欧盟MDR 2.0提案对AI医疗设备监管的改革与遗留挑战
The proposed reform of the Medical Device Regulation and its implications for AI-enabled medical devices in the EU
该文评估欧盟委员会2025年12月提出的医疗器械法规修订方案(MDR 2.0)对AI增强型医疗设备(AIeMDs)的实际影响。提案旨在减少MDR与《人工智能法案》之间的双重监管重叠,允许将MDR从AIA附录第A节移至第B节;修订软件分类规则并扩大临床证据来源;引入创新产品优先审查路径、监管沙盒以及院内设备的宽松条件。
The European Medical Device Regulations have been criticized for imposing disproportionate barriers to innovation due to excessive regulatory requirements. In response, the European Commission proposed amendments to key regulations in December 2025. The reforms aim to eliminate double regulation, reduce documentation requirements, and introduce new regulatory tools. In this Comment, we argue that these changes are generally well-targeted but leave specific challenges for innovative artificial intelligence-enabled medical devices unresolved.
In December 2025, the European Commission (EC) published a proposal to amend the European Medical Device Regulation (MDR)1, the In Vitro Diagnostic Medical Devices Regulation (IVDR)2, and the AI Act (AIA)3, in a single legislative package4, referred to as the ‘MDR 2.0 proposal’ in this comment. The MDR and IVDR are the EU’s central frameworks governing the safety and performance of medical devices (including medical device software (MDSW) and AI-enabled medical devices (AIeMDs)) and in vitro diagnostics (e.g., blood tests). This proposal takes the form of a targeted revision and responds to the core critiques of the MDR, which describe the legislation as overly complex, imposing overly strict requirements even for low-risk devices, and introducing a heavy documentation burden5,6,7. This is further exacerbated by overlaps with other legislation, such as the AIA, resulting in increased development and approval costs and longer review times for Notified Bodies, independent organizations designated to assess whether a device meets regulatory requirements, which oversee the ‘Conformity Assessment’ for devices above Class I5,6,8,9. Manufacturers reported negative effects on their operations, which led to the cancellation or delay of new product launches, the discontinuation of products, and the cessation of offering certain devices, especially orphan devices, in the EU9,10,11,12. Some manufacturers also report that they consider leaving the EU market9,11. Temporary solutions, such as extending deadlines for meeting MDR requirements, have not resolved these underlying problems. In its evaluation of the current MDR, the EC concluded that the aforementioned challenges limit patient access to innovative medical technologies within the EU and reduce competitiveness13,14.
In this Comment, we offer a multidisciplinary expert interpretation of the MDR 2.0 proposal, drawing on the authors’ experience in regulatory science, law, health policy, and industry. We ask whether the proposed reforms, if implemented, would materially improve the regulatory environment for AIeMDs in the EU. Accordingly, we assess whether the proposal reduces the regulatory burdens identified by the EC without weakening evidence requirements, regulatory oversight, or patient protection. A provision-level comparison is provided in Supplementary Table 1.
The Comment reflects the situation of August 26th, 2026, and includes the most recent Draft Report of the European Parliament’s (EP) Committee on Public Health, which contains suggested amendments to the EC’s proposal as part of the ongoing legislative process15.
The MDR-AIA overlap and its proposed solution
In 2025, the EU’s Artificial Intelligence Board (AIB) and the Medical Device Coordination Group (MDCG) issued non-binding guidance (AIB 2025-1/MDCG 2025-6) on the interplay between the AI Act and the MDR/IVDR. It describes how AI Act requirements for high-risk AIeMDs can be incorporated into the medical device conformity assessment process. However, it does not eliminate the existing overlap between the two regulatory regimes, since meeting the integrated conformity assessment does not remove the need to fulfill the requirements of both frameworks.
