Temporal dynamics in the content of wishes after the expression of suicidal ideation on Twitter: Focusing on changes in retweet content

Suicide prevention is an urgent issue, and examining the wishes of suicidal individuals is helpful for finding effective suicide prevention strategies. As shown in previous studies, Twitter can be used as a tool to express suicidal ideation, and the content of retweets (RTs), reflects users' interests or empathy. This study explored the wishes of suicidal individuals and the temporal dynamics of their wishes by focusing on how RTs change over time after individuals express suicidal ideation. Twitter data from Japan were collected via the Twitter API. Specifically, tweets containing “#shinitai (I want to die)”, RTs posted within the first hour of these tweets (targets), and RTs posted by users who did not tweet expressions of suicidal ideation (controls) were collected for four weeks. Approximately 60 % of the target RTs were negative, and negative emotional words occurred more frequently than did the control RTs (ps < .001). However, the correspondence analysis and co-occurrence network analysis revealed that frequent words changed over time, with fluctuations between words indicating the wish to die and words reflecting the wish for resolution. The results suggest that although users who tweeted expressions of suicidal ideation were primarily interested in the negative sentiments in others’ tweets, they oscillated between the wish to die and the wish to live. The potential contribution of focusing on RTs for a detailed understanding of temporal dynamics in the wishes of suicidal individuals is discussed.