Case Study 44.1 — A Hundred Years of Insulin: The Closest Precedent We Have

Why this case

Chapter 44 is full of questions about what happens when a genuinely effective peptide therapy meets a very large population over a very long time. Almost none of those questions has been answered for GLP-1 receptor agonists, because the drugs are young and the questions require decades.

But one peptide has already run the experiment. Insulin is the closest available precedent: a life-changing molecule, a large and growing population, an indefinite chronic therapy, and a price history now more than a century long. If you want to know what the arc of an effective peptide therapy looks like at the scale of generations, this is the only complete example anyone has.

It is also, in places, a discouraging one — and this case study is more useful for the ways the analogy fails than for the ways it holds.

The arc, compressed

The discovery. In the early 1920s, a research effort at a university in Canada isolated a pancreatic extract that reversed a condition that had until then been reliably fatal in its severe form. Children admitted to hospital in diabetic ketoacidosis, expected to die within days, recovered. There is very little in the history of medicine that matches the immediacy of that result, and essentially nothing in the history of peptide medicine.

The patent decision. The discoverers transferred the patent rights to their university for a nominal sum. The recorded intent was explicit: the discovery should not be used for private gain, and the therapy should be available to everyone who needed it. The patent was licensed broadly to manufacturers, with the university acting as steward. As a statement of values it is unambiguous, and it is quoted more often than almost any other sentence in pharmaceutical history.

The first several decades. Production scaled. The therapy reached progressively more people. Formulations improved substantially — purification advanced, animal-sourced insulin gave way to recombinant human insulin, and then to engineered analogs with modified time-action profiles that made the therapy far easier to live with. Chapter 11 covers this engineering in detail. Each improvement was real, and each improvement was also a new patentable product.

The later decades. In some markets, the price of insulin products rose substantially over a period during which the underlying molecule had been off patent for a very long time. In those markets, people rationed a therapy discovered before their grandparents were born — taking less than prescribed to make a supply last, with the predictable clinical consequences. The situation eventually attracted legislative and regulatory attention, and prices in several markets have since come down through a combination of policy intervention, negotiated pricing, and competitive entry.

Where it stands. Insulin today reaches vastly more people than it did in 1922, in vastly better formulations, at prices that vary enormously between health systems. Both of the following are true: it is one of the great triumphs of twentieth-century medicine, and its access history includes a long stretch that nobody defends.

What actually happened to the price, mechanically

The uncomfortable part of this story is that most of it happened without anyone doing anything obviously wrong.

Incremental improvement resets exclusivity. Each genuinely better formulation — longer-acting, more predictable, easier to dose — is a new product with new intellectual property. Prescribing migrates to the improved version, and the improved version is not the one whose patent expired. Nobody has to suppress the older product; the market simply moves.

The molecule is not the barrier to entry. A follow-on insulin is not a small-molecule generic. It is a complex biologic requiring cell-culture manufacturing, extensive characterization, and a regulatory pathway substantially more demanding than the generic route. Chapter 32 covers why. The consequence is that the number of firms capable of entering is small, and a market with few entrants does not behave like a market with many.

Devices carry their own patents. Delivery systems — pens, cartridges, pumps — are separately protected and separately improved. A therapy delivered by a device inherits the device's intellectual property landscape.

And the intermediaries are opaque. In several markets, the price a manufacturer lists, the price a payer negotiates, and the price a person hands over at a counter are three different numbers with a complicated relationship. Public argument about "the price of insulin" frequently involves people citing different ones of the three.

None of that requires villainy. It requires only that each participant behave normally within the structure they are in — which is precisely the point §44.3 makes about payer churn and §44.6 makes about the access curve.

What the case does predict for GLP-1 receptor agonists

That the shape of the access curve is real. Access to an effective, expensive peptide therapy did begin concentrated and did broaden. That is not speculation; it happened.

That "competition will eventually solve it" has an unspecified time constant. This is the strongest single lesson. The claim is not false. It is incomplete in the way that matters most, because "eventually" is doing work that nobody quantifies. For a person who needs a therapy this year, a correct prediction with a thirty-year horizon is not a plan.

That improvement and price decline are not the same process. Insulin got dramatically better and, for a long stretch in some markets, more expensive. A reader who assumes those move together will misread the next two decades.

That the biosimilar pathway is a genuine but slow mechanism. It works. It has produced real price reductions in several classes. It does not work like generic substitution, and expecting it to is the most common error in commentary on this subject.

What the case does NOT predict

This is the part that matters more, and it is the part the exercise sections push on.

The populations are different in size and in composition. Insulin's essential population is people who will die without it. The population for which GLP-1 receptor agonists are indicated is far larger and far more heterogeneous, spanning severe metabolic disease, cardiovascular risk reduction, and — at the boundary Chapter 43 examined — uses that are contested. A therapy whose population is enormous and partly elective faces different political and market dynamics than one whose population is defined by mortal necessity.

The competitive landscape is different. Insulin's price history unfolded partly because entry was hard and entrants were few. Whether the same holds here depends on manufacturing complexity, synthesis routes, and the number of firms with credible programs — variables that are not the same as insulin's and that are actively changing.

The political salience is different. Insulin pricing became a legislative issue only after decades. Metabolic drug pricing became a public argument almost immediately. Attention arriving early is not the same as attention arriving late, and it changes what is politically possible.

And the endpoint is different. Insulin's benefit is immediate and unmistakable: without it, a person with type 1 diabetes dies. The benefits at stake in Chapter 44 are largely long-horizon risk reductions, which are harder to defend in a budget argument precisely because nobody can point to the specific person who did not have the event.

Sharpest version of the disanalogy: insulin's price history is a story about a therapy nobody could argue was optional. If a long, ugly access period happened to that therapy, the argument that it cannot happen to a therapy some people consider optional is weaker, not stronger. Whether that makes insulin a warning or an outer bound is a genuine question, and this case study does not settle it.

Discussion questions

1. The patent transfer was an explicit act of values by the people who had the most to gain from doing otherwise. A century later, the outcome in some markets was rationing. What does that tell you about the relationship between the intentions of individuals and the behavior of systems? What would the discoverers have had to do differently — if anything could have worked?

2. Identify the three mechanisms in this case by which price stayed high without anyone violating a rule. For each, state whether an equivalent mechanism is available in the GLP-1 case, and how you would find out.

3. §44.6 rates the claim that these drugs will become substantially cheaper as ⚠️ for the direction, with timeline and magnitude unrated. After reading this case, would you keep that rating, raise it, or lower it? Name the specific observation that would move you.

4. "Insulin got better and more expensive at the same time." Explain the mechanism that makes those compatible. Then say what the equivalent would look like for GLP-1 receptor agonists over the next fifteen years, and what you would watch for.

5. Build the strongest argument that insulin is a poor analogy for this drug class. Then build the strongest argument that the disanalogies all cut in the pessimistic direction rather than the optimistic one. Which argument did you find harder to construct, and what does that suggest about your priors?

6. This case study describes a hundred-year arc and this chapter asks you to write a five-year prediction. What can a five-year observation actually tell you about a hundred-year process? State one thing it can establish and one thing it cannot, and then decide whether writing the prediction is still worth doing.