Free scoping read. A written assessment of your manuscript in two business days, at no cost.

Start one
Contents

Academy · Research Writing · Device regulatory challenges

Regulatory and Ethical Challenges for Medical Devices: Sources of Clinical Evidence under the 510(k) Pathway and EU MDR

The two largest device markets address the same problem in opposite ways, and the published evidence on each raises concerns. In the United States, equivalence to a predicate underpins most clearances, and only 4.4% of predicates in one JAMA analysis had any premarket clinical testing. In the European Union, EU MDR 2017/745 demands so much more clinical evidence that about 54% of high-risk manufacturers in one survey had planned to withdraw devices from the market.

Pubrica AcademyWritten by the medical writing and biostatistics teams Reviewed byPubrica biostatistics team Published9 October 2026 Last checked9 October 2026 Reading time15 minutes Sources26 peer-reviewed

Device regulation is neither a stricter nor a looser version of drug regulation; it is a distinct structure built on a different premise. Whereas a medicine is approved on trial evidence of its own effect, a device is frequently cleared on the basis of its similarity to a device already on the market, or certified on the basis of a clinical evaluation that may rest on literature concerning other devices. Both mechanisms have documented failure modes. This article sets out the requirements of each jurisdiction, the findings of published audits, and the ethical questions that neither framework resolves.

Learning objectives

  • How the US substantial-equivalence route and EU MDR conformity assessment differ in what they require manufacturers to demonstrate
  • What the published audits of the 510(k) pathway found about predicate devices and subsequent recalls
  • What EU MDR 2017/745 requires for clinical evaluation, when a clinical investigation is mandatory, and what the equivalence route requires
  • Why 62% of surveyed manufacturers cited lack of internal expertise, and the implications for resourcing a clinical evaluation report
  • Where the ethical gap sits for first-in-human and innovative surgical devices, and why it is not closed by either framework
  • How to write an equivalence justification that a notified body can assess

Common misconceptions

  • “A 510(k) device is FDA-approved.” It is cleared, on a determination of substantial equivalence to a legally marketed predicate. Approval is the term for the premarket approval pathway, and the evidentiary difference between the two is the central subject of this article.
  • “Substantial equivalence means the predicate was clinically proven.” Kadakia and colleagues, in JAMA, found that only 4.4% of the 229 unique predicate devices supporting a set of authorisations had undergone any premarket clinical testing.
  • “A recalled predicate cannot be used.” Of 156 devices authorised through the 510(k) pathway in that analysis, 56 (44.1%) relied on predicates that had been subject to a class 1 recall — the most serious category — and those devices carried a relative risk of subsequent recall of 6.40 (95% CI 3.59–11.40).
  • “The equivalence route under MDR is the low-cost option.” Demonstrating equivalence under EU MDR 2017/745 requires technical, biological and clinical equivalence and, where the equivalent device is not the manufacturer’s own, contractual access to its technical documentation. Such access is frequently impossible to obtain, which renders the route unavailable rather than low-cost.
  • “Ethics approval covers innovative surgical use.” Johnson and Rogers, in the Journal of Medical Ethics, and Muskens and colleagues, in the Journal of the American College of Surgeons, both identify a substantive gap: surgical innovation often falls between established practice and formal research, where neither the research ethics framework nor routine clinical governance clearly applies.

Contrasting premises of the US and EU systems

In the United States, most moderate-risk devices reach the market through premarket notification under section 510(k) of the Food, Drug, and Cosmetic Act, on a determination that the device is substantially equivalent to a legally marketed predicate in intended use and technological characteristics. Clinical data are required in a minority of these submissions 15. Higher-risk devices proceed through premarket approval, which does require clinical evidence of safety and effectiveness, and novel moderate-risk devices with no predicate can use the De Novo classification route. Significant-risk investigational studies are conducted under an investigational device exemption.

In the European Union, EU MDR 2017/745 replaced the preceding directives and requires conformity with the general safety and performance requirements in Annex I, assessed by a notified body against a class determined by the Annex VIII rules. Clinical evaluation is required under Article 61 and Annex XIV Part A for every device, and for class III and implantable devices a clinical investigation is required subject to the specific exceptions set out in Article 61 14. Post-market clinical follow-up under Annex XIV Part B is not optional. There is no European equivalent of substantial equivalence as a standalone route to market.

The consequence is that the same device faces structurally different evidence demands in the two markets, and may carry a different class in each. A European classification should not be inferred from a US one, or the reverse; that determination rests with the manufacturer’s regulatory affairs function and, in Europe, with its notified body. Pubrica describes what the instruments require and assists in evidencing it; Pubrica does not certify compliance.

