Crime & Justice Policy

Factors contributing to homicide-suicide: differences between firearm and non-firearm deaths

Schwab-Reese, L.M.; Peek-Asa, C.


The primary aim of this study is to determine the relationship between situational factors, method of death, and homicide-suicide deaths, specifically comparing method of death (firearm vs. nonfirearm) across these factors. We used data from the national violence death reporting system, a reporting system for violent deaths that links data from multiple sources. We included incidents that involved at least one homicide death followed by perpetrator suicide in the 42 states from 2013 to 2016. In addition to univariate analyses, we compared proportions of incidents that included a firearm to non-firearm incidents by incident, victim, and perpetrator characteristics. By far, firearm-related injuries were the most frequent cause of death for victims (85.6%) and for perpetrators (89.5%). Women, Hispanic individuals, individuals with a current mental health issue, and individuals with recent depression symptoms had lower odds of using a firearm, as did perpetrators who killed both an intimate partner and other family member. Individuals who had recently been in a fight had increased odds of using a firearm. Despite these differences, the overwhelming majority of homicide-suicide deaths involved firearms, which supports the need for adequate, appropriate firearm control measures to prevent these tragedies.

An estimated 11–17 homicide-suicide incidents, defined as at least one homicide death followed by a subsequent suicide by the perpetrator, occur each week in the United States, which is nearly twice the global average (Large, Smith, & Nielssen, 2009). Although low in number and in proportion to other violent deaths in the United States, homicide-suicide incidents often result in substantial trauma to the effected social networks and communities, which may be compounded by media coverage and social awareness of the homicide-suicide incidents (Marzuk, Tardiff, & Hirsch, 1992; Podlogar, Gai, Schneider, Hagan, & Joiner, 2018).

Despite the social impacts of homicide-suicide, research on these incidents has been limited, in part due to difficulty in case ascertainment and small sample sizes due to the relatively low number of deaths. Case identification is difficult in death certificate data because the homicide and suicide deaths are not linked. Many state domestic violence death review teams make this linkage, but few states have a sufficient number of cases for quantitative analysis. Homicide-suicide research is further complicated by the deaths of both perpetrators and victims, meaning individuals closest to the situation are no longer available to provide insight into the circumstances so vital information is often missing. Due to these limitations, most research on homicide-suicide is descriptive or theoretical (Bossarte, Simon, & Barker, 2006; Huguet & Lewis-Laietmark, 2015; Logan et al., 2008, 2013; Smith, Fowler, & Niolon, 2014).

This research suggests that homicide/suicide incidents are distinct from homicide and from suicide (Barber et al., 2008; Liem, 2010; Panczak et al., 2013; Podlogar et al., 2018; Reckdenwald & Simone, 2017; Starzomski & Nussbaum, 2000). Homicide perpetrators/suicide victims are, overall, predominantly middle-aged, married men with a history of domestic violence. Firearm injuries are the most frequent cause of death for both victims and perpetrators (Panczak et al., 2013). In early work on homicide-suicide, Marzuk and colleagues (1992) theorized these incidents could be categorized by the relationship between victim and perpetrator, and by primary motivating factor (e.g., financial/social stressors, “mercy killing”, retaliation, etc.). This classification system may be used to understand different types of homicide-suicide incidents and to better understand how risk and protective factors differ for each type (Marzuk et al., 1992). Some studies have found significant differences in circumstances around the time of death and the method of death, by victim-perpetrator relationship (Barber et al., 2008; Reckdenwald & Simone, 2017). For example, deaths involving intimate partners tend to have older victims and perpetrators, more female victims, and more male perpetrators than deaths compared with other victim-perpetrator relationships (Reckdenwald & Simone, 2017). However, it is not known how Marzuk’s overall typology applies to large samples of homicide-suicide deaths or how risk and protective factors vary across the types of incidents. This limited understanding of modifiable risk and protective factors for homicide-suicide hampers our ability to develop and implement effective intervention and prevention efforts.

The implementation of the National Violent Death Reporting System by the Centers for Disease Control, National Center for Injury Prevention and Control (NCIPC) overcomes some of the challenges associated with homicide-suicide research by creating a pooled, multi-source standard data collection process for all violent deaths (NCIPC, 2017). This dataset allows for systematic investigation into actionable factors contributing to homicide-suicide and how these factors differ across types of homicide-suicide incidents. Understanding how factors differ across firearm and non-firearm homicide-suicide incidents may inform prevention and intervention activities and enable targeting interventions based on specific risk characteristics. The primary aim of this manuscript is to determine the relationship between contributing factors, method of death, and homicide-suicide deaths, specifically determining differences in method of death (firearm vs. nonfirearm) by Marzuk classification for relationship type and motivating circumstances.