JMIR Publications

ABSTRACT

Background:

Unaddressed early stage mental health including stress, anxiety and mild depression can become burdens for individuals in the long term. Identifying milder symptoms before they become more serious is important and has motivated the use of digital phenotyping for that purpose. Digital phenotyping involves capturing continuous behavioural data via digital smartphone devices to monitor human behaviour and identify any abnormalities.

Objective:

This systematic literature review asked: What is the evidence of effectiveness for digital phenotyping using smartphones, in identifying behavioural patterns related to stress, anxiety, and mild depression? In particular, which smartphone sensors are found to be effective and what are associated challenges?

Methods:

We used the PRISMA process to identify 26 articles to assess the key smartphone sensors that are correlated with anxiety, stress and mild depression. Studies conducted with non-adult participants (e.g., teenagers, children) or with clinical populations were excluded. We also excluded personality and character measurement and phobia studies. As we were focused on the effectiveness of digital phenotyping with smartphones, results related to other wearable devices were excluded.

Results:

The study length varied from two weeks to 8.5 months. The Android operating system was more commonly used than iOS. A range of passive sensors were used in the studies including Global Positioning System (GPS), Bluetooth, ambient audio, light sensor, accelerometer, microphone, illuminance, and WIFI. These were used to assess locations visited, mobility, speech patterns, phone usage such as screen checking and time spent in bed, physical activity, sleep, and aspects of social interactions such as number of interactions and response time. Students, employees, and adults were the most-studied populations. Students who experienced depression and stress visited fewer locations, became less social, and accrued increased screen time. Depressed adults were less likely to leave home and were less physically active whereas anxious adults left their home more often and were more active. Anxious adults visited more locations but avoided going to places where they needed to socially interact. In contrast to students and adults, less mobility was seen as positive for employees because less mobility in workplaces was associated with more positivity and higher performance. We saw that location, physical activity, and social interaction data were highly related to participants’ mental health. Overall, sensing patterns through GPS, smartphone logs, audio, accelerometer, light, and keyboard were found to be significantly related to self-reported stress, anxiety and mild depression.

Conclusions:

The focus was to understand whether smartphone sensors could be effectively used to detect behavioural patterns associated to stress and anxiety in non-clinical participants. The reviewed studies provide evidence that smartphone sensors are effective in identifying behavioural patterns associated to anxiety, stress and mild depression.