The MDR 2.0 proposal aimed to reduce the overlap through an amendment to Annex I of the AIA. Currently, the MDR and IVDR are listed in Section A of that annex, so for AIeMDs classified as high-risk under the AIA, the full set of AIA requirements, including data governance, transparency, human oversight, accuracy, robustness, and cybersecurity, applies alongside the MDR’s general safety and performance requirements. This affects all medical devices that must involve a Notified Body in their Conformity Assessment, usually Class IIa or higher, where the AI system itself is a medical device or meets the AIA definition of a safety component. The proposed changes would move the MDR and IVDR from Section A to Section B of Annex I, resulting in the AIA Chapter III obligations not directly applying to medical devices.
This route has also been discussed by the European Parliament and Council in the context of the Digital Omnibus on AI (Regulation (EU) 2026/1744), which was adopted in July 202616. Here, legislators adopted a different route for simplification. With regard to the MDR and IVDR, the Digital Omnibus permits the Commission to limit AIA requirements where sectoral legislation provides an equivalent level of protection and delays the applicability of high-risk requirements to August 2028. The Commission must adopt the corresponding delegated acts by 2 August 2027 (Fig. 1). The move of the MDR from Section A to Section B, however, remains an option for targeted revisions.
Major legislative events since the adoption of the AIA (blue framed boxes for adopted policy, green boxes for relevant deadlines) and their timeline are shown. The figure distinguishes the double-regulation reduction route originally proposed in COM(2025) 1023 (red) from the subsequently enacted Digital Omnibus route (white with blue line), which retained the MDR and IVDR in Section A and now forms the legal baseline. From that baseline, it shows the possible outcomes of the ongoing negotiation processes, including delay or non-adoption, adoption aligned with the Digital Omnibus, and substantial revision (orange).
New classification rules
For MDSW and AIeMDs, the current MDR classification rules have been described as problematic by stakeholders and researchers17. Under MDR classification rule 11, software that provides information used to make decisions for diagnostic or therapeutic purposes is typically classified as at least Class IIa, leading to Notified Body involvement. The proposal revises rule 11 to make MDSW Class I by default, with up-classification when their output informs or drives clinical management in increasingly serious situations.
However, when considering the wording of the proposed version, informing clinical management in a serious or critical situation already results in Class IIa. Because non-serious situations also fall under Class IIa, little scope remains for the Class I default, which, we think, is therefore likely to stay exceptional or to require interpretive guidance. Rule 3.5, which assigns the highest applicable sub-rule, is left unchanged and reinforces this effect.
In our interpretation, the revision does not remove rule 11’s central difficulty. Instead, the distinction between informing and driving clinical management is relocated to the Class I boundary, where misclassification carries greater risk. The fault lines remain vague, inviting divergent interpretation and inconsistent decisions. To partially address this, the proposal codifies the ‘Helsinki procedure’ through changes in MDR Article 4, and new Articles 4a, 51a, and 51b4. Here, Notified Bodies and manufacturers can seek clarification from competent authorities regarding the classification of a medical device in the event of a dispute, while competent authorities can reach out to an expert panel to seek clarification.
Additionally, the rule also retains the two principal risk dimensions of the older International Medical Device Regulators Forum (IMDRF) N12 guidance (2014), namely, the significance of the software-provided information to the healthcare decision and the seriousness of the healthcare situation18. This leaves the rule anchored in a comparatively narrow and now dated conception of software risk. The more recent IMDRF N81 guidance reflects a broader understanding of software-specific risk by also considering factors such as autonomy, explainability, the software’s role in the clinical workflow, the intended user and use environment, and the management of software changes19. Thus, while the intention was to achieve a substantial improvement, we think that the current wording of the proposal and its interaction with rule 3.5 would have the opposite effect.
Expansion of clinical evidence
Also important for AIeMDs is the expansion of the definition of clinical data. The current MDR specifies that clinical data must come from clinical investigations, peer-reviewed literature, or post-market surveillance. The proposed revision broadens this to also include studies published in scientific literature that may not be peer-reviewed. Whether this also includes preprints remains unclear. The proposal further broadens the possibility to demonstrate a device’s safety and performance solely through non-clinical data, explicitly encouraging the adoption of New Approach Methodologies (NAMs), such as in-silico testing.
