Homicides of American Indians/Alaska Natives — National Violent Death Reporting System, United States, 2003–2018

Problem/Condition Homicide is a leading cause of death for American Indians/Alaska Natives (AI/ANs). Intimate partner violence (IPV) contributes to many homicides, particularly among AI/AN females. This report summarizes data from CDC’s National Violent Death Reporting System (NVDRS) on AI/AN homicides. Results include victim and suspect sex, age group, and race/ethnicity; method of injury; type of location where the homicide occurred; precipitating circumstances (i.e., events that contributed to the homicide); and other selected characteristics. Period Covered 2003–2018. Description of System NVDRS collects data regarding violent deaths obtained from death certificates, coroner/medical examiner reports, and law enforcement reports and links related deaths (e.g., multiple homicides and homicide followed by suicide) into a single incident. This report includes data on AI/AN homicides that were collected from 34 states (Alabama, Alaska, Arizona, California, Colorado, Georgia, Illinois, Indiana, Iowa, Kansas, Kentucky, Louisiana, Maine, Maryland, Massachusetts, Michigan, Minnesota, Missouri, Nebraska, Nevada, New Jersey, New Mexico, New York, North Carolina, Ohio, Oklahoma, Oregon, Pennsylvania, Rhode Island, South Carolina, Utah, Virginia, Washington, and Wisconsin) and the District of Columbia. Results NVDRS collected data on 2,226 homicides of AI/ANs in 34 states and the District of Columbia during 2003–2018. The age-adjusted AI/AN homicide rate was 8.0 per 100,000 population. The homicide rate was three times higher in AI/AN males than females (12.0 versus 3.9), and the median age of AI/AN victims was 32 years (interquartile range: 23–44 years). Approximately half of AI/AN homicide victims lived or were killed in metropolitan areas (48.2% and 52.7%, respectively). A firearm was used in nearly half (48.4%) of homicides and in a higher percentage of homicides of AI/AN males than females (51.5% versus 39.1%). More AI/AN females than males were killed in a house or apartment (61.8% versus 53.7%) or in their own home (47.7% versus 29.0%). Suspects were identified in 82.8% of AI/AN homicides. Most suspects were male (80.1%), and nearly one third (32.1%) of suspects were AI/ANs. For AI/AN male victims, the suspect was most often an acquaintance or friend (26.3%), a person known to the victim but the exact nature of the relationship was unclear (12.3%), or a relative (excluding intimate partners) (10.5%). For AI/AN female victims, the suspect was most often a current or former intimate partner (38.4%), an acquaintance or friend (11.5%), or a person known to the victim but the exact nature of the relationship was unclear (7.9%). A crime precipitated 24.6% of AI/AN homicides (i.e., the homicide occurred as the result of another serious crime). More AI/AN males were victims of homicides due to an argument or conflict than females (54.7% versus 37.3%), whereas more AI/AN females were victims of homicides due to IPV than males (45.0% versus 12.1%). For homicides related to IPV, 87.2% of AI/AN female victims were killed by a current or former intimate partner, whereas approximately half (51.5%) of AI/AN male victims were corollary victims (i.e., victims killed during an IPV-related incident who were not the intimate partners themselves). Interpretation This report provides a detailed summary of NVDRS data on AI/AN homicides during 2003–2018. Interpersonal conflict was a predominant circumstance, with nearly half of all AI/AN homicides precipitated by an argument and for female victims, 45.0% precipitated by IPV. Public Health Action NVDRS provides critical and ongoing data on AI/AN homicides that can be used to identify effective and early intervention strategies for preventing these deaths. When possible, violence prevention efforts should include community-developed, culturally relevant, and evidence-based strategies. These efforts should incorporate traditional native knowledge and solutions, implement and possibly adapt evidence-based IPV and other violence prevention strategies, and consider the influence of historical and larger societal factors that increase the likelihood of violence in AI/AN communities.


