Results from the Hawaii domestic violence fatality review, 2000-2009

Abstract: Background: Patterns of domestic violence fatalities and agency responses in Hawaii have not been explicated. Methods: Retrospective reviews of events leading up to domestic violence related fatalities in Hawaii were assessed from 45 adjudicated cases that resulted in 62 fatalities for the ten year period from 2000-2009. Results: Almost one-half of the fatalities were homicide/suicide combinations. Females were disproportionately more likely to be fatal victims of domestic violence relative to their proportion in the population. Those aged 21-40 years and those over 80 years were more likely to be fatal victims of domestic violence, relative to their proportion in the population. Filipinas and ‘Other” ethnic groups are disproportionately more likely to be fatal victims of domestic violence while Native Hawaiians and Japanese are less likely to be fatal victims, relative to their proportions in the population. In more than two-thirds of the cases, the victim had made some attempt to leave the relationship prior to the fatality. Conclusions: In the majority of cases there was agency involvement in some form: either the victim alone or the perpetrator alone, or both. However, less than one-third (31.1%) of the cases over the past ten years had documentation of prior violence from medical reports, so this may be an area to further document and address domestic violence.

Introduction omen in the United States experience more intimate partner violence than do men. Estimates from the National Violence Against Women Survey show that 22.1 percent of surveyed women, compared with 7.4 percent of surveyed men, reported they were physically assaulted by a current or former spouse, cohabiting partner, boyfriend or girlfriend, or date in their lifetime. 1 In Hawaii it is estimated that about 132,000 adults (15%) experienced intimate partner violence in their lifetime, 2 however, only a few of those adults re-ported such experiences to authorities and even fewer resulted in fatalities. A substantial proportion of all homicides in the U.S. are committed by intimate partners of the victims, 3, 4 with about 1 in 3 of female homicides committed by intimate/domestic partners. In the majority of cases of intimate partner/domestic violence fatalities (70-80%), no matter which partner was killed, the man physically abused the woman prior to the homicide. 4 Nationally in 2008, Hawaii ranked #23 with a rate of 1.26 per 100,000 population of females murdered by males in single victim/single offender homi-cides, compared to 2.96 per 100,000 in Nevada (the highest). 5 In 1997, Hawaii Revised Statutes, HRS §321-471 6 authorized the Department of Health (DOH) to conduct multidisciplinary and multiagency reviews of domestic violence fatalities to reduce the incidence of preventable fatalities. These retrospective reviews of events leading up to a domestic violence fatalities analyze (1) incident cases, their characteristics, risk factors and (2) the system responses to these cases by community agencies, institutions, and other organizations involved. The purpose of the Hawaii Domestic Violence Fatality Review (DVFR) is to conduct retrospective reviews of events leading up to a domestic violence fatality on a case by case basis. The review is nonjudgmental and is not to place blame on any of the agencies and entities involved in any of the cases prior to the fatality. The goal of reviewing domestic violence fatalities through fatality case reviews is similar to reviews conducted after airplane crashes: to help determine what went wrong and see what could have been done differently to prevent such fatalities. 7 The Hawaii Department of Health"s Domestic Violence Strategic Plan, 2007-2012 and the Hawaii Domestic Violence Action Center define domestic violence as a pattern of behavior which includes physical, sexual and/or emotional abuse between intimate partners (including verbal abuse and/or psychological tactics such as intimidation and/or degrading someone), including dating violence. Sometimes it involves other family members or friends. Domestic violence can include and can limit the ability of the victim to make personal choices, access family resources or assets and have selfdetermination.

Domestic Violence Fatality Review Council and Team
Selection and identification of cases for review is determined by the Domestic Violence Fatality Review (DVFR) Council. 8 The DVFR Council is a multidisciplinary, multiagency group consisting of representatives from public and private agencies and organizations. Case sources can include newspaper articles, police reports, referrals from participating members, community agencies and institutions, the medical examiner"s office, and other sources. Other related fatalities secondary to the domestic violence fatality may also be reviewed on a case-by-case basis (e.g., neighbors, police officers, and other emergency workers). For example, sometimes a child or other family member becomes an additional victim in a domestic violence fatality. However, the separate Child Fatality Review (CDR) team reviews child fatalities resulting from domestic violence situations, al-though this does not preclude joint reviews with the Domestic Violence Fatality Review (DVFR) team. All cases are reviewed by a core DVFR team, whose members include representatives from the Medical Examiner"s Office, law enforcement, Department of Human Services, Emergency Medical Services, domestic violence advocates/coalition, the Prosecuting Attorney"s Office, the Department of Health and the Judiciary. The DVFR team travels throughout the state. At each fatality review, the team invites community/county representatives to participate.