In our view, this represents a meaningful expansion of the evidentiary toolkit for types of AIeMDs, for which traditional randomized clinical investigations may be impractical or methodologically ill-suited to capturing algorithmic performance20. There is, however, a trade-off: Relaxing clinical evidence requirements could create space for innovation, but it also raises legitimate questions about whether the resulting evidence base will be robust enough to protect patients, particularly for higher-risk AIeMDs.
A new pathway for innovative products
The proposal outlines a new pathway for AIeMDs that introduce highly innovative diagnostic or therapeutic methods with a significant positive impact, either by performing substantially better than existing devices or by addressing unmet needs. After the review by a designated expert panel, these devices would receive priority and rolling reviews by Notified Bodies, as well as access to scientific, technical, and regulatory advice from experts.
Our critique is that this process could become uncoordinated, as the scientific advice from the panel and the conformity assessment are structured as largely separate steps rather than an integrated dialog among the manufacturer, the panel, and the Notified Body. The EP draft report addresses this partially by stating that the notified body shall be able to participate in the expert panel’s discussions15. Additionally, the pathway is confined to conformity assessment and is not linked to health technology assessment or reimbursement. Selected Class IIb and III devices meeting the criteria in Article 7 of Regulation (EU) 2021/2282 may nevertheless undergo EU joint clinical assessment, while Member States retain responsibility for conclusions on added clinical value, coverage, and reimbursement21. We think that the success of this approach in the EU depends on the practical implementation of the somewhat vague proposal text, on how broadly or cautiously expert panels interpret the designation criteria, and on whether Notified Bodies can deliver faster reviews given the existing capacity constraints.
Regulatory sandboxes
A regulatory tool that has already been widely discussed among researchers, policymakers, and manufacturers has now been included in the proposal: regulatory sandboxes22. These sandboxes are time-limited environments where manufacturers of innovative medical devices and regulators can collaborate to test new methods for evidence generation, regulatory requirements, and enforcement practices under conditions defined in the authority-approved sandbox plan. For AIeMDs, sandboxes can provide a controlled environment for testing adaptive algorithms and new human-AI interaction models that do not fit within the traditional frameworks or where standard requirements would impede or significantly delay device development and access. They are already part of the AIA, but AIB 2025-1/MDCG 2025-6 and the Digital Omnibus on AI describe no medical device-specific considerations.
At the same time, we see some potential issues. The proposal does not clarify how the MDR sandboxes will interact with the AIA’s sandbox provisions and those of the draft General Pharmaceutical Regulation (GPR)23, the draft EU Biotech Act24, or with local regulation on clinical research. This raises questions about harmonization and potential conflicts between sandbox approaches. Additionally, sandboxes are not instruments for large-scale authorization of new technologies, as they are time-limited and project-specific testing environments under regulatory supervision. Their effectiveness depends on whether policymakers apply the generated insights to adapt the regulatory framework for all devices.
In-house devices and hospital-developed AI systems
Another aspect of the reform concerns in-house devices, which are devices manufactured and used within healthcare institutions and not sold on an industrial scale, as defined in MDR Art. 5(5)1. Under current rules, these devices face limited MDR requirements, including exemptions from conformity assessment procedure obligations, provided certain conditions are met: there must be no equivalent commercial device available, and in-house devices cannot be sold themselves nor transferred to another legal entity1.
The proposal relaxes several of these conditions to reduce administrative burden on health institutions, to promote clinical research on in-house devices, and to widen patient access where no commercial alternative exists. In-house devices would be transferable to other health institutions when justified by the interests of public health, patient safety, or patient health, or to prepare for or respond to a public health emergency. If a commercial equivalent becomes available on the market, health institutions would have a ten-year transition period before the in-house exemption stops under the MDR. Under the IVDR, that requirement is removed altogether. The EP draft report proposes to lower the regulatory hurdles further by introducing a proportionate application of Annex I general safety and performance requirements15.
We think that these changes could reduce barriers to developing and sharing self-developed tools for hospitals and academic medical centers. The proposed 10-year transition would also increase the viability of investing in the local implementation of in-house development capabilities, as these institutions have a guaranteed period of use even if new commercial alternatives emerge.
