I

INTIMATE PARTNER VIOLENCE:HOMICIDES FOLLOWED BY

SUICIDES IN KENTUCKYSabrina Walsh, MPH; David Hemenway, PhD

January 20051 VOL 10310

Homicides followed by suicides are rare yet devas-tating events. This study, for the first time, detailsthe problem in Kentucky by linking coroner, medicalexaminer, vital statistics and administrative judicialdata. In the three-year period 1998-2000, there were492 fir,ea~rn homicides in Kentucky, of which 32(6. 5 "Ip).'JVerefollowed by a firearm suicide. All per-petrditorswho followed their homicide by killingthemselves were males. Just over 9% offirearm homi-cides (46/492) were intimate partner-related but themajority of these (54%) were followed by a suicide.Although we found that only a small percentage offirearm homicides are followed by firearm suicide,when women were shot and killed by their intimatepartners, the perpetrator shot himself in two thirds ofcases. Continued research about violent deaths mightprovide a better understanding of homicides followedby suicides-critical for surveillance and preventionefforts.

xtensive media coverage about homicides f01-{)Wedby suicides has created increased publicawareness-of these devastating"events. This

~ J-'

study details the problem in Kentucky and, forthe first time, links multiple data sources to de-termine whether victims attempted to get helpbefore problems escalated to lethality.

Homicides followed by suicides are rareevents. In most cases perpetrators are male andmost kill their female intimate partner."?Several studies report that the population in-volved in homicides followed by suicides isgenerally different from populations commit-ting single homicides .1,2,5,6According toFelthous and Hempel, in single homicideevents the victims are mostly male, "not so con-sistently closely tied to their killer,"! but themajority of victims in homicides followed bysuicides are female and intimately involvedwith their killers.I,2,6Surprisingly, perpetratorsof homicide followed by suicide more closelyresemble those who commit suicide thanperpetrators of homicide;I,5,6,sthe event is de-scribed by Palermo et al as an "extended sui-cide.:" Other studies report that where a men-tal disorder was mentioned, perpetrators ofhomicide followed by suicide were most oftendepressed, a known suicide risk factor.I,2,7,S1hisstudy examines the relationship between vic-tims and their killers in both single homicidesand those followed by suicides, and identifiescauses of both.

By carefully examining the data we canlearn more about how, to whom, where, when,and why these events occur. Populations maythen be identified and targeted through inter-ventions.

l METHODS

Data come from the Kentucky Firearm InjuryStatistics Program (KFISP), which includesdata from death certificates and coroner/med-ical examiner reports on all firearm homicidesand firearm suicides in Kentucky for the years1998-2000. Intimate partner violence (IPV)-re-lated data came from the narrative of theCoroner Investigation Report.

Information on the use of Community-Based Protective Services was providedthrough links to records of the Kentucky Adult

Protective Services andthe AdministrativeOffice of the Courts.These data indicatewhether there was a re-port of known or sus-pected adult domesticabuse or whether anemergency protectiveorder or a domestic vio-lence order had beenfiled. Data on court fil-ings for 2000 were notyet available.

RESULTS

Sabrina Walsh, is with theKentucky Injury Prevention and

Research Center, University ofKentucky, and Dr Hemenway is

with the Harvard Injury ControlResearch Center, Harvard School of

Public Health. Send correspondenceto: Sabrina Walsh, Kentucky Injury

Prevention and Research Center,College of Public Health,

University of Kentucky, 333 WallerAvenue, Suite 202, Lexington, KY40504; Phone: 859/257.6711; Fax:

859/257.3909; E-mail:slwalsO@uky.edu

In the three-year period 1998-2000, there were492 firearm homicides in Kentucky (TableI), ofwhich 32 (6.5%) were followed by a firearmsuicide. All perpetrators who followed theirhomicide by killing themselves were males.

Over 90% of firearm homicides (466/492)were not intimate partner violence (IPV)-related. Only 7 (1.5%) of these were followedby a firearm suicide. Four of the seven weremurders of family members and the other threewere suspects who killed themselves while be-ing chased by the police.

