Evaluating Factors for FDA Approval of Diabetes Devices

Dennis P. Scanlon, PhD: Let’s get into this a little bit more by talking about payer coverage and patient choice in insulin pump technology. First of all, in thinking about FDA-approval, what are your thoughts on the criteria the FDA uses to evaluate diabetes devices compared (with) what it does with medications?

Mary Ann Hodorowicz, RDN, MBA, CDE: I did a little research on this when I saw the question (and the physicians here would know better than me), but on devices, they use computer simulated models so they don’t have to put an actual human being at risk. And what I read last night is that, especially for insulin pumps, they use these computer models to simulate hypoglycemia and hyperglycemia to determine if the device, itself, is going to pose additional risk or reduce the risk to the patient. So, those computer simulated models are what the FDA is requiring for a lot of medical devices.

Robert Gabbay, MD, PhD, FACP: I think it’s also a rapidly evolving area. Automated insulin delivery, which now is on the verge of really becoming prime-time, was approved, by most standards, as a smaller study (or not as large a study as you would imagine such a revolutionary change, so to speak). I think the FDA is open to this area to approve devices, perhaps, more rapidly than they had been in the past. And I think the flood gates are about to open, because there is a whole series of iterative changes that can be made to push forward with semi-automated insulin delivery.

Dennis P. Scanlon, PhD: Right. So that gets you into the situation where the device might be approved but we still have to ask the question, who’s appropriate for it? Right? What are your thoughts on that? If we talk about an insulin pump, how do we determine appropriateness there?

Kenneth Snow, MD, MBA: That’s one of the great challenges, because there’s not a lot of good data that identifies who makes a good pump user. We can extrapolate of who we think might be, but there’s not really a lot of data that is going to predict, for the person sitting in front of you, whether they are likely to succeed or fail. We have criteria that people tend to use, but they’ve not necessarily been vetted or proven that they’re the right ones.

Dennis P. Scanlon, PhD: Does everyone agree with that?

Robert Gabbay, MD, PhD, FACP: Well, that’s changing now that we’re moving toward semi-automated insulin delivery. Prior to that, people needed to be relatively knowledgeable. Being able to count the grams of carbohydrates they’re going to eat, being able to check their blood sugar 4-times-a-day and make all the appropriate adjustments—that’s a pretty high bar. But you could take patients that have really poor glycemic control, maybe not terribly adherent, and put them on a semi-automated insulin delivery that would at least improve their blood sugars overnight, and they would benefit. Now, it sort of changes our thinking on this, I think.

Mary Ann Hodorowicz, RDN, MBA, CDE: We have to look at Medicare and CMS (Centers for Medicare & Medicaid Services) decisions. They decided to pay for the external insulin pump for type 1 and type 2 patients with diabetes who are on multiple insulin injections. Of course, they have strict criteria for approving coverage. But the cost benefit has to be there for CMS to approve coverage for such an expensive item, and, obviously, we’re talking about mostly a Medicare age of 65 and over. So, that tells us, right there, that older people are capable of doing this with a fair amount of accuracy and to prevent hypoglycemia events.

Dennis P. Scanlon, PhD: Are there differences across the devices? Are there different features that are important in selecting pumps?

Mary Ann Hodorowicz, RDN, MBA, CDE: From a diabetes-educator perspective, I would say, yes. Making it waterproof. There are also devices that are disposable. I don’t want to name names, but the wireless infusion of insulin versus manual control—the pump is combined with a CGM (continuous glucose monitor), integrated with a CGM, that does the wireless infusion. Alarms, ease of calibrating it, and setting your grams of carbohydrate versus your insulin bolus. The ease-factor, I think, is critical. And waterproof. But making it easy because the device, itself, can be so difficult. So, easy-peasy, I think, is what manufacturers really have to focus on.

Dennis P. Scanlon, PhD: As clinicians, hearing all this, is this what you think about when you’re considering it for a patient?

Robert Gabbay, MD, PhD, FACP: Honestly, yes. But it makes it complicated and, therefore, it limits the number of people that are really going to have access to pumps. I think it is as useful as pumps are. And I think they are very much so. The bigger opportunity (and I think what we’ll see in the future and very near future) is much wider use of continuous glucose monitors, because they’re on the verge of becoming a lot cheaper and miniaturized. And really, they are a relatively inexpensive commodity that could be placed on lots of people with diabetes. Pumps, I think, will still remain relatively expensive and complex. They’ll get easier, but it’s going to be for a small subset of the population.

Kenneth Snow, MD, MBA: One of the things that comes up, also, is the difference between population management and individual patient management. So, as a physician, when I had a patient in front of me, it was really the assessment of the patient—issues such as, were they motivated? Were they willing to do the work? Were they able to do the work? Did I think that they would gain value from this? That was crucial.

However, from a population management point of view, that person may be achieving an A1C (glycated hemoglobin) that’s relatively good. And while they may move downward a little bit, in terms of complications to a population of folks, that change in that benefit is going to be relatively small if that’s the population of folks that is selected for.

Whereas, the folks who are doing really poorly and have an A1C of 10% may not be as engaged, and they may not be what a physician would consider the best patient for the device or for the technology. On the other hand, if they’re able to lower their A1C a point-and-a-half, that’s going to keep the population of folks with diabetes healthier, overall (even though they may not be the ideal candidate on a one-on-one approach).

Neal Kaufman, MD: The one researchable hypothesis is that by giving someone with type 2 diabetes (who’s poorly controlled) a CGM, they could identify their own patterns. “Look at this. When I eat this, I really go postprandially, to very high blood sugar. Maybe I should change my behavior.” So, the question becomes, can it become a behavior change incentive to allow you to become even better at managing your own condition?

Robert Gabbay, MD, PhD, FACP: And as a tool.

Neal Kaufman, MD: As a tool. That would be a tool. You could have intervention around that because once the person has this awareness of their glucose not going low…

Robert Gabbay, MD, PhD, FACP: Now you educate them to understand the behaviors that they could do better with.

Neal Kaufman, MD: And you get it paid for. All of us have to be involved.