Byrne ML, Lind MN, Horn SR, Mills KL, Nelson BW, Barnes ML, Slavich GM, Allen NB. (2021). Using mobile sensing data to assess stress: Associations with perceived and lifetime stress, mental health, sleep, and inflammation. Digital Health. https://doi.org/10.1177/20552076211037227

Researchers conducted a pilot study to validate a mobile sensing collection tool called Effortless Assessment of Risk States with measures of stress, mental health, sleep duration and inflammation. The study collected affective text language from smartphones among 25 young adult participants at a university. Participants installed a custom keyboard on their phones that collect every third word typed across all apps and the researchers analyzed text sentiment using a software package. The study collected data at two timepoints: once during a relatively less academically demanding period and once during a final exam period when participants are likely to be more stressed. Measures of stress, mental health and sleep are self-reported surveys. Saliva samples were collected to assess inflammation by analyzing the level of sCRP protein. Results indicate that the total number of positive words, total of negative words, and total of emotion expression words were strongly associated with lifetime stress exposure. Total negative words were found to be associated with decreased hours of sleep. Affective language was also shown to be associated with higher levels of stress and lower sCRP protein levels. Findings support the potential of using a mobile sensing tool to identify high stress and stress-related problems. For future directions, it could be helpful to develop a tool that can collect and analyze phrases of text (rather than words) and use alternate mobile sensing tools outside of keyboard usage.