Loading
Old version — revision 1
This is a fixed snapshot of Access and inequality, saved by Import as part of the initial corpus import. It is not edited and it is not updated; the article may have changed since.
Edit summary: Initial import of content/access-and-inequality.md — the filesystem corpus, unchanged. Not an edit.
The problem of who can obtain expensive biomedical interventions, and whether biotechnology will convert economic inequality into biological inequality.
Access and inequality names the question of who actually receives a biomedical intervention once it works, and the associated worry that technologies capable of altering bodies and minds will translate existing economic stratification into biological stratification. The concern is old in the literature and recent in practice: until the 2010s no medicine cost enough to make it acute. As of 2026 several licensed therapies are priced above two million dollars per patient, and the mismatch between where the diseases are and where the treatments are delivered is the central unsolved problem of the genetic-medicine era.
Conventional pharmaceutical pricing spreads cost over years of dosing. A cure delivered once has to recover its development cost in a single transaction, and the manufacturer's incentive is to price against the lifetime cost of the disease it displaces rather than against the cost of goods. That logic produced the current price ladder: onasemnogene abeparvovec for spinal muscular atrophy at roughly $2.1 million on US launch in 2019, etranacogene dezaparvovec for haemophilia B at roughly $3.5 million in 2022, and atidarsagene autotemcel for metachromatic leukodystrophy at roughly $4.25 million in 2024 — the highest list price ever set for a medicine. Casgevy and a competing lentiviral product for sickle cell disease were approved in the United States in December 2023 at roughly $2.2 million and $3.1 million respectively.
Health economists have generally judged several of these prices defensible on cost-effectiveness grounds, because the alternative is decades of transfusions, factor concentrate, or intensive care. That defence does not answer the budget-impact problem. A public payer facing a hundred eligible patients in one year confronts a bill it cannot smooth, and the treatments are concentrated in exactly the populations — children with rare inherited disease, adults with sickle cell — least able to negotiate.
The commercial results have been poor for everyone. The first gene therapy licensed in Europe was withdrawn in 2017 after a single commercial course was sold. Two later products were pulled from European markets in 2021 when their maker could not agree prices with national payers, leaving patients in countries where the therapy was licensed with no route to receiving it. Approval and access are separate events, and the gap between them has widened.
Why price is not the only barrierAn autologous CRISPR therapy for sickle cell requires apheresis, a manufacturing round trip, and myeloablative conditioning with busulfan followed by weeks of inpatient care. Even at zero drug cost it needs a transplant-capable centre, a cell-processing facility, and a blood bank. Most of the world's sickle cell patients live where none of these exist.
Sickle cell disease is the clearest case because the mismatch is total. The overwhelming majority of affected births occur in sub-Saharan Africa, with India accounting for much of the remainder;1 the approved Somatic gene therapy products are administered almost entirely in North America, Europe, and the Gulf. A curative technology has been developed for a disease of the global South and deployed in the global North, and the constraint is not patents alone but the absence of the delivery infrastructure the therapy assumes.
The same structure recurs across the wiki's subject matter. Cochlear implant surgery is routine in high-income countries and rare in most low-income ones, where the majority of children with profound congenital deafness live. Organ shortage is managed by transplant systems that only some states can operate. The devices described in Brain–computer interface research require neurosurgical centres and years of technical support per patient. The largest instance by volume is not a one-time cure at all: incretin drugs for obesity and diabetes carry US list prices on the order of a thousand dollars a month, and obesity prevalence is rising fastest in the middle-income countries where they are furthest out of reach.
The optimistic case rests on precedent, and the precedent is real but partial.
Antiretroviral therapy for HIV cost well over ten thousand dollars per patient per year in 2000 and was, on standard analysis, permanently out of reach for African health systems. Within roughly a decade generic fixed-dose combinations manufactured in India, mass procurement through the Global Fund and PEPFAR, and voluntary licensing had brought first-line regimens to a small fraction of that figure, and tens of millions of people were on treatment.2 Direct-acting antivirals for hepatitis C followed a compressed version of the same path: a US list price of tens of thousands of dollars per course alongside voluntary licences that put courses in some low- and middle-income countries at a few hundred dollars.
Three features made those cases tractable, and all three are absent for cell and gene therapy. The products were small molecules that generic manufacturers could copy at scale. Administration required only a clinic and a pill. And the epidemic's size created political pressure sufficient to override ordinary intellectual-property arrangements.
Autologous cell therapy has none of those properties. Each dose is manufactured for one patient, so unit cost does not fall the way chemical synthesis does. The bottleneck is process and facility, not molecule. And the affected populations, though large in aggregate, are politically fragmented across dozens of rare conditions.
Against this, the cost of reading a genome fell from roughly a hundred million dollars in 2001 to a few hundred, one of the steepest cost declines recorded in any technology.3 That curve is often cited as evidence that therapy costs will follow. It is a poor analogy: sequencing is an instrument-and-reagent problem subject to semiconductor-like scaling, while manufacturing a cell product is closer to running a small hospital.