Case Ascertainment
The DVFR team uses a broad definition of domestic violence fatality. Reviews may include family or household members as referenced in HRS §321-471. The reviews may include both homicides and suicides relating to domestic violence. Domestic violence fatalities include, but are not limited to: all homicides in which the victim was a current or former intimate partner of the perpetrator; homicides of people other than the intimate partner, which occur in the context of domestic violence or during an attempt to kill the intimate partner; homicides occurring as an extension of or in response to ongoing abuse between intimate partners, e.g., when an ex-spouse kills the children; and suicides that may be a result of domestic violence.
The DVFR examines the facts and circumstances surrounding the fatality case under review. All information disclosed at the review is confidential. The formal review of a fatality is delayed until any criminal investigations or prosecutions connected with the fatality are completed. At the review, the facts and circumstances of the case are considered. This information is recorded on case review summary forms and includes: (1) background and basic demographics of the persons involved, (2) circumstances of the event, personal histories of the parties including any previous agency involvement: medical, mental health, financial, legal (civil and criminal complaints, e.g., arrests for assaults and/or abuse of family or household members, the existence of past or present protective orders), (3) services obtained by the victim, perpetrator and family prior to the fatal incident and services rendered after the fatality to family members and/or other affected persons and (4) outcomes. After the review, a determination is made by consensus as to whether or not the case is domestic violence related. The DVFR Team informs the DVFR Council on each review and the Council critically examines data and makes recommendations from each review to assure that system problems are addressed appropriately and timely, as well as follow up for improvement and change through training, policy development and/or legislative action.

Methods
Reviews of domestic violence fatalities occurring between the years 2000 and 2009 were conducted on all islands. There were 72 cases identified. Of these, 22 cases were not reviewed as these cases have not yet been adjudicated, which is a requirement of the DVFR. Of the remaining 50 cases, 45 were determined to be domestic violence related based on team determination after the review and 5 were found not to be domestic violence related.
Hard copies of all domestic violence related case fatality review summary forms (n=45 cases) were reviewed, variables created and data for each case were populated into a Statistical Package for the Social Sciences (SPSS) database in 2012. This database was specifically designed for this report and includes all variables on case review summary forms. The data were analyzed using SPSS and findings are summarized here. The data were analyzed first by total fatalities and then by cases, since almost one-half of the cases involved more than one fatality (homicide of the victim/suicide by perpetrator combinations as well as other fatalities). Data for all fatalities (n=62) were analyzed by characteristics of the victims and perpetrators (relationships), including sex, age ethnicity and county relative to the proportions in the population, and the circumstances of the fatalities (other additional fatalities and location of the fatalities). Data for the cases (n=45) were analyzed by characteristics, including demographics (sex, ethnicity, age at fatality, citizenship status and economic status), weapons used, attempts to leave by the fatality victim, agency involvement in the cases, fatality risk factors present, prior history of violence and the system responses to the cases. This information should inform and contribute to the analysis of system responses and recommend changes to prevent domestic violence fatalities. While this report represents 10 years of data, the number of fatalities is small. These small numbers limit statistical analysis and the ability to make definitive conclusions. Table 1 provides a summary of the total domestic violence related fatalities reviewed from 2000-2009. The 45 domestic violence related reviewed cases resulted in 62 fatalities, of which 33 fatalities (53.2%) were female domestic violence victims, 31 of whom (50.0%) were killed by a current or former husband/boyfriend. There were 28 total fatalities (14 cases) which were combined homicide/suicides (45.2%). There were 7 fatalities (11.3%) among friends or family members of the domestic violence victim (one was a child of the victim). There were 5 male domestic violence victim fatalities (8.1%), 4 of whom (6.5%) were killed by their current or former partners (wife/girlfriend) and one suicide. Of the total fatalities (N=62), 37 or 60% involved only one death, whereas 25 or 40% involved two deaths. Overall, of the 45 domestic violence cases, 37 or 82% resulted in single fatalities (60% of all fatalities). However, of