The Association Between X (Formerly Twitter) Content and Suicide Following a National Media Engagement Strategy in Canada.

Abstract

Objective: To identify associations between X (formerly Twitter) content and narratives and suicide deaths following Canadian media engagement efforts.

Method: We coded suicide-related tweets (N = 862) geolocated to Toronto originating from high-level influencers (July 2020-June 2021) for general, putatively harmful, and putatively protective content and overarching ("gestalt") narrative. We used multivariable logistic regression to examine associations between tweet characteristics and increases or decreases in suicide deaths in Toronto in the 7 days after posting, compared with a 7-day control window.

Results: Tweets most strongly associated with increased subsequent suicides included calls to action to address suicide (OR 3.59, 95%CI 1.97-6.55), those arising from celebrities (3.55, 95%CI 1.87-6.75), and gestalt narratives focused on suicide deaths and/or attempts in A-list celebrities (OR 2.85, 95%CI 1.24-6.53). Tweets most strongly associated with decreased subsequent suicides were those about specific at-risk groups (LGBTQI+ or indigenous peoples; OR 0.18, 95%CI 0.09-0.38) and artists/entertainers (OR 0.33, 95%CI 0.14-0.76), and that referenced the 988 helpline (OR 0.30, 95%CI 0.09-0.92). Neither putatively harmful nor protective elements were common in tweets.

Conclusions: These findings largely align with prior evidence of the potential harms and benefits of specific suicide-related media content and narratives. We found limited harmful content in tweets after Canadian media engagement efforts, but also little evidence of helpful content.