The same device, two regulatory structures

United StatesEuropean Union
Legal instrumentFederal Food, Drug, and Cosmetic Act and its device amendments, implemented through 21 CFREU MDR 2017/745 (Regulation (EU) 2017/745), directly applicable
How class is determinedClassification regulation for the device type, which sets the submission routeThe classification rules in Annex VIII; class determines the conformity assessment route
Routes to market510(k) premarket notification on substantial equivalence; De Novo classification where no predicate exists; premarket approval for the highest-risk devicesConformity assessment against Annex I general safety and performance requirements, by a notified body for all but the lowest class
Role of equivalenceCentral — substantial equivalence to a legally marketed predicate in intended use and technological characteristics is the basis of clearanceAvailable within clinical evaluation only: Annex XIV Part A requires technical, biological and clinical equivalence and, for another manufacturer’s device, contractual access to its technical documentation. It is not a route to market in itself
Clinical evidence expectationRequired for premarket approval; required in a minority of 510(k) submissionsClinical evaluation required for every device under Article 61 and Annex XIV Part A; a clinical investigation is required for class III and implantable devices subject to the exceptions in Article 61
Post-market obligationsAdverse event reporting, post-approval studies where imposed, recall frameworkPost-market surveillance plan; post-market clinical follow-up under Annex XIV Part B; periodic safety update reporting; summary of safety and clinical performance for class III and implantable devices under Article 32

Summarised from the instruments named. Provisions are cited to orient the reader, not as legal advice; classification, route and the applicable exceptions are determinations for the manufacturer’s regulatory affairs function and, in the EU, its notified body.

Measured evidence on predicate devices

The criticism of the 510(k) pathway was formerly theoretical. It is now empirical, and the findings are specific 9,15. Kadakia and colleagues, in JAMA, examined 156 devices authorised through the 510(k) pathway and found that 56 of them (44.1%) relied on predicate devices that had been subject to a class 1 recall — the category FDA uses where there is a reasonable probability that use will cause serious adverse health consequences or death 15. Devices with such a predicate carried a relative risk of subsequent recall of 6.40 (95% CI 3.59–11.40). In the same analysis, only 4.4% of the 229 unique predicate devices supporting those authorisations had undergone any premarket clinical testing 15.

Everhart and colleagues, also in JAMA, examined the question at scale: across more than 35,000 devices cleared via the 510(k) pathway between 2003 and 2020, 11% of descendant devices were later recalled 9. The word “descendant” is central to this finding. Equivalence chains propagate: a device is cleared against a predicate, that device becomes a predicate itself, and the characteristics that justified the original determination become progressively more distant from the device actually in use 9.

The appropriate inference for a manufacturer is not that the pathway is illegitimate (it clears the overwhelming majority of devices in use and is unlikely to be abolished) but that predicate selection is a risk decision with a measured effect size, not an administrative one 9,15. Selecting a predicate with a class 1 recall history because it simplifies the equivalence argument entails a materially higher probability of the manufacturer’s own recall. That is a decision for the risk file, and our device risk management page covers where it belongs.

44.1%of 156 devices authorised via 510(k) relied on a predicate that had been subject to a class 1 recall
6.40relative risk of subsequent recall for those devices (95% CI 3.59–11.40)
4.4%of the 229 unique predicate devices had undergone any premarket clinical testing
11%of descendant devices later recalled, across more than 35,000 devices cleared 2003–2020

First three cells: Kadakia KT, Dhruva SS, Caraballo C, et al. Use of recalled devices in new device authorizations under the US FDA’s 510(k) pathway and risk of subsequent recalls. JAMA. Fourth cell: Everhart AO, Sen S, Stern AD, et al. Association between regulatory submission characteristics and recalls of medical devices receiving 510(k) clearance. JAMA.

Clinical evidence under EU MDR and its cost to manufacturers

MDR raised the clinical evidence requirement substantially, and its effect on manufacturers has been measured. Kearney and McDermott, in Therapeutic Innovation & Regulatory Science, surveyed MedTech professionals (68 of 80 invited responded, a response rate of 85%) and found that about 54% of high-risk device manufacturers planned or had planned to remove devices from the EU market under MDR. This represents a regulatory requirement affecting market composition rather than documentation alone 14.

The same survey identified a capacity problem underlying the compliance problem. Half of small and medium-sized enterprises (50%) outsourced clinical evaluation report writing, against 18% of large companies, and lack of internal expertise was cited by 62% 14. Taken together, these figures suggest that the binding constraint for smaller manufacturers is not willingness but the availability of personnel able to construct a clinical evaluation argument; this is a different problem from the testing budget and requires a different solution 14.

The substance of what MDR requires merits precise statement, because it is often described loosely. Article 61 requires a clinical evaluation based on clinical data, planned and documented through a clinical evaluation plan and reported in a clinical evaluation report. For class III and implantable devices, Article 61(4) requires a clinical investigation, subject to the exceptions set out in the following paragraphs of that article, including, for devices modified from a manufacturer’s own already-marketed device or for certain well-established technologies, routes that rely on existing data under stated conditions. Where equivalence to another manufacturer’s device is claimed, Annex XIV Part A requires demonstration of technical, biological and clinical equivalence and sufficient access to that device’s technical documentation, normally by contract. Post-market clinical follow-up under Annex XIV Part B informs the evaluation continuously, and for class III and implantable devices a summary of safety and clinical performance is required under Article 32. The applicable provisions and any relevant guidance should be confirmed with the notified body before planning proceeds on that basis.

