EN 62366-1:2015+A1:2020 distinguishes three categories of user action that deviate from intended operation: use error, abnormal use, and misuse. The distinction is not academic. It determines which obligations the manufacturer carries under MDR Annex I §5 and §22, which controls are required, and which findings a notified body will raise. Most so-called user training problems are actually engineering failures wearing a different label.

By Tibor Zechmeister and Felix Lenhard.

TL;DR

  • EN 62366-1:2015+A1:2020 clause 3 defines use error, abnormal use, and the broader category of misuse with precision. Manufacturers must treat each correctly because each carries different risk control obligations.
  • Use error is a user action or lack of action that leads to a different result than intended by the manufacturer, where the user is acting within reasonably foreseeable use. The manufacturer is responsible for reducing these as far as possible through design.
  • Abnormal use is an act or omission characterised by intentional disregard for correct use, beyond any reasonable means of risk control through manufacturer design. It is outside the manufacturer's direct control obligation but still gets documented.
  • Tibor's Q7 insight is that "user training problem" is too simple a concept. Most use errors are engineering or user interface failures, and calling them training problems is how manufacturers dodge design obligations they legally carry.
  • Under MDR Annex I §5 and §22, reducing use error through safe design is a legal duty. Training is the last resort in the control hierarchy, not the first.

Why this matters

Tibor has audited enough risk files to recognise the pattern immediately. A startup submits a risk management file where every use-related hazard has the same control strategy: "user training". The rationale column says things like "user will be trained on correct use before first use" or "the IFU explains the correct procedure". Under the MDR, this is not an adequate risk control strategy for anything that could have been designed out or prevented with a protective measure.

The left-handed display story makes it concrete. If a team designs a handheld device that looks upside down to right-handed users, the resulting misread is not a training problem. No amount of training stops a tired user from reading a display that is oriented the wrong way for their grip. It is a use error, and under EN 62366-1:2015+A1:2020 the manufacturer is required to reduce its probability as far as possible through design. The right fix is an orientation sensor or a reshape of the grip, not a warning label in the IFU saying "hold device with left hand".

Felix coaches startups who cannot afford to make this mistake. The reason is not only regulatory. A device certified on the basis of "user will be trained" collapses in the field. Real users read the IFU once if at all. Real users at 3 a.m. in a dim bedroom do not remember the training. Real users are tired, impaired, distracted, or new. Reducing use errors through engineering is what makes a device actually safe, not just formally compliant. The formal compliance and the real safety happen to line up because the regulation was written by people who knew what they were asking for.

What MDR actually says

The MDR sets the obligation. EN 62366-1 supplies the vocabulary. The two have to be read together.

MDR Annex I Chapter I, General Safety and Performance Requirements. Risks arising from use must be reduced as far as possible.

MDR Annex I §5. Devices shall be designed and manufactured so that risks linked to ergonomic features of the device and the environment in which the device is intended to be used are reduced as far as possible. The text lists the user factors that must be considered: technical knowledge, experience, education, training, use environment, and medical and physical condition. Training is one of six factors, not the dominant one.

MDR Annex I §22. Devices for lay users shall be designed and manufactured to cope with variation in the lay user's technique and environment. The article presumes that lay users will deviate from ideal operation. The manufacturer's job is to design for the deviation, not to lecture it out of existence.

EN 62366-1:2015+A1:2020, clause 3 definitions. The standard defines the vocabulary manufacturers must use.

  • Use error is a user action or lack of user action while using the medical device that leads to a different result than that intended by the manufacturer or expected by the user. The definition covers slips, lapses, and mistakes that a foreseeable user could make while acting within normal use.
  • Abnormal use is a conscious, intentional act or intentional omission by the user that is counter to or violates normal use and that is also beyond any reasonable means of user-interface-related risk control by the manufacturer. The key phrases are "intentional" and "beyond any reasonable means of control by the manufacturer".
  • Misuse as a broader informal category in the literature can cover both use errors and abnormal use. EN 62366-1 itself frames manufacturer obligations primarily around use error, with abnormal use documented separately because it is not a design target.

The practical consequence: use error must be reduced as far as possible. Abnormal use is noted, analysed for residual risk, and usually closed with a warning or a rationale statement. If a manufacturer labels something as abnormal use in order to avoid designing a control, a notified body will push back and require the label to change to use error, which reopens the design obligation.

Plain-language translation: if a reasonable user could do it by accident, it is a use error, and the team is on the hook to engineer it out. If a user has to consciously and deliberately fight the device to do it, and there is no reasonable design control, it is abnormal use, and a warning or a rationale closes it out. Labelling matters.

A worked example

A Class IIb home-use injection device for a chronic condition, intended for self-administration by adult patients. The team runs use-related risk analysis and identifies three candidate deviations.

Deviation 1. A tired patient at night pushes the injector against the wrong body location because they misread a small pictogram on the device housing.

This is a use error. The user is acting within foreseeable use, the mistake is a slip, and the manufacturer has multiple design options to reduce it: larger pictogram, higher-contrast pictogram, haptic feedback when the device is against the correct anatomical region, a guard that only releases when a correct-orientation sensor is satisfied. Under MDR Annex I §5 and §22, "as far as possible" requires the team to take the most effective feasible control. Calling this a training problem would not survive notified body review.

Deviation 2. A patient reuses a single-use disposable injector by trying to refill it.

This is closer to abnormal use. The device is clearly single-use, marked with the single-use symbol under ISO 15223-1, and the refill action requires deliberate intent to bypass the device's design. However, the manufacturer still has design obligations: can the housing be made tamper-evident, can the internals be destroyed after first use, can a mechanical lockout prevent a second firing? The team investigates. An internal lockout is feasible and cheap. The team includes it. After the lockout, any remaining refill attempts require destruction of the device, and the residual risk is documented as abnormal use with a rationale.

