JMIR Publications

ABSTRACT

Background:

Many individuals who self-injure seek support and information through online communities and mobile peer-support apps. While researchers have identified risks and benefits of participation, empirical work linking participation in these online spaces to self-injury behaviors and thoughts is limited.

Objective:

The objective of our study was to investigate the relationship between behavioral and linguistic traces on a mobile peer support app and self-injury outcomes.

Methods:

Natural use data and online surveys (n = 697) assessing self-injury outcomes were collected from 268 members (aged 13 – 38, median 19, 55.6% female) of a mobile peer-support app (aged 13 – 38, median 19, 55.6% female), over the course of 4 months. Participants were identified as having posted self-injury content by an internal classifier. Natural log data were used to predict self-injury outcomes in a series of multilevel logistic and linear regressions.

Results:

We found that greater engagement on an mobile peer-support app was associated with decreased likelihood of self-injury thoughts (OR = .25, CI = .09, .73) and fewer intentions to injure (b = -0.37, SE = .09), whereas posting triggering content was associated with increased likelihood of engaging in behaviors (OR = 5.37, CI = 1.25 – 23.05) and having self-injury thoughts (OR = 17.87, CI = 1.64, 194.15). Moreover, viewing triggering content was related to both a greater ability to resist (b = 1.39, SE = .66) and a greater intention to injure (b = 1.50, SE = .06).

Conclusions:

This is the first study to our knowledge to connect naturally occurring log data to survey data assessing self-injury outcomes over time. This work provides needed empirical support for the relationship between participation in online forums and self-injury outcomes and begins to articulate mechanisms contributing to this relationship.