For those preparing these documents, the regulatory writing and medical device and IVD pages describe the document set; where the clinical evaluation rests on published literature, that literature must be identified by a documented, reproducible search, which is a systematic review discipline rather than a citation exercise.

54%of high-risk device manufacturers planned or had planned to remove devices from the EU market under MDR
50% vs 18%of SMEs versus large companies outsourcing clinical evaluation report writing
62%cited lack of internal expertise

Survey of MedTech professionals; 68 of 80 invited responded (85%). Source: Kearney B, McDermott O. The Challenges for Manufacturers of the Increased Clinical Evaluation in the European Medical Device Regulations: A Quantitative Study. Therapeutic Innovation & Regulatory Science. These are reported plans and practices, not observed market outcomes.

Post-market clinical evidence

Both frameworks lean on post-market evidence to carry risk that premarket evidence did not resolve, and the quality of that post-market evidence is where the system is weakest. Hoogervorst and colleagues, in the International Journal of Health Policy and Management, systematically reviewed 20 cardiovascular and 26 orthopaedic registries for their usefulness in regulatory evaluation of device safety and performance across the implant lifecycle 12. Of 33 structure and methodology items assessed, a median of 33% (IQR 14–71%) were reported for cardiovascular registries and 60% (IQR 28–100%) for orthopaedic registries.

These medians are the central finding. A registry that does not report its own structure and methods cannot be appraised, which limits what regulators can conclude from it and what a manufacturer can claim on the basis of it 12. The spread (interquartile ranges from 14% to 71% in one domain) also means that generalisations about “registry evidence” are of limited value: some registries are well documented and some are not, and the specific registry on which reliance is intended must be appraised.

This has a practical implication for planning. If a post-market clinical follow-up plan relies on an existing registry, it should be confirmed that the registry captures the device at the required level of granularity, records the relevant failure mode, and documents its own methods sufficiently well to be cited. Discovering otherwise after certification is a substantial and avoidable problem.

The ethical gap unaddressed by either framework

Regulatory compliance and ethical adequacy are not the same test, and for devices the divergence is widest at first use in humans. Johnson and Rogers, in the Journal of Medical Ethics, examine innovative surgery and identify the structural problem: a surgeon modifying a technique or using a device in a new way is often doing something that is neither established practice nor formally constituted research, so neither research ethics review nor ordinary clinical governance clearly applies 25. Muskens and colleagues, in the Journal of the American College of Surgeons, reach a compatible conclusion about the introduction of novel devices in surgery, arguing that current oversight leaves ethically significant decisions to individual clinical judgement 22.

Three specific issues follow. The first is consent: a patient consenting to a procedure involving a device used in a novel way must be informed of that novelty and of what remains unknown about it; such information is harder to write than a standard consent form and is often omitted entirely 22,25. The second is the learning curve. Device outcomes depend on operator experience in a way that drug outcomes generally do not, so early results reflect the learning curve rather than the device, and both favourable and unfavourable early series are misleading for that reason. The third is that the manufacturer frequently has a commercial relationship with the early adopters generating the evidence, which should be declared rather than managed informally.

None of these issues is resolved by a compliant technical file. They merit explicit treatment in the clinical development plan, and where human participants are involved they intersect with the research governance framework set out in regulation of research involving human participants.

Borderline products and comparison with advanced therapies

Some products sit on the boundary between device, medicinal product and advanced therapy, and the classification decision is sufficiently consequential to be made deliberately rather than discovered. The advanced therapy experience illustrates the effect of a demanding evidence regime on approval rates: Yu and colleagues, in Frontiers in Bioengineering and Biotechnology, report that only 10 of 19 advanced therapy medicinal products submitted to the European Medicines Agency between 2009 and March 2018 were approved 20. The Committee for Advanced Therapies, writing in Nature Reviews Drug Discovery, set out the development and regulatory challenges that account for much of that attrition 26.

For a device manufacturer the relevance is twofold. If a product combines a device with cells, tissue or a gene therapy component, its route may not be the device route at all, and discovering this late is costly. If a product is a device delivering a biological, the drug-device combination rules apply, and the component that determines the primary mode of action determines the lead framework. Both are regulatory affairs determinations, and both should be settled before the clinical plan is written.

Construction of the equivalence argument

Whether a manufacturer is making a substantial equivalence case to FDA or an equivalence claim under Annex XIV of MDR, the failure mode is the same: a statement that the devices are similar, supported by a table of specifications and no reasoning about what the differences mean. A reviewer or notified body is assessing whether the differences could affect safety or clinical performance. That is the question the text must answer, difference by difference.

