Important question. I guess the attached slide deck may in part provide an answer. It is freely available from the web.

Best

Thank you for this interesting slides on FDA position.I noticed the following statement which requires further explanation: "FDA approval of an Investigational Device Exemption (IDE) is required for US human study of a significant risk device which is not approved for the indication being studied." How do you interpret this positioning?

My interpretation is if a significant risk device is approved for one indication and another indication is pursued with the same device an IDE could be requested to the FDA.

Do you think the same goes for a new medical device with significant risks if a predicate already exists on the market (in the United States or elsewhere)?

About this I'm not really sure.

It may works if the predicate is on the market in US but surely not if it is on the market elsewhere.

An SR determination (and the risk to human subjects) isn't based on the device per se, so a change in indication could certainly change a study from SR to NSR. For example, an extended wear contact lens (ie, a lens worn overnight) would be considered as SR, whereas if used for daily wear the very same device would be NSR. Or a device used in a coronary artery considered as SR might be considered as NSR when indicated for use in a peripheral vein.

As noted by the FDA, "The risk determination should be based on the proposed use of a device in an investigation, and not on the device alone.

Studies where the potential harm to subjects could be life threatening, could result in permanent impairment of a body function or permanent damage to body structure, or could necessitate medical or surgical intervention to preclude permanent impairment of a body function or permanent damage to body structure should be considered SR." [https://www.fda.gov/regulatory-information/search-fda-guidance-documents/significant-risk-and-nonsignificant-risk-medical-device-studies]

With regards to a predicate device - implying a device would be cleared by a premarket notification (ie, FD&CA section 510k) - such comparison must be based, in part, on the same Indications for Use, hence more than likely an SR/NSR determination would follow suit.

In answer to the original titled question, clinical evaluation might be based in part on the risk, but also on the scientific question or clinical objectives that are posed. A NSR device might very well have a very comprehensive clinical evaluation (eg, statistical analysis), yet not present a significant high risk profile.

@Jack Douglas, Thank you for that answer. It is very interesting to take into account the indication which can transform the level of risk of the same medical device according to its use. Could the risk also vary according to the technological environment of the place where the new device would be used, or according to the operator's expertise or qualification? and how to take it into account ?

Risk management should take into account all sources of potential hazard to the health and safety of a patient, including user expertise (eg, education), training (eg, company device training) and environment (eg, is the clinical lab adequately equipped?). Normally this is required to qualify a user prior to participating in a clinical study, or as a condition for device purchase after market approval. Of course this hasn't been ideal as many adverse events occur due to user error. It should be noted here that another obvious source of risk is the state of disease itself. Devices and training might be optimal, yet the patient's anatomy and disease can present treatment challenges.

To fully clarify the question " Do you think it is necessary to distinguish different clinical study methods for the authorization of new medical devices to be marketed according to their level of risk? ", note that most medical devices do not require clinical study data, ie, most can be safely marketed using other means of evaluation.