MedTech, August 2026: everything that mattered happened after approval

Regulatory approval no longer determines whether a medical device sells in Europe. Reimbursement bodies, vigilance authorities, and procurement evidence standards each apply their own test on their own timetable, and a device can clear approval in every market while failing all three.

About this publication

Vital Signs: MedTech is a briefing on devices, diagnostics, and surgical innovation from HealthSeed AG, published every two weeks. This is the first edition.

It is written for people carrying commercial responsibility in medical technology, expert in one corner of the field and generalist everywhere else. Someone building diagnostics has rarely filed a French reimbursement dossier. Where a term needs explaining, we explain it.

Every claim carries a source you can open. The HealthSeed Perspective blocks are our own read rather than reporting.

This edition covers 11 to 24 August 2026 in six movements: reimbursement, money, vigilance, two device categories, what counts as evidence, and the argument.

Reimbursement

Most device companies budget carefully for approval and treat reimbursement as the administrative step that follows it. The sequence runs the other way. Approval is the predictable part, with published requirements, a known cost, and a defined end. Reimbursement is a negotiation with a counterparty who has a fixed budget, no obligation to you, and a different question in mind.

This cycle produced two of those decisions in nine days, in two countries, pointing in opposite directions.

Eight decisions in one day. France’s reimbursement authority is the Haute Autorité de Santé. Its device committee asks whether a product improves on what the country already pays for, and grades the answer on a five-point scale where the top grade is a major improvement and level five is none at all.

On 11 August the committee published eight opinions together. Seven products were accepted, and every one of the seven was graded level five. They ranged from Medtronic’s ONYX embolisation implant and Abbott’s TRICLIP G5 tricuspid valve repair system down to wound dressings from Smith & Nephew and Paul Hartmann. The same test, applied the same way, to an implant and to a dressing.

The eighth was refused. According to the HAS record for that date, Edwards Lifesciences’ SAPIEN 3 Ultra RESILIA, the newest generation of the transcatheter aortic valve family Edwards leads worldwide, was found to have insufficient expected benefit for the listing it requested. HAS publishes the outcome without the reasoning, so the basis for the refusal is not on the public record and is recorded here as absent.

Why France grades this way. A national reimbursement budget is finite and largely fixed in advance. Every euro committed to a new device is a euro unavailable elsewhere in the same system, which makes the relevant question narrower than it first appears. The committee is not asking whether a product works. It is asking whether paying more for it than for the incumbent buys the health system anything.

That framing explains the seven level-five gradings. Each of those products almost certainly works. None of them demonstrated that it works better than what France already funds, so France will fund them at the rate it already pays.

It also explains why market position offers no protection. A large installed base, a strong brand, and years of registry data all answer the question of whether a product performs. None of them answers the question of whether the new generation outperforms the old one, and that is the only question being asked.

Nine days later, Britain opened a category France had just closed. NICE published guidance HTG780 on 20 August supporting routine NHS funding for ex-situ machine perfusion, the technique of keeping a donor liver alive outside the body before transplant instead of packing it in ice. Four systems are named: OrganOx metra, XVIVO Liver Assist, Aferetica PerLife Pro, and Bridge to Life VitaSmart. The evaluation found the technology clinically beneficial and a cost-effective use of NHS resources.

France measured seven devices against products it already funds and found no improvement in any of them. Britain measured a category it had never funded and opened it. The distinction is not national temperament. A device entering an established category is measured against an incumbent, and a device creating a category has no incumbent to be measured against.

What the EU mechanism does not yet cover. The Joint Clinical Assessment is the EU process built to replace duplicated national evaluations with one shared assessment. On 10 August its tracker listed fifteen assessments under way. All fifteen are medicines, and the tracker carries no medical devices and no in-vitro diagnostics.

Any plan resting on European HTA harmonisation reaching devices should be read against that. The mechanism is real and running, and for now it goes around you. Every market you enter is a separate assessment, on a separate timetable, against a separate comparator.

Money

Two calendars decide what a medtech company can do over the next eighteen months. One is the funding calendar, which a founder and a board control and can accelerate or delay. The other is the assessment calendar, which belongs to national reimbursement bodies working through their own queues and publishing when they are ready. Companies plan against the first and are governed by the second.

This cycle produced a clean example of the two colliding, to nobody’s harm and nobody’s design.

A raise for one market, a decision in another. Bridge to Life closed 110 million dollars in Series C equity and debt on 12 August. Soleus Capital led the equity with Lauxera Capital Partners. The money is for building the United States field organisation behind VitaSmart, which has held FDA De Novo clearance since January 2026 and is the only hypothermic oxygenated perfusion system authorised in that country for liver transplantation.

Eight days later NICE named VitaSmart in the guidance covered in section one. The capital was raised for America. The route that opened was British.