An equivalence claim, asserted versus reasoned

Asserts similarity

The subject device is equivalent to the comparator device, which has been on the market since 2015 with a well-established safety profile. Both devices are intended for the same clinical indication, use the same operating principle, and are manufactured from the same grade of titanium alloy. Minor differences in geometry and surface treatment do not affect safety or performance. Clinical data from the comparator device are therefore applicable to the subject device.

Which comparator, and can you access its documentation? Not identified; access is a condition of the equivalence route under MDR where the device is not the manufacturer’s own.

What are the differences, exactly? “Minor” is the claim under assessment, not a premise.

Why would each difference not affect safety or performance? Asserted in one clause for two unrelated differences.

Which endpoints are being read across? “Clinical data are therefore applicable” claims everything at once.

What if the assessor disagrees? With no partitioning, one rejected difference invalidates the whole claim.

Reasons from each difference

The comparator device is identified in Table 4 with its certificate and the basis of our access to its technical documentation. Technical equivalence: both devices have the same intended purpose, operating principle and principles of operation; they differ in stem geometry (lateralised offset, 4 mm greater) and in surface treatment (plasma-sprayed hydroxyapatite on the subject device, grit-blasted on the comparator). Each difference is assessed in Section 5.3. The geometry difference alters joint loading and is addressed by finite element analysis and by the fatigue testing in Section 6.1, run on the subject device geometry; it is not addressed by the comparator’s clinical data, and we do not claim that it is. The surface treatment difference affects the bone-implant interface and the biological evaluation, so hydroxyapatite-specific implantation and degradation data were generated rather than read across. Biological equivalence: the base alloy and contact category are identical; the coating is not, so the biological evaluation for the subject device was conducted on the coated finished device in its own right. Clinical equivalence: the indicated population, anatomical site and duration of use are the same. Conclusion: equivalence is claimed for the base material and intended purpose, and clinical data from the comparator are relied upon only for the endpoints listed in Table 6. For survivorship at the coated interface, equivalence is not claimed and the post-market clinical follow-up plan in Section 9 specifies prospective collection.

The comparator and the basis of documentation access are stated, so the route’s precondition is visibly met.

Differences are enumerated with magnitudes, and each is assessed separately against technical, biological and clinical equivalence.

Where a difference cannot be bridged by read-across, device-specific testing is presented instead and the limit is stated.

The read-across is partitioned by endpoint, so a disagreement about one does not collapse the argument.

The unresolved question is moved into post-market clinical follow-up explicitly rather than left implicit.

The second version concedes more and claims less, and is considerably more likely to be accepted. Partitioning the claim by endpoint is the most useful single step: it shows the assessor precisely what they are and are not being asked to accept, and it leaves the remainder of the argument intact if one element is challenged.

A single evidence programme for two submissions

The practical response to divergent requirements is not two programmes but a single programme designed against the stricter requirement and reported twice. In most cases MDR’s clinical evaluation expectation is the stricter, so a study designed to satisfy Article 61 and Annex XIV will generally also serve a US submission, whereas the reverse is frequently not the case. The programme should be designed once, at protocol stage, with both audiences identified.

The economic question should be incorporated at the same time. The purchaser’s question (whether the health gained is worth the additional cost against local practice) requires resource-use and outcome data that are inexpensive to collect prospectively and costly to reconstruct later, as set out in cost-effectiveness evidence for device manufacturers. The biological evaluation should also be settled separately and early, because the contact categorisation drives a testing programme with its own lead time; this is covered in selecting materials for a medical device.

On resourcing, the survey finding warrants serious attention. If lack of internal expertise was cited by 62% of respondents and half of smaller manufacturers outsource clinical evaluation report writing, the question for a small manufacturer is not whether to acquire that capability but when; acquiring it before the clinical evaluation plan is written represents considerably better value than acquiring it to correct a report that a notified body has already queried 14.

Limitations of the available evidence

Three limitations should be noted. First, the 510(k) audits establish association, not mechanism: Kadakia and colleagues show that devices with recalled predicates are recalled more often, which is a strong signal, but the studies do not establish that the predicate relationship caused the later recall rather than marking a class of devices that is intrinsically riskier 15. The practical implication is the same either way; the causal claim should not be overstated.

Second, the MDR impact figures describe intentions and early experience. The 54% of high-risk manufacturers planning or having planned EU withdrawal is a survey finding about plans, from 68 respondents, and what actually happened to market availability is a separate empirical question that will be answered over years 14. It should be treated as a measure of how the requirement was experienced, not of how many devices left the market.

Third, no study has established what level of premarket clinical evidence is optimal for devices. More evidence reduces the chance of a harmful device reaching patients but delays beneficial ones, and the two markets have made that trade-off differently without either having demonstrated that its setting is correct. Claims to the contrary, in either direction, rest on position rather than on data.

