(PS13-65) User Engagement in Online Suicide Forums: Implications for Peer Support Interventions

Introduction: Suicide is a leading cause of death globally. It ranks as the second leading cause of death among children aged 10-14 and young adults aged 25-34 in the United States (CDC, 2021). A significant portion of individuals at risk of suicide do not engage with formal mental health services, highlighting barriers such as stigma and financial constraints (Stene-Larsen & Reneflot, 2019; Walby et al., 2018). Online forums dedicated to mental health and suicide have emerged as crucial platforms for those seeking support outside traditional mental health services. These digital spaces offer anonymity and accessibility, which can reduce the stigma associated with discussing suicidal thoughts. They can also provide a sense of belonging, allowing individuals to share their experiences, receive support, and access information in a non-judgmental environment. Previous research has underscored the therapeutic potential of such forums, their role in destigmatization, and their capacity to deliver peer support (Webb et al.; Sagar et al., 2020; Nokling, 2016; Gibson, 2014). Despite these benefits, fostering active and sustained user engagement remains a challenge. Recent findings by Low et al. (2021) indicate that a higher frequency of comments on users' posts can significantly enhance engagement, encouraging continued interaction within these online communities. In this study, we explored predictors of user engagement on the suicide-specific forum Sanctioned Suicide by examining variables such as the total number of responses to a post, the emotional tone of the posts (positive versus negative), and linguistic features of posts.

Method: Our study leverages data from Sanctioned Suicide, an online forum focused on discussions about suicide, aiming to offer a space for open dialogue and support for those feeling isolated or seeking advice. We employed data scraping techniques and analyzed 10,646 posts (representing 4,087 unique users).

Results: Using negative binomial regression analysis, results revealed that the more responses a user’s post received, the higher the likelihood was that the user would make a future post on the site (B = 0.0096, SE = 0.0001, z = 65.64, p < .001). Further, emotional tone analysis showed that posts with a positive tone (e.g., posts containing words such as good, well, new, love) were associated with more responses from peers (B = 0.0498, p < .001). In contrast, posts with a negative tone (e.g., posts containing words such as bad, wrong, too much, hate) tended to receive fewer peer responses (B = -0.01612, p = .00792). We will conduct additional analyses using Linguistic Inquiry and Word Count (LIWC-22) to characterize linguistic features that may predict peer response and user engagement.

Conclusion: These findings highlight that the number of responses and the emotional tone of users’ posts are crucial factors influencing peer response and users' continued engagement in online forums. This research emphasizes the importance of understanding user engagement in digital communities, particularly those centered around sensitive topics like suicide, which has implications for the integration of peer support interventions on these platforms.