From claim to confirmation: how to properly define clinical benefits for SaMD.

Defining clinical benefits for SaMD will inform the clinical evaluation process and enable your clinical performance data to validate the intended purpose

When developing a software medical device, the first thing you should do is define its intended purpose, which describes how the device works and provides information on:

• The users (e.g. doctors)
• The environment you want them to use it in (e.g primary care)
• The medical purpose for which you want it to be used (e.g. disease management)
• The associated clinical conditions (e.g. cardiovascular disease); and
• The people you want it to be used on (adults)

Armed with your intended purpose, there’s now an important next step: define one or more clinical benefits. This means identifying the positive impact that your device can have on the health of an individual, or a population of people.

Doing so will allow you to properly plan the clinical evaluation process, ensuring that relevant clinical performance data can be generated to confirm the validity of your clinical benefit claim. This process is key to meeting regulatory requirements and demonstrating your device's impact.

Making a claim

Almost all medical devices, software or otherwise, need to define at least one clinical benefit. In the early stages of device development, manufacturers will make a clinical benefit claim. Through a clinical evaluation process, they will look to provide the necessary data to demonstrate that, when the device is used according to its intended purpose, that claim can actually be realised.

At Scarlet, we see manufacturers making their lives harder by claiming clinical benefits that:

• don’t meet the definition of a clinical benefit;
• lack the necessary supporting data; or
• are not defined by an appropriate outcome measure

When claiming a clinical benefit for individual patients, it’s important to ensure the benefit is meaningful, measurable and includes a patient-relevant clinical outcome.

Getting the basics right

Let’s consider a device that provides personalised exercise and nutrition recommendations to patients with type 2 diabetes for the purpose of optimising the management of that condition. The device uses inputs like weight measurements, blood biomarkers, and daily questionnaire responses to generate the recommendations.

Here are some examples of low-quality clinical benefit claims for our example device:

1. “The device will improve the patient’s management of their diabetes.”
2. “The device will increase the patient’s awareness of their blood sugar control.”
3. “The device will provide a better patient experience.”
4. “The device will free up resources for patients with more complicated health needs.”

What’s wrong with these claims?

Number 1 is not measurable - while improving the management of diabetes is an admirable goal, by what measure are they going to determine whether diabetic management has been improved?

Number 2 is not meaningful - patient awareness of their blood sugar control might be important, but it’s not explicitly clear why that might be beneficial when considering the intended purpose (diabetes management).

Number 3 and 4 do not include a patient-relevant clinical outcome - of course, improving patient experience and freeing up healthcare resources is laudable, but it might be pointless if the actual condition for which the device is intended cannot be improved.

With all these factors in mind, a higher-quality clinical benefit claim might look something like:

• “Patients, on average, will achieve their target HbA1c level within 6 months”

This statement is meaningful (HbA1c is a biomarker for blood sugar control), measurable (HbA1c levels can be measured via a blood test, and specific timeframe is provided), and includes a patient-relevant clinical outcome (HbA1c is commonly used to monitor long-term blood sugar control).

Clinical data support

Clinical benefits need to be backed up by clinical data - clinical performance data to be exact. This blog further explores the different types of clinical data. In short, clinical performance data is generated from using the device under evaluation (or an equivalent device), and demonstrates how a clinical benefit is achieved when the device is used as intended.

Manufacturers need to consider what clinical performance data they can realistically generate in order to support any of their clinical benefit claims, particularly when a device is seeking its initial certification.

For our example above (“Patients, on average, will achieve their target HbA1c level within 6 months”), it is plausible that a manufacturer might be able to generate clinical performance data through an observational study, completed in the pre-market phase.

Post-market data

Manufacturers should continue collecting real-world clinical performance data once the device is on the market and use this to refine or expand their clinical benefit claims. There may be clinical performance data available to support more extensive clinical benefits, for example:

“Patients using the device have a reduced risk of heart attack or stroke.”

Again, this claim is measurable and clinically relevant. In fact, it is probably more meaningful than our HbA1c example as avoiding things like heart attacks and strokes is the ultimate aim of good diabetes management, rather than just optimising the value of a particular biomarker.

However, measuring grander benefits like reduced risk of heart attack or stroke requires longer time horizons than those typically available in the pre-market phase.

Start strong

Taking the time to define one or more high-quality clinical benefits up front is incredibly important. It will inform the planning of the clinical evaluation process, and it will ensure relevant clinical performance data can be generated to confirm that the clinical benefit(s) can be achieved.

Failing to properly define the benefits at the outset can result in the collection of data that is irrelevant, or insufficient to demonstrate the device's effectiveness. This will always make it harder to get it to market.

Clear, measurable clinical benefit claims can enable both regulatory approval and market success, so manufacturers should start early, think critically, and plan for the long term.