An ideal medical device

Imagining an ideal medical device as a benchmark for evaluating benefit-risk in a more structured and quantitative framework.

Benefits of a medical device must outweigh the risks from its use, but by how much?

A core tenet of medical practice is to first do no harm, primum non nocere. The essence of this principle is to avoid an intervention if it would cause more harm to the patient than otherwise.

If we consider the use of a medical device to be an intervention to diagnose, cure, mitigate, treat or prevent a disease¹ , than it must not cause any more harm than what a patient would experience if the device were not used at all.

In other words, benefits of the intended use of a medical device must outweigh the risk(s) of harm to patients and users. This is the essence of the regulatory requirement of safety and effectiveness², which a medical device must meet with valid scientific data before it is allowed to enter the market.

But by how much should the benefit(s) of a medical device outweigh its risk(s)? What level of benefit(s) in comparison to the risk(s) would be considered sufficient to provide a reasonable level of assurance of safety and effectiveness³?

There are no clear answers to these questions. There are no benchmarks or standardized quantitative models for determining benefit-risk of medical devices.

Although FDA has issued several guidance documents, the process by which FDA evaluates benefit-risk remains largely opaque. Here is a recent LTR case study, which attempted to reverse engineer the benefit-risk determination for an AI/ML device using FDA’s guidance and publicly available information. However, this exercise is not without considerable uncertainty and it is not easily replicated.

There continues to be a need to treat medical device benefit-risk determinations in a more structured, and preferably quantitative, way to reduce the amount of uncertainty and judgment involved in this exercise.

What if we could imagine an ideal medical device?

An ideal medical device, similar to the concept of an ideal gas⁴ or an absolute temperature⁵, can be helpful in developing a structured framework for analyzing the benefit-risk of real medical devices.

A key challenge, however, is to figure out a common numerical scale that could facilitate a direct comparison of benefits and risks. Unlike the risk-benefit analysis often used to make financial decisions, it is not feasible to assign a monetary value to health benefits or harms resulting from the use of a medical device. This has been a main reason for the use of a qualitative and checklist-based approach prevalent in the medical device industry.

It may be useful to consider some measure of a health status⁶ for an individual, or a population, to quantify the effect of a medical device. If an intervention using a medical device results in a net positive change in the health status, then we can conclude that benefits of its use outweigh the risks. Further, the magnitude of the net gain can be used to compare a device against a benchmark or other similar devices.