From Clippy the Microsoft Word assistant to iPhone’s Siri, digital avatars have provided helpful advice in peoples’ lives for years. And the margin for error in these software-based personalities may be declining as their role shifts from providing Yelp lookups to medical services. A recent MIT Technology Review article explored the possibilities surrounding a new physical therapy program that uses a digital guide rather than a human therapist to coach patients through exercises.
A platform developed by California startup Sense.ly is currently being tested in a pilot program in the San Mateo Medical Center, the publication reported. The platform uses an avatar named Molly, who interviews patients about their pain levels while guiding them through exercises and tracking their movements with 3-D cameras and other medical devices. The avatar can ask simple questions and answer basic queries, giving diet recommendations to those with diabetes, for instance.
The system then sends reports to physicians with graphs and analytics tracking patient performance, transcripts of voice interactions and red flag notifications if a problem needs to be addressed right away. Such technology can help improve care at hospitals with overextended staff by allowing patients to perform rehabilitation exercises from home.
“A physician’s time is always limited,” Benjamin Kanter, chief medical information officer at Palomar Health in San Diego, told MIT Technology Review. “For a long time, we’ve had the challenge of just getting information into the system. Now the system is starting to actually help me.”
Getting digital care right
Efforts to apply new technology advances to healthcare have expanded rapidly in recent years with developments in telemedicine programs and robotic care. Robots could soon fill the role of in-home aides by performing household tasks and administering prescriptions, a recent New Yorks Bits blog article noted. And a therapeutic robot called Paro is being used to calm dementia and Alzheimer’s patients.
Researchers at Georgia Tech have explored the ways a robot’s appearance can affect what tasks people will trust it to perform, the publication noted. People want more anthropomorphic robots for tasks that require intelligence, such as medicine recommendations, while they prefer more mechanical appearances for robots that perform manual labor.
Appearance is also a factor with building trust in Molly, the Sense.ly avatar. She can modulate her voice and facial expressions. According to company cofounder Ivana Schnur, a clinical psychologist who has spent years developing virtual reality tools in medicine and mental health, patients are often more willing to share honestly with a digital avatar than a human doctor. The company is working to develop more advanced features, such as the ability to respond to peoples’ facial expressions.
For now, though, there is a trust battle to be won in demonstrating to doctors and patients that such systems are error-free. Kanter told MIT Technology Review that while electronic systems can be helpful in reducing errors, they also make it easier for mistakes to propagate more rapidly than with paper systems. Ensuring that a system such as Sense.ly’s won’t misinterpret patient interactions is critical.
As such technologies become increasingly central to healthcare practices, tools like static analysis software will be essential for helping to catch errors, minimize mistakes and build trust. With glitch-free code, digital avatars may someday become as trusted as many other medical innovations.
Software news brought to you by Klocwork Inc., dedicated to helping software developers create better code with every keystroke.