Just over 9% of firearm homicides (46/492)were intimate partner-related but the majority

Table 1. Firearm Homicides in Kentucky, 1998-2000

Number FollowedHomicides Number by Suicide

% Followedby Suicide

All 492 32*

Non-Intimate Partner 466 7tMale Victim 330 7FemaleVictim 66 0

Intimate Partner 46 25MaleVictim 11 2§Female Victim 35 23

6.5%

1.5%1.8%0.0%

54.3%18.1%65.7%

'In all 32 murder-suicides, the perpetrator was a maletOf all these 7 murder-suicide cases, 4 were murders of a familymember (son, father, father-in-law) and 3 were suspects who killedthemselves while being chased by police§The two cases of male victims involved males murdered while trying toprotect females (sister, mother)

January 20051 VOL 103 11

I

INTIMATEPARTNERVIOLENCE:HOMICIDESFOLLOWEDBYSUICIDESINKENTUCKY

12

Table 2. Firearm Suicides in Kentucky, 1998-2000

Number Preceded % PrecededSuicides Number by Homicide by Homicide

All 1001 32 3.2%

Non-Intimate Partner 871 7 0.8%MaleVictim 746 7 0.94%Female Victim 125 0 0.0%

Intimate Partner 130 25 19.2%MaleVictim 122 25 20.5%Female Victim 8 0 0.0%

'In all 32 murder-suicides, the perpetrator was a maletOf all these 7 murder-suicide cases, 4 were murders of a familymember (son, father, father-in-law) and 3 were suspects who killedthemselves while being chased by police§The two cases of male victims involved males murdered while trying toprotect females (sister, mother)

of these (54%)were followed by a suicide. Overthree quarters of the IPV-related firearm homi-cide victims were women, and in 66% of thesecases, the perpetrator then killed himself witha firearm. In the 2 cases of murder-suicide inwhich a male was murdered during an IPV-related incident, the murder victim had beentrying to protect a female relative.

In 23 of the 35 cases of female victims ofIPV-related firearm homicide the perpetratorthen killed himself with a firearm. Two of these23 cases involved ex-spouses; one victim wasa woman who was four months pregnant; andsix occurred during a separation, recent filingfor divorce, or threat of the woman leaving therelationship; one occurred in the presence ofchildren. Most of the women were killed intheir own homes.

In the period 1998-2000, there were 1001firearm suicides in Kentucky (Table2), or abouttwice as many firearm suicides as firearmhomicides. Intimate partner violence wasnoted in coroners' narrative reports for 130(16%).Of all firearm suicides, 32 (3%)were pre-ceded by a homicide. Only 7 (0.8%)of non-IPVrelated suicides followed a homicide, whereas25 IPV-related suicides (19%)followed a homi-cide. Of the 130 IPV-related firearm suicides,only 8 of the shooters were females.

Linkage with court records is limited to the23 women killed in IPV firearm homicides in

January 200S/VOL 703

Table 3. IPV-Related Firearm Homicides inKentucky 1998-1999

IPV-Related % Who Used AdultFemale Homicides Number Protective Services

HomicidesTotalUsed ProtectiveServices

239 39.1%

Homicide followedby suicideTotalUsed ProtectiveServices

142 14.2%

Data provided by the Intimate Partner Violence Sur-veillance Project: Data Linkage Report, KentuckyInjury Prevention and Research Center

1998and 1999.Nine had either filed DomesticViolence Orders or Emergency ProtectiveOrders, or had been seen by or reported toAdult Protective Services (Table 3). In 2 of the14 IPV-related firearm homicide-suicide casesin 1998-1999,the female homicide victim wasknown to the Department for CommunityBased Services.

DISCUSSION

In Kentucky for the period 1998-2000,over onethird (35/101) of female firearm homicideswere IPV-related; by contrast, only 3% (11/341)of the male firearm homicides were IPV-re-lated. National studies also find that the mainthreat of firearm homicide for women is theirintimate partners."

Nine percent of all firearm-related homi-cide cases were IPV-related. Only 1.5% ofhomicides without IPV factors were followedby a suicide, compared with 54%of homicideswith IPV noted in coroner reports. In IPV-re-lated cases, when a female was murdered witha firearm, almost two thirds of the time the per-petrator then shot and killed himself.