Deviation 3. A patient deliberately injects the device into another person's arm.

This is abnormal use. No reasonable manufacturer design control prevents a determined user from performing a physical action on another person. The team documents it in the risk file with a rationale, a warning in the IFU, and a reference to the EN 62366-1 definition of abnormal use. A notified body auditor is unlikely to push back on this classification because the criteria are met: intentional, deliberate, counter to normal use, and beyond reasonable design control.

Three deviations, three classifications, three different control obligations. The labels are not free. Each label determines what work the team has to do and what evidence the risk file has to contain.

The Subtract to Ship playbook

Felix's Subtract to Ship approach treats the use error versus abnormal use question as a forcing function for honest engineering.

  1. Default every deviation to use error first. Tibor's rule: a deviation is a use error until the team can prove, with reference to the EN 62366-1:2015+A1:2020 clause 3 definition, that it is abnormal use. This reverses the burden. Teams that default to abnormal use end up dodging design obligations. Teams that default to use error end up engineering safer devices.

  2. For every use error, walk the control hierarchy in order. Inherent safety by design is the first option. Protective measures (locks, sensors, automatic shutdowns) are the second option. Information for safety (labels, IFU, training) is the last resort. Under MDR Annex I §5, the hierarchy is not a suggestion. A control strategy that skips straight to information when design changes were feasible is a notified body finding waiting to happen.

  3. Retire the phrase "user training" as a control strategy. Training is a factor the MDR requires the team to consider, not a control strategy on its own. If "user training" is the only control for a use-related hazard, the team has not actually done risk control. They have written a note.

  4. Document the reasoning when classifying a deviation as abnormal use. The EN 62366-1 definition has four criteria: conscious, intentional, counter to normal use, and beyond reasonable means of control. The risk file entry must show that all four are met, and must show that the team investigated possible design controls and found none feasible. Without this documentation, a notified body will reclassify the deviation as use error and reopen the design work.

  5. Re-check the classification when new information arrives. Post-market feedback showing that a supposed abnormal use is actually happening to reasonable users means it was never abnormal use. It was a missed use error. The classification flips and the design obligation reappears.

  6. Teach the team the definitions once, properly. Tibor's experience is that the confusion between use error, abnormal use, and misuse is a training issue for the team itself (ironic, given the topic). One clear session on EN 62366-1:2015+A1:2020 clause 3 prevents hundreds of misclassifications later.

  7. Audit the risk file for the phrase "user will be trained". Every occurrence is a candidate for escalation to a design control. In Tibor's experience, fewer than ten percent of such entries survive a serious review unchanged.

Reality Check

  1. Does the team know the EN 62366-1:2015+A1:2020 clause 3 definitions of use error and abnormal use well enough to explain them without looking?
  2. How many hazards in the risk management file have "user training" or "IFU warning" as their sole control strategy?
  3. For each deviation currently classified as abnormal use, can the team point to documented evidence that all four EN 62366-1 criteria are met?
  4. For each use-related hazard, has the team walked the control hierarchy in the order inherent safety, protective measures, information, and documented why each option was accepted or rejected?
  5. Has any use error been re-opened after post-market feedback showed a classification was wrong?
  6. When a design review meeting discusses a use-related deviation, is the first question "can we design this out" or "can we warn about it"?
  7. Does the team default to use error for ambiguous cases, or to abnormal use?
  8. Could the risk file withstand a notified body audit that specifically targeted the use error versus abnormal use classification?

Frequently Asked Questions

What is the difference between use error and user error? EN 62366-1:2015+A1:2020 deliberately uses "use error" rather than "user error". The shift is intentional. "User error" implies the user is at fault. "Use error" frames the same event as a property of the interaction between the user and the device, which is what the manufacturer is obliged to engineer. The vocabulary choice is a regulatory stance.

Can a use error ever be controlled by training alone? Rarely and only when inherent safety and protective measures have both been investigated and found infeasible with documented rationale. Training alone as a first-choice control strategy for a use error is an MDR Annex I §5 violation waiting to be flagged.

What is the relationship between misuse and abnormal use? Misuse is the broader informal category that literature sometimes uses to cover both use errors and intentional deviations. EN 62366-1:2015+A1:2020 prefers the two precise terms, use error and abnormal use, and does not rely on "misuse" as a standalone regulatory category. The clean classification is use error and abnormal use. Misuse is a word to use carefully or not at all in the risk file.

Who decides if something is abnormal use or use error? The manufacturer classifies first, documents the reasoning, and the notified body audits. Ultimately, if there is disagreement, the notified body has authority during conformity assessment. In Tibor's experience, auditors reclassify abnormal use to use error far more often than the other way around.

Does EN ISO 14971:2019+A11:2021 handle abnormal use differently from use error? Both feed the risk management process, but the control obligations differ. Use error drives design action under the hierarchy of controls. Abnormal use is documented and closed with a rationale or warning where design control is not feasible. Both must appear in the risk management file with traceability back to the usability engineering file.

Sources

  1. Regulation (EU) 2017/745 on medical devices, consolidated text. Annex I Chapter I, Annex I §5, Annex I §22.
  2. EN 62366-1:2015+A1:2020, Medical devices – Application of usability engineering to medical devices. Clause 3 (definitions), Clauses 5.3, 5.5, 5.7.
  3. EN ISO 14971:2019+A11:2021, Medical devices – Application of risk management to medical devices. Clauses 5, 6, 7, 8.