Answers are grounded in this paper and cite their sources.
Crisis counseling aims to reduce suicide risk in a single conversation. Although recent studies have suggested that suicide risk can change nonlinearly over time, few have examined moment-to-moment changes in brief conversations. In this study, we investigated sudden shifts in the expressed suicide risk during crisis chats. Using 2,110,114 messages from 81,569 text-based crisis-line chats, we adapted sudden-gain and -loss criteria from psychotherapy research to identify within-conversations sudden changes in suicide risk. A machine-learning model assessed message-level suicide-risk expression. We compared expressed-risk changes between the conversations with and without sudden shifts. Overall, sudden reduction in expressed risk occurred in 12.5% of chats, and 13.9% showed sudden elevation. Chats with sudden reduction demonstrated greater improvement at the end of the chat than those with sudden elevation or no sudden changes. This study supports and extends findings on the nonlinearity of suicide risk to a moment-to-moment scale.
No takes yet. Share an insight, caveat, or question.
Zhang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: