When a medical device performs more than one function, more than one Annex VIII classification rule applies to it, and under the implementing rules in Annex VIII Section 2 of Regulation (EU) 2017/745, the strictest rule and sub-rule producing the highest classification governs the device as a whole. This is the "highest class wins" principle. It is not a tie-breaker to use when convenient — it is a mandatory implementing rule that applies every time multiple functions, multiple rules, or multiple sub-rules point to different classes. MDCG 2021-24 (October 2021) is the authoritative guidance for applying it correctly, and it is the rule a Notified Body will check first when a multi-function device lands on their desk.

By Tibor Zechmeister and Felix Lenhard. Last updated 10 April 2026.

TL;DR

  • A multi-function device is any medical device whose intended purpose covers more than one distinct medical function — for example, a device that both monitors a physiological parameter and delivers a therapeutic intervention.
  • Under Annex VIII Section 2 of Regulation (EU) 2017/745, the strictest rule and sub-rule producing the highest classification applies when multiple rules are triggered by the same device.
  • "Highest class wins" applies both across different rules (Rule 9 vs Rule 10) and within sub-rules of a single rule (Rule 11 IIa vs Rule 11 IIb).
  • The practical question is never "which rule applies" but "which rules apply" — the plural is mandatory, and stopping at the first rule that seems to fit is the single most common classification error.
  • When functions are separable into independent devices with independent intended purposes, labels, and conformity assessments, each device is classified on its own. When they are inseparable inside one product placed on the market as one device, the highest class wins.
  • MDCG 2021-24 (October 2021) and the Borderline Manual v4 (September 2025) are the interpretation layer for edge cases.

Why multi-function devices are where classification mistakes multiply

Classification goes wrong most often on devices that do more than one thing. A pure wound dressing is easy. A pure ECG monitor is easy. A wound dressing with an embedded sensor that streams wound-bed data to a clinician dashboard is not easy, because it is simultaneously a Rule 4 device (non-invasive in contact with injured skin) and a Rule 10 or Rule 11 device (active diagnostic monitoring or software providing information for clinical decisions). Two rules. Two candidate classes. One device going to market.

Founders default to one of two wrong instincts. Either they pick the rule they like best and document that one, quietly ignoring the other. Or they pick the rule that looks most obvious to an outsider and never check whether a second rule catches the device at a higher class. Both instincts fail at the Notified Body gate. The correct move is written into Annex VIII: identify every rule that applies, determine the class each one produces, and take the highest.

This post walks through how the implementing rule actually works, when multiple functions can be separated into independent devices, when they cannot, and how to document the reasoning so a Notified Body auditor can follow it. It is the spoke on multi-rule application in the classification cluster, and it assumes you have already read the hub post on MDR device classification and the rule-by-rule walk in MDR Annex VIII classification rules.

The implementing rule for multiple functions — what Annex VIII actually says

Classification is governed by Article 51 of Regulation (EU) 2017/745, which refers the reader to Annex VIII for the rules themselves. Annex VIII has three sections: Section 1 contains definitions, Section 2 contains implementing rules, and Section 3 contains the twenty-two classification rules.

The implementing rule for multi-function devices lives in Section 2. In substance, it reads: if several rules, or if, within the same rule, several sub-rules, apply to the same device based on the device's intended purpose, the strictest rule and sub-rule resulting in the higher classification shall apply (Regulation (EU) 2017/745, Annex VIII, Section 2).

Three things are worth noting about how this is written. First, it is framed around intended purpose, not around what the device physically is. If a device has a secondary function that is not part of the intended purpose — for example, a technical capability that exists in the firmware but is never claimed, labelled, or used — then the rules that would be triggered by that function do not apply. Intended purpose is the filter. Second, the word "applies" is plural by design. Annex VIII assumes multiple rules will frequently apply to the same device. Third, the escalation is mandatory. The rule does not say "may apply" or "should be considered." It says "shall apply." A Notified Body will treat this as a hard obligation.

