Identifying Suicide-Related Language in Smartphone Keyboard Entries Among High-Risk Adolescents

Adolescent suicide rates have risen over the past two decades, underscoring the need for improved strategies to detect risk. This study leverages passively collected smartphone data to identify suicide-related language in adolescents’ keyboard usage using natural language processing. We developed a youth suicide lexicon for adolescent language and validated it with labeled data (N=121,515 entries), demonstrating higher sensitivity and precision than lexicons not designed for youth. Across two independent cohorts at elevated suicide risk (Ns=208 and 211; >6 million text entries), both lifetime suicidal thoughts and behaviors and current suicidal ideation were associated with increased frequency of smartphone suicide-related language. Human coding indicated varied language—e.g., serious expressions of active suicidal ideation, jokes, hyperbole, and expressing support for others. Most suicide-related entries did not express serious current first-person suicidal ideation, underscoring the need for improved approaches to distinguish intent. Findings highlight both the promise and limitations of NLP approaches for suicide prevention.