Would You Want Your Phone to Monitor Your Mental Health?

Fred Dufour/AFP/Getty Images

Fred Dufour/AFP/Getty Images

When you carry your smartphone with you all day, it invariably collects a lot of data about where you are and what you do. Research shows that all of that data may offer some important insights into your mental health — but gaining those insights, of course, requires some careful thought about privacy and ethics.

A study published by the Journal of Medical Internet Research and written by researchers at Northwestern University details how the GPS and usage sensors already built into your smartphone can detect “daily-life behavioral markers” and determine whether or not you’re exhibiting signs of depression. Depression is a very common but very treatable mental health issue, but imposes a high societal burden because few people who need treatment receive it, since it takes the healthcare system months or even years to identify and treat it.

The researchers behind the study, titled “Mobile Phone Sensor Correlates of Depressive Symptom Severity in Daily-Life Behavior: An Exploratory Study,” administered a standard self-reported depression survey, and used a sensor data acquisition app called Purple Robot, which was installed on participants’ phones at the start of the two-week-long experiment. The app enabled the researchers to measure how mobile users were, what their daily routine was and how much they deviated from it, and how often they used their phones. The software distinguished participants with depressive symptoms from those without depressive symptoms with an accuracy of 86.5%. The researchers explained:

Features extracted from mobile phone sensor data, including GPS and phone usage, provided behavioral markers that were strongly related to depressive symptom severity. While these findings must be replicated in a larger study among participants with confirmed clinical symptoms, they suggest that phone sensors offer numerous clinical opportunities, including continuous monitoring of at-risk populations with little patient burden and interventions that can provide just-in-time outreach.

In the past, other groups of researchers have found that the array of sensors integrated into a smartphone can effectively detect social and sleep behaviors among patients with depression, gathering data on patterns that correlated significantly with the severity of depressive symptoms.

The current study aimed to extend previous work by focusing specifically on behavioral markers that are related to movement through geographic space because depression results in decreased motivation and activity. Additionally, the study looked at phone usage because the excessive use of a phone is considered compulsive behavior and has been linked to some symptoms of depression.

They posit that because “Depression is associated with several behavioral components (eg, reduction in activity, psychomotor retardation, changes in sleep) and motivational states (eg, anhedonia), some of which may be detectable using mobile phone sensors,” the mobile phones that we already carry with us daily “hold significant promise as a platform to monitor behavioral and environmental indicators of risk and resilience and to improve long-term management and treatment delivery to people suffering from depression.”

The study is an interesting read, but thinking about its implications raises a difficult question: would you really want your phone to monitor your mental health? The open-source Purple Robot app used in the study has full access to the Android sensor framework. That means that users grant it access to all of the sensors integrated into their phones, which, depending on the phone, can include the accelerometer, gravity sensor, gyroscope, rotational vector sensor, barometer, photometer, thermometer, orientation sensor, and magnetometer. Purple Robot sampled the GPS location sensor once every 5 minutes, and collected phone usage data by detecting the screen on and off events. The researchers explain the app’s full capabilities:

The Purple Robot mobile app and supporting server infrastructure is capable of collecting information about the user’s physical context (eg, motion), social settings (eg, number of Facebook friends), and phone usage behavior (eg, screen state). It also enables us to craft a complete data collection strategy configured for analyzing the relationship between depression and behavior data features of daily life.

The study had many limitations, which its authors acknowledge, but is likely to spur more studies that plumb the depths of smartphone-collected data for insights about the device’s owner. It’s nothing new that we want services that depend on deep troves of personal data, but at the same time, have reservations about making the privacy sacrifices necessary to use them. Nowhere is the struggle more significant than among the plethora of health-related apps that have sprung up on every smartphone and smartwatch platform.

When Apple introduced the first research apps for HealthKit, a platform that could eventually form the framework of a mental health monitoring app, many noted that these apps could be ethical liabilities if researchers weren’t careful. Institutions conducting studies need new ways to ensure that participants are eligible for studies, that they’re knowledgeable about the risk, and that their data is secure. Privacy with a research app like Purple Robot depends on the quality of the anonymization of users’ data. It’s difficult for researchers to promise perfect anonymity, but if there’s any way for users’ data to be reidentified with them, then the app isn’t adequately protecting users’ privacy.

An app that passively detects behavioral factors related to depression and even to other mental health conditions sounds like a promising application of all of the data that our smartphone sensors collect on a daily basis. But medical studies involving apps are still uncharted territory, and there are many ethical, technical, and logistical issues to be worked out before they go mainstream.

More from Gear & Style Cheat Sheet: