Can your device go paperless? An eIFU decision tree

Instructions for use must in principle be supplied with a device. The MDR excepts class I and class IIa devices that can be used safely without them. The exception is narrow, and other provisions of that Section may still require information to be supplied with the device.

This gate is first because it can remove part of a portfolio at no cost: where no instructions are required, there is nothing to publish.

“Device” means medical devices, accessories for medical devices, and the products listed in Annex XVI . Devices supplied under the Article 120 transitional provisions are covered too (recital 3 of 2025/1234).

In vitro diagnostics fall outside this regulation. The IVDR carries its own permission, on similar but not identical terms.

The term is defined: professional users are persons using the device in the course of their work in the framework of a professional healthcare activity. The question is therefore the capacity in which the device is handled, rather than the identity of the purchaser.

The provision operates on the instructions rather than on the device, so foreseeable lay use no longer removes the device from scope.

It only assists where the clinical document and the lay-facing document exist separately. Where a single instruction serves both audiences, there is nothing to separate, and the document is printed in full.

Gates 1 to 4 establish that electronic instructions are permitted. They do not establish that the electronic form is appropriate for the device in question — that is what the documented risk assessment is for, and it is where an eligibility review does most of its work.

It has to demonstrate that the electronic form maintains or improves the level of safety obtained on paper, and to cover at least the eleven elements listed at Article 4(1). Three of them tend to decide the answer for a given device.

• The environment the device is used in What has to be described is the setting at the point of use, not the conditions the document was written in. A shared terminal in an operating theatre, a sterile field where nobody can touch a screen, an ambulance outside network coverage and a patient’s living room give four different answers for the same instruction — and where one device is used in several of them, the assessment is only as strong as the least favourable of them. Art. 4(1)(b)
• Access to the resources needed at the time of use “At the time of use” is what makes this element bite. A document that is faultlessly available from an office desk can still be out of reach in the room where the device is opened, so what counts is the hardware the user can reasonably be expected to have to hand, the connection in that room, and anything standing in between — a login, an app, a file that will not open on the screen that is actually there. Art. 4(1)(d)
• Foreseeable medical emergencies This element is not a judgement on the electronic system; it identifies the situations in which nobody will be looking anything up. Where the assessment finds one, that information goes on the device itself or on a leaflet with it ( Article 5(4) ) — which is a carve-out from the electronic route rather than a disqualification from it, since the rest of the instructions can still be electronic. Art. 4(1)(g)

The assessment also fixes the period within which paper will be supplied on request ( Article 4(1)(i) ). That period is for the manufacturer to justify, and Article 5(3) caps it at 7 calendar days from the request — or delivery with the device, where the user asks at the time of order.

The final gate is operational rather than legal. The conditions are not difficult to read; the question is whether they can be held for as long as the retention periods run.

Article 7(2)(e) is the condition least likely to be satisfied by a document hosted on a general marketing website, because the address is fixed for the retention period rather than for the life of the current site.