In May 2019, the US Food and Drug Administration approved Zolgensma, a one-time gene therapy for spinal muscular atrophy, a disease that kills many affected infants before their second birthday. Novartis set the launch price at US$2.125 million, at the time the most expensive medicine ever sold. The company framed it as a bargain relative to years of an existing treatment.
Both things can be true at once. Zolgensma can be a scientific marvel and a sign that the system designed to produce rare-disease drugs has drifted a long way from its original purpose.
The original bargain
Rare diseases are individually uncommon and collectively not rare at all. The US National Institutes of Health estimates there are around 7,000 of them. According to the US Government Accountability Office, more than 90 per cent have no FDA-approved treatment.
The logic of the US Orphan Drug Act of 1983, copied in various forms by the European Union and others, was simple. If a disease affects too few people for a company to recoup development costs, the state will sweeten the deal: tax credits, fee waivers, regulatory support and, most valuable of all, seven years of market exclusivity in the US. The EU offers ten.
It worked, in the sense that rare-disease drug development grew enormously. It also created an incentive structure that its architects did not fully anticipate.
When “orphan” becomes a business model
In 2017, a Kaiser Health News investigation found that more than 70 drugs originally approved for mass-market use were later approved as orphans for a rare indication, including well-known products such as Crestor, Abilify and Humira. Around 80 drugs had received multiple exclusivity periods for two or more rare conditions. The investigation noted that companies were not breaking the law. They were using it in ways its architects said they had not intended.
The pattern points to something structural. Orphan status was designed as a subsidy for commercially unattractive diseases. Once a small, well-defined patient population, high willingness to pay and strong exclusivity are combined, rare diseases can become commercially attractive in their own right. Pricing is then set not by development cost but by what payers will tolerate for a desperate patient with no alternative.
That is not a moral failure of individual companies. It is what the incentives reward. If a policy pays more for a behaviour, it should not be surprised to get more of it, including the parts nobody intended to buy.
Why the price floats free of cost
It is worth being precise about the mechanism, because the usual complaint, that orphan drugs are priced far above what they cost to make, misses the point. Almost no branded drug is priced on cost. The question is what disciplines price, and for most medicines the answer is some mix of competition, a payer’s willingness to walk away, and the size of the market. Orphan status weakens all three at once.
Exclusivity removes direct competition for the approved indication. A tiny population means a manufacturer must recover development and commercial costs from few patients, which justifies a high per-patient price in principle and makes it very hard for a payer to judge in practice. And refusal is politically costly. A payer that declines a drug for a common condition is making a statistical choice; a payer that declines the only treatment for a named child with a fatal disease is making a headline. Companies understand this, and so do patient groups, some of which receive industry funding.
A simple back-of-envelope comparison shows how the incentives can tilt. Suppose a company can spend a similar amount developing either a drug for a common condition with many competitors, where it might capture a modest price from a large population, or a drug for a rare condition with 2,000 treatable patients in rich markets, no competitor and seven to ten years of exclusivity. If payers will pay several hundred thousand dollars a year for the rare-disease drug, 2,000 patients generate revenue in the hundreds of millions annually, with low marketing costs and little price pressure. The rare-disease bet can easily look better. That was the intention for genuinely neglected diseases. The problem is that the same arithmetic applies to slicing a common disease into smaller, biomarker-defined subgroups, each of which can qualify as rare.
The thresholds we invented to cope
Health technology assessment agencies have had to bend their rules to accommodate this. In England, NICE normally judges value against a standard cost-effectiveness threshold. For highly specialised technologies for very rare conditions, it applies a threshold of £100,000 per quality-adjusted life year, rising to £300,000 in specific circumstances.
There are principled reasons for this. Trials in tiny populations are hard, evidence is uncertain, and societies may reasonably decide that people with rare conditions should not be excluded simply because of how few they are.
But the trade-off is real and rarely stated plainly. A health budget is finite. Every pound that buys a year of life at £300,000 is a pound not buying a year of life at a small fraction of that price somewhere else in the system, often for patients who are older, poorer and less organised as an advocacy group. Rare-disease patients have compelling stories and effective charities. The people displaced by a high-cost approval usually have neither. They are statistical, not visible.
There is a compounding effect. Each approval at a very high price becomes a reference point for the next. When a new therapy arrives, it is priced relative to the existing standard of care, and a very expensive incumbent makes the newcomer look reasonable. Zolgensma’s launch was itself justified by comparison with years of an existing SMA treatment. Over time, the ceiling does not stay fixed; it ratchets, one reasonable-looking comparison at a time.