the multiple fatalities, a total of 25 deaths resulted from as few as 8 domestic violence cases. Four women were pregnant at the time of the fatality. One unborn child was delivered by cesarean section and survived, so there were an additional 3 fetal fatalities not included in total number of fatalities. Table 2 provides a summary of the total domestic violence fatalities by sex and age along with the average population in the State of Hawaii from 2000-2009, the population proportion, the number and proportion of fatalities. Females are disproportionately more likely to be fatal victims of domestic violence (61.3%) and men are less likely (38.7%), relative to their proportion in the population. In the analysis of all fatalities, we are including all male fatalities as domestic violence related, since 15 of the 25 male fatalities (60%) were suicides; there were no female suicides. By age groups, the proportions in the age groups, 21-30, 31-40 and over 80 years (8.1%) are disproportionately represented as victims of domestic violence, relative to their Hawaii state population age distribution. The differences from the state"s population proportion were statistically significant for both the 21-30 and the 31-40 age groups (p<.05) (see Table 2). Table 3 provides a summary of total domestic violence fatalities by ethnicity and county, along with the average population from 2000-2009, the proportion of the population, the number and proportion of fatalities. Filipinos (24.2%) and other ethnic groups (19.4%) (Which includes Samoan, other Pacific Islander, Black, Mexican and 6 with multiple ethnicities) were more likely to be fatal victims of domestic violence. These ethnic groups were disproportionately represented relative to their proportions in the population (24.2% vs. 15% and 19.4% vs. 9.6% respectively) and the proportion differences were statistically significant (p<.05). Native Hawaiians (9.7%) and Japanese (14.5%) were disproportionately less likely to be fatal victims of domestic violence, relative to their proportion in the population, and this was statistically significant for Native Hawaiians (p<.05).

Domestic Violence Related Fatalities (N=62)
As seen in Table 3, neighbor island counties have higher proportions more likely to be victims of domestic violence relative to their numbers in the population (especially Kauai), while the city and county of Honolulu and Maui county have lower proportions (64.5%) compared to the total proportion in the population (70.7%).
More than one in ten cases included innocent bystanders, not just a domestic violence pair (11.3% of total fatalities). Forty six of the 62 domestic violence fatalities (74.2%) occurred in a home. Twenty five occurred in the mutual home of the victim and perpetrator (40.3% of all fatalities), 14 in the victim"s home (22.6%), 6 in the DV perpetrator"s home (9.7%), and 1 in the home of a family or friend (1.6%). Three occurred at the domestic violence victim"s workplace (4.8%), while 13 occurred in a public area (21.0%) such as a street, park or public building. Among the fatalities occurring in a home, 20 were in the bedroom  Table 4 examines the demographic characteristics of the domestic violence victims and perpetrators in the relationship that led to the fatality (N=45). The majority of the victims were female (86.7%) and the majority of the perpetrators were male (86.7%). Filipinas had the highest proportion (28.9%) being the victims of domestic violence relationships, followed by Europeans/Whites (20.0%) and those with multiple ethnicities (13.3%). Among perpetrators, the majority were Filipino (22.2%) followed by whites, other ethnic groups, multiple ethnicities (17.8% each) and Japanese (15.6%). In just over one-half of the cases (53.2%), the victim was between 21-40 years of age at the time of death. The majority of both victims and perpetrators were US citizens (71.1%), although there were some documented immigrant/refugee and some tourist cases. Most of the victims were employed (60%), as were the perpetrators (40%), yet many of the perpetrators were unemployed (17.8%). Educational status was not presented because more than half of all the cases had missing information for educational status. Chi-square analysis found that only sex had a significant association with domestic violence status (victim and perpetrator) (p<.05).

Domestic Violence Related Fatality Cases (N=45)
In addition to the characteristics noted above, there were children in the family unit in 88.9% of the cases (n=40) and 2 women were pregnant at the time of the fatality. There were 31 cases (68.9%) where children were present at the location of the fatality event. Two children were involved as attempted homicides (one was a homicide as noted in Table 4). In 11 cases multiple weapons were utilized during the fatality event, so that 59 weapons were used in 45 cases. A knife was utilized 20 times (33.9%), a gun 14 times (23.7%). In addition, blunt objects (11.9%), body parts (11.9%), and motor vehicles (10.3%) were used as the weapon. Other weapons use included tools, flammable liquid/light and caregiver neglect.
