Risk Factors for Femicide-Suicide in Abusive Relationships: Results from a Multisite Case Control Study

Article excerpt

The killing of women by men who then take their own lives (femicide-suicide) is the most common form of homicide-suicide. This study identified femicide-suicide risk factors in an 11-city case-control study of femicide in the United States. Perpetrator, victim, relationship, and incident characteristics were analyzed for femicide-suicide cases (n = 67) and controls (n = 356, women living in the community with nonfatal physical abuse) using logistic regression modeling. Two risk factors emerged that were unique to femicide-suicides cases compared to overall femicide risk analyses: prior perpetrator suicide threats and victims having ever been married to the perpetrator.

Keywords: murder-suicide; homicide-suicide; partner violence; domestic violence

It is estimated that 1,000 to 1,500 homicide-suicide deaths occur annually in the United States (Brock, 2002; Marzuk, Tardiff, & Hirsch, 1992). Our understanding of the epidemiology of homicide followed by suicide, however, is hampered by the lack of a national surveillance system (Hannah, Turf, & Fierro, 1998; Malphurs & Cohen, 2002; Paulozzi, Mercy, Frazier, & Annest, 2004). Databases relied on for homicide and suicide rates, such as the Supplemental Homicide Report and National Vital Statistics System, are unable to link homicide to suicide events. The CDC National Violent Death Reporting System (NVDRS) is purported to correct this surveillance inadequacy (Paulozzi et al., 2004), but the first report from the NVDRS did not include homicide-suicide data (National Center for Injury Prevention and Control CDC, 2005). Until linked data is available, researchers have had to rely on police and medical examiner record review and follow-up interviews (a reasonable task only for small studies) or search newspaper clippings for case identification (Brock, 2002; Cohen, Llorente, & Eisdorfer, 1998; Lund & Smorodinsky, 2001; Malphurs & Cohen, 2002). Despite these methodologic limitations, a growing body of international literature confirms that homicide-suicide is patterned. Homicide is more likely to be followed by suicide when there is a close bond between the victim and perpetrator, with intimate partners-followed by children-most commonly killed; and following male perpetrators killing female partners (Aderibigbe, 1997; Alien, 1983; Barraclough & Harris, 2002; Brock, 2002; Buteau, Lesage, & Kiely, 1993; Campanelli & Gilson, 2002; Chan, Beh, & Broadburst, 2004; Currens et al., 1991 ; Gillespie, Hearn, & Silverman, 1998; Hannah et al., 1998; Lecomte & Fornes, 1998; Malphurs & Cohen, 2002; Milroy, Dratsas, & Ranson, 1997; Palermo et al., 1997; Polk, 1994; Rosenbaum, 1990; Stack, 1997; Websdale, 1999; Wolfgang, 1958). For example, a review of firearm homicides in Kentucky (1998-2000) found that when an intimate partner was killed (46), 70% involved a male killing his partner followed by killing himself (Walsh & Hemenway, 2005), confirming an earlier Kentucky homicide review (Currens et al., 1991). Across studies, approximately 25% of intimate partner femicides (homicide of women) in the US, Australia, Canada, and Sweden are followed by suicide, compared to less than 5% of nonintimate killings (Belfrage & Rying, 2004; Cooper & Eaves, 1996; Dawson & Gartner, 1998; P. Easteal, 1994; P. W. Easteal, 1993; Johnson & hotton, 2003; Lund & Smorodinsky, 2001; Morton, Runyan, Moracco, & Butts, 1998; Rosenbaum, 1990).

One theoretical explanation for femicide-suicide is that the perpetrator becomes remorseful after killing his source of nurturance and commits suicide (Stack, 1997; Wolfgang, 1958). This explanation, however, is challenged by the premeditated nature of the majority of femicide-suicides (Cooper & Eaves, 1996; Dawson, 2005; Dawson & Gartner, 1998; Marzuk et al., 1992) and the immediacy between the two acts (Cooper & Eaves, 1996). Several homicide-suicide researchers have developed and used homicide-suicide typologies (Belfrage & Rying, 2004; Campanelli & Gilson, 2002; Dawson, 2005; Felthous & Hempel, 1995; Hannah et al. …