However, we also see three challenges. First, the eligibility and transfer conditions are weak gatekeepers for software. Because any medical device is by definition intended to serve patient health, the “patient health” ground for transfer imposes little real constraint. As software can be differentiated by adding or altering features, the requirement that no equivalent device be available on the market is easily satisfied. Second, the ten-year transition could leave in-house AIeMDs in routine clinical use for extended periods without the Notified Body oversight applied to commercial devices. Third, the reform may open an alternative route to the market as a commercial developer could supply software placed into service under a health institution’s responsibility and quality management system, reaching patients without undergoing its own conformity assessment. Another problem we see is the reduced transparency in the EP draft report, which proposes to remove the public declaration requirement under Article 5(5)(e)15.
Post-market changes to AIeMDs
A characteristic of many innovative AI systems is adaptability, meaning they can continue to change during deployment in response to new data, feedback, and user interactions20. This allows them to compensate for shifts in patient populations, clinical practice, and measurement conditions that gradually erode the accuracy of a static model25. Under the current regulatory framework, these updates may be considered a significant change (under the MDR, respectively, a substantial modification under the AIA), triggering a new conformity assessment1. The resulting documentation burden and costs often lead manufacturers to market their devices with a static AI model after placement on the market.
The MDR proposal addresses this issue to some extent by introducing predetermined change control plans (PCCPs), a mechanism already allowed under the AIA3. PCCPs are submitted with the medical device application and allow predicted and predefined changes to be made to an authorized, on-market device without requiring a new authorization process26. For AIeMDs, this would permit some post-market modifications, but only for changes that can already be anticipated at the time of submission, which is often not the case for innovative AIeMDs26. Consequently, experiences from manufacturers using PCCPs under other legislation, such as in the US, show that the actual adaptability and flexibility these plans allow is dependent on what the regulatory body is willing to accept.
Thus, we think PCCPs are a positive development, but they do not resolve the inherent conflict between a regulatory framework designed for devices and AI systems with an architecture or model intended to evolve during deployment.
Characteristics of advanced AI systems are not addressed
In our view, several known challenges related to advanced AIeMDs are not addressed in this proposal. The reform does not fundamentally reconsider how the MDR manages continuous learning algorithms. While PCCPs offer a mechanism for anticipated changes, they do not accommodate the open-ended adaptation that defines many advanced AI systems, such as diagnostic models that continuously learn from new data27, or agentic systems that query updatable databases (e.g., PubMed)28. The proposal also does not address broad, orchestrating, or agentic AI systems, which are increasingly prevalent in medical research29 and moving toward clinical deployment30. Table 1 summarizes the proposed changes and their impact.
Conclusion
In our assessment, the proposed MDR reform would improve the regulatory environment for AIeMDs, but only partially. While it is the most significant attempt to address the structural challenges for AIeMDs so far, several known problems remain unaddressed. Especially advanced characteristics of AI systems, such as open-ended learning, agentic AI, and the unprecedented versatility of adaptive, broad-purpose AI able to handle previously unseen tasks, all remain outside the scope of these reforms. Additionally, many of the operational details have been delegated to future legal instruments that do not yet exist.
The practical effectiveness of the reforms will depend on the MDCG, national competent authorities, and Notified Bodies, as the MDR is a regulatory framework whose real impact is shaped downstream by MDCG guidance documents, common specifications, standards, and day-to-day regulatory practice. We think that well-designed delegated or implementing acts and consistent interpretation across Member States could deliver the intended simplification; inconsistent or overly cautious implementation could reproduce the fragmentation the proposal aims to correct.
Currently, the reforms are only proposals and have not become law. The legislative package must still be adopted by the European Parliament and Council, where amendments and revisions are probable. This process is expected to take at least a year, making adoption in 2027 or 2028 likely. In the meantime, the Digital Omnibus on AI16, has established the current route to resolve the MDR-AIA overlap. AIeMDs remain in Annex I, Section A, while the Commission may limit equivalent AIA obligations through delegated acts. The further MDR 2.0 reforms may be delayed or not adopted, adopted after alignment with the Omnibus, or more substantially revised, strengthening or weakening their effects on innovation and patient protection. Until the final MDR reforms and the delegated acts are fixed, the landscape in which manufacturers are currently making decisions remains uncertain.