In 20 cases (44.4%) the domestic violence victim had successfully left the relationship (both physically and emotionally). In 15 of the cases (33.3%) the victim had made some attempt to leave. [Of these 15 cases (data not shown), in 6 cases (13.3%) the victim had emotionally left (broken-up, stated intent to leave) but was still physically present. In 9 cases (20.0%) the domestic violence fatality victim had tried to leave previously but was physically and emotionally present when the fatality event occurred]. Therefore, in 35 cases (77.8%) the victim had made some attempt to they leave or try to leave. In two thirds of the cases (73.3%), there was agency involvement with both the victim and perpetrator. However, in the vast majority of cases (95.6%) there was agency involvement in some form: either the victim alone, the perpetrator alone or with both. In almost one-half of the cases (44.4%) there were 5 or more agencies involved in the case, including law enforcement (29 cases), the prosecutors" office (20 cases), family court/judge (22 cases), district court/judge (4 cases), circuit court/judge (10 cases), substance abuse treatment and assessment (13 cases), mental health (8 cases), health care providers (24 cases), local hospitals (9 cases), religious organizations/churches (4 cases), protection order advocacy program (4 cases), and DHS (6 cases). Other agencies involved (22 cases) include: community or court based legal advocacy organizations, culturally specific organi-zation, homeless shelter, sexual assault program, EMS, other non-specified social service agency involvement, and other non-specified domestic violence victim support services. Health care system involvement (health care providers, mental health providers, local hospitals and EMS) was present in 41 of the 45 cases (91.1%). Table 5 shows risk factors related to agency documents (reports from law enforcement, court, medical and counseling/advocacy) or verbal reports/accounts (perpetrator made threats to kill or harm victim and/or children and/or family members) by anyone (the victim or others) about prior violence or abuse event(s). Thirty one agency reports (68.9% of cases) and 34 verbal accounts by victims or others (75.6% of cases) provide evidence of prior physical abuse of the victim or others by the perpetrator. Thirty one victim or other verbal  Table 6 shows the risk factors related to a prior history of violence based on agency documented and/or verbally reported prior violence. Less than one-third (31.1%) had documentation of prior violence from medical reports and less than one-fourth (22.2%) from counseling or advocacy reports. In more than one-half of the cases there was documented prior history of violence from law enforcement reports (71.1%), and/or court documents (62.2%), and in the majority of cases, verbal reports by the victim (88.9%) or others (93.3%). In the vast majority of cases (95.6%) or 43 of 45 cases, there was documentation and/or verbal reports of prior violence by the perpetrator. Table 7 shows the risk factors related to disability, language and substance abuse for the 45 cases. In some cases there was evidence of physical, cognitive or mental disability, but no discernible pattern. However, English proficiency/language other than English was an issue. English was not the first language for 9 victims (20%) and 11 perpetrators (25%) and these included Filipino and other languages. A substantial number of victims (11.1%) and perpetrators (17.8%) could not speak or verbalize in English without a translator. More than one-half of the perpetrators (28, or 62.2%) had a history of substance abuse (drugs and/or alcohol) as did 13 of the victims (28.9%). Further, 22 of the perpetrators (48.9%) and 8 of the victims (17.8%) were affected by drugs or alcohol at the time of the fatality. Eighteen of the perpetrators (40.0%) and 11.1% of victims had a history of mental illness. Table 8 addresses the relationship of known homicide cases from the Uniform Crime Reports (UCR) to the Domestic Violence Fatality Review (DVFR). The UCR  In any form 43 95.6% *More than one possible. are also homicide victims. Child fatalities are reviewed by a separate process and would also not be included in IPV counts. Therefore, it is possible that the percentage of domestic violence related homicides is slightly higher than 20-22%. Table 9 lists identified barriers in the fatality cases reviewed. Multiple barriers were identified for each fatality case. Power and control, lethality or stalking issues were found to be present in 24 cases (53.3%). These are also known as "red flags" or warnings that domestic violence may be present. Though these early warning signs were identified they did not prevent the fatalities from occurring. The need for agency protocol improvement or change was identified in 19 cases (42.2%) as well as law or policy change in 16 cases (35.6%). Lack of involvement from significant others (family, friends, co-workers, etc.) was found in 18 cases (40.0%). Cultural, language or religious involvement and sensitivity regarding domestic violence identification and intervention was identified in 17 cases (37.8%). Table 10 lists recommendations identified by the DVFR teams that could prevent or assist with domestic violence fatality prevention. Improving and increasing training for professionals in domestic violence was recommended in every case (100.0%). Agency protocol improvement or change was recommended in 43 cases (95.6%) and systems change in 19 cases 942.2%). Increasing public awareness regarding domestic violence prevention and assistance for those in need was recommended in 42 cases (93.3%). The need to increase referrals to existing organizations or agencies was recommended in 30 cases (66.7%). Increasing cultural, language and religious involvement and sensitivity was rec-ommended in 20 cases (44.4%). In addition, there is also national recommended screening for intimate partner violence for girls beginning at age 14 (due to increases in teen dating violence) and at all visits of pregnant women to healthcare providers (due to escalation of intimate partner violence during pregnancy).

