BACKGROUND: Multivariate studies of specific suicide means are relatively rare, given the logistical challenges associated with the low base rate of suicide in the general population. Thus, information on individual characteristics associated with specific suicide means remains relatively wanting. The current study provided the largest examination to date of sociodemographic characteristics associated with different means of lethality among suicide decedents, using data from a multi-state population-based surveillance system. METHODS: Multivariate logistic regression was used with data for 20,577 suicide decedents in the National Violent Death Reporting System from 2003 to 2005. RESULTS: Firearm decedents were more likely male, elderly, non-Hispanic white, married, veterans, and born in the U.S. Hanging and suffocation decedents were more likely male, young, racial/ethnic minorities, never married, non-veterans, and foreign-born. Decedents that jumped from heights were more likely female, older, non-Hispanic black, never married, non-veterans, and foreign-born. Decedents who used sharp instruments were more likely older, never married, and foreign-born. Self-poisoned decedents were more likely female, middle-age, non-Hispanic white, and not married. Regarding specific poisons, alcohol was more likely to be used by middle-age decedents; gas by males, elderly, and married individuals; over-the-counter drugs by females, adolescents, and foreign-born decedents; prescription drugs by females, middle-aged, and U.S. born individuals; and street drugs by males and racial/ethnic minorities. LIMITATIONS: The data were drawn from 18 states and so cannot be regarded as nationally representative. CONCLUSIONS: Substantial sociodemographic variability exists across different suicide means. Recognition of this variability may help to tailor prevention efforts involving means restriction.
BACKGROUND: Multivariate studies of specific suicide means are relatively rare, given the logistical challenges associated with the low base rate of suicide in the general population. Thus, information on individual characteristics associated with specific suicide means remains relatively wanting. The current study provided the largest examination to date of sociodemographic characteristics associated with different means of lethality among suicide decedents, using data from a multi-state population-based surveillance system. METHODS: Multivariate logistic regression was used with data for 20,577 suicide decedents in the National Violent Death Reporting System from 2003 to 2005. RESULTS: Firearm decedents were more likely male, elderly, non-Hispanic white, married, veterans, and born in the U.S. Hanging and suffocation decedents were more likely male, young, racial/ethnic minorities, never married, non-veterans, and foreign-born. Decedents that jumped from heights were more likely female, older, non-Hispanic black, never married, non-veterans, and foreign-born. Decedents who used sharp instruments were more likely older, never married, and foreign-born. Self-poisoned decedents were more likely female, middle-age, non-Hispanic white, and not married. Regarding specific poisons, alcohol was more likely to be used by middle-age decedents; gas by males, elderly, and married individuals; over-the-counter drugs by females, adolescents, and foreign-born decedents; prescription drugs by females, middle-aged, and U.S. born individuals; and street drugs by males and racial/ethnic minorities. LIMITATIONS: The data were drawn from 18 states and so cannot be regarded as nationally representative. CONCLUSIONS: Substantial sociodemographic variability exists across different suicide means. Recognition of this variability may help to tailor prevention efforts involving means restriction.