Jewels Doskicz is a registered nurse, freelance writer, patient advocate, health coach, and long-distance cyclist. Jewels is the moderator of Diabetic Connect’s weekly #DCDE Twitter chat, and she and her daughter both live healthfully with type 1 diabetes.
Diabetes alert dogs are capable of sensing blood sugar changes prior to our glucometers; now a mathematical model has caught up with the canines.
Research out of Penn State University has created a mathematical model that does in fact predict changes in blood glucose levels up to thirty minutes in advance—and with stunning 90 percent accuracy.
Meg Flynn of Power Paws in Scottsdale, Arizona, trains diabetes alert dogs. She shared with me that Power Paws puppies are trained to be 80 percent accurate and can predict blood sugar changes fifteen to twenty minutes before they happen.
I personally don't have a diabetes alert dog, but I do have a Dexcom continuous glucose monitor (CGM), and they have predictive value as well. Think about this scenario: If you check a blood sugar on a glucometer you may get the number 98 and think you're ready to tuck yourself into bed, right? In contrast, if you see the same result on a CGM with slanted or straight arrows downward—you'll probably be heading to the kitchen instead.
Here's a curve ball from Professor Peter Molenaar in Penn State News: "Glucose levels under the skin trail blood glucose levels from anywhere between eight and 15 minutes." Perhaps this is why we don't believe our glucometers on occasion—it's reading 100 but you feel like you're 60.
With either example, predictive cues give a person with diabetes time to react and possibly avert the low blood sugar from happening. So how does this research impact our current situation?
Here's where this data is headed: straight into an algorithm for an artificial pancreas, serving to tighten up safety measures so that one day we can rely more fully on technological advances to help us live between the lines without constant human intervention.