Introduction
Missing and Murdered Indigenous People (MMIP)* is an issue that has gained federal attention (1)(2)(3). In 2019, homicide was the fifth leading cause of death for American Indian/Alaska Native (AI/AN) males and the seventh leading cause of death for AI/AN females aged 1-54 years (4). The Presidential Task Force on Missing and Murdered American Indians and Alaska Natives, also known as Operation Lady Justice, was established in 2019 to enhance the operation of the criminal justice system and to address the concerns of AI/AN communities regarding MMIP (1). In 2020, Savanna's Act (2) was passed to increase U.S. governmental agency coordination to reduce violent crimes within tribal lands and against AI/ANs, and the Not Invisible Act (3) directed the U.S. Department of Justice to review, revise, and develop law enforcement and justice protocols to address MMIP. According to the U.S. Department of Justice, the rate of violent crime against AI/ANs (101 violent crimes per 1,000 AI/ANs) was more than twice the rate of the general U.S. resident population (41 per 1,000 persons) during 1992-2002 (5). AI/ANs also experience higher rates of adverse childhood experiences, including child abuse and neglect and family and community violence, than other racial/ethnic groups (6)(7)(8), which increase their risk for other forms of violence, such as homicide. The risk factors for violence among AI/ANs are compounded by multiple and multilayered traumas, including historical (e.g., war and loss of land, language, access to traditional ways, and cultural identity), intergenerational (e.g., child and elder abuse and neglect), and ongoing (e.g., racism and structural inequities) traumas (9)(10)(11)(12).
AI/ANs also have reported high levels of intimate partner violence (IPV) (13), defined as physical or sexual violence, stalking, or psychological harm by a current or former intimate partner or spouse (14). According to National Intimate Partner and Sexual Violence Survey data collected during 2010-2012, 47.5% of non-Hispanic AI/AN women and 40.5% of non-Hispanic AI/AN men have experienced contact sexual violence, physical violence, or stalking by an intimate partner during their lifetime (13). Similar to women of other racial/ethnic groups, approximately half of AI/AN female homicides are related to IPV (15). National Violent Death Reporting System (NVDRS) data during [2003][2004][2005][2006][2007][2008][2009][2010][2011][2012][2013][2014] indicate that IPV contributed to 55.4% of homicides among non-Hispanic AI/AN women (15).
In 2019, an estimated 6.9 million persons, representing approximately 2% of the U.S. population, reported AI/AN * Other terms commonly used might include but are not limited to Missing and Murdered Indigenous Women, Missing and Murdered Indigenous Women, Girls, and Two-Spirit People.
ancestry either alone or in combination with one or more other races (16). Approximately 75% of AI/ANs live in urban, suburban, and rural settings outside of reservations (17). The 10 states with the largest AI/AN populations in 2019 were Arizona, California, Oklahoma, New Mexico, Texas, North Carolina, Alaska, Washington, South Dakota, and New York (17). Currently, there are 574 federally recognized Indian tribes (18) and approximately 100 state-recognized tribes in the United States (17), each representing distinct communities with diverse histories and cultures, and 324 federally recognized American Indian reservations, including federal reservations and off-reservation trust land (16). Tribes are sovereign nations with distinct political communities and territorial boundaries, within which their authority is exclusive, and are subordinate to the federal government but not to state governments (19). The first step in preventing violence against AI/ANs is defining the problem (20). Accurate and comprehensive surveillance data help identify risk factors for violence and can be used to guide culturally relevant public health interventions. Further, when developing prevention programs, it is important to consider the circumstances of AI/AN homicides and how they might differ across tribes and from other populations. Tribal communities have requested studies on AI/AN homicide to guide prevention efforts (21), and a key activity for the Presidential Task Force on Missing and Murdered American Indians and Alaska Natives is to review AI/AN homicide data (1). However, previous studies have been limited by racial misclassification, which can underestimate violence and homicide prevalence in AI/AN populations (22), and the diverse population sizes of AI/AN persons are often too small to provide estimates, which can lead to the "statistical invisibility" of these groups (23). These limitations make it difficult to define fatal interpersonal violence among AI/ANs and, in turn, might restrict attention and resources to prevent these deaths.
This report presents data from NVDRS to better understand the circumstances surrounding AI/AN homicides and to highlight the contextual factors that place AI/ANs at risk for homicide. Because of the high percentage of homicides related to IPV among AI/AN women, this report further examines data on IPV-related AI/AN homicides. This report also recommends community approaches to protect AI/AN persons from homicide and IPV using a violence prevention framework that incorporates complex and contextual knowledge of AI/AN communities (20). The findings in this report can be used by tribes; federal, state, and local governments; public health and health services sectors; the criminal justice system; and victim services to guide violence prevention efforts. US Department of Health and Human Services/Centers for Disease Control and Prevention