Discussion
This report highlighted the fact that domestic violence may lead up to premature deaths and can include not only one victim death per case but multiple victims. The overall number of domestic fatalities may be considered small at less than .01% of the total deaths during this period (62 period deaths due to domestic violence/89,393 total period deaths). 10 Yet this number is greater than CDC"s standard of 50 for a minimum sample size. Regardless of the small numbers, premature deaths due to violence have a substantial burden and impact on Hawaii"s population. This report is a beginning to provide better insight on the characteristics, circumstances and background to such events and thus may be able to suggest ways to prevent or minimize the occurrence of fatalities attributable to domestic violence. This report provides useful data that could be compared with other states and the nation on domestic violence fatalities. On the other hand, given Hawaii"s multicultural environment the generalizability of the results may be limited. In Hawaii, over the past ten years, Filipinos and Other ethnic groups are disproportionately more likely to be fatal victims of domestic violence, relative to their proportion in the population. It is possible that tolerance for domestic violence is influenced by socio-cultural factors and length of immigration and acculturation. Steinberg, 11 who spent more than 40 years studying the Philippines, suggested that: "Personalism, smooth interpersonal relationships, and hierarchical structures are safety lids for the Philippine society. Filipinos are both friendly and tolerant, but the society also tolerates moments of violence. To run amok is an understandable behavior if an individual has been wronged or provoked sufficiently. Crimes of passion abound and revenge is, in Philippine terms, often an acceptable explanation of criminal behavior. Temporary, explosive anger at a personal affront is a way Filipinos expresses existential rage. Political violence, especially just prior to elections, is widely accepted and usually unpunished".
On the other hand, regardless of cultural issues the statistically significant differences found by age are also instructive, whereby younger people 21-30 years and 31-40 years are disproportionately more likely to be fatal victims of domestic violence suggests something about the nature of intimate partner relationships at this age (sexual reproductive age), along with sexual jealousies, possessiveness and power and control issues. A known history of prior violence is the key element in all these cases.
The sociological and criminological literature provides a conceptual framework for explaining criminal or deviant behavior such as domestic violence. That is, deviant behavior is the result of the same social processes that result in conformity: these are learned behaviors. 12,13 People learn the values, norms and beliefs of conformity, deviance or criminality through socialization and interaction with others. People learn that is acceptable or not acceptable to act violently toward family members, spouses, girlfriends or boyfriends Victim refused services 3 6.7% *More than one could be selected. through exposure to their own immediate families, and among other families in their communities as well as the tolerance for acts of violence in the community at large. Suggesting that violent behaviors learned could imply that such behaviors can be un-learned and changed as well.
There is a definite need for (1) increased awareness of the problem of domestic violence, (2) actions to prevent such fatalities from occurring and (3) the promotion of healthy non-violent relationships. Less than one-third (31.1%) of the Hawaii cases over the past ten years had documentation of prior violence from medical reports, so this may be an area to further document domestic violence. In Hawaii, there is information on domestic violence assistance available in Filipino languages (Ilocano and Tagalog) because of the disproportionate impact this has had on the Filipino community in Hawaii. This is due both to advocacy in various Filipino communities and public health efforts on different islands to educate the community. Information/translation in other non-English languages is also available by law upon request at all health facilities. In addition, more data on intimate partner/domestic violence injuries from emergency department and hospital discharges is being examined to better ascertain the extent of the problem of domestic violence in Hawaii.