The IVDR in Plain Language

An in vitro diagnostic medical device examines specimens taken from the human body. It may detect a pathogen, measure a biomarker, match a medicine to a tumor characteristic, determine blood-group compatibility, or help estimate risk. The device can be a reagent, instrument, or kit. It can be a software product, specimen receptacle, calibrator, or control material.

In the European Union, the central legal framework is Regulation (EU) 2017/746 on in vitro diagnostic medical devices, known as the IVDR. It became applicable on 26 May 2022, replacing the older directive-based system for most purposes.

The IVDR is not simply a larger premarket filing. It creates a lifecycle system. A manufacturer must define what the device is for, classify its risk, generate evidence for that purpose, use the correct conformity-assessment route, maintain a quality system, monitor performance after market entry, report certain events, and update the evidence as knowledge changes.

Your exact obligations depend on the product, intended purpose, and economic-operator role. They depend on market history, contracts, and current European Union and national requirements.

Start with intended purpose#

The intended purpose states what the device detects or measures, the function it performs, the specimen, the target population, the intended user, the setting, and how the information will be used. It appears in the label, instructions, promotional material, performance evaluation, and technical documentation.

This is not decorative wording. Intended purpose determines whether a product qualifies as an IVD and strongly shapes its risk class, evidence needs, performance claims, usability work, and post-market plan.

A narrow intended purpose can sometimes be supported by focused evidence. A broad claim across specimen types, populations, settings, or clinical roles requires evidence that covers that breadth. Marketing cannot safely outrun the validated purpose.

Software needs the same disciplined definition. A program may be an IVD if, on its own or in combination, it provides information from examination of human specimens for a medical purpose listed in the regulation. Storage, transfer, or simple search functions may fall outside that definition. The actual function and claimed purpose matter more than calling the product a platform, calculator, laboratory tool, or decision aid.

The four risk classes#

Annex VIII creates four classes, from A to D. Classification follows specific rules, not a free ranking by the manufacturer.

Class A covers lower-risk devices. General laboratory products intended by the manufacturer specifically for in vitro diagnostic procedures, instruments intended specifically for IVD procedures, and specimen receptacles may fall here. Sterility changes the assessment route.

Class B is the default class for devices that are not captured by another rule. Many devices with individual-level implications but less severe public-health or patient consequences are class B.

Class C covers many tests where an incorrect result could create a serious risk for an individual, including specified infectious-disease testing, companion diagnostics, genetic testing, certain screening, and other uses defined by the rules.

Class D covers the highest public-health and patient risks, including specified testing for transmissible agents in blood or tissue intended for transfusion, transplantation, or cell administration, and certain agents associated with life-threatening disease and high propagation risk.

MDCG 2020-16, revised in March 2025 as revision 4, explains how the rules apply. Guidance helps interpretation, but the regulation remains the law. Your classification memo should connect every element of intended purpose to the relevant rule and document why competing rules do or do not apply.

When more than one rule applies, the highest resulting class governs. New scientific information can also affect an old conclusion. Classification should therefore be controlled and revisited when intended purpose, technology, epidemiology, or guidance changes.

What a notified body does#

A notified body is an organization designated by a European Union member state to conduct specified conformity-assessment work. The scope of designation matters: a body must be designated for the relevant device and codes.

Class A nonsterile devices can generally be placed on the market through manufacturer self-declaration after the manufacturer fulfills applicable requirements. For class A sterile devices, notified-body involvement is limited to aspects that establish, secure, and maintain sterility.

Class B, C, and D devices generally require notified-body assessment. The review can cover the quality-management system, technical documentation, performance evaluation, post-market processes, and device-specific sampling or assessment rules. Successful assessment leads to the relevant certificate, after which the manufacturer draws up the European Union declaration of conformity and affixes the CE marking as applicable.

Notified-body review does not transfer the manufacturer's responsibility. The manufacturer remains accountable for compliance, performance, and vigilance. It remains accountable for updates, suppliers, and the device placed on the market.

Class D devices can have added controls. European Union reference laboratories may verify claimed performance and batch testing where the legal conditions apply. Expert panels can provide views in specified circumstances. These mechanisms reflect the consequences of a missed or incorrect result.