Data availability
No datasets were generated or analyzed during the current study.
References
European Parliament, Council of the European Union. Regulation (EU) 2017/745 of the European Parliament and of the Council of 5 April 2017 on Medical Devices, Amending Directive 2001/83/EC, Regulation (EC) No 178/2002 and Regulation (EC) No 1223/2009 and Repealing Council Directives 90/385/EEC and 93/42/EEC. OJ L vol. 117 (2017).
European Parliament, Council of the European Union. Regulation (EU) 2017/746 of the European Parliament and of the Council of 5 April 2017 on in vitro diagnostic medical devices and repealing Directive 98/79/EC and Commission Decision 2010/227/EU (Text with EEA relevance). http://data.europa.eu/eli/reg/2017/746/oj/eng (2017).
European Parliament, Council of the European Union. Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 Laying down Harmonised Rules on Artificial Intelligence and Amending Regulations (EC) No 300/2008, (EU) No 167/2013, (EU) No 168/2013, (EU) 2018/858, (EU) 2018/1139 and (EU) 2019/2144 and Directives 2014/90/EU, (EU) 2016/797 and (EU) 2020/1828 (Artificial Intelligence Act). (2024).
European Commission. Proposal for Amending Regulations (EU) 2017/745 and (EU) 2017/746. COM(2025) 1023 (2025).
Ben-Menahem, S. M., Nistor-Gallo, R., Macia, G., von Krogh, G. & Goldhahn, J. How the new European regulation on medical devices will affect innovation. Nat. Biomed. Eng. 4, 585–590 (2020).
Carl, A.-K. & Hochmann, D. Impact of the new European medical device regulation: a two-year comparison. Biomed. Tech. (Berl.) 69, 317–326 (2024).
Freyer, O. The regulation of medical devices in developing EU frameworks. (TU Dresden, Dresden, Germany, 2025).
Melvin, T. The European Medical Device Regulation-what biomedical engineers need to know. IEEE J. Transl. Eng. Health Med. 10, 4800105 (2022).
Svempe, L. Impact of the new European medical device regulation on digital health manufacturers. NPJ Digit. Med. 9, 638 (2026).
McDermott, O. & Kearney, B. A review of the literature on the new European Medical Device Regulations requirements for increased clinical evaluation. Int. J. Pharm. Healthc. Mark. 19, 1–21 (2025).
Widmann, M. Mangel an Medizinprodukten: Plötzlich ein Loch im Kinderherz – weil ein Teil für wenige Euro fehlt. Die Zeit (2024).
MedTech Europe. MedTech Europe IVDR & MDR Survey Results 2024. https://www.medtecheurope.org/wp-content/uploads/2025/01/mte_report_ivdr_mdr_2024-v7.pdf (2024).
Draghi, M. The Draghi Report on EU Competitiveness. https://commission.europa.eu/topics/eu-competitiveness/draghi-report_en (2024).
European Commission. COMMISSION STAFF WORKING DOCUMENT EVALUATION on the Targeted Evaluation of Regulation (EU)2017/745 on Medical Devices and Regulation (EU)2017/746 on In Vitro Diagnostic Medical Devices Accompanying the Document Proposal for a REGULATION OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL Amending Regulations (EU) 2017/745 and (EU) 2017/746 as Regards Simplifying and Reducing the Burden of the Rules on Medical Devices and in Vitro Diagnostic Medical Devices, and Amending Regulation (EU) 2022/123 as Regards the Support of the European Medicines Agency for the Expert Panels on Medical Devices and Regulation (EU) 2024/1689 as Regards the List of Union Harmonisation Legislation Referred to in Its Annex I. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex:52025SC1051 (2025).
