Although suicide and homicide are two of the leading causes of death, theoretical understanding and rigorous quantitative examination of homicide-suicideuthe rare combination of suicide and homicideuare sparse. We ground homicide-suicide in the stream analogy of lethal violence and use three analytical techniques to examine the shared and unique correlates of suicide, homicide-suicide, and homicide. Data come from the National Violent Death Reporting System (NVDRS), a pooled, cross-sectional time-series database of US victims across 46 states from 2003 to 2013. A logistic regression model with robust standard errors clustered across states and year fixed effects on 105,243 suicide and homicide-suicide cases predicts the odds of homicide-suicide relative to suicide across a range of theoretically informed covariates. A multinomial logistic regression model investigates the factors that distinguish 26,243 appended homicide cases (from the NVDRS) from suicide and homicide-suicide cases. And, an item-response-based statistical approach examines the extent to which homicide-suicide perpetrators are primarily homicidal versus primarily suicidal. Results indicated that interpersonal stressors and criminal history increased, while physical and mental health stressors decreased, the odds of homicide-suicide relative to suicide. In addition, the etiology of homicide-suicide more closely resembled homicide than suicide. Finally, almost one-half of the homicide-suicide cases in our sample were classified as primarily homicidal, while less than one-quarter were classified as primarily suicidal. Findings suggest that homicide-suicide can be conceptualized as a current in the stream analogy of lethal violence, and that the prevention of homicide-suicide would be better facilitated via screening of violence prevention than suicide prevention programs.
