Abstract
OBJECTIVE: Understanding demographic differences in planned suicide methods is key to prevention efforts, including means restriction. Although prior studies have identified gender differences in suicide methods, the stability vs emergence of those gender differences across age bands has not previously been examined. We aimed to identity stability or emergence of gender differences in planned suicide methods across development by testing for age by gender interactions, in a large cohort of crisis conversations. METHOD: We analyzed an archival dataset of 34,465 deidentified conversations from Crisis Text Line between January 5, 2017, and January 4, 2024. A total of 5,136 texters selected boy/man as the only applicable gender option, 23,387 selected only girl/woman, and 5,942 selected any other combination of gender identity options (transgender and gender diverse category [TGD]). Planned suicide methods were evaluated using a multi-step transformers-based natural language processing pipeline to extract and classify methods. RESULTS: There were significant effects of age for each of the most common planned suicide methods (overdose, sharp objects, hanging, firearms, jumping, vehicles, poisoning, drowning, asphyxiation, and "other"). There were significant effects of gender for overdose, hanging, firearms, jumping, vehicles, poison, asphyxiation, and other. There were age by gender interactions for overdose: χ(2) (6) = 33.097, p = .00001, firearms, χ(2) (6) = 30.48, p = .00003, and sharp objects, χ(2) (6) = 21.82, p = .001305. CONCLUSION: Gender effects differed across age groups for several of the most common suicide methods. These gender-specific age patterns present possibilities for targeted prevention efforts.