Methods National Violent Death Reporting System
NVDRS is an active, state-based surveillance system that collects information from death certificates, coroner/medical examiner reports, and law enforcement reports on the characteristics and circumstances of violent deaths, including homicides (24). NVDRS combines information for each death and links deaths that are related (e.g., multiple homicides and homicide followed by suicide) into a single incident. Trained state-level data abstractors code up to 600 variables in NVDRS using standardized guidance from CDC and enter the variables into an NVDRS web-based system. These variables include characteristics of victims and suspected perpetrators (suspects), incidents (e.g., when and where the incident occurred), weapons that inflicted fatal injuries, and circumstances that directly contributed to the death (e.g., IPV). Data on these precipitating circumstances often originate from investigators' interviews with informants who knew the victim, witnessed the incident, or both.
State participation in NVDRS has expanded over time (25)

Definitions
NVDRS defines homicide as a death resulting from the use of physical force or power, threatened or actual, against another person, group, or community when a preponderance of evidence indicates that the use of force was intentional (26). Homicide was classified by the International Classification of Diseases, 10th Revision cause-of-death codes X85-X99, Y00-Y09, Y87.1, and U01-U02 (27). Victims and suspects were classified as AI/AN if they had ancestries of the original inhabitants of North America who maintained their cultural identification through tribal affiliation or community recognition † (26). Precipitating circumstances are defined as the events that contributed to the infliction of a fatal injury and are reported on the basis of the content of coroner/ medical examiner and law enforcement investigative reports (Appendix) (26). Data abstractors select from a list of potential circumstances and are required to code all circumstances that are known to relate to each incident; therefore, circumstances are not mutually exclusive. If either the coroner/medical examiner report or law enforcement report indicates the presence of a circumstance, then the abstractor endorses the circumstance (e.g., if the law enforcement report indicated that a physical fight between two persons resulted in the death of a victim, then the circumstance variable "physical fight" is endorsed). Certain circumstances are coded to a specific manner of death (e.g., "drug involvement" is collected for homicides); other circumstances are coded across all manners of death (e.g., "argument or conflict" led to the victim's death). If circumstances are unknown (e.g., a body was found in the woods with no other details reported), the data abstractor does not endorse circumstances, and these deaths are then excluded from the denominator for circumstance values.
IPV-related deaths were defined as those involving intimate partner homicide (i.e., victim was an intimate partner or spouse [current, former, or unspecified] of the suspect); corollary victims of IPV-related homicide (i.e., other deaths associated with IPV, including homicides of victims who were not the intimate partner, such as family, friends, others who intervened in IPV, first responders, and bystanders); or homicides precipitated by jealousy or distress over an intimate partner's relationship or suspected relationship with another person. Detailed descriptions of all variables collected by NVDRS are available in the NVDRS Coding Manual (26). † Information on race and ethnicity are recorded as separate items in NVDRS consistent with U.S. Department of Health and Human Services (HHS) and Office of Management and Budget standards for race/ethnicity categorization. HHS guidance on race/ethnicity is available at https://aspe.hhs.gov/basic-report/ hhs-implementation-guidance-data-collection-standards-race-ethnicity-sexprimary-language-and-disability-status.

Analysis
This report summarizes AI/AN homicide data from 34 states and the District of Columbia that participated in NVDRS during 2003-2018 (all available years). Six states (Connecticut, Delaware, Hawaii, New Hampshire, Vermont, and West Virginia) had no AI/AN homicides during 2003-2018 and were not included in this analysis. Therefore, the AI/AN homicide data in this report are from 30 states that collected statewide data, four states that collected data from a subset of counties (Illinois, Pennsylvania, Washington, and California), and the District of Columbia. Analyses were conducted for all AI/AN homicides and also the subset of IPV-related AI/AN homicides. Homicide rates were calculated using intercensal and postcensal bridged-race population estimates compiled by CDC's National Center for Health Statistics and were age-adjusted to the 2010 standard U.S. population (28). Rural-Urban Commuting Area codes were used to classify geographic areas into metropolitan and nonmetropolitan categories. § From all data captured by NVDRS, variables relevant to homicides were selected for analysis. Descriptive analyses of sociodemographic characteristics of victims and suspects (i.e., age, race/ethnicity, education, metropolitan status, and pregnancy status), mechanisms used to inflict fatal injuries, and incident characteristics (i.e., location of injury and victim's relationship to the suspect) were conducted. Categories of precipitating circumstances included victim's mental health and substance use, interpersonal problems and conflict (e.g., IPV and family relationship problem), life stressors (e.g., crisis during previous or upcoming 2 weeks), crime and criminal activity (e.g., drug involvement), and other events of the homicide that were relevant to the death (e.g., victim used a weapon). All descriptive analyses were conducted using SAS (SAS Institute).

Characteristics of Homicide Victims
NVDRS collected data on 2,226 AI/AN homicides (1,681 male victims and 545 female victims) in 34 states (30 states collecting statewide data and a subset of California, Illinois, Pennsylvania, and Washington counties) and the § Zip Code Rural-Urban Commuting Area (RUCA) codes (2010) were used to determine whether decedents lived in nonmetropolitan versus metropolitan areas. RUCA codes measure daily commuting flows, population density, and urbanization levels to classify subcounty level geographic areas. Victim residential Zip codes and Zip codes where injury occurred were dichotomized as "metro" (RUCA codes 1-3) and "nonmetro" (RUCA codes 4-10). Descriptions of the RUCA classification codes 1-10 are available at https://www.ers.usda.gov/dataproducts/rural-urban-commuting-area-codes/documentation.  Table 1). A firearm was used in nearly half (48.4%) of homicides and in a higher percentage of homicides among AI/AN males than females (51.5% versus 39.1%). A sharp instrument was used in 21.0% of homicides. More AI/AN females than males were killed in a house or apartment (61.8% versus 53.7%) or a natural area (e.g., field or river) (9.0% versus 5.4%), whereas more AI/AN males than females were killed on a street or highway (16.2% versus 7.2%). More AI/AN females than males were killed in their own home (47.7% versus 29.0%).

Characteristics of Homicide Suspects
A suspect was identified in 82.8% of AI/AN homicides (81.1% of homicides of males and 87.9% of females) ( Table 2). The age, sex, and race/ethnicity of the suspect were known in 69.6%, 90.8%, and 71.7% of cases, respectively. Nearly one fourth (21.6%) of suspects were aged 25-34 years and 20.1% were aged 18-24 years. Most suspects were male (80.1%). Nearly one third (32.1%) were AI/AN, 39.6% were non-AI/AN, and the race/ethnicity of the remaining 28.3% were unknown. Half (51.8%) of non-AI/AN suspects were non-Hispanic White, 28.4% were non-Hispanic Black, 16.3% were Hispanic (any race except AI/AN), and 3.6% were non-Hispanic Asian or NHOPI. The victim's relationship to the suspect was known in 76.4% of homicides (73.8% of male victims and 83.9% of female victims). For AI/AN male victims, the suspect was most often an acquaintance or friend (26.3%), a person known to the victim but the exact nature of  (2010) were used to determine whether decedents resided in nonmetropolitan versus metropolitan areas. RUCA codes measure daily commuting flows, population density, and urbanization levels to classify subcounty level geographic areas. Victim residential Zip codes were dichotomized as "metro" (RUCA codes 1-3) and "nonmetro" (RUCA codes 4-10).
Descriptions of the RUCA classification codes 1-10 are available at https://www.ers.usda.gov/data-products/rural-urban-commuting-area-codes/documentation. † † Other method includes (in descending order): motor vehicles (e.g., buses, motorcycles, other transport vehicles), fire/burns, drowning, fall, poisoning, intentional neglect, and other (single method). § § Other location includes (in descending order): commercial/retail area, hotel/motel, park/playground/sports or athletic area, bar/nightclub, jail/prison, abandoned house/building/warehouse, synagogue/church/temple, hospital or medical facility, office building, industrial or construction area, farm, preschool/school/ college/ school bus, bridge, supervised residential facility, railroad tracks, and other unspecified location. . § The following sentence can be used as a general guide for interpreting victim-suspect relationship: "The victim is the ____________ of the suspect. " For example, when a parent kills a child, the relationship is "Child" not "Parent" ("The victim is the child of the suspect"). The sentence is a general guide and some relationships might not be captured by this sentence (e.g., other person known to victim or victim was law enforcement officer killed in the line of duty). ¶ Other relationship includes (in descending order): parent, child of suspect's boyfriend/girlfriend (e.g., child killed by mother's boyfriend), rival gang member, intimate partner of suspect's parent (e.g., teenager kills his mother's boyfriend), and victim was a law enforcement officer injured in the line of duty.
the relationship was unclear (12.3%), or a relative (excluding intimate partners [10.5%]). For AI/AN female victims, the suspect was most often a current or former intimate partner (38.4%), an acquaintance or friend (11.5%), or a person known to the victim but the exact nature of the relationship was unclear (7.9%).

Precipitating Circumstances of Homicides
Precipitating circumstances were identified in 83.2% of homicides (Table 3). An argument or conflict precipitated half (50.3%) of AI/AN homicides and preceded a larger percentage of homicides of AI/AN males than females (54.7% versus 37.3%). A crime precipitated 24.6% of homicides (i.e., the homicide occurred as the result of another serious crime), and the crime was in progress at the time of the fatal injury in 66.6% of these incidents. The type of crime most frequently precipitating homicides of AI/AN males was assault or homicide (35.3%), robbery (32.4%), burglary (16.5%), and drug trade (15.3%) and for AI/AN females was assault or homicide (35.8%), rape or sexual assault (25.7%), robbery (19.3%), burglary (9.2%), and  drug trade (9.2%). More homicides of AI/AN females than males were related to IPV (45.0% versus 12.1%). Other common precipitating circumstances included a physical fight between two persons (23.3%), which precipitated more homicides of AI/AN males than females (27.7% versus 10.9%), and drug involvement (10.9%). Homicide victims were reported to have had a problem with substance use, alcohol, or both in 27.8% of cases; 17.8% of victims had a noted problem with substances other than alcohol and 14.1% with alcohol. Although less common, 4.3% of homicide victims were reported to have been currently diagnosed with a mental health problem and 1.2% had a current depressed mood at the time of death; 1.8% of victims were reported to be currently receiving mental health treatment.

Characteristics of IPV-Related
Homicide Victims Nearly half of AI/AN victims of IPV-related homicide lived or were killed in metropolitan areas (44.4% and 46.1%, respectively) ( Table 4). A firearm was the most common mechanism of IPV-related homicide injury overall (42.6%) and for both AI/AN males and females (46.1% and 39.9%, respectively). More IPV-related homicides of AI/AN males than females were perpetrated using a sharp instrument (33.5% versus 16.9%), and more IPV-related homicides of AI/AN females than males were perpetrated using personal weapons (e.g., hands, feet, or fists) (13.6% versus 7.2%); blunt instrument (11.7% versus 6.0%); and hanging, strangulation, or suffocation (9.9% versus 2.4%). Most (72.6%) AI/AN victims of IPV-related homicide were killed in a house or apartment. More AI/AN females than males were killed in a natural area (e.g., field or river) (6.1% versus 2.4%), whereas more AI/AN males than females were killed on a street or highway (12.0% versus 6.6%). More AI/AN females than males were killed in their own home (58.2% versus 47.3%).

Characteristics of IPV-Related Homicide Suspects
A suspect was identified in nearly all (98.9%) IPV-related AI/AN homicides ( Table 5). The age, sex, and race/ethnicity of the suspect were known in 80.6%, 98.9%, and 83.2% of cases, respectively. One fourth (25.5%) of suspects were aged 25-34 years. Most (95.7%) IPV-related homicides of AI/AN females were perpetrated by male suspects compared with 57.6% of IPV-related homicides of AI/AN males. More than one third (35.4%) of suspects were AI/AN, 47.9% were non-AI/AN, and the race/ethnicity of the remaining 16.8% was unknown. Almost two thirds (65.0%) of non-AI/AN suspects were non-Hispanic White, 16.7% were non-Hispanic Black, 12.8% were Hispanic (any race except AI/AN), and 5.6% were non-Hispanic Asian or NHOPI. The victim's relationship to the suspect for a higher proportion of AI/AN female victims than male victims was a current intimate partner (72.0% versus 37.6%), former intimate partner (10.0% versus 4.2%), or intimate partner but the status of the relationship was unclear (i.e., unknown whether a current or former intimate partner; 5.2% versus 1.2%), whereas a higher proportion of AI/AN male victims than female victims were corollary victims of IPV-related homicide (51.5% versus 10.9%). In particular, a higher percentage of suspects for AI/AN male victims than female victims were an acquaintance or friend (21.8% versus 3.3%), a person known to the victim but the exact nature of the relationship was unclear (11.5% versus 3.3%), or other nonintimate partner (18.2% versus 4.3%).