Key points

Key points

Device regulation rests on equivalence in the United States and on clinical evaluation in the European Union, and the published audits of both identify substantial weaknesses. Kadakia and colleagues found that 44.1% of 156 devices authorised through 510(k) relied on predicates subject to a class 1 recall, with a relative risk of subsequent recall of 6.40, and that only 4.4% of 229 unique predicates had any premarket clinical testing; Everhart and colleagues found 11% of descendant devices later recalled across more than 35,000 clearances. Under MDR the burden shifted the other way, with about 54% of high-risk manufacturers in one survey planning EU withdrawal and 62% citing a shortage of internal expertise. Manufacturers are advised to design one evidence programme against the stricter requirement, to choose predicates and comparators as risk decisions, to partition every equivalence claim by endpoint, and to settle classification and route with their regulatory affairs function and notified body rather than inferring them.

Application to your study

Application to your study

  1. Treat predicate or comparator selection as a risk decision and record it in the risk file. Check the recall history of any candidate predicate; the measured relative risk of subsequent recall for devices with a class 1 recalled predicate was 6.40.
  2. Determine class and route separately in each target market. US classification regulation and EU MDR Annex VIII rules apply respectively. One should never be inferred from the other, and regulatory affairs should document the determination.
  3. Verify the equivalence precondition before relying on it. Under Annex XIV, equivalence to another manufacturer’s device requires contractual access to its technical documentation. Without such access the route is unavailable, and this should be established before the plan depends on it.
  4. Enumerate every difference with its magnitude, and assess each one separately. Differences may be technical, biological or clinical. Any difference that cannot be bridged requires its own data, and this should be stated.
  5. Partition the read-across by endpoint. State which endpoints the comparator’s data support and which it does not, and move the remainder into post-market clinical follow-up explicitly.
  6. Appraise the specific registry to be relied upon, not “registry data” in general. Reporting of structure and methodology ran at a median of 33% of 33 items for cardiovascular registries and 60% for orthopaedic ones; it should therefore be confirmed that the chosen registry captures the device and the relevant failure mode.
  7. Identify literature for the clinical evaluation by a documented, reproducible search. A notified body will ask how the evidence was found; a citation list does not answer that.
  8. Address the novelty in the consent process where a device is used in a new way. The unknowns, and the operator’s experience with the device, should be addressed both in the consent discussion and in the consent document.
  9. Secure clinical evaluation expertise before writing the plan rather than after a query is raised. Lack of internal expertise was cited by 62% of surveyed professionals; as a known constraint, it should be budgeted for from the outset.

Frequently asked questions

No, and the distinction is substantive. A 510(k) clearance is a determination that a device is substantially equivalent to a legally marketed predicate in intended use and technological characteristics; clinical data are required in only a minority of such submissions. Approval is the term used for the premarket approval pathway, which applies to the highest-risk devices and does require evidence of safety and effectiveness for that device. A third route, De Novo classification, exists for novel moderate-risk devices with no suitable predicate. The practical point for anyone reading device literature is that “FDA cleared” and “FDA approved” describe different evidentiary bases, and marketing material frequently blurs them.

The findings are specific and consistent, though associational. Kadakia and colleagues, in JAMA, examined 156 devices authorised through the pathway: 56 (44.1%) relied on predicates that had been subject to a class 1 recall, and those devices had a relative risk of subsequent recall of 6.40 (95% CI 3.59–11.40). Only 4.4% of the 229 unique predicates involved had undergone any premarket clinical testing. Everhart and colleagues, also in JAMA, found that across more than 35,000 devices cleared between 2003 and 2020, 11% of descendant devices were later recalled. These studies do not establish that the predicate relationship caused the later recalls rather than marking intrinsically riskier device classes; in either case, predicate choice is a measurable risk decision.

Article 61 of EU MDR 2017/745 requires a clinical evaluation based on clinical data for every device, planned in a clinical evaluation plan and reported in a clinical evaluation report. For class III and implantable devices, Article 61(4) requires a clinical investigation, subject to the exceptions set out in the subsequent paragraphs of that article, which include routes for devices modified from a manufacturer’s own already-marketed device and for certain well-established technologies, each under stated conditions. Whether a particular exception applies to a given device is a determination to be made with the notified body, with reference to the relevant guidance, rather than an inference drawn from a summary such as this one.

Only through the equivalence provisions, which are demanding. Annex XIV Part A requires demonstration of technical, biological and clinical equivalence, and where the equivalent device is not the manufacturer’s own it requires sufficient access to that device’s technical documentation, which in practice means a contract with that manufacturer. Competitors rarely grant such access, which makes the route unavailable rather than merely difficult. Where the route can be used, the claim should be partitioned: the endpoints that the comparator’s data support should be stated, the manufacturer’s own data presented for differences that cannot be bridged, and any unresolved question moved into the post-market clinical follow-up plan.

The best-attributed figures available describe manufacturers’ experience and plans. Kearney and McDermott, in Therapeutic Innovation & Regulatory Science, surveyed MedTech professionals, with 68 of 80 invited responding (85%), and found that about 54% of high-risk device manufacturers planned or had planned to remove devices from the EU market under MDR. Half of SMEs outsourced clinical evaluation report writing against 18% of large companies, and 62% cited lack of internal expertise. These are reported intentions and practices rather than observed market outcomes; they should be read as a measure of how the requirement was experienced, particularly by smaller manufacturers, rather than as a count of devices withdrawn.

