EU MDR vs FDA: medical device clinical investigations compared

Companies running device studies on both sides of the Atlantic work under two structurally different regimes. In the United States, a clinical investigation of an investigational device sits under the Investigational Device Exemption (IDE) framework at 21 CFR Part 812, backed by the informed consent rules at 21 CFR Part 50 and the IRB rules at 21 CFR Part 56 ^1347916. In the European Union, device clinical investigations are governed by Chapter VI and Annex XV of the Medical Device Regulation (Regulation (EU) 2017/745, "MDR"), with the application/authorisation mechanics in Articles 62, 70, 74 and 82 and detailed operational expectations set out in MDCG guidance ²¹²⁶²⁷.

The two systems converge on the same goals (protect subjects, generate valid clinical evidence, capture safety signals promptly) but diverge sharply on who authorises a study, how risk is triaged, and how safety events flow to regulators. The sections below give a side-by-side view across the four stages the question asks about.

At a glance

Dimension	United States (FDA, 21 CFR 812/50/56)	European Union (MDR 2017/745, Chapter VI + Annex XV)
Legal basis for the study	IDE regulation 21 CFR Part 812; consent Part 50; IRB Part 56 1347916	MDR Articles 62 to 82; documentation per Annex XV; MDCG guidance 212627
Risk triage	Sponsor classifies device study as significant risk (SR) or nonsignificant risk (NSR); FDA is final arbiter 371518	Procedure keyed to the purpose of the study (Article 62/70 pre-market vs Article 74 post-market vs Article 82) rather than an SR/NSR label 2126
Who authorises	SR: FDA approval of the IDE plus IRB approval before start; NSR: IRB approval only, no FDA IDE 478	Member State competent authority (application/notification) plus national ethics committee review 252730
Clock to decision	30-day FDA review; deemed approved if no adverse notice in 30 days 25	Validation within 10 days; competent authority decision within 45 days of validation; coordinated assessment also on a 45-day period 22
Clinical evidence standard	Valid scientific evidence and totality of the evidence; benefit-risk of probable benefits vs probable risks 75767882	Clinical evaluation (plan and report) demonstrating conformity with the general safety and performance requirements in Annex I 9798111
Informed consent	Eight basic elements at 21 CFR 50.25(a) plus additional elements; IRB-approved signed form; narrow exceptions (50.23, 50.24) 72736064	Articles 63 to 68; consent process defined in the clinical investigation plan; specific rules for minors (Art 65), incapacitated subjects and emergencies (Art 68) 129126127
Safety reporting	Unanticipated adverse device effects (UADEs): investigator to sponsor/IRB within 10 working days; sponsor to FDA/IRBs/investigators within 10 working days 313334	Article 80: sponsor reports SAEs and device deficiencies "without delay" to all Member States where the study runs, via the Article 73 electronic system (Eudamed) 48

1. Getting the study authorised: IDE vs MDR Articles 62/70

United States: the SR/NSR fork drives everything

The defining feature of the US system is that the sponsor first classifies the study, and that classification determines whether FDA is even in the loop before enrollment.

• The sponsor makes the initial significant risk (SR) versus nonsignificant risk (NSR) determination and presents it to the IRB, but FDA is the final arbiter if it makes or is asked to make the call ³⁷¹⁵¹⁸. Once FDA has made a risk determination, that determination is final ³.
• An SR device study is broadly one where the device is an implant, is life-supporting/sustaining, is of substantial importance in diagnosing/treating disease, or otherwise presents a potential for serious risk to subject health, safety or welfare ³⁴⁶. An NSR study is simply one that does not meet the SR definition ³.
• SR studies require an FDA-approved IDE before they may proceed, and the study cannot start until both FDA and the IRB have approved it ⁴⁷.
• NSR studies do not need an FDA IDE. They may begin as soon as the IRB approves them, with the IRB acting as FDA's surrogate for review and continuing review ⁷⁸. NSR studies run under the abbreviated requirements at 21 CFR 812.2(b), covering labeling, IRB approval, informed consent, monitoring, records, reports, and the prohibition on promotion, and they carry no obligation to file progress or final reports with FDA ⁷.