Schenk, O. DRAFT REPORT on the Proposal for a Regulation of the European Parliament and of the Council Amending Regulations (EU) 2017/745 and (EU) 2017/746 as Regards Simplifying and Reducing the Burden of the Rules on Medical Devices and in Vitro Diagnostic Medical Devices, and Amending Regulation (EU) 2022/123 as Regards the Support of the European Medicines Agency for the Expert Panels on Medical Devices and Regulation (EU) 2024/1689 as Regards the List of Union Harmonisation Legislation Referred to in Its Annex I (COM(2025)1023 – C10-0360/2025 – 2025/0404(COD)). https://www.europarl.europa.eu/doceo/document/SANT-PR-787987_EN.html (2026).
European Parliament, Council of the European Union. Regulation (EU) 2026/1744 of the European Parliament and of the Council of 8 July 2026 Amending Regulations (EU) 2024/1689, (EU) 2018/1139 and (EU) 2023/1230 as Regards the Simplification of the Implementation of Harmonised Rules on Artificial Intelligence (Digital Omnibus on AI) (2026).
Gerhart, M. MDR Rule 11: The Classification Nightmare. Johner Institute. https://blog.johner-institute.com/regulatory-affairs/mdr-rule-11/ (2026).
IMDRF Software as a Medical Device (SaMD) Working Group. “Software as a Medical Device”: Possible Framework for Risk Categorization and Corresponding Considerations. https://www.imdrf.org/sites/default/files/docs/imdrf/final/technical/imdrf-tech-140918-samd-framework-risk-categorization-141013.pdf (2014).
IMDRF Software as a Medical Device Working Group. Characterization Considerations for Medical Device Software and Software-Specific Risk. https://www.imdrf.org/sites/default/files/2025-01/IMDRF_SaMD%20WG_Software-Specific%20Risk_N81%20Final_0.pdf (2025).
Rosenthal, J. T., Beecy, A. & Sabuncu, M. R. Rethinking clinical trials for medical AI with dynamic deployments of adaptive systems. NPJ Digit. Med. 8, 252 (2025).
European Parliament, Council of the European Union. Regulation (EU) 2021/2282 of the European Parliament and of the Council of 15 December 2021 on Health Technology Assessment and Amending Directive 2011/24/EU (2021).
Gilbert, S. Digital medicine’s international race for regulatory sandboxes and voluntary alternative pathways picks up tempo. NPJ Digit. Med. 9, 190 (2026).
European Commission. Proposal for a REGULATION OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL laying down Union procedures for the authorisation and supervision of medicinal products for human use and establishing rules governing the European Medicines Agency, amending Regulation (EC) No 1394/2007 and Regulation (EU) No 536/2014 and repealing Regulation (EC) No 726/2004, Regulation (EC) No 141/2000 and Regulation (EC) No 1901/2006. https://eur-lex.europa.eu/resource.html?uri=cellar:e3f40e76-e437-11ed-a05c-01aa75ed71a1.0001.02/DOC_1&format=PDF (2023).
European Commission. Proposal for a Regulation to Establish Measures to Strengthen the Union’s Biotechnology and Biomanufacturing Sectors (European Biotech Act). (2025).
Subbaswamy, A. & Saria, S. From development to deployment: dataset shift, causality, and shift-stable models in health AI. Biostatistics 21, 345–352 (2020).
Freyer, O., Jayabalan, S., Kather, J. N. & Gilbert, S. Overcoming regulatory barriers to the implementation of AI agents in healthcare. Nat. Med. 31, 3239–3243 (2025).
Kiyasseh, D., Zhu, T. & Clifton, D. A clinical deep learning framework for continually learning from cardiac signals across diseases, time, modalities, and institutions. Nat. Commun. 12, 4221 (2021).
Ferber, D. et al. Development and validation of an autonomous artificial intelligence agent for clinical decision-making in oncology. Nat. Cancer 6, 1337–1349 (2025).
Freyer, O. et al. The regulation of artificial intelligence in intensive care units: from narrow tools to generalist systems. NPJ Digit. Med. 9, 246 (2026).
Gallifant, J. et al. An agentic AI system enhances clinical detection of immunotherapy toxicities: a multi-phase validation study. medRxiv (2026) https://doi.org/10.64898/2026.02.26.26347179.