The specific registry should be appraised rather than the category. Hoogervorst and colleagues, in the International Journal of Health Policy and Management, reviewed 20 cardiovascular and 26 orthopaedic registries against 33 structure and methodology items and found a median of 33% reported (IQR 14–71%) for cardiovascular registries and 60% (IQR 28–100%) for orthopaedic ones. The wide interquartile ranges indicate that quality varies considerably, so a plan stating “we will use the national registry” is incomplete. It should be confirmed that the registry identifies the device at the required granularity, records the relevant failure mode, and documents its own methods sufficiently well to be cited.

Three, none of which is resolved by regulatory compliance. First, first-in-human and innovative use frequently falls between established practice and formal research: Johnson and Rogers, in the Journal of Medical Ethics, and Muskens and colleagues, in the Journal of the American College of Surgeons, both argue that current oversight leaves ethically significant decisions to individual clinical judgement. Second, consent must convey not only the procedure but also that the device use is novel and what remains unknown about it. Third, device outcomes depend on operator experience, so early series reflect a learning curve rather than the device; this effect operates in both directions and should be stated in any publication of early results. Commercial relationships with early adopters belong in the declarations.

This depends on the primary mode of action, and the determination should be obtained early rather than assumed. In the United States such a product is a combination product, with the primary mode of action determining which FDA centre leads. In the European Union, a product with cells, tissue or a gene therapy component may fall under the advanced therapy framework rather than the device framework, and an integral drug-device combination brings a notified body opinion requirement into a medicinal application. The advanced therapy route is demanding: Yu and colleagues, in Frontiers in Bioengineering and Biotechnology, report that only 10 of 19 advanced therapy medicinal products submitted to the EMA between 2009 and March 2018 were approved. The route should be settled before the clinical plan is written.

