin, though, can rapidly precipitate unconsciousness and, potentially, death.

Many Type II patients can treat their condition with oral medications that either cause their pancreases to produce more insulin or enhance the sensitivity of their body tissues to the insulin they naturally produce. Some Type II and all Type I patients, however, must take daily injections of insulin to survive. Insulin cannot be taken orally because it is a protein and would be broken down by the digestive system.

All Type I patients must begin insulin injections immediately upon diagnosis. Most Type II patients, however, progress through several stages. Typically, their diabetes initially is so mild (they can use most of the insulin they produce) that they aren’t symptomatic1—their illness remains undiagnosed until it is discovered during a routine physical exam or in the course of treatment for some other disease. Upon initial diagnosis, many can reduce their levels of blood glucose to normal through a combination of diet, exercise, and weight loss. Many Type II patients subsequently reach a point, however, when they require oral medications. As the disease progresses the oral agents often fail. At that point, Type II patients join the Type I’s in having to take insulin injections.

In 1995 Europe and North America accounted for over 80% of the world insulin market because the rates of diagnosis in those regions were high relative to other areas; because the incidence of obesity was higher; and because a larger proportion of Type II patients tended to be treated with insulin. Approximately 2% of the world population had diabetes, although many remained undiagnosed with Type II diabetes. Of the diagnosed diabetic population, 10% were Type I (this population was increasing by 2% to 3% a year) and 90% were Type II (increasing at 4% to 5% a year).

Early Discovery and Development of Insulin

Until 1921 there was no effective treatment for diabetes. Type I patients could expect to live approximately one year from the time of diagnosis. Primary treatment was a starvation diet, based upon the theory that less food generated less blood glucose and, therefore, prolonged life—slightly. Diabetes wards in hospitals were populated by emaciated bodies, dying from a combination of untreated diabetes and malnutrition. The physical appearance of patients in diabetes wards was later likened to that of prisoners in the Nazi concentration camps of World War II.

1 Three symptoms typically induce a patient with diabetes to seek medical attention: 1) A rapid loss of weight, caused by their inability to metabolize the food they eat (those with diabetes quite literally are starving at this stage, even though they eat larger-than-normal amounts of food). 2) Constant thirst, and abnormally high frequency and amount of urination. This happens because the body’s rate of urination is driven not only by how much water is in the body, but by how much glucose the kidneys secrete. Hence, the body dehydrates. 3) Blurry vision, caused by the osmotic effects of high levels of glucose in the eye on the lens of the eye.

For the exclusive use of c. ma, 2019.

This document is authorized for use only by congyao ma in SPRING 2019 (UNGER) THE INNOVATION PROCESS-1-1 taught by BARRY UNGER, Boston University from Jan 2019 to May 2019.

Eli Lilly and Company: Innovation in Diabetes Care 696-077

3

In 1921, four researchers from the University of Toronto—F.G. Banting, J.J.R. MacLeod, C.H. Best, and J.B. Collip—had begun experimenting on dogs with pancreatic extractions of insulin. By 1922, these researchers were injecting insulin extracted from animals into human patients with miraculous results. Patients with diabetes could finally metabolize their food. Those who watched their nearly instant recovery from starvation likened what they saw to the resurrection of the dead (Exhibit 3). Banting and Best received the 1923 Nobel prize for this work.

The researchers from Toronto needed capital to begin consistent production in large quantities and offered the Indianapolis-based drug maker, Eli Lilly and Company, an exclusive license to produce and sell insulin in the United States. Lilly immediately began commercial development of the product, and by the fall of 1923, 25,000 Americans were receiving insulin Lilly extracted from pancreases of cows and pigs. Half of all Eli Lilly’s profits soon were derived from insulin sales. A number of other companies subsequently began producing insulin in other regions of the world. These included two Scandinavian companies, Nordisk and Novo, and the German chemical giant, Hoechst. By 1995, Lilly and Novo-Nordisk (the two companies merged in 1989) dominated the world market. Hoechst had a significant market position only in Germany.2 The sizes of the world insulin market by region and the major competitors’ market shares, are detailed in Exhibit 4.

Subsequent Improvements in Insulin

Over the next 60 years, Lilly and its competitors improved their insulins along two dimensions. The first was its purity. The second was in its “time profile”—matching the rate at which injected insulin was absorbed into the blood3 with the rate at which glucose was absorbed into the bloodstream.

Purity and the Development of Humulin

Parts per million (ppm) of proinsulin, the impurity which caused the majority of side effects from insulin therapy, dropped from 50,000 ppm in 1921, to 10,000 ppm in 1970, and 10 ppm in 1980. Still, though, all insulin products were derived from the pancreases of either cows or pigs (pork insulin was closer in its molecular structure to human insulin than was beef-derived insulin). While similar to that of humans, animal-derived insulin could never be molecularly equivalent to human insulin. A fraction of a percent of the population with diabetes became resistant to insulin as a result.

In addition to this problem with animal insulins, Lilly feared an insulin shortage caused by a combination of decreased red meat consumption and increased insulin usage. In response to these concerns, Eli Lilly teamed with the biotechnology company Genentech to genetically engineer bacteria that could synthesize and secrete human insulin. The result of this effort, branded in 1980 as Humulin,® represented the supreme breakthrough in insulin product and process technology—a 100% pure insulin that was structurally identical to the insulin healthy humans produced. Lilly invested over $700 million to build the first large-scale biotechnology plant in the world to produce its Humulin.®4 The market responded poorly to Humulin®, though. Consumers resisted its 2 In 1995 there were also a few small producers of insulin in economically less developed countries as well. Because their process technology for extracting insulin from animal pancreases was not sophisticated, many of their insulins were notoriously impure.

