Go to main content
Type 1
Type 2

AID Technology Improves Diabetes Care No Matter How You Get It

3 Minute Read
Doctor and patient visit by video appointment on a tablet.

Key takeaways:

  • Obtaining automated insulin delivery during a virtual doctor's visit is just as effective as in-person training with an endocrinologist or primary care physician, a new study finds.
  • Researchers tested out the iLet Bionic Pancreas and saw average glucose levels drop and time in range increase for all study participants, regardless of what type of provider they saw and whether they were seen remotely or in person. 
  • Researchers highlighted the need for standardized education and treatment plans for primary care physicians on using diabetes technology starting at diagnosis.

Finding a great diabetes specialist can be hard. And unless you live in a major city, finding one who understands the latest diabetes technology can be even more challenging. 

But a new study suggests people with diabetes benefit from using an automated insulin delivery (AID) system, regardless of whether their care was managed by an endocrinologist, primary care physician, in-person or remote. 

About the study

The small 40-person feasibility study conducted by the University of Colorado School of Medicine was designed to test how successful a future larger study on AID systems would be. 

Adults with type 1 diabetes used the iLet Bionic Pancreas AID system with the Dexcom G6 for two weeks, and then their usual method of care for another two weeks. For half the group consulting with a family physician, their usual care meant multiple daily injections with or without the use of a continuous glucose monitor (CGM). The other half were seeing an endocrinologist and used a Dexcom G6 CGM with a sensor-augmented pump, but not as part of an AID system, as their usual care. 

The study also split the number of in-person and telehealth visits. The iLet Bionic Pancreas, while not FDA approved at the time the study was conducted, was chosen because it requires very little input – just the person’s weight – to start. The device was mailed to those receiving training and ongoing care via telehealth. 

What did the study find?

All study participants on the iLet Bionic Pancreas had average glucose levels between 145-156 mg/dL, saw a decrease of 17 mg/dL, and an 11% increase in time in range

Notably, the improved results were nearly identical  across all groups using the Bionic Pancreas.

“That trial used only academic endocrinology sites that are really experienced at doing this and have, in theory, the most expertise in doing this, and we had neophytes who hadn't touched this technology before, who did the same,” said Dr. Sean Oser, family medicine professor at the University of Colorado. 

“There was no difference between endocrinology and primary care.”

In addition, study results indicated no significant difference in outcomes between participants who had in-person appointments and those who had telehealth visits. 

Why this research is important

The study found that being introduced to the iLet Bionic Pancreas by a primary care provider during a virtual visit is just as effective as in-person training with an endocrinologist, which suggests that access to AID systems could assist people with diabetes who are limited geographically or can’t access subspecialty care such as an endocrinology team.

“The study showed that non-expert doctors could put a person on the technology and they didn't even have to be in person. The patient would do just as well as with the experts,” said Dr. Eden Miller, an endocrinologist in Bend, Oregon whose expertise includes diabetes technology in primary care. She urges primary care physicians to be more proactive when dealing with diabetes treatment and technology. 

Since 90% of people with type 2 diabetes and 50% of people with type 1 diabetes are cared for by primary care physicians, the study results point to the need for more training on CGM and AID systems, researchers said. 

Learn more about AID systems for diabetes management here: