Amazon’s Alexa is a lot of fun. However, it may be more that just a toy or household helper. What if voice activation could help you with diabetes control or call 911 if you fall and can’t reach your pendant?
From Sandy Johanson who found this in MedCity News: “There were five finalists presenting at the Alexa Diabetes Challenge enlisting Amazon’s voice-activated, machine learning-enabled robot in New York City this week at Merck’s event. But one application that resonated strongly with attendees involves the potential to improve the mood of diabetes patients, many of whom struggle with depression that comes with coping with being diagnosed with a chronic condition associated with potentially life-threatening complications.
DiaBetty was the brainchild of the University of Illinois at Chicago associate professors Dr. Olusola Ajilore, with a focus on psychiatry, and Dr. Alex Leow, who teaches psychiatry and bioengineering. They both work in the Collaborative Neuroimaging Environment for Connectomics at the university. Perhaps coincidentally, the project shares a name with a character on The Simpsons — Cletus Spuckler’s overweight diabetic cousin Dia-Betty.
The goal is to help patients better manage their condition with addressing their depression a key part of its approach. Social support, mindfulness, and care coordination are among the components of the project’s program.
In his presentation, Ajilore observed that emotions and stress can affect blood sugar levels. Depression accompanying diabetes can make the condition 2-3 times more costly to manage, Ajilore said. He offered up the potential for Alexa to use biomarkers in a diabetes patient’s voice to do mood analysis. That assessment could trigger Alexa to offer different ways to lift the user’s mood when they need it from playing a song that resonates with them to playing back a recording from a grandchild.
When asked by a panel of judges what the company’s vision is, Ajilore answered this way:
“We are helping to integrate data from a smartphone to assess mood but we also want to take advantage of what voice detection can do.”
Although there are those who are excited by Alexa’s potential, it’s unclear how long it would take Alexa to become a mainstream patient engagement tool for a few different reasons. The device is also still learning to understand accents and languages which could have an impact on how voice assessment tools figure out someone’s mood. Despite the potential to make Alexa part of a patient engagement approach, Alexa is not yet HIPAA-compliant although that’s rumored to be in the pipeline.
The complications from diabetes are another pain point that healthcare innovators are keen to address. Team Wellpepper developed Sugarpod, a device that combines Alexa with a WiFi connected scale and snaps pictures of feet to flag diabetic foot ulcers caused by diabetic neuropathy, a condition whereby people lose sensation in their feet. The danger of ulcers is that if they’re not treated in time they can lead to costly foot amputations. The goal of the device is to reduce the cost of treatment from these kinds of complications.
Mike Van Snellenberg of Wellpepper said the idea is that the team’s technology could be used to do earlier intervention to detect hot spots on feet before they become ulcers and even identify early signs of diabetic neuropathy. MIT spinout Podimetrics developed a similar device for early foot ulcer and hotspot detection for people with diabetic neuropathy, sans Alexa.
One potential downside of Sugarpod flagged by judges is helping users understand when the camera is on or off.
A factor that makes Alexa of such interest to health tech developers is also what has the potential to makes it a bit creepy. HCL America and Ayogo developed appointment management tool Glucoach. It is also intended to get a sense of the user’s confidence or “perceived self-efficacy”, as Ayogo CEO and founder Michael Fergusson phrased it, in areas such as diet, exercise and medication management.
“We can embed ourselves into their lives,” Fergusson said. “We can know the precise content they need in the way they need it.”
Other approaches to the diabetes management conundrum involved a remote monitoring technology licensed from NASA called Personal Intelligent agents from Team Ejenta. A Columbia University team from its biomedical informatics lab developed T2D2, a virtual nutrition assistant to support meal planning and personalized recommendations based on the user’s food and glucose log.
Events centered on identifying myriad ways for how consumer tech devices can be applied to healthcare pain points provide a useful exercise. Not only do participants identify different ways to address chronic disease care for patient engagement, but these events flag the blind spots that developers run into early on and offers a way for them to succeed in the long-term.”