On 2 October 2026 the European Commission selected the first medical devices for a joint clinical assessment under the EU's Health Technology Assessment Regulation. Both are transcatheter mitral valve replacement systems. One comes from Edwards Lifesciences. The other comes from HighLife, a privately held, venture-backed French company whose system is not approved in the United States.
A joint clinical assessment, or JCA, is a new step for high-risk devices. It happens after CE marking and before most national reimbursement decisions. For a company planning to enter the EU with a Class III implant, it changes what evidence you need, when you need it, and who gets to read it. This note explains the mechanism from the instruments themselves, works out the timeline, and then uses our registration data to ask where these devices are sold outside Europe.
What changed
Commission Implementing Decision C(2026) 7113 of 2 October 2026 selects two devices for joint clinical assessment at Union level:
| Device | Manufacturer | EU certificate | Assessor / co-assessor |
|---|---|---|---|
| SAPIEN M3 Transcatheter Mitral Valve Replacement System | Edwards Lifesciences LLC | 11 April 2025 | Netherlands (ZIN) / Norway (DMP) |
| HighLife Trans-septal Mitral Valve Replacement System | HighLife SAS | 23 January 2026 | France (HAS) / Poland (AOTMiT) |
The Member State Coordination Group on HTA recommended both on 22 June 2026. The decision says each device meets five of the six selection criteria in the Regulation: unmet medical need, first in class, potential impact on patients and health systems, a significant cross-border dimension, and major Union-wide added value. The sixth criterion, software using artificial intelligence or algorithms, was not cited for either.
The decision also records a problem. Both devices already held their certificates when the recommendation was made, so national HTA bodies could have started their own assessments. The Coordination Group judged the risk of duplication acceptable because only one Member State had declared an ongoing national assessment of each device. The Commission says it expects to select devices quarterly. The Coordination Group's 2026 work programme planned for about five device JCAs this year, with the first expected to start in June. Two have been selected so far.
What a JCA is, and what it is not
A JCA is a scientific report on the relative effects of a device: how it performs on health outcomes against the comparators that Member States ask about, and how certain that evidence is (Regulation (EU) 2021/2282, Article 9(1)). It is written by HTA bodies from two Member States, reviewed by the others, and published.
It is not a price, a reimbursement decision or a verdict on value. The Regulation says the report must contain no value judgement on the device's overall clinical added value, and that pricing and reimbursement stay with each Member State (Articles 1(2) and 9(1)). What it does bind is process. A Member State carrying out its own HTA on the same device must give "due consideration" to the joint report, must annex the joint report and your dossier to its own report, and must not ask you again for evidence you already submitted at EU level (Article 13(1)). You, in turn, may not submit that same evidence nationally (Article 10(3)).
In practice, then, the JCA becomes the common evidence base for reimbursement across the EU. Each country still makes its own decision on it, but none of them makes that decision on a different file.
Which devices can be selected
This is the part most summaries get loose, and it decides whether any of this applies to you. Article 7(1)(c) of the Regulation covers Class IIb and Class III devices for which an expert panel has given a scientific opinion under the MDR's clinical evaluation consultation procedure (MDR Article 54). That procedure applies only to Class III implantable devices and to Class IIb active devices intended to administer or remove a medicinal product (MDR Article 54(1)). It does not apply to a certificate renewal, to a modification of your own marketed device that the notified body accepts does not worsen the benefit-risk balance, or where common specifications cover the clinical evaluation (Article 54(2)). Even within scope, the expert panel may decline to give an opinion, and it gives one mainly for novel devices with a possible major clinical impact (MDR Annex IX, section 5.1).
So there are three filters before the selection criteria are even applied: the device has to be the right kind, the consultation has to happen, and the panel has to issue an opinion. The European Medicines Agency's public list of expert panel opinions held 22 entries on 19 October 2026. Eleven of those are dated 2024 or later. Opinions are often listed long after they are given, so treat that count as a floor.
For software and AI companies, this matters. One of the selection criteria is the incorporation of AI, machine learning or algorithms. But the gate is still the expert panel consultation, and that consultation does not cover a stand-alone Class IIb or Class III software device. On our reading of Article 7(1)(c) with MDR Article 54(1), a stand-alone diagnostic AI cannot be selected for a JCA, however novel it is. The AI criterion has teeth for algorithms inside devices that are within the consultation's scope: closed-loop drug-delivery systems (Class IIb under Rule 12) and AI-driven implants such as neurostimulators. For in-vitro diagnostics, the equivalent route is Class D with an expert panel's views under IVDR Article 48(6).
The timeline, relative to CE marking
A device can be selected only after it is certified, but the process starts before that. When the Coordination Group recommends a device, the HTA secretariat tells the manufacturer and the notified body (Implementing Regulation (EU) 2025/2086, Articles 2 and 4). That can happen while certification is still in progress. The clocks after that are set out in Implementing Regulation (EU) 2025/2086, in force since 9 November 2025:
| Step | Deadline | Source |
|---|---|---|
| Send certificate and IFU to the HTA secretariat | 7 days after certification or the request, whichever is later | Art. 3 |
| Assessment scope (PICO) finalised | 60 days after that information, or 10 days after the Commission decision, whichever is later | Art. 12(2) |
| Scope explanation meeting, if you ask for one | Within 20 days of the scope being finalised | Art. 13 |
| Your dossier | 100 days from the Commission's request, extendable by up to 30 days in justified cases | Art. 14(2)–(3) |
| Commission checks the dossier | 15 working days | Art. 15 |
| Your answer to a request for missing evidence | 15 days (7 if minor). An incomplete dossier ends the JCA, and that is published. | Art. 14(4); Reg. Art. 10(6) |
| Revised draft reports finalised | 165 days after the dossier is confirmed | Art. 17(2) |
| Your factual-accuracy and confidentiality check | 7 days | Art. 16(4) |
| Coordination Group endorses | 30 days | Art. 17(5) |
For these two devices, the scope was due by 12 October 2026, ten days after the decision. Suppose the Commission's request went out in mid-October and every step takes its full time. That puts the dossier in late January 2027 and endorsement in late August 2027. The Commission then has ten working days for a procedural review before it publishes. That is our arithmetic from the statutory maxima, not a published schedule. On that basis, a published report lands about 20 months after HighLife's certificate and about 29 months after Edwards's.
What the company has to submit
The dossier contents are fixed by Annex II of the Regulation, in the template in Annex I of Implementing Regulation 2025/2086. It includes the notified body's clinical evaluation assessment report, your clinical evaluation documentation, the expert panel's opinion, and all studies you sponsored, published or not, including protocols and analysis plans. It also includes ongoing or discontinued studies you sponsor or fund, registry data, and the research questions answered against each PICO. If you had a joint scientific consultation, you must explain any departure from its advice.
Two features change the job compared with the MDR file. First, the questions are not yours. The assessor drafts a scope that has to reflect every Member State's needs on population, intervention, comparators and outcomes (Regulation Article 8(6); Implementing Regulation Articles 11–12). Your clinical evaluation proved safety and performance against the state of the art. The JCA asks how your device compares with what each health system uses today. A single-arm pivotal study answers the first question but not the second. That is our reading, and it is where the "degree of certainty" language in the report will bite. Second, it is published. Once the report is out, the Commission publishes the dossier with it, minus whatever it accepts as commercially confidential (Regulation Article 30(3)(d); Implementing Regulation Article 21).
What it costs a startup, and the one lever you have
There is no fee for a JCA, and joint scientific consultations are offered free in 2026. The cost is time and evidence. You need a team that can build an HTA-grade dossier against an externally set PICO within 100 days, at the same moment you are launching commercially. You also need to have designed your studies with a comparator in mind years earlier. We have not found a reliable published figure for what dossier preparation costs, and we will not guess one.
The lever is the joint scientific consultation (JSC). A device is eligible if it is likely to be the subject of a JCA and its clinical studies are still in the planning stage (Regulation Article 16(2)). A device JSC can run in parallel with the early expert panel consultation that MDR Article 61(2) offers for Class III and Rule 12 devices (Regulation Articles 16(5) and 17(2)). The outcome is not binding on anyone, but the JCA assessors must read it when they draft the scope (Implementing Regulation Article 11(1)). Capacity is small. The 2026 work programme plans two to five device JSCs for the whole year, and the last 2026 request window closes on 21 October 2026. If you are oversubscribed, selection uses criteria such as unmet need and first in class (Article 17(3)), so a startup is not at a structural disadvantage. But you have to ask while your trial is still a plan.
HighLife's position shows what the early decision is worth. By the time it learned of the recommendation, its device was certified and its evidence was what it was. A company two years from its pivotal study can still shape its comparator and endpoints to the questions the JCA will ask.
Where else these devices are registered: what our data shows
A JCA is an EU process, but the companies selected sell globally, and the evidence they generate will travel. We asked our registration data, as at 19 October 2026, where transcatheter mitral devices are registered outside the EU and US.
Method. We first searched by device type, using the GMDN term for a mitral transcatheter heart valve. That found only 9 records in two markets, because the same valve is coded differently from one registry to the next. We then searched by product and brand name across all 30 markets, and checked each record by hand.
| Device | United States | Elsewhere in our data |
|---|---|---|
| SAPIEN M3 (Edwards) | PMA approved 22 December 2025, with five supplements to July 2026 | No record in any other market |
| HighLife TSMVR | Investigational only | No device record anywhere. HighLife appears only as a registered EU manufacturer. |
| Tendyne (Abbott), the other CE-marked replacement system | PMA approved 23 May 2025 | Listed as active in Costa Rica and Saudi Arabia, and listed in Israel, whose registry publishes no status. Cancelled or lapsed in Hong Kong, Taiwan and Colombia. Accessories only in Brazil. |
For contrast, the established transcatheter mitral repair devices are everywhere. Abbott's MitraClip appears in 17 of our markets and Edwards's PASCAL in 12. Mitral valve replacement is still a first-market product. Outside the US and EU it has reached a handful of markets, and one system's registrations have already ended in three of them.
What the data could not answer is just as useful. Our EU device records do not yet include either selected device. EU device registration only became mandatory this year, so we do not read that absence as evidence of anything. We cannot tell why Tendyne's registrations ended in Hong Kong, Taiwan and Colombia. A cancelled registration is a fact; the reason for it is not in the record. And we do not hold HTA or reimbursement status as data for any market, so we cannot yet tell you which countries outside Europe will treat a published JCA as a reference.
Who it affects
- Makers of Class III implants and Rule 12 drug-delivery devices that are novel enough to draw an expert panel opinion. That is the pool.
- Startups with a first-in-class implant, where "first in class" is itself a selection criterion and where single-arm evidence is most common.
- Software and AI companies, but only where the algorithm sits inside an in-scope device. On our reading, stand-alone software cannot reach a JCA.
- Anyone sequencing EU reimbursement country by country, because a JCA freezes the evidence base you can use with every national HTA body.
What to do
- Work out whether your device will go through the expert panel consultation. Check whether it is a Class III implant or a Rule 12 device, and whether any Article 54(2) exemption applies. If it will not go through the consultation, the JCA cannot reach it.
- If it will, and your pivotal study is still a plan, request a JSC, ideally in parallel with an early expert panel consultation under MDR Article 61(2). The current window closes 21 October 2026. The 2027 windows will be in the Coordination Group's next work programme.
- Design for a comparator. Assume several Member States will ask about different standards of care, and plan evidence that can answer more than one PICO.
- Plan a 100-day dossier sprint for the months right after certification, and set aside people for it before launch, not after.
- Write everything as if it will be published. Decide early which information you will claim as commercially confidential, and be ready to justify each item.
The calendar
- 21 October 2026: the last 2026 request window for joint scientific consultations closes
- Quarterly: further device selections, as the expert panel opinions come through
- About January–February 2027: dossiers due for the first two devices (our arithmetic)
- About August–October 2027: first device JCA reports endorsed and published (our arithmetic)
Sources
- Commission Implementing Decision C(2026) 7113 of 2 October 2026 on the selection of medical devices for joint clinical assessment, and the Commission's JCA tracker for medical devices (data extracted 5 October 2026)
- Regulation (EU) 2021/2282 on health technology assessment: Articles 1, 7–13, 16–17 and 30, and Annex II
- Commission Implementing Regulation (EU) 2025/2086 (OJ L, 20 October 2025): Articles 2–4 and 11–21, and Annex I
- Regulation (EU) 2017/745 (MDR): Article 54, Article 61(2), and Annex IX section 5.1
- HTA Coordination Group, Annual Work Programme 2026 (adopted 28 November 2025)
- DG SANTE newsletter, 6 October 2026; EMA, list of expert panel opinions (consulted 19 October 2026); FDA PMA database, P250019
Which of your markets will want comparative evidence, which accept your CE file as it stands, and in what order to file are cross-market questions. Answering them takes more than one regulator's rulebook, and our market assessments are built for exactly that.