Clinical Evaluation 4 answers

Does software that provides glasses prescriptions perform a diagnosis under EU MDR?

Anonymous · Published February 25, 2026 · 2 comments
Our company has developed a device for eye exams that calculates refractive measurements (myopia, hyperopia, astigmatism) and provides values (+ or -) for glasses or contacts. Currently, the results are reviewed by an optometrist, who then issues a prescription, but we are considering automating this step.
Under the Medical Device Regulation (MDR), would providing these values be considered a diagnosis? Or is it simply a measurement, similar to how a thermometer provides temperature? How does this affect the software's classification, especially in light of Rule 11 and Article 2(1)?
We are interested in both the classification and the distinction between measurement and diagnosis in this context.

Join the discussion. Leave a comment. Guest comments are welcome — add your email to get reply notifications.

Anonymous 4 months ago
Is the main concern the device qualification or its risk classification?
Reply to this comment
Anonymous 4 months ago
Is the question specific to MDR, or are other regulations or regions also relevant?
Reply to this comment

Discussion

4 Answers

Accepted answer Dr. Oliver Eidel · Founder & CEO, OpenRegulatory ·
The distinction between diagnosis and measurement is important for MDR classification. If your software strictly measures refractive error and reports values (like a thermometer), and a healthcare professional interprets those results, then it's a measurement tool. But if the software goes further and generates a prescription or makes a recommendation for glasses based on those values, it's likely to be seen as providing information that is used for a therapeutic decision (even if it's compensatory rather than curative).
Rule 11 doesn't require the software to perform the diagnosis itself; it only needs to provide information intended to be used for diagnosis or therapy. Therefore, automating the prescription step would probably increase the risk classification, as the software is making or supporting a therapeutic decision. This is still a gray area, and interpretations can vary by country and authority.
In summary: if the software only measures and reports, it's closer to Class I; if it automates prescription or recommendation, it's likely to be at least Class IIa under MDR.

Join the discussion. Leave a comment. Guest comments are welcome — add your email to get reply notifications.

No comments yet. Be the first to share your thoughts.

A
Anonymous ·
The device itself doesn't perform a diagnosis—that's something only a healthcare professional can do. However, the device provides information that can be used as a basis for a diagnosis. If you automate the step that the optometrist currently handles, then your software might be taking on a diagnostic role.
Regarding MDR Rule 11: if your software is providing information used for diagnostic or therapeutic decisions, it may fall under a higher risk class. If it's just measuring and presenting data (like a thermometer), and a professional interprets it, the classification could be lower. But once you automate the decision-making, your software is more likely to be considered as supporting or making diagnostic or therapeutic decisions, which affects classification.

Join the discussion. Leave a comment. Guest comments are welcome — add your email to get reply notifications.

No comments yet. Be the first to share your thoughts.

A
Anonymous ·
There's an argument that your device is purely measuring refractive error and not diagnosing a disease. For example, myopia is listed as a disease in ICD-10 (H52.1), but in many countries, opticians (not medical professionals) can measure eyes and provide glasses without a medical diagnosis. If the device only outputs the refractive values and doesn't interpret them as a diagnosis, some would say it's just a measurement tool, similar to a thermometer.
However, MDR Rule 11 says that software intended to provide information used for diagnostic or therapeutic purposes is classified higher, regardless of whether it provides the diagnosis itself. If your device outputs values that are used to decide on glasses (a kind of therapy, even if it's just compensation), that may still bring it under a higher classification. Whether or not glasses count as a therapy is debatable, as they're compensatory rather than curative.

Join the discussion. Leave a comment. Guest comments are welcome — add your email to get reply notifications.

No comments yet. Be the first to share your thoughts.

A
Anonymous ·
Rule 11 of the MDR states that software intended to provide information used for diagnostic or therapeutic purposes is usually at least Class IIa. But for software used only by laypeople that does not drive diagnosis or therapy, some argue it could be classified as Class I. There are examples of Class I software being accepted by some competent authorities, especially for prevention or pure measurement purposes, but this depends on the country and the risk aversion of the notified body or authority.
If your software is just providing measurements and not being used by a professional to make a diagnosis or prescribe, you might be able to argue for Class I. But as soon as it replaces or automates the optometrist's decision, it will likely be classified higher.

Join the discussion. Leave a comment. Guest comments are welcome — add your email to get reply notifications.

No comments yet. Be the first to share your thoughts.

Want to add your answer to this question?
Write an answer under your name by logging in or signing up, or post anonymously.

Still have a question? Ask a question here publicly — for free.

Or talk to one of our consultants — first calls are free. Check out our services and prices.

Looking to automate your regulatory work? Check out our eQMS, Formwork. Built for lean, founder-led companies. There’s a free version too.