References

  1. U.S. Food and Drug Administration. Container Closure Systems for Human Drugs and Biological Products (draft guidance). U.S. Food and Drug Administration. 2026. read 9 October 2026. packagingdigest.com
  2. U.S. Food and Drug Administration, CDER Office of Generic Drugs. FDA's Inactive Ingredient Database (IID) Program & AI Consideration. FDA FY2026 Generic Drug Science and Research Initiatives Public Workshop. 2026. read 9 October 2026. fda.gov
  3. U.S. Food and Drug Administration (CDRH, CBER). Use of International Standard ISO 10993-1, "Biological evaluation of medical devices - Part 1: Evaluation and testing within a risk management process". U.S. Food and Drug Administration. 2023. read 9 October 2026. fda.gov
  4. U.S. Food and Drug Administration. Questions and Answers Regarding Enhancements to the Inactive Ingredient Database (IID). U.S. Food and Drug Administration (CDER). 2019. read 9 October 2026. fda.gov
  5. International Council for Harmonisation; U.S. Food and Drug Administration (CDER/CBER). Guidance for Industry Q8(R2) Pharmaceutical Development. U.S. Department of Health and Human Services, Food and Drug Administration. 2009;Revision 2. read 9 October 2026. fda.gov
  6. Daniel SA, Suya Prem Anand P, Jesuarockiam N, Tabrej K, Shabir Hussain K. Advancement in biomedical implant materials-a mini review. Frontiers in Bioengineering and Biotechnology. 2024;12. doi:10.3389/fbioe.2024.1400918 read 9 October 2026. doi.org/10.3389/fbioe.2024.1400918
  7. Ginn SL, Mandwie M, Alexander IE, Edelstein M, Abedi MR. Gene therapy clinical trials worldwide to 2023-an update. The Journal of Gene Medicine. 2024;26(8):e3721. doi:10.1002/jgm.3721 read 9 October 2026. doi.org/10.1002/jgm.3721
  8. Metuh P, Petersen PM, Ou Y. Recent Advances in Wireless Optoelectronic Biomedical Implants. Laser & Photonics Reviews. 2024;18(12):2300752. doi:10.1002/lpor.202300752 read 9 October 2026. doi.org/10.1002/lpor.202300752
  9. Everhart AO, Sen S, Stern AD, et al. Association between regulatory submission characteristics and recalls of medical devices receiving 510(k) clearance. JAMA. 2023;329:144-156. read 9 October 2026. tctmd.com
  10. Al-Zyoud W, Haddadin D, Hasan SA, Jaradat H, Kanoun O. Biocompatibility Testing for Implants: A Novel Tool for Selection and Characterization. Materials. 2023;16(21):6881. doi:10.3390/ma16216881 read 9 October 2026. doi.org/10.3390/ma16216881
  11. Shchaslyvyi AY, Antonenko SV, Tesliuk MG, Telegeev GD. Current State of Human Gene Therapy: Approved Products and Vectors. Pharmaceuticals. 2023;16(10):1416. doi:10.3390/ph16101416 read 9 October 2026. doi.org/10.3390/ph16101416
  12. Hoogervorst LA, Geurkink TH, Lubbeke A, et al. Quality and Utility of European Cardiovascular and Orthopaedic Registries for the Regulatory Evaluation of Medical Device Safety and Performance Across the Implant Lifecycle: A Systematic Review. International Journal of Health Policy and Management. 2023;12(1):1-11. doi:10.34172/ijhpm.2023.7648 read 9 October 2026. doi.org/10.34172/ijhpm.2023.7648
  13. Kohn DB, Chen YY, Spencer MJ. Successes and challenges in clinical gene therapy. Gene Therapy. 2023;30(10-11):738-746. doi:10.1038/s41434-023-00390-5 read 9 October 2026. doi.org/10.1038/s41434-023-00390-5
  14. Kearney B, McDermott O. The Challenges for Manufacturers of the Increased Clinical Evaluation in the European Medical Device Regulations: A Quantitative Study. Therapeutic Innovation & Regulatory Science. 2023. doi:10.1007/s43441-023-00527-z read 9 October 2026. doi.org/10.1007/s43441-023-00527-z
  15. Kadakia KT, Dhruva SS, Caraballo C, et al. Use of recalled devices in new device authorizations under the US Food and Drug Administration's 510(k) pathway and risk of subsequent recalls. JAMA. 2023;329:136-143. read 9 October 2026. tctmd.com
  16. Husereau D, Drummond M, Augustovski F, et al. Consolidated Health Economic Evaluation Reporting Standards 2022 (CHEERS 2022) Explanation and Elaboration: A report of the ISPOR CHEERS II Good Practices Task Force. Value in Health. 2022;25(1):10-31. read 9 October 2026. ispor.org
  17. Husereau D, Drummond M, Augustovski F, et al. Consolidated Health Economic Evaluation Reporting Standards 2022 (CHEERS 2022) statement: Updated reporting guidance for health economic evaluations. BMJ. 2022;376:e067975. read 9 October 2026. cris.iucc.ac.il
  18. Alhakamy NA, Curiel DT, Berkland CJ. The era of gene therapy: From preclinical development to clinical application. Drug Discovery Today. 2021;26(7):1602-1619. doi:10.1016/j.drudis.2021.03.021 read 9 October 2026. doi.org/10.1016/j.drudis.2021.03.021
  19. Wouters OJ, McKee M, Luyten J. Estimated Research and Development Investment Needed to Bring a New Medicine to Market, 2009-2018. JAMA. 2020;323(9):844-853. doi:10.1001/jama.2020.1166 read 9 October 2026. doi.org/10.1001/jama.2020.1166
  20. Yu TTL, Gupta P, Ronfard V, Vertes AA, Bayon Y. Recent Progress in European Advanced Therapy Medicinal Products and Beyond. Frontiers in Bioengineering and Biotechnology. 2018;6:130. doi:10.3389/fbioe.2018.00130 read 9 October 2026. doi.org/10.3389/fbioe.2018.00130
  21. Kang CW, Fang FZ. State of the art of bioimplants manufacturing: part I. Advances in Manufacturing. 2018;6:20-40. doi:10.1007/s40436-017-0207-4 read 9 October 2026. doi.org/10.1007/s40436-017-0207-4
  22. Muskens IS, Gupta S, Hulsbergen A, Moojen WA, Broekman MLD. Introduction of Novel Medical Devices in Surgery: Ethical Challenges of Current Oversight and Regulation. Journal of American College of Surgeons. 2017;225:558-565. read 9 October 2026. scholarlypublications.universiteitleiden.nl
  23. Atun R, Jaffray DA, Barton MB, et al. Expanding global access to radiotherapy. The Lancet Oncology. 2015;16(10). read 9 October 2026. oncology-central.com
  24. Mitera G, Swaminath A, Rudoler D, Seereeram C, Giuliani M, Leighl N, Gutierrez E, Dobrow MJ, Coyte PC, Yung T. Cost-Effectiveness Analysis Comparing Conventional Versus Stereotactic Body Radiotherapy for Surgically Ineligible Stage I Non-Small-Cell Lung Cancer. JCO Oncology Practice. 2014;10(3):e130-e136. doi:10.1200/jop.2013.001206 read 9 October 2026. doi.org/10.1200/jop.2013.001206
  25. Johnson J, Rogers W. Innovative surgery: The ethical challenges. Journal of Medical Ethics. 2012;38(1):9-12. doi:10.1136/jme.2010.042150 read 9 October 2026. doi.org/10.1136/jme.2010.042150
  26. Committee for Advanced Therapies (CAT) and CAT Scientific Secretariat. Challenges with advanced therapy medicinal products and how to meet them. Nature Reviews Drug Discovery. 2010;9(3):195-201. doi:10.1038/nrd3052 read 9 October 2026. doi.org/10.1038/nrd3052

Every source above is a peer-reviewed article, a publisher’s record of one, or a standard issued by the body that owns it. Each carries the date it was last read. Where a standard moves, this article is revised rather than left up with a note.