For a companion diagnostic, the notified body must consult a competent medicinal-products authority or, where applicable, the European Medicines Agency about suitability in relation to the medicine concerned. The device and medicine remain governed through linked but distinct frameworks.

Performance evaluation is a chain of evidence#

Article 56 and Annex XIII organize performance evaluation around three connected elements.

Scientific validity asks whether the analyte or marker has a valid association with the clinical condition or physiological state. A technically precise assay is not useful for a claimed purpose if the measured signal lacks a defensible clinical relationship.

Analytical performance asks whether the device measures or detects the analyte as intended. Relevant characteristics can include analytical sensitivity and specificity, accuracy, and precision. They can include repeatability, reproducibility, and measuring range. They can include linearity and limits of detection and quantitation. They can include interference, cross-reactivity, specimen stability, and metrological traceability. Not every characteristic applies in the same way to every qualitative or quantitative device.

Clinical performance asks whether the device produces results correlated with the clinical condition or process for the target population and intended purpose. Measures may include diagnostic sensitivity, diagnostic specificity, and predictive values. They may include likelihood ratios, expected values in affected and unaffected groups, or other justified clinical metrics.

The three are summarized in a performance evaluation report. Literature, existing data, analytical studies, clinical performance studies, external-quality information, and post-market evidence may contribute, but each source must fit the claim you are making and be appraised for quality and relevance.

Evidence sufficiency is proportional to device characteristics and risk. “State of the art” does not mean only the newest technology. It includes the accepted body of scientific, technical, clinical, and regulatory knowledge against which safety and performance are judged.

Clinical performance studies#

Some claims require a prospective clinical performance study. Your protocol should specify objectives, endpoints, and participant and specimen selection. It should specify reference methods, sample-size logic, and statistical analysis. It should specify missing data, discordant results, and sites. It should specify operators and handling of deviations.

Representativeness matters. A test intended for decentralized use should not rely only on ideal specimens handled by expert central-laboratory staff. A test intended across age groups, variants, stages, or comorbidities needs data that justify those populations or clearly limited labeling.

Reference standards need critical appraisal. If the benchmark misses disease or is applied only to selected participants, estimates can be biased. Blinding, order, and timing belong in the report. So do specimen flow, exclusions, indeterminate results, and all analysis sets. None of that data transfers automatically between technologies, because a biomarker's scientific validity does not validate every assay that measures it, and one assay's thresholds cannot simply be assigned to another platform.

Quality management and technical documentation#

The manufacturer's quality-management system covers regulatory strategy, management responsibility, and resource management. It covers suppliers, risk management, and performance evaluation. It covers product realization, unique device identification, and post-market surveillance. It covers vigilance, corrective action, and improvement.

Technical documentation should let a reviewer reconstruct the device and the case for conformity. It includes device description and specification, labeling, design and manufacturing information, general safety and performance requirements, benefit-risk and risk management, product verification and validation, software evidence where relevant, performance evaluation, and post-market planning.

These records must agree. If your instructions describe one specimen type but your analytical study used another, or the risk file assumes professional laboratory use while promotional material addresses home users, the problem is not solved by placing each statement in a different document.

Change control is critical. A software update, reagent reformulation, new specimen type, threshold change, added population, manufacturing-site change, or revised claim can affect validation, the notified-body relationship, certification, labeling, registration, and eligibility for a transitional provision.

Post-market surveillance continues the evaluation#

After market entry, manufacturers systematically collect and review experience with the device. Post-market surveillance can include complaints, vigilance data, and trend reports. It can include literature, user feedback, and external-quality results. It can include returned-product analysis, registry information, and post-market performance follow-up.

Post-market performance follow-up, or PMPF, is a continuous process that updates the performance evaluation. It can confirm performance in broader practice, evaluate emerging risks, identify systematic misuse, examine new populations or variants, and test whether the benefit-risk conclusion still holds. If PMPF is not appropriate, the justification must be documented rather than assumed.

Class C and D devices generally require a periodic safety update report, updated at least annually, with the applicable availability to the notified body and authorities. Class A and B devices use a post-market surveillance report, updated when necessary. Specific duties differ, so the legal text and current guidance should control implementation.

Serious incidents and field safety corrective actions enter the vigilance system. Trend reporting can apply to statistically significant increases in frequency or severity of non-serious incidents or expected undesirable side effects that could affect benefit-risk. Corrective and preventive action should connect signals to root cause, containment, and effectiveness checks. It should connect them to risk files, performance evidence, and communication.