For SR studies, the IDE mechanics are tightly time-boxed. The signed application goes to FDA's Document Control Center; FDA acknowledges receipt with a date and an IDE number, and the 30-day review clock runs from the date in that acknowledgement letter ². FDA notifies the sponsor of its decision within 30 days, and if it does not notify the sponsor that the study may not begin, the IDE is deemed approved (unless the device is banned) ²⁵. Approval with conditions lets enrollment begin on IRB approval provided the sponsor addresses FDA's issues within 45 days; on disapproval, enrollment cannot start until the sponsor resolves the deficiencies, with the option to respond by amendment or request a Part 16 regulatory hearing ²⁵. Notably, IRB approval may precede FDA approval ²⁵.

European Union: one procedure, three purposes, decentralised authorisation

The MDR does not use an SR/NSR label. Instead, the applicable route depends on the purpose of the investigation, and authorisation is granted at Member State level rather than by a single central agency.

• Investigations run to support conformity assessment and CE-marking follow Article 62(1), with the documentation set out in Chapter II of Annex XV, and are submitted as an application under Article 70 ²¹²⁶. The Investigator's Brochure is part of that package, and MDCG guidance offers an IB checklist to help meet the minimum requirements for validation under Article 70 ²⁶²⁹.
• Certain post-market investigations follow Article 74(1) as a notification, but the underlying Annex XV Chapter II documentation burden is the same as an application ²¹.
• Investigations not performed for any Article 62(1) purpose fall under Article 82 and must still comply with selected MDR provisions, but their submission requirements are governed by national law, so sponsors must check each competent authority's and ethics committee's rules ²¹.

Submissions go to the relevant Member State competent authority, and until EUDAMED is mandatory for these submissions sponsors file with the national competent authorities unless a Member State specifies otherwise ²⁵. Ethics review runs in parallel through national arrangements; the MDR deliberately does not prescribe the detailed ethics-review procedure, so sponsors must follow national rules and ensure the ethics committee and competent authority work from the same document versions ²⁷³⁰.

On timing, the MDCG Q&A sets a validation period of 10 days, requires the competent authority to notify the sponsor of its decision within 45 days of validation, and applies a 45-day period to the coordinated assessment procedure under Article 78 ²².

The practical contrast: FDA offers a single 30-day federal clock with a self-executing "deemed approved" default for the SR minority (and no federal filing at all for NSR studies), whereas the MDR routes every CE-marking study through a Member State competent authority plus a national ethics committee, on a 10-day validation and 45-day decision cadence, with the exact procedure keyed to the study's regulatory purpose rather than a risk label.

2. Clinical evaluation and when a study is even required

United States: valid scientific evidence and the totality of the evidence

FDA frames the clinical evidence question around valid scientific evidence and the totality of the evidence, sized to the device, its intended use, and its conditions of use ^75767882. Valid scientific evidence includes well-controlled investigations, partially controlled studies, studies without matched controls, well-documented case histories by qualified experts, and reports of significant human experience with a marketed device; isolated case reports, random experience, and unsubstantiated opinion do not qualify ^768182869075.

The determination weighs probable benefits to health against probable risks of injury or illness, interpreted in light of the target population and the device labeling, and accepts that uncertainty is part of the calculus ^7578828693. Clinical data may come from pre- and post-market investigations, comparable-device studies, published or unpublished clinical experience, registries, adverse-event databases, and medical records, and FDA may also accept peer-reviewed literature, outside-US data, and real-world evidence where appropriate ^79757787. Clinical evidence need not always come from randomized controlled trials, and FDA may reduce premarket data collection by relying on postmarket controls such as post-approval studies, postmarket surveillance and registries ^{93947780899196}. In short, whether a US device even needs a clinical investigation is a pathway- and device-specific judgement rather than a categorical rule.