Circumstances of IPV-Related Homicide
All but two of the 380 IPV-related homicides (99.5%) had known circumstance information (Table 6). Nearly half (45.8%) of IPV-related homicides were precipitated by an argument or conflict. Jealousy over an actual or perceived relationship precipitated more IPV-related homicides of AI/AN males than females (30.5% versus 13.7%). Jealousy also was more common in IPV-related homicides in which the suspect killed a nonintimate partner (i.e., corollary victim) versus an intimate partner, particularly for male victims (45.8% versus 9.9% for male victims and 17.2% versus 13.0% for female victims, respectively). A crime precipitated 15.3% of IPV-related homicides (i.e., the homicide occurred as the result of another serious crime), and the crime was in progress at the time of the fatal injury in 56.9% of these incidents. The type of crime most frequently precipitating IPV-related homicides of AI/AN males was assault or homicide  RUCA codes measure daily commuting flows, population density, and urbanization levels to classify sub-county level geographic areas. Victim residential Zip codes were dichotomized as "metro" (RUCA codes 1-3) and "nonmetro" (RUCA codes 4-10). Descriptions of the RUCA classification codes 1-10 are available at https://www.ers.usda.gov/data-products/rural-urban-commuting-area-codes/documentation. § § Other method includes (in descending order): motor vehicles (e.g., buses, motorcycles, other transport vehicles), fire/burns, poisoning, fall, and other (single method). ¶ ¶ Other location includes (in descending order): hotel/motel, motor vehicle, parking lot/public garage/public transport, park/playground/sports or athletic area, commercial/retail area, synagogue/church/temple, hospital or medical facility, office building, and other unspecified location. 1%]). ¶ The following sentence can be used as a general guide for interpreting victim-suspect relationship: "The victim is the ____________ of the suspect. " For example, when a parent kills a child, the relationship is "Child" not "Parent" ("The victim is the child of the suspect"). The sentence is a general guide and some relationships might not be captured by this sentence (e.g., other person known to victim or victim was a law enforcement officer killed in the line of duty). ** Other relationship includes (in descending order): other relative, child, child of suspect's boyfriend/girlfriend (e.g., child killed by mother's boyfriend), stranger, parent, and intimate partner of suspect's parent (e.g., teenager kills his mother's boyfriend).

TABLE 5. Number and percentage of American Indian/Alaska Native intimate partner violence-related homicides,* by victim's sex, selected characteristics of homicide suspects, and the victim's relationship to the suspect -National Violent Death Reporting System, † 2003-2018
(68.8%), burglary (12.5%), rape or sexual assault (9.4%), and robbery (9.4%) and for females was assault or homicide (46.2%), rape or sexual assault (34.6%), arson (7.7%), and burglary (7.7%). More IPV-related homicides of AI/AN males than females were precipitated by a physical fight between two persons (34.2% versus 11.2%). More AI/AN female than male victims of IPV-related homicide were known to have experienced interpersonal violence during the month preceding their death (15.9% versus 4.5%).
IPV-related homicide victims were reported to have had a problem with substance use, alcohol, or both in 25.9% of cases; 16.5% of victims had a noted problem with substances other than alcohol and 16.1% with alcohol. Although less common, 4.3% of IPV-related homicide victims were reported to have a current diagnosis of a mental health problem, and 2.4% with a current depressed mood at the time of death; 1.6% of victims were reported to currently be receiving mental health treatment.

