IE 11 is not supported. For an optimal experience visit our site on another browser.

Smartphones can safeguard your health in some surprising ways

Data collected by your device could soon spot medical and psychological problems before you do.
Dr. Sarah Timmapuri, Gary Wilhelm
Hackensack University Medical Center cardiologist Dr. Sarah Timmapuri, left, looks at data on a smartphone synchronized to a new Fitbit Surge worn by patient Gary Wilhelm, 51, during an examination in Hackensack, New Jersey.Mel Evans / AP file

Smartphones can be literal lifesavers, but they've also been linked to health problems including depression and sleep disruption.

But now scientists are starting to recognize that the data collected by our iPhones and Android devices — location, the frequency of our calls and texts, and so on — can be used to detect and even predict certain health issues.

So even if they may raise the risk of some problems, our phones may soon help us avoid others — and become a sort of digital guardian angel watching over us from our own pockets.

“They’re a very good proxy for capturing how we interact with our environment and other people,” Dr. John Torous, co-director of the digital psychiatry program at Beth Israel Deaconess Medical Center in Boston, says of smartphones. "Things like how active people are, how much people are sleeping, how far people are traveling each day. That basic health information is important across anything from just general wellness to heart conditions to mental health conditions to diabetes."

Spotting trouble in small changes

Torous, who has built an app that collects smartphone data from people with depression, schizophrenia, and Alzheimer’s disease, says even subtle changes in a person’s behavior and daily routine can provide early warning signs of mental health problems — and that these changes are reflected in phone data.

A phone’s motion sensors and GPS system, for example, might show that someone isn’t getting enough exercise or sleep. Early signs of cognitive problems like dementia might show up as delayed reaction times as seen in the way someone interacts with a smartphone’s touchscreen.

Call and message logs can reveal social isolation, which could be evidence of depression. A study conducted recently at Northwestern University linked depression to frequent smartphone use and GPS data showing a lack of regular routines — like leaving for work at different times each day.

Researchers at the University of Michigan have built a smartphone app that monitors the voices of bipolar disorder patients to detect mood swings. Loud or rapid speech can be indicative of mania, while long pauses between utterances can suggest depression.

The Michigan researchers hope the app will eventually be able to predict mood swings and then notify the user and his/her doctors that an episode may be imminent.

At Aston University in England, Dr. Max Little, an associate professor of mathematics, is using smartphones to detect hand tremors, walking abnormalities, and speech problems suggestive of the progressive neurological disorder Parkinson’s disease.

Preliminary tests show his prototype system can use this information to determine that someone has the disease with up to 99 percent accuracy. If the system can be commercialized, it might help doctors better tailor treatments for Parkinson’s patients. Especially promising, says Little, is the system’s potential for spotting Parkinson’s sooner than is now possible.

“This is the excitement with this sort of tool,” he says. “It allows us to scale up measurement of behavior that could be relevant to diseases like this.”

Beyond academia

Academic researchers aren’t the only ones looking for ways to harness smartphone data for better health. Mindstrong, a startup in Palo Alto, California, is testing an app that collects location, activity, and social interaction data as well as keyboard use and even word choice. The hope is that the app will be able to help detect various mental illnesses; unusually slow typing or misspellings could indicate cognitive impairment, while certain language patterns — frequent use of the word “I,” for instance — might mean depression.

Chicago-based Triggr Health is already selling an addiction recovery app that it says can tell when users are likely to relapse. It relies on smartphone data as well as drug use history and other personal information. When the app spots a potential relapse, a member of the company’s chat support team steps in to offer advice or alert the individual’s real-world care team.

Not ready for prime time?

For all the promise of smartphone health data, there’s scant evidence proving its effectiveness in the real world. Then there’s the matter of privacy. “People may be surprised to find that personal data entered into a mental health app is forever out of their control,” says Torous.

Reliability is another concern. Little says insufficiently tested services could wind up diagnosing people with problems they don’t have. That could cause needless distress and waste doctors’ time.

"Human behavior is infinitely complex,” Little says. "Trying to measure a hand tremor from a smartphone is great until you realize someone's using a lawnmower.”

Running in the background

At present, Little says the most reliable disease monitoring systems involve the use of so-called “structured” tests, in which people carrying smartphones perform certain specific tasks. For instance, researchers might ask an individual to walk 20 steps to generate movement data, which is then analyzed — or to repeat a specific sentence so the same can be done with voice data.

But the key going forward will be to create smartphone systems that work in the background, monitoring an individual’s health without his or her awareness. That’s a tricky prospect, experts agree, but the intimate connection between us and our phones means the potential is there.

“All mental illnesses have some behavioral, social, or cognitive component that is likely already being recorded or manifest in the ways we use smartphones,” says Torous. “The challenge is identifying what those signals are and separating them from what is noise.”