Why these clocks never align. Assessment bodies work from published workplans that queue technologies by clinical priority and evidence readiness, not by any company’s commercial timing. NICE does not consult a manufacturer’s cash position when scheduling an evaluation, and it does not publish a decision date the way a court publishes a hearing. A company learns the outcome when the guidance appears.

Funding works the opposite way. A round closes when investors commit, which is a date the company influences directly and often controls within a quarter.

So the two events that most determine a device company’s next year are set by different people, and only one of them is yours. Nothing about that is dysfunctional. It is simply how the system is built, and it means the plan you present at a board meeting always contains at least one date you cannot move.

Two owners changed hands in diagnostics. Astorg completed the carve-out of Thermo Fisher’s microbiology diagnostics business into a standalone company, at a value reported near 1.075 billion dollars. iRhythm agreed to acquire VitalConnect, which makes wearable cardiac monitors, for approximately 287.5 million dollars in cash and stock.

These three transactions run in two directions at once. Capital is arriving to scale newly cleared platforms, and a large strategic is shedding a diagnostics business it no longer wants inside the group. Both moves make sense from where each party sits, because a specialist owner will fund a diagnostics business through a rebuild that a diversified parent judges a poor use of group capital.

Correction to the record. Siemens Healthineers was reported on 12, 14, and 19 August to be in preliminary discussions with private equity about selling its Diagnostics division, with Chinese procurement pressure and a laboratory platform migration given as the reasons.

No primary source confirmed it in any of those three reports, across a full two-week window.

If you are building a comparable set for a diagnostics valuation, this transaction is not evidence, and at least one published analysis has already treated it as though it were.

Vigilance

Getting a device approved is a project. It has a start, an end, a budget, and someone whose job finishes when it is done. Keeping a device on the market is not a project. It runs for as long as the product exists, it has no completion date, and in most companies under two hundred people nobody owns it exclusively.

This cycle showed what that asymmetry costs when a fault appears.

One fault, three regulators, seventeen days. ResMed’s Astral 100 and Astral 150 are ventilators used at home, by people who depend on them to breathe through the night. An internal supercapacitor can leak, and the leak can stop therapy.

The FDA classified the action as a Class I recall on 31 July, its most serious category, meaning the agency judged that continued use could cause serious injury or death. Spain’s AEMPS published its own safety notice on 5 August. The MHRA escalated to a National Patient Safety Alert on 17 August, the British top tier, which obliges every NHS organisation to act by a stated deadline and report completion centrally.

Same device, same root cause, three regulators, seventeen days between the first action and the last. Nothing in the three notices indicates they were coordinated.

Why they did not move together. European device regulation harmonised approval. A CE mark means the same thing in Lisbon and Helsinki, and that took two decades of work.

Vigilance was never harmonised to the same degree. Each national competent authority retains its own reporting formats, its own escalation tiers, its own deadlines, and its own decision about whether a fault warrants a notice, an alert, or nothing at all. EUDAMED will eventually carry some of this load, and it does not carry it yet.

So a company that built its regulatory function around getting products approved has built a function shaped for the harmonised problem, and vigilance is the unharmonised one. The mismatch stays invisible until a signal arrives, at which point one engineering fault becomes three separate remediation obligations, each with its own format, deadline, and evidence of completion, running on three clocks that nobody is synchronising for you.

Six further Class I recalls landed in the same window, from Boston Scientific, Abiomed and Oscor, Hamilton Medical, Baxter, Becton Dickinson, and Medline. Seven Class I actions from seven manufacturers in two weeks is a heavy cycle rather than a pattern about any one company. It is also seven reminders that this is routine rather than exceptional.

Two device categories that did not exist last month

When a new kind of device reaches the American market, something happens that has nothing to do with that device. The FDA writes the rules for the category it has just created, and every product that follows is measured against rules shaped to fit the first one through. That makes a first authorisation a competitive event as much as a regulatory one. It happened twice this cycle.

Why De Novo is different from the other two routes. The FDA has three main routes to market. Most devices clear through 510(k) by showing they are substantially equivalent to something already on the market, which is fast and cheap because the comparison does the work. High-risk devices go through premarket approval, which requires clinical evidence and takes years.

De Novo exists for devices with no equivalent product to compare against. It does something the other two routes do not: it creates a new device classification, and it writes the special controls that every subsequent product in that category must meet. Those controls are drafted around the device in front of the agency at the time.

For the first company through, that is a durable advantage. For everyone following, it is a specification written to someone else’s design.

Two of them, one month apart. On 21 July the FDA granted De Novo authorisation to Johnson and Johnson’s OTTAVA, a surgical robot integrated into the operating table, cleared for nine named upper abdominal procedures from gastric bypass to hiatal hernia repair.

On 19 August it authorised Vitestro’s Aletta, the first standalone robotic blood-draw device, for supervised use with adults in outpatient settings. Blood draw is among the most frequently performed procedures in medicine and among the least automated, and Aletta is the first machine cleared to perform it without a phlebotomist’s hands.

What counts as evidence

The question a hospital asks about new technology has changed. It used to be whether the thing works. Enough health systems have now been through two or three rounds of technology procurement, and enough of those purchases underdelivered, that the question has moved to whether the number in your deck can be trusted and who produced it.

Three numbers circulated this cycle, all describing technology working in hospitals, all sitting on different evidentiary footings, and all being used in the same procurement conversations.

The three, and where each one comes from. Apella reported a 7 percent rise in monthly surgical case volume at Houston Methodist, roughly 25 additional cases, achieved without adding operating rooms, staff, or hours. The study was peer-reviewed and authored independently of the company.

Oura’s chief medical officer, speaking to Politico, cited 86 percent of physicians seeing clinical value in wearable data against 6 percent who have integrated it into clinical workflows, across six surveyed countries.

Ochsner Health and Paradigm reported a 41 percent increase in trial screening capacity across 47 hospitals.

The first is independent research. The second is a survey finding about the gap between belief and practice. The third is an organisational self-report, published by the party that ran the deployment.

Why the gap between 86 and 6 is the whole story. Those two figures describe a market where clinical belief has stopped being the constraint. Physicians have largely been persuaded, and what stands between that persuasion and actual use is integration into a workflow, a procurement process, a budget holder, and evidence a finance director will accept.

That reframes what a vendor is actually selling into. The buyer is rarely unconvinced about the technology. They are unconvinced that the number in front of them will survive contact with their own institution, because they have previously bought a number that did not.

A self-reported figure carries no dishonesty and no weakness. It is simply a different kind of claim from an independently authored one, and a second-time buyer knows the difference even when the deck does not mark it.

The argument

Every decision in this edition happened after approval.

France graded seven devices against products it already funds and found no improvement in any of them, then refused an eighth from the company that leads its category worldwide. Britain funded a category it had never funded before. A single ventilator fault produced three separate regulatory obligations across three countries in seventeen days. Two FDA authorisations created device classifications that did not exist and wrote the requirements everyone following will be measured against. Hospital technology numbers circulated in procurement conversations on entirely different evidentiary footings.

None of these was a regulator deciding whether a device works. All of them decided whether it sells, whether it stays on the market, what it costs to build, and whether a buyer believes the claim.

There is a shared reason. Approval was the part of this system that got harmonised, standardised, and made predictable, because that is where two decades of European regulatory effort went. Everything downstream of approval stayed national and discretionary, with no mechanism synchronising one country’s decision to another’s. A company organised around the harmonised problem is organised around the part that no longer decides outcomes.

For a company selling devices in Europe over the next eighteen months, the work that gets you approved and the work that gets you paid are different work, drawn from different evidence, assessed by different people, on calendars you do not control. Both have to be built. Only one of them has a published deadline, which is precisely why the other one is usually late.

What each proxy does not tell you:

• A CE mark does not tell a health system what it will pay.
• An added-benefit grade does not tell you why, because France publishes the outcome without the reasoning.
• A clearance in one market does not tell you what a safety signal will cost you in the other three.
• A published percentage does not tell you who measured it.

Critical deadlines. Five dates that reach a European medtech company this cycle.

The CMS RAPID proposal is the one to read first. If finalised, a device holding FDA Breakthrough designation could receive a proposed national Medicare coverage decision on the same day as its FDA authorisation, compressing a wait that currently runs years into a matter of months. For a company whose American revenue model depends on Medicare coverage, that single change would move the point at which a device starts earning by more than any efficiency programme could. In-vitro diagnostics are excluded from the proposal as drafted, which is itself worth a comment if you make one.

International signals. Three developments outside Europe and the United States carry consequences dated 2027 and later.

Saudi Arabia’s SFDA issued its first authorisation under a new innovative-device pathway, giving a route to market that did not previously exist for products without an established comparator in the Kingdom. For a company holding a first-in-class device, a market that previously had no way to assess it now has one.

Abu Dhabi’s Department of Health published new device reporting standards alongside a governance framework for digital health products, part of an emirate-wide digital health programme that also includes a new intelligent surgical network.

China’s NMPA released 46 device standards with implementation staged between 2027 and 2029. Any manufacturer with a China plan should map which of the 46 touch their product now, because the earliest compliance dates arrive in a little over a year and standards work of this kind consumes engineering time that has to be scheduled rather than found.

Vital Signs: MedTech is produced by HealthSeed AG, Weidmannstrasse 5, 8046 Zurich, CH-020.3.056.106-6. HealthSeed holds interests in companies operating in this sector; where an edition touches a company we are invested in or working with, the relationship is declared in the section where it appears.