What the transitional dates actually mean#

European Union legislators extended certain transition periods because certification capacity and market readiness created supply concerns. Regulation (EU) 2024/1860 provides the current staged dates for eligible legacy IVDs:

• Class D devices may transition until 31 December 2027.
• Class C devices may transition until 31 December 2028.
• Class B devices and class A devices placed on the market in sterile condition may transition until 31 December 2029.

These dates do not grant a blanket delay. Eligibility depends on cumulative conditions. The device must continue complying with the prior directive, must not undergo a significant change in design or intended purpose, and must not present an unacceptable risk to health or safety. The manufacturer had to put an IVDR-compliant quality-management system in place by 26 May 2025.

The manufacturer also needed to lodge a formal application with a notified body by the class-specific deadline: 26 May 2025 for class D, 26 May 2026 for class C, and 26 May 2027 for class B and class A sterile devices. A written agreement follows within the legal interval.

New devices do not inherit legacy status. Nor did legacy class A nonsterile devices receive the same staged dates; they generally needed IVDR compliance from 26 May 2022. A certificate or transition pathway also does not excuse ongoing prior-law compliance, surveillance, vigilance, registration, or IVDR provisions that already apply. A transition date is the end of a conditional path, not the beginning of a plan.

EUDAMED and the 2026 change#

EUDAMED is the European database for medical devices. It supports actor registration, device information, and certificates. It supports clinical investigations and performance studies, vigilance, and market surveillance through linked modules.

Regulation (EU) 2024/1860 allowed individual modules to become mandatory after audit and a Commission notice, without waiting for the entire database. From 28 May 2026, use of the first four modules became mandatory:

• Actor registration
• Unique device identification and device registration
• Notified bodies and certificates
• Market surveillance

The clinical investigations and performance studies module and the vigilance module follow their own readiness and transition path. Teams should use current Commission implementation material because responsibilities can depend on module status, data migration, actor role, and whether a device is new or already on the market. Readiness is not merely data entry either. Names, addresses, and basic UDI-DI structure all have to reconcile with your controlled records. So do nomenclature, intended purpose, and risk class. So do certificate data, economic-operator relationships, and market status.

Supply interruption and discontinuation#

Article 10a has applied since 10 January 2025. When a manufacturer anticipates interruption or discontinuation of supply of certain devices and it is reasonably foreseeable that this could cause serious harm or a risk of serious harm to patients or public health, it must provide notice.

The manufacturer informs the competent authority where it or its authorized representative is established and the economic operators, health institutions, or healthcare professionals it directly supplies. Notice is generally required at least six months before the anticipated interruption or discontinuation, except in exceptional circumstances. Recipients in the supply chain have duties to pass the information onward without undue delay. Behind that notice sits a structured assessment of clinical dependency and alternatives, because inventory inconvenience alone is not the legal threshold, while a device's role in transfusion, transplant, urgent diagnosis, treatment selection, or a scarce testing pathway can make an interruption consequential.

The plain-language conclusion#

The IVDR connects a claim to a risk class, a risk class to an assessment route, and the assessment route to evidence and ongoing controls. Its core question is not whether a document exists. It is whether a device, used for its stated purpose, has a coherent and current case for safety and performance.

That case begins with precise intended purpose. It continues through classification, scientific validity, analytical and clinical performance, quality management, technical documentation, conformity assessment, registration, surveillance, and transparent response to change. Transition dates, database milestones, and guidance revisions matter, but they sit inside that lifecycle rather than replacing it, so a durable IVDR program treats each as controlled regulatory intelligence, maps it to named owners and records, and checks that the device on the market still matches the one that was evaluated.

References#

• European Parliament and Council. Regulation (EU) 2017/746 on in vitro diagnostic medical devices.
• European Parliament and Council. Regulation (EU) 2024/1860, including gradual EUDAMED rollout and supply provisions.
• European Commission Medical Device Coordination Group. MDCG 2020-16 revision 4: Guidance on classification rules for in vitro diagnostic medical devices. March 2025.
• European Commission. Overview of EUDAMED.
• European Commission. Information for IVD manufacturers.
• European Commission. Medical Devices Coordination Group guidance and endorsed documents.