The Doctor Who Built a Research Machine Because He Never Had Time to Do Research

Ask most founders what pushed them to start a company and you will hear something about a market gap or a spreadsheet. Dr Matt Wilson describes hunger. "When on shifts in the emergency department, we would sometimes go 14 hours without eating anything," the former NHS anaesthetist has said, "so the idea of doing research in these situations is simply unrealistic." That single observation - that the people best placed to run clinical research are the people with the least time to do it - is the whole reason uMed exists.

uMed's answer was not to hire more researchers or nag more doctors. It was to take the humans out of the loop. The company builds software that lives inside networks of GP practices, health systems, and care providers, plugs directly into their electronic health records, and quietly does the work of finding, contacting, and enrolling the right patients. What used to take months of manual chart-hunting can happen in days, and the clinician barely notices it running.

What uMed actually does

Strip away the category jargon - real-world evidence, decentralized trials, patient cohorts - and the product is simpler than it sounds. uMed connects to a provider's health record, identifies patients who fit a study's criteria, gets their informed consent, and then keeps collecting data from them over time: routine EHR entries, patient-reported outcomes, wearable readings, even biosamples mailed from home. The result is a living registry that researchers can query and re-use.

The flagship version of this is called ACCESS Cohorts: reusable, longitudinal research registries that link routine clinical records with patient-reported data. The pitch to a pharmaceutical company is pointed - the evidence you need usually does not exist in any database you can buy, so stop searching and start building the cohort that generates it. uMed has spun that engine up into disease-specific registries too, including ACCESSPD for Parkinson's disease, an interstitial lung disease cohort developed with the Three Lakes Foundation, and a cardiometabolic registry.

Who is paying for it

uMed gives the software to providers for free, which raises the obvious question of who writes the checks. The answer is the demand side of research: pharmaceutical and life-sciences companies, contract research organizations, and disease foundations that fund studies and desperately need patients. Reported partners and customers span names like AstraZeneca, Moderna, Sanofi, Neurocrine, Teva and BIAL, alongside the Michael J. Fox Foundation on the Parkinson's side and academic centers including King's College London, the University of Oxford and Queen Mary University of London.

That is the business model in a sentence: embed on the supply side of patient data, monetize on the demand side of research. Providers get research capability without extra staff; sponsors get faster, cleaner cohorts; uMed sits in the middle taking the friction out.

The problem it is built for

uMed's timing is not an accident. When it raised its Series A in July 2023, UK press was openly describing a "collapse" of clinical trials, with the country sliding down global rankings for how many studies it could actually run. Recruiting patients is slow, expensive, and skewed toward people who live near big research hospitals. Rural patients, older patients, and underserved communities routinely get left out - which means the evidence itself is skewed.

By living inside the primary-care record instead of the research hospital, uMed can reach the patients traditional trials skip. That thesis is now being tested at national scale in the US.

Into the living room

In December 2025, uMed announced a partnership with WellSky, whose software runs across more than 10,000 home-based care organizations. Pairing that footprint with uMed's registries is a bet that clinical research can be brought into people's homes nationwide - reaching rural, aging and underserved populations across conditions like Parkinson's, cardiometabolic disease and autoimmune disorders. It followed earlier North American moves: a US expansion with Innovaccer that opened access to over seven million patients, and a first Canadian partnership with MCI Onehealth.

How it differs from the giants

Real-world evidence is a crowded space, ruled by heavyweights like IQVIA, Optum, Parexel and Medidata, plus a wave of newer players such as Aetion, TriNetX and Castor. Most of them work by buying, licensing or aggregating data that already exists. uMed's distinction is architectural: it embeds inside the provider network and builds the registry itself, with consent captured up front and prospective data collected on purpose.

Approach	Typical RWE platform	uMed
Data source	Buys / licenses existing datasets	Builds consented cohorts at the source
Consent	Often retrospective / de-identified	Prospective, patient opt-in
Reach	Concentrated near research sites	Primary care & home-based care
Re-use	Study-by-study	Reusable longitudinal registries

None of that works without trust, which is why uMed's least glamorous feature may be its most important: ISO 27001, ISO 9001, Cyber Essentials Plus, HIPAA and NHS Data Security certification, with patient consent baked in from the first click. Moving 10 million people's records requires the boring stuff to be airtight.

The founder, and the moat

Dr Matt Wilson's path is not the standard founder résumé. He earned his medical degree with distinction from Cardiff University, served as a Medical Officer in the Royal Navy and later with the Royal Marines, worked as an NHS anaesthetist and emergency medicine doctor, and was selected as an NHS England Clinical Entrepreneur Fellow. The company he founded is legally named Umedeor Ltd - which is why its LinkedIn handle reads "umedeor" rather than "umed."

The quiet moat is the registry itself. Any competitor can write recruitment software; far fewer can point to a reusable, consented, longitudinal cohort that took years of provider relationships to assemble. Build the cohort once, and every future study runs on rails. That is the asset compounding underneath uMed's roughly $17-25M in total funding and its ~$3M in reported revenue.

What you can take from it

There is a repeatable pattern here for anyone building in a slow, regulated industry. Find the workflow everyone hates - in this case, recruiting patients for trials. Automate the tedious 80% of it. Give the tool away to the side that has the data but no budget, and charge the side that has the budget but no access. Then build a durable asset (the cohort) that gets more valuable every year rather than a feature that gets copied.

It does not work everywhere. uMed's model depends on deep, trusted integration with health records, a regulatory environment that permits consented data re-use, and provider networks willing to host software they do not pay for. In a market without those conditions - weak data infrastructure, low trust, or fragmented consent law - the same playbook stalls. But where those pieces line up, running research from inside the doctor's office, rather than dragging patients to a research center, starts to look less like a startup pitch and more like plumbing.