Three technical shifts would alter the economics rather than the pricing negotiation.
In vivo editing. Delivering an editor directly to the patient removes cell collection, manufacturing, and conditioning at a stroke. Lipid nanoparticles carrying CRISPR components have been used systemically in humans, and the first bespoke therapy for a single infant with a urea-cycle disorder was designed and dosed within months in 2025 — a demonstration that the regulatory and manufacturing path can be compressed, though at a cost per patient that no system could generalise. Whether in vivo editing reaches the tissues that matter for sickle cell is unresolved; the target is the haematopoietic stem cell, and reaching it in the marrow is harder than reaching the liver.
Allogeneic and off-the-shelf products. A cell therapy manufactured in batches from donor material amortises across many patients. This is the main cost lever in the CAR-T field and applies in principle to regenerative products built on Induced pluripotent stem cells.
Cheap geroprotection. If the interventions that come out of the Geroscience hypothesis turn out to be repurposed generics, the distributive picture inverts. Metformin and the TAME trial costs pennies. Exercise as a geroprotector describes the best-evidenced intervention available, which costs nothing and is nonetheless distributed with a steep social gradient — a reminder that price is not the binding constraint on health inequality, and never has been.
Payers have experimented with instruments that spread or condition payment. Outcomes-based agreements make part of the price contingent on durable benefit, which fits one-time therapies well in principle and requires long-term follow-up data that health systems are poorly equipped to collect. Subscription or "Netflix" arrangements, in which a payer buys unlimited access for a fixed annual fee, were adopted by two US states for hepatitis C treatment in 2019 and sharply increased the number of people treated in their prison and Medicaid populations. In 2024 the US Centers for Medicare and Medicaid Services announced a multi-state outcomes-based purchasing model specifically for sickle cell gene therapy, with participation beginning in 2025 — the first attempt to negotiate a gene therapy price on behalf of a public programme at national scale.
Compulsory licensing under the TRIPS agreement and the 2001 Doha declaration remains available and is almost never used for biologics, partly because a licence to the patent does not transfer the manufacturing know-how, which is the actual barrier. Regional manufacturing initiatives — mRNA technology transfer hubs, African vaccine production capacity — address that barrier directly and operate on a decadal timescale.
The specifically futurist worry is not about access to cures but about heritable advantage. Lee Silver's 1997 scenario of a society splitting into genetically enhanced and unenhanced lineages remains the reference version, and it depends on premises that have aged unevenly.4
The premise that has held is that expensive medicine reaches the wealthy first. The premise that has not held is that genetic enhancement would deliver large, heritable, purchasable advantage. Genetic enhancement of cognition explains why: the traits people would pay for are highly polygenic, and the achievable shift from Polygenic embryo screening is on the order of a fraction of a standard deviation under favourable assumptions, smaller across ancestries not represented in the training data, and not obviously larger than the effect of the environment money already buys. Human germline editing could in principle do more, but not for polygenic traits, and its clinical use remains prohibited in most jurisdictions.
Which inequality is the live oneSome argue the genetic-divide scenario distracts from the inequality that exists now: a difference of a decade or more in life expectancy between rich and poor districts of the same city, produced by ordinary causes and requiring no new technology. Others reply that a stratification which becomes heritable is categorically different from one that resets each generation, and is worth preventing before it is cheap.
Both readings agree on the near-term implication. The interventions plausibly capable of stratifying a population — screening, Embryo selection, In vitro gametogenesis if it matures — are lightly regulated in most countries precisely because they involve selection rather than modification, and therefore fall outside the instruments described in Governance of human genome editing. The technologies that ethics has watched most closely are not the ones distributing advantage.
No health system has yet solved the problem of paying for a large cohort of one-time cures arriving at once, and the cohort is growing: dozens of gene therapy programmes are in late-stage development for conditions with patient populations in the tens of thousands. Nor has any mechanism been established for delivering a curative therapy in a country without transplant infrastructure. The 2023 international summit on genome editing devoted an unusual share of its attention to this, on the argument that a governance regime concerned only with restraint has answered half the question. Nothing binding followed, and the affected populations still have no route to the treatments licensed for their disease.
paperPiel, F.B., Steinberg, M.H., Rees, D.C. "Sickle Cell Disease." New England Journal of Medicine, 2017. ↩
reportMédecins Sans Frontières. Untangling the Web of Antiretroviral Price Reductions. MSF Access Campaign, successive editions from 2001.↩Compiled by an advocacy organization from prices it and other buyers were quoted, so it records transaction prices rather than an independent audit.
reportWetterstrand, K.A. "DNA Sequencing Costs: Data from the NHGRI Genome Sequencing Program." National Human Genome Research Institute.↩Tracks production cost at funded sequencing centres, not the price a customer pays and not the cost of interpreting the result.
bookSilver, L. Remaking Eden: Cloning and Beyond in a Brave New World. Avon Books, 1997. ↩