3 Patients had to be careful to inject insulin into subcutaneous tissue (under the skin) and not directly into the bloodstream. Doing so would suddenly lower blood sugar, precipitating unconsciousness, brain damage, and possibly death.

4 The pharmaceutical industry’s cost of developing a new drug, and of managing its progress through clinical trials to receive regulatory approval, averaged over $300 million per approved drug in 1995.

For the exclusive use of c. ma, 2019.

This document is authorized for use only by congyao ma in SPRING 2019 (UNGER) THE INNOVATION PROCESS-1-1 taught by BARRY UNGER, Boston University from Jan 2019 to May 2019.

696-077 Eli Lilly and Company: Innovation in Diabetes Care

4

premium price and retailers were reluctant to add yet another set of SKUs to their already crowded refrigerators of insulin products. The substitution of Humulin® for animal insulins occurred at a slow pace, as a result, and it was not until the early 1990s that Humulin® accounted for 80% of Lilly’s insulin volume. “In some developing countries the animal insulins they make have all sorts of impurities in them,” noted Kathy Wishner, one of Lilly’s senior medical executives. “But all markets are getting more sophisticated. Levels of impurities that were accepted a generation ago would not be tolerated today. “Nonetheless,” added researcher Bruce Frank, “In retrospect the market was not all that dissatisfied with highly purified pork insulin.”

The market’s sluggish response to Humulin®, its cost, and Lilly’s high share of the North American insulin market (which often exceeded 85% and made it difficult to generate additional volume through a new product such as Humulin®), all contributed to decreasing support for continued investment in the insulin business at Lilly. Ron Chance, another Lilly researcher, recalled, “At that point in time people were saying, ‘We can’t get any better than Humulin®. And we can’t grow the business. It’s time to do something else.’ As a result, many of us went to other projects, like the human growth hormone.”

Through the 1980s, particularly after Novo-Nordisk introduced its version of bio-engineered human insulin in 1984, insulin became viewed essentially as a commodity product—the products of Lilly, Novo and Hoechst were essentially identical in purity and efficacy. Nonetheless, because of the high cost of clinical trials for new biotechnology products and the cost of an efficient manufacturing facility, entry to the industry was limited, and as of 1995 the industry had not been afflicted by the sorts of price battles that characterize the markets for many commodity products.

Action Profile

Glucose flows into the blood as food is digested. This flow rate, if graphed, resembles a bell- shaped curve which reaches its peak flow rate about 1.5 hours after a meal. Some glucose is taken into the liver, converted into a substance called glycogen, and stored there. To provide the body with between-meal energy, the liver converts this stored glycogen back into glucose and secretes it at a steady rate back into the bloodstream.

In nondiabetic persons, the body senses the amount of glucose flowing into the blood from these two sources and signals the pancreas to secrete an offsetting amount of insulin to maintain about 100 milligrams of glucose per deciliter of blood. To achieve the same level of control over blood sugar (and thereby avoid the complications of excessive or insufficient blood glucose), patients with diabetes needed insulins that are absorbed from the subcutaneous injection site at two different rates—one to be absorbed into the blood at a slow, constant rate, to offset the flow of glucose from the liver, and the other to be absorbed into the blood at a faster, bell-shaped pace, to offset the flow of glucose from digestion. Hence, many patients mixed a regular-acting and slow-acting insulin together in their syringes for most injections. These rates of flow are depicted in Exhibit 5.

Because of these different flow patterns, patients who wanted to control their blood glucose levels carefully had to take injections of regular insulin before every meal. Unfortunately, many found injections to be inconvenient or uncomfortable, and took only one or two daily injections of long- acting insulin.5 The resultant unhealthy pattern of insulin and glucose flows in the bloodstream of these patients gave them high levels of blood glucose in the morning and low levels in the afternoon—causing a high incidence of near-term emergencies and long-term complications.

5 Experts believed that only 20% of insulin-injecting patients actually took a shot before each meal.

For the exclusive use of c. ma, 2019.

This document is authorized for use only by congyao ma in SPRING 2019 (UNGER) THE INNOVATION PROCESS-1-1 taught by BARRY UNGER, Boston University from Jan 2019 to May 2019.

Eli Lilly and Company: Innovation in Diabetes Care 696-077

5

Unfortunately for the small proportion of patients taking pre-meal shots to keep better glucose control, the fastest-acting available insulins followed a flow or action profile that was slightly slower than the rate of glucose flow from digestion, as depicted in Exhibit 5.6 They therefore suffered a temporary high level of blood glucose after meals unless they took their injection about 40 minutes prior to eating. For most patients, however, taking an injection 40 minutes prior to each meal was risky and inconvenient. For example, a diabetic could take an injection just prior to leaving work in the evening, planning to drive home and start eating dinner an hour later. If caught in a long traffic jam, however, the result could be disastrous. Similarly, if a diabetic took an advance injection prior to eating out, but then was unable to eat the type or amount of food he or she had planned upon in the injection, high or low blood glucose could result. Hence, many of even the most careful patients simply injected themselves just before their meals, and lived with the mismatch in flow rates.

To respond to this problem, in the late 1980s Lilly launched an effort to develop an insulin that could mimic more closely the normal physiologic secretion of insulin in people without diabetes. The result, code-named Match, was successful. By 1994 it was clear that Match was absorbed into the blood after injection at a rate that was much closer to the rate at which glucose was absorbed into the blood after a meal. Consequently, patients using Match insulin in the clinical trials required for regulatory approval were able to achieve better control of blood glucose levels after meals, reporting fewer incidents of high and low blood glucose, compared to patients using regular insulin.