How to cite this article

Pick your referencing style and copy it. Your university or target journal may vary the punctuation — check its guide if it has one.

Pubrica Academy. (2026, October 9). Regulatory and Ethical Challenges for Medical Devices: Sources of Clinical Evidence under the 510(k) Pathway and EU MDR. Pubrica. https://pubrica.com/academy/research-services/how-to-overcome-regulatory-and-ethical-challenges-regarding-medical-device/

This article is written to be learned from. Cite it if you are referring to what it says; if you are using the evidence behind it, cite the sources listed above instead.

Regulatory and ethical readiness check

This checklist should be completed when the clinical plan is drafted, for each target market. Lines that cannot be completed are those that cause delay later. It is a planning aid, not a compliance determination; such determinations rest with the regulatory affairs function and the notified body.

REGULATORY AND ETHICAL READINESS CHECK
Device: ______________________   Intended purpose (as it will be labelled): __________

1. CLASSIFICATION AND ROUTE  (complete once per market; do not infer across markets)
   Market: __________
   Class: __________   Rule / regulation relied on: ______________________
   Route: [ ] 510(k)  [ ] De Novo  [ ] PMA  [ ] MDR conformity assessment  [ ] other
   Notified body (EU) engaged?  [ ] no  [ ] yes  __________
   Regulatory affairs sign-off on this determination: __________  date: __________

2. PREDICATE / COMPARATOR SELECTION  (a risk decision)
   Candidate: ______________________
   [ ] Recall history checked.  Class 1 recall in history?  [ ] no  [ ] yes
   [ ] If yes, the elevated recall risk is recorded in the risk file at: __________
   [ ] Technical documentation access secured (required for MDR equivalence to a
       device that is not ours).  Basis: ______________________

3. DIFFERENCE ANALYSIS  (one row per difference; no row says "minor")
   Difference            Magnitude   Technical / biological / clinical impact   Addressed by
   ____________________  _________   ________________________________________  ____________
   ____________________  _________   ________________________________________  ____________
   ____________________  _________   ________________________________________  ____________

4. READ-ACROSS PARTITION
   Endpoints we rely on comparator data for: ______________________
   Endpoints we generate our own data for: ______________________
   Endpoints deferred to post-market clinical follow-up: ______________________

5. CLINICAL EVIDENCE
   [ ] Clinical investigation required?  basis / exception relied on: ____________
   [ ] Literature identified by a documented, reproducible search  protocol ref: ____
   [ ] Clinical evaluation plan written BEFORE the report
   [ ] Author of the clinical evaluation report identified, with their qualification

6. POST-MARKET
   PMCF method: [ ] registry  [ ] prospective study  [ ] survey  [ ] other: ______
   If registry: name: __________
     [ ] identifies our device at the needed granularity
     [ ] captures the failure mode of interest
     [ ] documents its own structure and methods adequately to cite

7. ETHICS  (complete even where no formal research is planned)
   [ ] Is any use first-in-human or a novel use of an existing device?
   [ ] Consent documentation states the novelty and what is not yet known
   [ ] Operator experience / learning curve addressed in the analysis plan
   [ ] Research ethics or institutional review obtained, or the reason it does not apply
   [ ] Commercial relationships with investigators declared

8. BOUNDARY CHECK
   Any biological, cellular, tissue or gene therapy component?  [ ] no  [ ] yes
   If yes: primary mode of action determination obtained?  [ ] no  [ ] yes  ref: ______

Pubrica AcademyWritten and maintained by Pubrica’s subject matter experts. Articles are revised when the underlying standards change, and each article states the date on which it was last reviewed.

How Pubrica can help

Pubrica writes clinical evaluation plans and reports, equivalence justifications and post-market clinical follow-up plans to the structure against which notified bodies and reviewers assess them — difference by difference, with the read-across partitioned by endpoint. The team conducts the documented literature searches on which a clinical evaluation must rest, and appraises the registries on which a manufacturer plans to rely before any commitment is made. Pubrica describes what the instruments require and assists in evidencing it; classification, route and compliance remain determinations for the manufacturer’s regulatory affairs function and notified body.

Find something else

Search the Pubrica Academy

Access is free and requires no account. Enter a topic, or open the full library to filter by category, sub-topic or resource type.

Pubrica can review the intended purpose and candidate comparator of a device and identify where the equivalence argument is likely to be challenged.

A scoping assessment is free and takes two business days. It provides a written assessment of what the work requires and which specialist would undertake it.

Get a free quote

Regulations

Check your project against this rule

  1. 1Your work
  2. 2Your details
  3. 3Send
What do you need?

Material you send is held under signed confidentiality, with access limited to the assigned team and encrypted transfer. It is not used to train anything and is not passed to third-party tools. Ask for the agreement before you send anything you would rather cover first.