European Union: clinical evaluation as a lifecycle obligation tied to the GSPRs

The MDR treats clinical evaluation as a systematic and planned process that runs across the whole device lifecycle to generate, collect, analyse and assess clinical data, with "clinical evidence" being data plus evaluation results of sufficient amount and quality to support the device's safety and clinical benefit ¹⁰⁵. The manufacturer must maintain a clinical evaluation plan and a clinical evaluation report and keep both updated ⁹⁸¹¹¹. The plan covers intended purpose, novelty, state of the art, patient risk, users, disease state and device class ¹¹⁰, and the report compiles the evidence, benefit-risk analysis, equivalence/state-of-the-art considerations, literature, clinical investigations, and PMS/PMCF outputs ^106111112.

Crucially, clinical evaluation exists to verify conformity with the relevant general safety and performance requirements (GSPRs) in Annex I, and the evidence must be sufficient to show the benefit-risk ratio is acceptable in light of the state of the art ^{9798111110118119}. Clinical investigations are a primary source of pre-market clinical data feeding this evaluation, and the process is iterative, with PMCF/PMS data feeding back and potentially triggering further investigations ^{1051069899113}.

Unlike the US case-by-case approach, the MDR sets a categorical trigger in Article 61: for implantable and class III devices, clinical investigations must be performed unless a listed exemption applies (for example a demonstrated equivalence to a device already on the market, subject to notified-body endorsement and full data access, or certain well-established legacy device types) ¹⁰⁴¹¹⁶. Where available clinical data are insufficient to demonstrate MDR compliance, the manufacturer must generate more through clinical investigation ⁹⁸.

3. Informed consent and subject protection

United States: prescriptive elements at 21 CFR 50.25

FDA requires informed consent for FDA-regulated investigations except in limited circumstances (21 CFR 50.23 and 50.24) ⁶⁴. Consent must contain the eight basic elements of 21 CFR 50.25(a), including a statement that the study involves research and its purpose, foreseeable risks, benefits, alternatives, confidentiality, injury compensation for greater-than-minimal-risk research, contacts, and the voluntary nature of participation with the right to withdraw ⁷²⁷³. Six additional elements at 50.25(b) apply when appropriate, such as a statement that the procedure may involve currently unforeseeable risks ⁷³.

Consent is normally documented by a written, IRB-approved form signed and dated by the subject or their legally authorized representative, with a copy given to the signer ⁶⁰. An IRB may waive documentation under 21 CFR 56.109(c)(1) for minimal-risk research that involves no procedures normally requiring written consent, and a short form is permitted where elements are presented orally ⁶⁰⁷². The investigator may not begin the consent process until the IRB has reviewed and approved the study, the consent form, and the associated materials ⁶⁸. The two exceptions are tightly conditioned: emergency use under 50.23 requires the investigator's documentation to reach the IRB within 5 working days of use, while emergency research under 50.24 permits a waiver only where the IRB (with a licensed physician's concurrence) finds a life-threatening situation, unproven or unsatisfactory alternatives, infeasibility of consent, prospect of direct benefit, and impracticability without the waiver, backed by community consultation, public disclosure and independent data monitoring ^596562616771.

European Union: consent embedded in the CIP, with detailed vulnerable-population rules

Under MDR Articles 63 to 68, the informed consent process is described in the clinical investigation plan (CIP), including how new information is conveyed to subjects and how the process handles subjects unable to consent ¹²⁹. For minors, the CIP must show how Article 65 is met, including consent from the legally designated representative and information adapted to the minor's age and maturity ¹²⁹. For emergencies, the CIP must justify that the Article 68 conditions for inclusion and first intervention without prior consent are fulfilled, and the sponsor must plan how a legally designated representative is identified so consent can be obtained without undue delay ¹²⁹.

Subject-protection provisions are explicit about vulnerable populations. The CIP must state whether children, pregnant or breastfeeding women, incapacitated, immunocompromised or elderly subjects are involved, and for incapacitated persons, minors, and pregnant/breastfeeding women the benefit-risk rationale must address whether there is an expected direct benefit ¹²⁶¹²⁷. The study must be designed to minimise pain, discomfort, fear and foreseeable risk, and emergency inclusion must be justified as minimal risk and minimal burden relative to standard treatment ¹²⁷. The investigation cannot begin until the required regulatory and ethical assessments conclude with non-negative outcomes, and the CIP must state compliance with the ethical principles of the Declaration of Helsinki ¹²⁹¹³⁶.

