Most cancer drugs granted accelerated approval by the US Food and Drug Administration between 2013 and 2017 still had no demonstrated survival or quality-of-life benefit by mid-2023 (Liu et al., JAMA, 2024). Through Project Orbis the MHRA now reviews many of the same applications concurrently with the FDA, so drugs arrive in England and Scotland with the same immature evidence they carried at the point of US approval (Jenei et al., The Lancet Oncology, 2024). In a new editorial in PharmacoEconomics,we describe the resulting reimbursement model as ‘pay-and-hope’, in which the health system pays the full price at launch and hopes the benefit appears later, and we argue that routinely collected health data now make an alternative workable from the first day of a drug’s commercial life. Scotland is well placed to demonstrate it.
Why administrative data rather than companion diagnostics?
A companion diagnostic tests the patient first so that only likely responders are treated, and we remain enthusiasts for that future. Validating a predictive biomarker, however, takes years of prospective evidence, and new drugs are arriving faster than biomarkers can be validated. Every health system already records who received a drug, how long they stayed on it, when they moved to the next line of treatment, and whether they are still alive, and although none of this was collected for research, as a basis for sharing financial risk it is available now and costs almost nothing extra to use.
In our proposal, reimbursement at launch is tied to pre-agreed administrative endpoints, with rebates when the benefit fails to appear, and later, as biomarker evidence matures, the same data are used to check that predicted responders respond in routine care. The limits are set out plainly: treatment persistence is a proxy, discontinuation has many causes, and an outcomes-based contract shares financial risk rather than proving a drug works. Contracts have to be written with all of that in mind.
Why Scotland?
The CHI number links every patient in Scotland across the cancer registry, national prescribing data, hospital admissions, and chemotherapy e-prescribing, and this linkage has been proven at scale. When our group built a national colorectal cancer resource we joined 32 separate administrative datasets (Hanna et al., International Journal of Population Data Science, 2021), while colleagues in Glasgow used record linkage to report real-world survival on abiraterone and enzalutamide across the whole population, including the majority of patients who would never have met trial eligibility criteria (Baillie et al., Pharmacoepidemiology and Drug Safety, 2020). The capability is now institutional, because the Cancer Medicines Outcomes Programme within Public Health Scotland feeds real-world evidence directly into SMC assessments and supplied reports for 8 of the 13 requested between February and July 2025 (Mueller et al., Value in Health, 2026).
The policy foundations exist as well. England’s Cancer Drugs Fund collects data under managed access for up to five years (NICE), and the SMC’s ultra-orphan pathway funds medicines for up to three years ahead of reassessment (SMC), but both schemes gather data to inform a later decision rather than linking payment to what the data show. Taking that further step is more feasible in Scotland than almost anywhere else, because one payer, one HTA body, and one linked data system sit within a single national conversation.
A pilot would attach outcomes-based contracts to a handful of newly launched cancer medicines, draw the endpoints from national data, and publish the methods openly. No other country has yet run outcomes-based reimbursement as the operating standard from launch, and the blueprint would transfer to any system with comparable data.
Credit to Giovanni Tramonti, who led the editorial. Read it in full: Tramonti and Hall, PharmacoEconomics, 2026, The End of ‘Pay-and-Hope’: Managing Accelerated Approvals with Outcomes-Based Reimbursement in the Era of Routinely Collected Health Data.