Hard plastic chairs, a reception desk with software that hasn’t significantly changed in ten years, and a stack of forms requesting information that the system already has somewhere are all familiar sights when you walk into most hospital waiting rooms these days. Given what’s going on in labs, government agencies, and research hospitals only a few miles from any one of these waiting rooms, it’s an odd contrast. Seldom has the gap between the medication being developed and the medication being administered felt so large or worthy of careful examination.
It turns out that there is more than one invention that will shape health technology in the future. Genetic therapies are being produced in closed sterile cartridges at the point of care rather than in centralized factories, wearable technology is continuously tracking biomarkers in ways that would have required a hospital admission twenty years ago, and artificial intelligence is able to read medical imaging with accuracy that surprises seasoned radiologists. On its own, each of these is important. Together, they’re starting to change both what medicine can truly accomplish and who has the authority to determine how.
| Entity / Initiative | Role / Focus | Key Development |
|---|---|---|
| MHRA (UK) | Medicines and Healthcare products Regulatory Agency — UK regulator for medicines, medical devices, and AI in healthcare | Launched Regulatory Innovation Corridor pilot (Dec 2025) with Singapore’s HSA; signed refreshed MOU March 2026 |
| HSA (Singapore) | Health Sciences Authority — Singapore’s national regulator for health products | Published updated AI in Healthcare Guidelines (AIHGle 2.0); CEO sits on UK’s National Commission on AI in Healthcare |
| Access Consortium | Multilateral regulatory collaboration: UK, Australia, Canada, Switzerland, Singapore | ~40 new active substances approved via work-sharing; faster median patient access times across member nations |
| WHO / Europe | World Health Organization — European Region health policy and digital health agenda | Hosting hearings for Second European Programme of Work (EPW2) 2026–2030; focus on AI, equity, and person-centred care |
| Agentic AI in Medicine | Autonomous AI systems making clinical decisions — diagnoses, treatment protocol adjustments, patient data analysis | Described by regulators as a “paradigm shift” beyond traditional medical device regulation; no universal governance framework yet |
| Cell & Gene Therapies (CTGTP) | Advanced personalised treatments moving to point-of-care manufacturing | Emerging “smart factory” platforms using sterile closed cartridges, barcode tracing, and cloud-based batch documentation |
| Precision / Personalised Medicine | Treatments tailored to individual genetic, lifestyle, and environmental data | AI and big data enabling predictive modelling at population scale; focus areas include cancer, neurodegenerative disease, obesity, rare conditions |
| Imperial College London Report | Academic projection: “The Future State of Health and Healthcare in 2035” (June 2025) | Projects healthcare becoming more personalised, predictive, and preventative by 2035; driven by tech integration |
| NHS England | National Health Service — England’s public health system | 10-Year Health Plan for England integrating new technologies, wearables, and digital health tools into patient care delivery |
| Key Equity Concern | Access gap between technology-enabled and underserved populations | WHO warns AI/precision medicine advances risk deepening health inequalities if affordability and access gaps aren’t addressed |
| Reference | https://www.gov.uk/government/news/shaping-the-future-of-healthcare-through-global-regulatory-innovation | |
In some ways, the most fascinating and underreported aspect of this story is the regulatory discourse. A revised agreement formalizing their cooperation was signed in March 2026 by Singapore’s Health Sciences Authority and the UK’s Medicines and Healthcare Products Regulatory Agency. The document appears formal until you read what they are actually attempting to govern together. For example, Professor Raymond Chua, CEO of the HSA, defines agentic AI systems as “autonomous software agents that can analyse patient data, make clinical recommendations, and even adjust treatment protocols independently.”
He compares it to having an adaptable and learning digital doctor. The real question that follows is whether something that can independently modify a patient’s course of treatment should be governed as an evolving service that needs ongoing supervision or as a product with fixed and predictable capabilities.
This hasn’t been resolved yet. In an effort to anticipate that question rather than rushing to provide an answer after the fact, the Regulatory Innovation Corridor was established in December 2025 and provides both agencies with early visibility into emerging products prior to their formal review processes.
The picture of gene and cell therapy is equally striking and underreported. Advanced therapy manufacturing is shifting from the batch-production model, which involves producing large quantities in a single, specialized facility and shipping them all over the world, to something more akin to a distributed network. Sterile, closed cartridges with integrated in-process quality controls, barcode tracking, cloud documentation, and the ability to run several therapy batches concurrently.
Science fiction is not what this is. This infrastructure, which is currently under construction, would enable the availability of rare disease therapies and customized cancer treatments at a speed and scale that is just not possible with the current systems. With significantly quicker rollout times than national review processes run independently, the Access Consortium, which consists of the UK, Australia, Canada, Switzerland, and Singapore, has already approved almost forty new active substances through shared regulatory work. There are actual patients behind that number.
However, it’s difficult to ignore the awkward coexistence of these technologies’ excitement with a number of issues that technology cannot solve on its own. The WHO has been clear about this: AI and precision medicine have the potential to transform health planning, forecast patient flows, maximize resource allocation, and provide population-level individualized care.
Additionally, they may worsen already-existing disparities rather than lessen them if the access question is not handled carefully. At best, a genomic medicine program is only partially successful if it reaches patients in university hospitals in Singapore or London but not rural communities in central Asia or sub-Saharan Africa. No algorithm can overcome these limitations, which include the cost of these technologies, the infrastructure needed to operate them, and the health workforce needed to interpret their results.
Health technology coverage has a propensity to reduce all of this to a level of optimism that ignores the difficulties. The technology functions, the AI advances, the treatments progress, and the system adjusts. However, systems do not automatically adjust.
The NHS’s ten-year plan for England, for instance, is explicitly building around digital tools and new medicines, but the health workforce it’s building toward those technologies is already under strain, shrinking in some specialties, and aging across the board. Because the technical solutions are outpacing the institutional capacity to implement them equitably, WHO Europe has been soliciting feedback from Member States, civil society organizations, and clinicians during its hearings for its 2026–2030 program.
The most ambitious predictions, such as the predictive, preventative, and truly personalized medicine that Imperial College London projected for 2035, are still up in the air. The fact that the investment is flowing, the trajectory is real, and the regulatory frameworks are being rebuilt in real time to accommodate something none of them were intended for seems more certain.
Depending on how much trust one has in the organizations carrying out the reconstruction, that can either be reassuring or frightening. Maybe both responses make sense at the same time.