The vast majority (87%)of firearm suicidesin Kentucky were men. Sixteen per cent of sui-cides were IPV-related; these were even morelikely to be males (94%).The leading causes ofsuch suicides were designated as depression,

p

separation, recent filing for divorce, or threatof the woman leaving the relationship (notshown). Of the IPV-related male suicides, over20% were preceded by homicide. Successfulmental health treatment of males might saveboth intimate partners.

Few of the female IPV-related homicidevictims in Kentucky were known to the De-partment of Community Based Services. Theymay not have had contact with law enforce-ment, a physician, or another third party aboutdomestic violence, or IPV issues may havegone undetected or unreported during suchcontacts. Better outreach by social service agen-cies could facilitate better communication be-tween potential victims and those that can helpwithin their communities. For the victim, thatwould mean knowing who is safe to talk to andwhat services are available, and teaching thosethat interface with domestic violence victimstheir responsibilities and available services.

The data for this study are only for 3 yearsfor one state, and thus may not be generaliz-able to the USpopulation. In addition, the dataare only about firearm homicide and firearmsuicide; non-firearm methods of violent deathwere not examined. IPV cases undercountedthe deaths that involve domestic affairs be-cause they did not include incidents involvingother family members.'

The strengths of this study are that wewere able to link coroner,medical examiner, vi-tal statistics, and administrative judicial data.Wecould determine which homicides were fol-lowed by suicides, which were IPV-related,andwhich had a documented (with CommunityBased Services) history of PIV prior to thedeadly incident. In this study we found thatonly a small percentage of firearm homicidesare followed by firearm suicide, but whenwomen were shot and killed by their intimatepartners, the perpetrator shot himself in twothirds of cases.Most female IPV-relatedfirearm

homicide victims in Kentucky had not usedprotective services.

Until recently, information related to vio-lent death in Kentucky remained dormant,sketchy, scattered, and unusable. The coro-ner/medical examiner system is not central-ized and while police and forensic laboratorydata are centralized and available, they havenot been collected for research purposes. A cen-tralized system that includes investigation re-ports from police and forensic laboratories, aswell as more complete coroner and medical ex-aminer reports, allows the analysis of previ-ously fragmented data to tell a more completestory of violent death that is critical for sur-veillance and prevention efforts.

REFERENCES

1. Felthous AR, Hempel A. Combined homicide-suicides:a review. J Forensic Sci. 1995;40:846-857.

2. Rosenbaum M. The role of depression in couples in-volved in murder-suicide and homicide. Am J Psychi-atry. 1990;147:1036-1039.

3. Felthous AR, Hempel AG, Heredia A, et al. Combinedhomicide-suicide in Galveston County. J Forensic Sci.2001;46:586-592.

4. CDC Current trends: homicide followed by suicide-Kentucky, 1985-1990. MMWR 1991;40:652-653,659.

5. Lund L, Smorodinsky S. Violent death among intimatepartners: a comparison of homicide and homicide fol-lowed by suicide in California. Suicide Life Threat Behav.2001;31:451-459.

6. Palmer S, Humphrey JA. Offender-victim relationshipsin criminal homicide followed by offender's suicide,North Carolina, 1972-1977. Suicide Life Threat Behav.1980;10(2):106-118.

7. Buteau J, Lesage AD, Kiely MC Homicide followed bysuicide: a Quebec case series 1988-1990.Can J Psychiatnj.1993;38:552-556.

8. Palermo GB, Smith MB, Jenzten JM, et al. Murder-sui-cide of the jealous paranoid type: a multicenter pilotstudy. Am J Forensic Med Pathol. 1997;18(4):374-383.

9. Kellermann AL, Mercy JA. Men, women and murder:gender-specific differences in rates of fatal violence andvictimization. J Trauma. 1992;33:1-5.

ACKNOWLEDGMENTS:Shannon Beaven, Data Coordinator,Intimate Partner Violence Surveillance Project: DataLinkage Report, Kentucky Injury Prevention and ResearchCenter, University of Kentucky.

Reprinted from pages 10-13 of the January, 2005, JOURNAL OF THE KENTUCKY MEDICAL AsSOCIATIONCopyright 2005, by Kentucky Medical Association

January 200S/VOL 103 13