Data gathered by a smartphone's sensors might one day be able to identify whether someone suffers from symptoms of depression, a study published today in the Journal of Medical Internet Research shows. By tracking average daily phone use and recording GPS data, scientists say they were able to identify people with depressive symptoms with a high level of accuracy. Unfortunately, the small size of the study and missing data points suggests that this method isn't exactly ready for prime time.
Major depression is one of the most common mental health issues in the US. In 2012, about 16 million adults — or 6.9 percent of all US adults — experienced at least one major depressive episode. That's why finding new ways of keeping track of patients is so important. Harnessing a phone's sensors could help identify people who are at risk for depression, and ensure that patients receive treatments more promptly, says David Mohr, a behavioral scientist at Northwestern University and co-author of the study.
People who suffer from depression tend to frequent fewer locations
In the study, the researchers used a Craigslist ad to recruit 40 people between the ages of 19 and 58. The researchers asked the participants to fill out a common depression survey. Then, they tracked the participants' movements and phone usage for a period of two weeks through an Android app called "Purple Robot."
People who suffer from depression tend to frequent fewer locations than people who don't, and from a social standpoint, they tend to be more withdrawn. So, the researchers analyzed the data of 28 participants to see if there was a relationship between high levels of phone usage — a proxy for being withdrawn — or GPS data, and depression. (Because of certain technical and adherence issues, 12 participants couldn't be included in the study.)
Some variables gathered by the phones were strongly related to participants' depression scores, the researchers found. For instance, "the more depressed people were, the more irregular their movements were," meaning that they didn't leave the house or return home at regular times, explains Sohrob Saeb, a computer scientist at Northwestern University and a co-author of the study. The researchers also found that people who were more depressed used their phones more, as expected.
The model identified people with depressive symptoms with 87 percent accuracy
Overall, the model that the scientists developed was able to identify people with depressive symptoms with 87 percent accuracy. The finding "suggests that phone sensor data can be used to provide objective behavioral markers of depression," says David Mohr, a behavioral scientist at Northwestern University and co-author of the study.
"It's a very small study, and they didn't get data over a long period of time, but those things aside, it definitely advances our knowledge base," says Ethan Berke, an epidemiologist at Dartmouth College who didn't work on the study. Justin Baker, a psychiatrist at Harvard University, agrees. "The study's novelty is in showing that tracking this information across many individuals is possible — and does a decent job at predicting depression scores."
Unfortunately, the researchers didn't break down the phone usage data to elucidate whether or not a participant was using the phone to browse the internet, play games, or communicate with others through instant message, for instance. Given that avoiding social interactions with others is often a symptom of depression, it's possible that figuring out if the phone was used to talk to others would have yielded a more accurate result.
Mohr thinks that the phone usage that was captured in the study was "not communication," however. "Rather it was using apps, browsing, etc. — activities that people engage in to distract themselves from emotional pain, or that people to do avoid situations."
Scientists didn't look at age or whether participants used their phones for work
The scientists also failed to ask the participants if they used their phones for work. "We've learned since this publication that understanding work context is important," Mohr says. Having information about whether a participants works in one location or many, for instance, "could improve the accuracy of the classifiers."
Finally, the age of the participants wasn't taken into account — even though studies have shown that people under the age of 45 spend more time using apps compared with older individuals. The fact that these data points are missing from the study "doesn't make the study invalid, but it does mean that you need to use caution when interpreting it," Berke says. And "this certainly isn't the end of the question — it's the beginning."
Mohr and his team are already in the process of enrolling 120 people into a new study. Those participants will be "more carefully defined in terms of depression, anxiety, and other variables," Mohr says. Because the scientists looked at participants' movements, it's possible that people with certain health conditions, like fatigue and mobility issues, "might look depressed when they in fact are not," he says. So increasing the information gathered about the participants and their environment will be part of the research team's next steps.
"Imagine a world where the devices we carry around in our pockets can help us track our own mental health."
The study's limitations can be overcome in future experiments, Baker says. A lot of doctors have been looking into ways of harnessing smartphone sensors to gather health data — Apple's ResearchKit is a good example — so it isn't surprising that mental health researchers have joined in. The ability to passively detect behavioral factors related to depression could open up a ton of possibilities for treatment.
"Imagine a world where the devices we carry around in our pockets can help us track our own mental health and catch ourselves before we end up needing psychiatric care," Baker says. This study "takes us one step closer" to that, "by showing how data that our cell phones already keep track of may be able to help spot depression early on."