Using a Smartphone to Predict the Onset of Depression
News
In early October 2016, rapper Kid Cudi made the headlines after writing on Facebook that he entered rehab for treatment for depression and suicidal thoughts. While he’s not as famous as other musicians, his heartfelt message garnered a lot attention and support from celebrities and mental health professionals.
“Yesterday I checked myself into rehab for depression and suicidal urges. I am not at peace. I haven't been since you've known me. If I didn't come here, I would’ve done something to myself,” he wrote.
Kid Cudi’s story highlights what mental health professionals and physicians have long known; it is hard to detect depression, and it is even harder to provide treatment to people experiencing it. His message indicated that he struggled with it a long time before he finally realized he needed treatment.
Kid Cudi is just one of the estimated 16 million people in America who are suffering from depression. While the rapper recognized it and sought treatment, many never get to that point.
Leveraging the Power of our Smartphones
While the problem of identifying and treating depression stymies experts, Jason Hong and John Zimmerman wondered if the solution to detecting it was in our pockets.
“Depression is a very common kind of issue; it is the leading mental health issue in the world and a leading cause of sick days and disabilities,” says Hong, associate professor and member of the Human Computer Interaction Institute. “Our smartphones are with us almost all the time. Can we use them to predict the onset or exit of depression?”
Hong talked with physicians and mental health professionals about depression—what signs and symptoms best predict whether someone is experiencing it—and used what he found to design an app. Experts agreed that three categories of symptoms, social, physical, and sleep, could accurately determine whether someone is experiencing depression. All Hong needed to do was harness the power of the smartphone to record data about each of these symptoms.
He started with sleep.
“How much you sleep in terms of quality and quantity is a good predictor of depression,” Hong says.
Looking at the phone, he wondered what he could use on it to understand sleep habits. He decided the accelerometer and microphone could help. The accelerometer measured motion and whether the lights were on while the microphone recorded how loud the surroundings were. It there was too much noise, movement, or light, it indicated a person was tossing and turning rather than sleeping, for example. That information could be sent to doctors so they could see a picture of sleep habits.
“[Doctors] just need to see the trends over time … as evidence,” he says.
Next, they tackled the social aspect.
He wondered whether GPS could provide an accurate estimate of when people left their homes. If it is less than normal that could be a sign of potential problem. And, checking phone and SMS logs could show how often people connect with social supports.
“Using those factors we can create a predictor pretty well,” he says.
While the phone provides a good estimate of how social someone is, it does present an incomplete picture. While GPS might track when people leave the house, it can’t tell if people are being social when they are out.
“Face-to-face, it turns out, that is really hard to do,” he says.
Finally, they examined how physical activity might indicate depression.
Loads of researchers looked at how smartphones can capture that data so they spent the least amount of time on that. But they knew that if a person’s activity level dropped suddenly that could be a sign of a problem.
But knowing how to capture the data only represented one part of the research. How could the data be used? Therapists and psychiatrists thought it might be good to receive regular updates about patients so that they had a better understanding of what patients did between visits.
But also, the app could feature a notification type element—a reminder of sorts to encourage someone to do something they enjoy or call a friend. Say a person likes walking in the park and the therapist recommends it. A person might forget but the phone could send a reminder. Something like “Frick Park is only five minutes away and it is known for its great walking paths” could be an indirect nudge encouraging people to follow treatment.
What's Next for our Phones?
While the research lost its funding because of the government sequestration and Hong could not develop the app, he used some of the lessons he learned with that project for a new one involving health coaches, professional who basically help others transform an unhealthy behavior to healthy one. Again, Hong and Zimmerman wondered if a smartphone could help people.
“If the smartphone is with you all the time can it get you to change behaviors?” Hong said.
The research exists in its early phases but Hong, Zimmerman and their colleagues are looking at how smartphone notification systems could provide incentives for healthy behaviors, much like he thought the notification system could work with the depression app.
Take someone who wants to lose weight. A lifestyle coach would create a plan for the person to follow. Notifications could use social pressure or turn it into a game. The app could alert people that a coworker walked nine miles this week and encourage the other person to try to beat the coworker. Or perhaps it could work like Pokemon Go and people could earn points for engaging in certain activities.
“This is where the smartphone, because it is with you all the time, you can do lots of things with it,” he says.
But the team will need to understand whether the smartphone incentives can truly lead to changed behavior, Hong says. If someone trying to quit smoking fails, her friends and family might be disappointed. This human element can contribute to greater adherence. Few people would care if their phones are disappointed in them.
“Having that human element is really important. You still need a person,” he says.
As they think of the ways to make the smartphone work best for health coaches, Hong believes that an app could help them reach more people. Maybe instead of seeing 20 or 30 people a week they’ll be able to work with 60 or 70.
“Maybe they can double their reach without inducing stress.”
Read the full paper from Jason Hong or find more research from the HCII's in healthcare.