The substance of consent is similar across both regimes, but the drafting home differs: FDA codifies a fixed list of consent elements in the regulation itself, while the MDR pushes the consent process and vulnerable-population justifications into the sponsor's CIP under a framework of enumerated articles, with the operational detail (and the legally designated representative concept) frequently governed by national law.

4. Safety reporting during the study

United States: the UADE, on a 10-working-day clock

The core safety-reporting duty under 21 CFR 812 is the unanticipated adverse device effect (UADE), defined as any serious adverse effect on health or safety, or life-threatening problem or death, caused by or associated with the device and not previously identified in nature, severity or incidence, or any other unanticipated serious problem relating to subjects' rights, safety or welfare ³¹. The flow is two-tiered:

• The clinical investigator must report a UADE to the sponsor and the reviewing IRB as soon as possible, and no later than 10 working days after first learning of it ³¹³³.
• The sponsor must immediately evaluate the UADE and report the results to FDA, all reviewing IRBs, and all participating investigators within 10 working days of receiving notice, consistent with 21 CFR 812.150(b)(1) ³¹³⁴³⁷.

Beyond UADEs, sponsors and investigators owe other reports during the study: progress reports to the IRB (and, for approved-but-not-started studies, before IRB approval lapses), a final report or notice at completion, annual progress reports for minor changes under 812.150(b)(5), and a 5-working-day notice to FDA for certain device or protocol changes made without prior approval ^155154153.

European Union: SAEs and device deficiencies to every Member State, without delay

Under MDR Article 80(2), the sponsor must report, without delay, to all Member States in which the investigation is being conducted, via the electronic system referred to in Article 73 (Eudamed): any serious adverse event (SAE) with a causal relationship (or reasonably possible causal relationship) to the device, comparator or procedure; any device deficiency that might have led to an SAE absent intervention; and any new findings on those events ⁴⁸. An SAE is an event leading to death, life-threatening illness/injury, permanent impairment, hospitalisation or its prolongation, medical/surgical intervention to prevent such outcomes, chronic disease, or foetal distress/death or congenital abnormality; a device deficiency is any inadequacy in the identity, quality, reliability, safety or performance of the investigational device, including malfunctions and use errors ⁵³⁵⁶.

The sponsor is the responsible reporter and files to the national competent authorities of the countries where the study runs, reporting to only those Member States running an investigation under the same CIP code ⁴⁸⁵⁵. Article 80 sets the standard as "without delay," with the reporting period taking account of the severity of the event; where necessary the sponsor may file an initial incomplete report and follow up with a complete one, updating the record as new findings emerge ⁴⁸⁵²⁵³. Pending a fully functional Eudamed, MDCG guidance provides a tabular clinical investigation summary safety report form for the transition period, with reporting moving to the Eudamed web form once mandatory ^4851525354.

A note on precision: the retrievable MDCG material states the Article 80 standard as "without delay," proportionate to severity, rather than a single fixed day-count for clinical-investigation SAEs. The 15-calendar-day figure that appears in the guidance corpus attaches to Article 87 post-market serious incident reporting, which is a separate vigilance regime for CE-marked devices, not the in-study Article 80 obligation ¹⁵⁶. Sponsors should confirm the current SAE reporting timetable and form against the latest MDCG safety-reporting guidance and national requirements, since Member States may also require separate reporting to the ethics committee ¹⁴⁹.

Bottom line for cross-border programs

The clearest structural differences a sponsor must plan around are: (1) authorisation locus, one 30-day FDA clock (only for SR studies) versus a Member-State competent authority plus national ethics committee on a 10-day/45-day cadence for every CE-marking study; (2) the trigger for pre-market clinical data, FDA's device- and pathway-specific "valid scientific evidence" judgement versus the MDR's categorical Article 61 requirement for implantable and class III devices; and (3) safety-report routing, a single national channel to FDA and IRBs on a 10-working-day UADE clock versus parallel "without delay" reporting to every Member State running the study through Eudamed. Consent substance is broadly aligned, but the MDR's vulnerable-population and legally-designated-representative rules are more codified at the article level and are frequently completed by national law.