The highest-class-wins principle — how it actually works in practice

Highest class wins is straightforward in the abstract and easy to get wrong in the concrete. The correct procedure:

  1. Write down the intended purpose in one paragraph, grounded in Article 2(12) of the MDR and consistent with the label, IFU, promotional materials, and planned clinical evaluation.
  2. Walk through every rule family in Annex VIII Section 3 — non-invasive (Rules 1-4), invasive with respect to body orifices (Rule 5), surgically invasive (Rules 6-8), active (Rules 9-13), and special (Rules 14-22).
  3. For every rule and every sub-rule that could apply to any function in the intended purpose, write down the class it produces and the reasoning.
  4. Take the highest class from the list. That is the class of the device.
  5. Document the classification with the rule that produced the highest class cited first, and the other applicable rules listed alongside with their respective classes.

The devices that trip founders up are the ones where the highest-class rule is not the most obvious rule. A wound-care product with a sensor might feel like a Rule 4 wound dressing, but Rule 11 for the integrated software may produce Class IIa while Rule 4 produces Class I. Rule 11 wins. A non-invasive device that administers a substance by inhalation might feel like a Rule 2 channelling device, but Rule 20 escalates inhalation delivery devices to Class IIa or IIb depending on the mode of action and the condition treated. Rule 20 wins. The discipline is not to stop at the first rule that fits.

MDCG 2021-24 reinforces this explicitly in its section on the implementing rules and returns to the principle repeatedly in the worked examples. When you are not sure whether a second rule applies, MDCG 2021-24 is the next document to read.

When functions are separable

Not every multi-function product is a multi-function device. Sometimes what looks like one product is actually two devices placed on the market together, each with its own intended purpose, its own label, its own technical documentation, and its own conformity assessment. The classification question for a separable-functions product is asked twice, once per device, not once for the bundle.

The test for separability is practical, not cosmetic. The functions are separable when:

  • Each function has its own distinct intended purpose that can be written independently and stand on its own regulatory feet.
  • Each function can be placed on the market, used, and discontinued independently.
  • The user can operate one function without depending on the other for its medical claim to hold.
  • The technical documentation for each function can be assembled without relying on the other.

A pulse oximeter packaged in the same box as a separate patient-diary application that does not process the oximetry data is two devices, not one. Each is classified on its own intended purpose. The oximeter sits in Rule 10. The diary software sits where Rule 11 lands it based on its own intended purpose.

When functions are separable and you classify them independently, the conformity assessment for each must reference the correct class, and the labelling must make clear which CE marking covers what. A shared commercial name does not make two devices into one device if their intended purposes are independent — and it does not merge two intended purposes into one if the underlying use is distinct.

When functions are not separable

When the functions cannot be pulled apart into independent devices — when the user's ability to benefit from function A depends on function B being there, or when the intended purpose of the product as a whole is the combination — the device is one device with multiple functions, and the highest-class-wins principle applies.

A cardiac monitor that also delivers closed-loop therapy based on what it monitors is one device, because the therapeutic function depends on the monitoring function and the product is marketed for the combined use. Rule 22 would apply for the integrated therapeutic-diagnostic function, and Rule 10 would apply for the monitoring function. Rule 22 puts the device at Class III. Rule 10 alone would put it lower. Rule 22 wins.

A dressing-plus-sensor product where the sensor's readings drive decisions about whether to change the dressing, and the dressing is sold and used only in combination with the sensor, is one device. Rule 4 applies for the dressing function, and Rule 11 applies for the decision-support software in the sensor. The software class is typically higher, and it governs the whole product.

A combination surgical instrument that cuts and coagulates tissue in the same action is one device. Rule 6 applies for the surgically invasive function, and Rule 9 applies for the active-therapeutic energy delivery function. Whichever produces the higher class governs. For most electrosurgical tools this is Rule 9 with a IIb result.

The separability test matters because the wrong answer in either direction is expensive. Treat two separable devices as one and you pay for a higher class than the lower-risk device actually needs. Treat one inseparable device as two and the Notified Body will catch it and require reclassification of the combined product.

Worked examples

Example 1 — Software-driven infusion pump. Intended purpose: deliver a defined dose of medicinal product intravenously under programmed control. Applicable rules: Rule 12 (active device for administering medicinal products) and Rule 11 (software providing information for therapeutic decisions, through the dosing algorithm). Rule 12 produces Class IIb because the mode of administration is potentially hazardous. Rule 11 produces Class IIb or Class III depending on the consequences of a wrong decision. The software, because it drives the hardware, falls in the same class as the hardware under the Section 2 software-driving-hardware implementing rule. Highest class wins. The device is classified at the higher of the two, with the reasoning documented against both rules.

Example 2 — Diagnostic imaging platform with AI triage. Intended purpose: acquire diagnostic images and provide AI-based prioritisation of findings for radiologist review. Applicable rules: Rule 10 (active device for diagnosis) for the imaging hardware, and Rule 11 for the AI triage software providing information for diagnostic decisions. Rule 10 typically produces Class IIa or IIb depending on the imaging modality. Rule 11 typically produces Class IIa, escalating to IIb if the triage decisions could cause serious deterioration of health. The highest of the two applicable results governs the combined product, and the rule citation in the technical documentation names both.

Example 3 — Wound dressing with embedded biosensor. Intended purpose: maintain wound environment and provide clinicians with continuous readings of wound-bed parameters to inform dressing-change decisions. Applicable rules: Rule 4 (non-invasive device in contact with injured skin) for the dressing and Rule 11 for the software component producing clinical decision information. Rule 4 typically produces Class IIa for devices managing the wound microenvironment. Rule 11 typically produces Class IIa for decision-support information. Both rules produce IIa in this example, so the outcome is IIa — but the reasoning has to name both rules, because if the clinical claim escalates, Rule 11 can escalate first and the class can move before the physical dressing changes.

Example 4 — Hardware-software combination with disputed measurement function. We have worked on exactly this kind of product — a combined hardware-software device where the software side was clearly Class IIa under Rule 11, no dispute, and the hardware side was where the argument lived. The question on the hardware was whether it carried a measurement function that would trigger the Im sub-class. The rules allow judgment here: what the device actually measured, how the measurement result was used in the intended purpose, whether the quantitative output met the Annex VIII definition of measurement function. A rigorous walk through the evidence defended Class I without measurement function for the hardware. Combined with the Class IIa software, the device was classified at Class IIa as a single combined product — because the software drove the hardware and the Section 2 implementing rule merged them into one classification. The Notified Body accepted the reasoning. Highest class wins, and the highest class was IIa.

The Subtract to Ship angle on multi-function classification

The Subtract to Ship framework for MDR compliance treats classification as the second of four passes. On multi-function devices, the Classification Pass has a specific discipline: separate the functions whenever it is legitimate to separate them, and accept the combined class whenever it is not.

Subtracting here means two things. First, do not bundle functions into a single device just because they share a product name. If the regulatory-correct move is to ship two devices with two intended purposes, the lower-risk device gets a lower-risk conformity assessment path and the runway lasts longer. Second, do not argue for a lower class on the combined device when the highest-class-wins rule makes the combined class inevitable. A class you cannot defend at the Notified Body is not a saving — it is a delay waiting to happen.

The test is always the same: every activity in the project, every page in the technical file, every clinical evidence decision traces back to the class the rules produce. For multi-function devices, the class the rules produce is the highest class among all applicable rules. No less. No more.

Reality Check — Where do you stand on multi-function classification?

