Estimating suicide risk from text

Scientists at MIT's McGovern Institute have created a tool to assess suicide risk from text conversations, providing a valuable resource for counselors and clinicians. The tool, developed by Daniel Low and Satra Ghosh, analyzes language patterns to estimate the likelihood of suicidal thoughts and behaviors. The researchers used Crisis Text Line's extensive database of crisis conversations, comprising approximately 16,000 interactions, to train their language-processing model.

The tool focuses on identifying key risk factors, such as depression, substance use, and active suicidal ideation, which are more prevalent in the most critical risk group. By analyzing the language used in crisis conversations, the tool can help clinicians prioritize intervention efforts and better understand the factors contributing to suicide risk.

The researchers emphasize that their lexicon, which includes about 60 words or phrases for each of 49 risk factors, must be used in conjunction with other assessment methods due to its limitations in considering context and missing similar but unlisted terms. Overall, this innovative tool represents a significant step forward in suicide risk assessment, offering a valuable resource for mental health professionals in crisis situations.