Acknowledgments
This work was supported by the European Commission under the Horizon Europe Program, as part of the project CYMEDSEC (101094218). The views and opinions expressed are those of the authors only and do not necessarily reflect those of the European Union. Neither the European Union nor the granting authorities can be held responsible for them. Responsibility for the information and views expressed therein lies entirely with the authors. This work has been supported by the BRIDGE project, funded by the Innovative Health Initiative Joint Undertaking (IHI JU) under grant agreement No 101165912. The IHI JU receives support from the European Union’s Horizon Europe research and innovation programme and from COCIR, EFPIA, EuropaBio, MedTech Europe, and Vaccines Europe. This publication reflects only the views of the authors and does not necessarily represent the views of the BRIDGE consortium. This work was supported by the German Federal Ministry of Education, Research and Space (BMFTR) as part of the Zukunftscluster SEMECO (03ZU2210BA). The views and opinions expressed are those of the authors only and do not necessarily reflect those of their employer. During the preparation of this work, the authors used DeepL (DeepL SE), Grammarly (Grammarly, Inc.), and ChatGPT (version GPT-5.5; OpenAI, Inc.) to improve the manuscript's grammar, spelling, and readability. After using these tools and services, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.
Author information
Authors and Affiliations
Else Kröner Fresenius Center for Digital Health, TUD Dresden University of Technology, Dresden, Germany
Oscar Freyer, Anett Schönfelder & Stephen Gilbert
Philips, Brussels, Belgium
Koen Cobbaert
Axon Lawyers, Amsterdam, The Netherlands
Erik Vollebregt
Faculty of Business and Economics, TUD Dresden University of Technology, Dresden, Germany
Stephen Gilbert
Authors
- Oscar Freyer
- Koen Cobbaert
- Anett Schönfelder
- Erik Vollebregt
- Stephen Gilbert
Contributions
OF: Conceptualization, Methodology, Investigation, Writing - Original Draft, Visualization, Project Administration. KC: Writing - Review & Editing, Visualization. AS: Writing - Review & Editing. EV: Writing - Review & Editing. SG: Conceptualization, Writing - Review & Editing, Visualization, Supervision, Funding Acquisition. All authors have read and approved the manuscript.
Corresponding author
Correspondence to Oscar Freyer.
Ethics declarations
Competing interests
OF has a leadership role and holds stock in WhalesDontFly GmbH, has had consulting relationships with Prova Health Ltd, and has a subcontractual consulting relationship with the Saudi FDA. OF is an Associate Editor for npj Digital Medicine. OF was not involved in the journal’s review of, or decisions related to, this manuscript. KC is employed by Philips and represents industry positions through COCIR and DITTA in European and international regulatory and standardization working groups. KC also teaches professional training courses through IPI Academy. SG declares the following competing financial interests: he has or has had consulting relationships with Ada Health GmbH, AMBOSS SE, European Commission Directorate General for Research & Innovation, Flo Ltd., FORUM Institut für Management GmbH, High-Tech Gründerfonds Management GmbH, Haleon Plc., ICURA ApS, Lindus Health Ltd., Prova Health Ltd., Rock Health Inc., Saudi Arabia Food and Drug Authority (FDA) funded through the United Nations Development Programme (UNDP), Thymia Ltd., Una Health GmbH and holds share options in Ada Health GmbH. SG is a News&Views editor for npj Digital Medicine. SG was not involved in the journal’s review of, or decisions related to, this manuscript. EV holds stock in Raiana B.V. and works for a number of medical device companies with AI systems that qualify as medical devices. The other authors do not have a competing interest.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Freyer, O., Cobbaert, K., Schönfelder, A. et al. The proposed reform of the Medical Device Regulation and its implications for AI-enabled medical devices in the EU. npj Digit. Med. 9, 907 (2026). https://doi.org/10.1038/s41746-026-03475-8
Received: 23 June 2026
Accepted: 20 October 2026
Published: 18 November 2026
Version of record: 18 November 2026
DOI: https://doi.org/10.1038/s41746-026-03475-8
来源:Nature 健康护理主题 · nature.com