  1. Have you written the intended purpose for your device as a single paragraph that lists every medical function the product actually performs?
  2. Have you walked through every Annex VIII rule family and identified every rule that could apply to any of those functions, not just the most obvious one?
  3. For each rule you identified, have you written down the class it produces and the reasoning paragraph?
  4. Have you applied the Annex VIII Section 2 implementing rule — strictest rule and sub-rule resulting in the higher classification shall apply — explicitly in your classification document?
  5. Have you asked whether your multi-function product is actually separable into two or more independent devices with independent intended purposes, and can you defend your answer either way?
  6. If the functions are inseparable, have you documented why — with reference to the shared intended purpose, the dependency between functions, and the way the product is placed on the market?
  7. Have you cross-checked your reasoning against MDCG 2021-24 (October 2021), especially its worked examples for multi-rule application?
  8. If a Notified Body auditor asked why this multi-function device is the class you say it is, could you name every rule that applies, cite the highest-class-wins implementing rule, and explain your separability decision in one paragraph?

Frequently Asked Questions

What is a multi-function device under the MDR? A multi-function device is a medical device whose intended purpose covers more than one distinct medical function — for example, a device that both monitors a physiological parameter and delivers a therapeutic intervention, or a combination of a wound dressing and an embedded diagnostic sensor. The term is not formally defined in the MDR, but Annex VIII Section 2 explicitly anticipates the situation by stating that when multiple rules apply to the same device, the strictest rule producing the higher classification governs.

What does "highest class wins" mean in Annex VIII? It means that when more than one classification rule in Annex VIII Section 3, or more than one sub-rule within a single rule, is triggered by the intended purpose of a device, the rule or sub-rule producing the highest class is the one that governs the device's classification. It is an implementing rule in Section 2 of Annex VIII, not a discretionary tie-breaker, and it is mandatory under Regulation (EU) 2017/745.

Can I separate a multi-function product into two devices to avoid the highest class? Only when the separation reflects the regulatory reality. Functions are separable when each has its own distinct intended purpose, can be placed on the market and used independently, and has a clinical claim that stands on its own. A shared commercial name does not merge two genuinely independent devices, and a cosmetic separation does not split one inseparable device. The test is the intended purpose and the way the device is placed on the market, not the packaging.

Which MDCG guidance covers multi-rule application? MDCG 2021-24 (Guidance on classification of medical devices, October 2021) is the authoritative guidance on Annex VIII and includes the interpretation of the implementing rules, the highest-class-wins principle, and worked examples of multi-rule application. For software specifically, MDCG 2019-11 Rev.1 (June 2025) is the companion guidance. Both should be consulted for multi-function products that include software.

What happens if I classify a multi-function device under only one rule? If the Notified Body identifies a second applicable rule during conformity assessment and that rule produces a higher class, they will require reclassification. The downstream consequence is that any technical documentation, QMS scope, and clinical evidence based on the lower class has to be re-done to match the higher class. Catching the multi-rule issue before the Notified Body application is much cheaper than catching it during the audit.

Does the software-driving-hardware implementing rule replace highest class wins? No — it works alongside it. The software-driving-hardware rule in Annex VIII Section 2 says software that drives or influences a hardware device falls in the same class as the hardware. That rule determines which class the software inherits. Highest class wins then determines which class the combined product sits in when multiple rules apply across the hardware and any standalone software components. Both implementing rules are part of the same section and both apply.

Sources

  1. Regulation (EU) 2017/745 of the European Parliament and of the Council of 5 April 2017 on medical devices, Article 51 (classification), Annex VIII Section 2 (implementing rules, including highest-class-wins and software-driving-hardware), Annex VIII Section 3 (classification rules 1-22). Official Journal L 117, 5.5.2017.
  2. MDCG 2021-24 — Guidance on classification of medical devices, October 2021.

This post is part of the Device Classification & Conformity Assessment series in the Subtract to Ship: MDR blog. Authored by Felix Lenhard and Tibor Zechmeister. Multi-function devices are where classification errors multiply, and they are also where rigorous application of the Annex VIII Section 2 implementing rules produces the biggest savings in downstream conformity assessment work. When the combination of functions in a specific device exceeds what a blog post can resolve, a classification working session with a sparring partner who has run this exercise across fifty-plus certifications is usually the fastest path to a defensible answer.