Rare disease has become the most attractive place in pharma to launch a drug, and one of the hardest to size. Both facts have the same cause. The industry moved its growth to rare disease because the patient population is small, genetically defined, and concentrated, which delivers premium prices, thin competition, and better odds. The same population that delivers those advantages is what makes the market impossible to measure with tools built for larger ones. The property that makes rare disease attractive is the property that makes it unknowable, and a team that sees the first without the second will build its plan on a number it cannot defend.
Why rare disease became the prize
Most rare diseases are genetic, begin in childhood, and are scattered thin: on the Orphanet database, 72 percent are genetic in origin, 70 percent begin in childhood, and 85 percent affect fewer than one person in a million, across more than 6,000 catalogued diseases. The commercial profile is the inverse of the blockbuster model. Newly approved orphan drugs carry a median annual price near $219,000 per patient against roughly $12,800 for non-orphan drugs, a seventeen-fold gap, and the regulatory system underwrites it with seven years of US market exclusivity and a clinical-trial tax credit under the Orphan Drug Act, ten years in Europe.
Rare-disease programs are also likelier to succeed and cheaper to run. L.E.K. puts the likelihood of approval from Phase I at about 30 percent for them against 20 percent for the rest, on pivotal trials that enrol roughly 60 percent fewer patients. Competition is thin, because fewer than 5 percent of rare diseases have any approved treatment. With more than $390 billion of branded revenue exposed to patent expiry this decade, orphan drugs are now on track to be a fifth of the global prescription market by 2030, a share that has doubled in ten years. Every one of those advantages rests on the population being small and hard to find.
Measuring a market versus building one
Commercial analytics was built to measure markets that are large, coded, and visible in the data, and a blockbuster is sized against a population that prescription and claims data track directly. Rare disease offers no such population to measure. Most patients are undiagnosed, most diseases carry no specific code, and the numbers are too small for the methods that need volume. The discipline has to change at the root, from reading a market off the data to assembling it.
The four shifts rare disease forces on commercial analytics
Building the number rather than looking it up takes four shifts, each replacing a blockbuster habit with a rare-disease discipline.
1. From prevalence to the patient funnel
You cannot start from prevalence, because the diagnosed population sits far below the true one. When researchers screened 37,104 male newborns for Fabry disease, they found it at roughly 1 in 3,100, against a textbook estimate near 1 in 50,000. The defensible number is built from the bottom, prevalence to diagnosed to eligible to treated, with the diagnosis rate carried as an explicit, sensitivity-tested variable rather than a buried assumption, because it moves the answer more than anything else. The undiagnosed pool is the addressable market, not a caveat on it.
2. From one indication to several markets
A single indication is usually several markets that cannot be added together, and the split happens at the point of prescribing. In Fabry, the oral therapy migalastat treats only the 35 to 50 percent of patients with an amenable variant. In ATTR amyloidosis, one genotype is cardiac and lands in cardiology while another is neuropathic and lands in neurology. In myasthenia gravis, the newer therapies are written against specific antibody subtypes. Size any of these as one figure and the number is wrong before it starts, so you segment first and size second, the reverse of the usual order.
3. From syndicated data to registries and centres
The tools that power ordinary analytics assume a scale rare disease never supplies. Syndicated and claims data thin to noise, which is why patient-finding moves to unstructured clinical records and why registries, not big datasets, carry the evidence. Forecasting inherits the weakness: a McKinsey review of 1,700 forecasts found consensus off by more than 40 percent in most cases even in common disease, with orphan products singled out as harder. Even the value tools break, which is why NICE runs a separate threshold for very rare conditions and ICER reports ultra-rare cost-effectiveness out to $500,000 per quality-adjusted life year. Intelligence here is registry- and centre-led, not dashboard-led.
4. From a global number to a country build
No single global number survives, because the three things that decide how many patients exist commercially, diagnosis rate, screening coverage, and reimbursement, all differ by country. Newborn-screening panels vary, national definitions of rare range from 5 to 80 per 100,000, and the same asset has a different market in each geography. The Gulf makes the point, where high consanguinity enlarges the recessive-disease pool and mandatory screening changes who is found, so the number is built market by market rather than scaled from one.
The discipline: build the number, do not look it up
Most analysis measures a market. In rare disease, the work is to build one. Skip that and the errors are fixed in the plan before launch, when they cost the most to correct. Start from prevalence and the opportunity can be overstated by an order of magnitude. Size a field force to that number and you staff for a market that does not exist. Price against a threshold the therapy was never built to meet and the access negotiation is lost before it opens. The blockbuster playbook does not simply lose precision in rare disease. It produces a number that is confident, defensible, and wrong, and for a one-time therapy it mistakes a depleting bolus for an annuity.
The right question of any rare-disease forecast is not how big the market is. It is whether the number was looked up or built, and whether it will hold, disease by disease and country by country, in front of a brand lead or a payer who will test it.
AXLRx builds that number across the US, UK and GCC, sizing the funnel from prevalence to the found, eligible, and treated pool with the diagnosis rate made explicit rather than assumed. Browse the rare-disease briefs, or commission a model scoped to your asset.