MDCG 2023-3: Vigilance terms and concepts, explained
What counts as an incident, a serious incident, and a trend — the vocabulary of MDR vigilance reporting.
MDCG 2023-3 explains the MDR and IVDR vocabulary that determines vigilance reporting: incidents, serious incidents, serious public health threats, expected undesirable side-effects, device deficiencies, use error, abnormal use, and trend reporting. The practical lesson is to assess the event and its possible outcome, including what might have happened without intervention, rather than reporting only events that caused actual harm.
Who this applies to
This Q&A applies to manufacturers, authorised representatives, importers, distributors, healthcare institutions, users, notified bodies, and authorities working with vigilance under the MDR and IVDR. Rev.2 covers both medical devices and IVDs even where examples use the language of one regulation.
Manufacturers should build these definitions into complaint intake and vigilance procedures. Front-line staff do not need to make the final regulatory decision, but they do need to capture enough facts and escalate quickly enough for the manufacturer to meet the reporting timeline.
Incident versus serious incident
An incident can be a malfunction or deterioration in device characteristics or performance, including use error caused by ergonomic features; inadequacy in information supplied by the manufacturer; or an undesirable side-effect. For IVDs, erroneous results and related performance failures require the same structured assessment.
A serious incident is an incident that directly or indirectly led, might have led, or might lead to one of the serious outcomes defined in the regulations: death; temporary or permanent serious deterioration in health; or a serious public health threat.
The words might have led are crucial. Actual harm is not required. If a clinician, alarm, backup system, or chance prevented the serious outcome, assess the outcome reasonably possible without that intervention. The assessment should consider the device, failure mode, clinical context, exposed population, detectability, and effectiveness of measures already in place.
Situation |
Vigilance question |
|---|---|
Device malfunction with no injury |
Could it have caused a serious outcome if the event recurred or intervention had not succeeded? |
Known side-effect |
Is it documented, predictable, clinically acceptable, correctly reflected in product information, and subject to trend reporting? |
Use error |
Did device design, interface, labelling, or instructions cause or contribute to the error? |
Abnormal use |
Was the act or omission outside reasonable risk control by the manufacturer, or was it reasonably foreseeable misuse? |
Problem found before use |
Is it a device deficiency that could have led to a serious incident had it not been detected? |
Do not close an event as “user error” before assessing ergonomics and foreseeable use. A confusing display, ambiguous unit, poor alarm, difficult connector, or inadequate instruction may make the event a device-related incident.
Serious deterioration and indirect harm
Serious deterioration is wider than permanent injury. It can include life-threatening illness or injury, permanent impairment, hospitalisation or prolonged hospitalisation, medical or surgical intervention to prevent life-threatening illness or permanent impairment, chronic disease, and foetal distress, death, or congenital impairment.
Indirect harm is especially important for diagnostic devices and software. A false result, missing result, delayed result, or incorrect information can lead to delayed treatment, unnecessary intervention, failure to treat, or exposure to a contraindicated therapy. Assess the clinical decision and possible patient outcome, not only whether the device physically contacted the patient.
A serious public health threat is an event that may result in imminent risk of death, serious deterioration, or serious illness requiring prompt remedial action and may cause significant morbidity or mortality or is unusual or unexpected for the place and time. This is a high threshold with an accelerated reporting route, not a synonym for every issue affecting multiple devices.
Expected undesirable side-effects
An undesirable side-effect may be known and documented and still be an incident. The regulations provide a reporting exception for certain expected undesirable side-effects that are clearly documented in product information, quantified in the technical documentation, and subject to trend reporting. Apply every condition before using the exception.
This means the risk file and clinical or performance evaluation need usable frequency and severity information. “Known risk” by itself is not enough. An event that is more severe, more frequent, clinically different, or outside the documented conditions needs a fresh reportability assessment.
Device deficiencies and events before market placement
A device deficiency is an inadequacy in identity, quality, durability, reliability, safety, performance, or information supplied. A problem may be caught before use and still matter for vigilance if use could have led to a serious incident. Record how it was detected, whether affected devices reached users, and the reasonably possible outcome.
The vigilance system concerns devices made available on the Union market, but information arising during testing, servicing, distribution, or stock checks may expose the same defect in marketed devices. Route it into complaint, risk, CAPA, and field-action processes rather than discarding it because the particular unit was never used.
Trend reporting is a separate decision
Not every non-serious incident is individually reportable, but statistically significant increases in the frequency or severity of incidents that could significantly affect benefit-risk and have led or may lead to unacceptable risks require trend reporting. Expected undesirable side-effects covered by the individual-reporting exception are also monitored through trends.
Predefine methods, observation periods, denominators, alert levels, and escalation rules in the PMS plan. A count without exposure data can mislead; a sales denominator may also be poor where use intensity varies. Use a denominator and stratification appropriate to how the device creates risk.
What this means for you, practically
- Capture the counterfactual. Ask what reasonably might have happened without clinical action, a backup, an alarm, early detection, or luck, and document the evidence for that outcome.
- Separate intake from reportability. Let staff record every complaint and deficiency consistently; give trained reviewers a controlled decision tree for incident, seriousness, timing, and trend assessment.
- Investigate “use error” as a design signal. Review interface, labelling, training, environment, foreseeable misuse, and recurrence before allocating responsibility.
- Control expected side-effects quantitatively. Link the IFU, risk file, clinical evidence, frequency limits, and trend rules so the reporting exception can actually be justified.
- Connect vigilance to action. Feed findings into CAPA, risk management, clinical or performance evaluation, PMS/PMCF or PMPF, labelling, SSCP, PSUR, and any field safety corrective action.
Related MDCG guidance
Related resources
Turn templates into working QMS documents.
Start from OpenRegulatory templates, fill them out with AI assistance, and keep them connected to your QMS in Formwork.