Today at CES 2016 in Las Vegas, IBM Watson Health announced a new partnership with Under Armour, as well as sharing new information about its partnership with Medtronic, announced last April.
The Medtronic and IBM teams have been working on using Watson’s cognitive computing alongside data from Medtronic’s Carelink Connect system to predict hypo- and hyperglycemic events and give people with diabetes coaching and guidance to avoid those events. In a research study, the companies say they were able to predict a hypoglycemic event three hours in advance 80 percent of the time.
“That’s a lot different than just tracking realtime glucose data on the sensor that they can do today,” IBM VP of Partnerships Kathy McGroddy told MobiHealthNews. “If you only know a few minutes in advance, there’s not much time to do anything about it. Insulin takes about 20 minutes to really have an effect, so being able to alert a patient and have the opportunity to really give them insight as to what’s going on and what patterns have they seen before and so forth can be extremely powerful.”
Regulatory delays could slow the timetable, but Medtronic hopes to have an initial launch this summer, and to continue to iterate on the system after that, McGroddy said.
Even when patients have a continuous glucose monitor and collect a wealth of information about their lifestyle and symptoms, they often only see a doctor every three months, and then for 10 minutes at a time. So there’s not really time to go through that data and mine it for insights. This is where McGroddy believes IBM and Medtronic can make a difference.
“If you think of what’s important for a diabetic patient, it’s things like activity, exercise, going for a walk after dinner, as well as the dietary aspect,” she said. “Starting with the data from their sugar and their insulin levels and then being able to incorporate other datasets and understand the personal responses is extremely powerful.”
In the future the companies see potential to crunch data from additional sources — everything from activity trackers and nutrition apps, to accessing the phone’s calendar to get a sense of stressful events that might be coming up (stress can also be a factor that impacts blood sugar).
“It’s a very powerful thought that the system can learn and get to know you better and get to know your patterns and then again understand what are the true factors that are important, which of those are controllable, and what can you do about it. It’s like this personal advisor to some extent,” she said.
The Under Armour partnership is similar in some ways. It will apply Watson computing to all the data collected in Under Armour’s UA Record app, and come up with personalized coaching and training recommendations. It also could use cognitive computing to recognize foods and do automated food logging.
McGroddy explained the broad strokes of the system with a personal example.
“Recently I had a concussion and I couldn’t do anything for a few weeks,” she said. “So imagine I’m coming back, I’m ready to start sprinting again and I want to run the NYC marathon in March — what’s the right training regimen for someone of my age, my height, my weight, and my fitness level, knowing I haven’t run in a few months? How should I get started?”
The system could also take weather and time of day into account to give context-aware workout suggestions. McGroddy said to expect those to be incorporated into the UA Record app at some future date.