The strongest counterargument
The defence of the current system is not trivial. Before 1983, very few drugs for rare diseases were developed at all. The incentives, including the generous ones, are the reason many patients with conditions like cystic fibrosis, haemophilia or SMA have treatments that change the course of their disease. Some of the most important biotech platforms, including gene therapy and antisense drugs, were developed and proven in rare diseases first, and their later use in broader populations depended on that early commercial foothold. Cut the reward, the argument goes, and investors will move their money elsewhere, and the next generation of therapies will not be made.
That argument partly holds. The investor calculation is real, and trimming incentives bluntly would reduce investment in precisely the high-risk, small-population science that has delivered. Where it fails is in its treatment of all orphan designations as alike. A first-in-class gene therapy for a fatal infant disease and a second indication for a drug already earning billions share a label but not a risk profile. A policy that cannot tell them apart overpays for one to protect the other. The defence is a case for keeping strong incentives where the risk is genuine, not for leaving them untouched everywhere.
Price is not the only lever, and not the best one
Debates about orphan drugs tend to become debates about whether a single price is outrageous. That misses the design questions underneath, and each one belongs to a different actor.
Legislators and regulators own the question of proportion. A drug repurposed from a blockbuster did not face the risk the orphan incentive was meant to compensate, and exclusivity for it could be shorter, or tapered once revenue from an orphan indication passes a set level. They also own the question of the neglected majority. Incentives have concentrated activity in diseases with identifiable, reimbursable markets, and many rare diseases, especially those common in poorer countries, remain untouched. Priority-review-style rewards aimed specifically at those gaps would be more targeted than blanket exclusivity.
Payers and health technology assessment bodies own the terms of payment. For one-time therapies, paying over several years, conditional on the treatment continuing to work, shifts some risk back to the manufacturer. It also forces something useful: a registry that tracks what happens to patients after approval, which is where the evidence for small-population drugs is thinnest. Payers can also be more explicit about opportunity cost, publishing what else a given budget would have bought, so that the displaced patients become visible in the decision.
Public research funders own the question of contribution. When basic research behind a therapy was publicly funded, that contribution deserves to be reflected in the price or in access terms, and funders can write those conditions into grants and licences before the science becomes a product rather than arguing about them afterwards.
Hospital leaders feel the problem in their budgets. A single infusion or gene therapy can consume a meaningful share of a specialist centre’s drug spending, and where reimbursement lags approval, the hospital carries the risk. Pharmacy and finance teams increasingly need the same contracting sophistication as national payers.
Founders and investors, finally, should read the direction of travel. A business model that depends on stacking exclusivity periods or carving common diseases into rare subsets is exposed to exactly the reforms most likely to pass. One built on genuinely unmet, high-risk indications is the one policymakers are trying to protect.
The global blind spot
Almost the entire orphan-drug conversation takes place in the US, Europe and Japan. Yet people with rare diseases live everywhere. A price calibrated to what the US or German health system will pay places therapies effectively out of reach for most of the world, including middle-income countries with growing capacity to diagnose rare conditions. A diagnosis without an affordable treatment is a particular kind of cruelty.
There is also an evidence problem that travels. Countries without large HTA agencies often rely on decisions made in England, Germany or the US. When those decisions stretch thresholds for rare diseases, poorer systems inherit the conclusion without the budget that made it bearable. A threshold chosen in London becomes, in effect, a price signal in Jakarta or Nairobi.
What to watch in the next two years
Several signals will show whether the system is being adjusted or merely argued about. In Europe, the revision of the EU’s pharmaceutical legislation, including proposals to modulate orphan market exclusivity, is the most concrete test of whether exclusivity can be tied to unmet need. In the US, watch how drug-price negotiation rules treat orphan drugs, since exemptions there shape which indications companies pursue first. Watch, too, for outcomes-based contracts for one-time gene therapies being renewed or abandoned, which will reveal whether they actually transfer risk. And watch whether any middle-income country secures a gene therapy at a price tied to its own ability to pay, rather than a discount on a US list price.
The takeaway
The Orphan Drug Act solved the problem it was built for: companies now develop medicines for rare diseases. The new problem is that the incentives keep paying out long after the market failure they were meant to correct has disappeared. Reform is not about being against rare-disease patients. It is about making sure the subsidy goes to the diseases that still need it, at prices that do not quietly take care away from someone else.
Should rare-disease drugs be judged by a different value threshold, and if so, where should the line be? We would like to hear your view.