Discussion
MMIP is an issue of urgent concern as indicated by the formation of the Presidential Task Force on Missing and Murdered American Indians and Alaska Natives and the passage and signing of Savanna's Act and the Not Invisible Act (1-3). NVDRS data can be used to characterize and monitor AI/AN homicides and identify effective and early intervention strategies for preventing such deaths. This report presents the most comprehensive analysis to date using NVDRS data to examine homicides of AI/ANs and the circumstances surrounding these deaths. Consistent with the homicide rates for other racial/ethnic groups (25), the homicide rate for AI/AN males was higher than the rate for females. A firearm was used in nearly half of AI/AN homicides, and one fourth of homicides were precipitated by another serious crime (e.g., robbery and sexual assault). More than half of AI/AN homicides occurred in a residence, and one third occurred in the victim's home. Interpersonal conflict was a predominant circumstance, with nearly half of AI/AN homicides precipitated by an argument; for female victims, 45.0% were precipitated by IPV. Findings from this report can be used to guide the work of the Presidential Task Force on Missing and Murdered American Indians and Alaska Natives and others to address AI/AN homicide (1).
NVDRS programs have used their local Violent Death Reporting System (VDRS) data to examine AI/AN homicides. For example, Arizona VDRS examined AI/AN homicides in their state during 2015-2017 (3.1% of all homicides) and found that the homicide rates for AI/ANs were more than double the rates for non-AI/ANs among both males (20.0 versus 8.7 per 100,000 population) and females (5.8 versus 8.7) (29). Further examination revealed that the characteristics and circumstances of AI/AN victims and homicides differed from those of non-AI/ANs. Notably, AI/AN homicide victims had completed substantially less education than non-AI/AN victims; 9% of AI/AN male victims and few, if any, AI/AN female victims had earned, at minimum, some college credit or degree compared with 22.8% of non-AI/AN male victims and 44.8% of non-AI/AN female victims. AI/AN male and female homicide victims were more likely than non-AI/AN male and female victims to never have been married (for males, 88.2% versus 59.7%; for females, 63.2% versus 34.6%, respectively). Female victims, both AI/AN (31.3%) and non-AI/AN (45.0%), were at significantly greater risk for homicide by current or former intimate partners than their male counterparts (fewer than 5 victims and 3.9%, respectively). AI/AN male and female homicide victims were significantly less likely than non-AI/AN victims to have been killed with a firearm (for males, 42.9% versus 73.9%; for females, 40.9% versus 65.4%, respectively). Oklahoma VDRS examined AI/AN homicides in their state during 2004-2008 and found that the overall homicide rate among AI/ANs was 20% higher than the overall rate among non-AI/ANs, and the homicide rate among AI/ANs aged 25-44 years was nearly double that of non-AI/ANs of the same age group (30). Arguments and IPV were the leading circumstances of homicide among AI/AN victims (49% and 22%, respectively). These details from examination of local VDRS data increase the knowledge base about the circumstances associated with AI/AN homicide and can assist local public health authorities and their partners in developing and guiding effective, datadriven approaches to violence prevention.
From 2000 to 2010, the AI/AN population increased almost twice as fast as the total U.S. population (31), with the percentage of AI/AN persons aged <18 years higher than the percentage of those aged <18 years for the total U.S. population (29.0% versus 21.9%) (32). AI/AN youths, particularly young adult males, experience disproportionate rates of violent injury and homicide (7,22,25). Socioeconomic factors (e.g., lack of economic opportunity, income inequality, and poverty) contribute to violence and have been associated with higher rates of homicide (33). According to the 2000 U.S. Census, more AI/ANs lived below the federal poverty level than those from all other racial/ethnic groups (25.7% versus 12.4%) (34). Thirteen percent of AI/AN males aged ≥16 years were unemployed compared with 5.7% of males aged ≥16 years in all other racial/ethnic groups; unemployment for AI/AN females aged ≥16 years was 11.7% compared with 5.8% of females in all other racial/ethnic groups (34). In 2019, AI/ANs were less likely to have high school diplomas or higher-level education than non-Hispanic Whites (84.4% versus 93.3%) (17). To improve outcomes and reduce violence in AI/AN communities, exploring economic opportunity and addressing other social determinants of health, such as education, should be considered (35).
Several findings are relevant to the understanding of AI/AN homicides and have implications for violence prevention efforts. Approximately half of AI/AN homicide victims lived or were killed in metropolitan areas. Misperceptions exist about where AI/ANs live and how they access resources (20). Although more than half of AI/ANs lived in rural areas in 1970, approximately 70% now live in urban areas, which affects their access to health care and other services (36,37). AI/AN populations face persistent disparities in health and health care, including high uninsured rates, significant barriers to obtaining care, and poor health status (38). AI/ANs who are members of federally recognized tribes and some descendants have legal rights to and are eligible for health care from the federal government through the Indian Health Service (IHS) (34,39). However, IHS has facilities in only 35 states, primarily on or near rural reservation lands and in few cities with large AI/AN populations (40), and IHS provides health services to only 2.6 million of the 6.9 million persons reporting AI/AN ancestry in the United States (16,17). In addition, the overall IHS budget meets just over half of the demonstrated health care needs of the eligible AI/AN population (41). IHS focuses on providing primary health care at no cost to patients, with limited specialty care and inpatient services, which might not meet the needs of victims of violence or primary prevention programs working to address violence as a public health problem (40,41). AI/ANs in rural areas might experience further challenges related to social isolation and cultural and transportation barriers (42) that hinder their ability to access health care and other resources (43). Thus, violence prevention efforts in health care-based settings, such as IHS, might be limited, and community approaches considering the needs of AI/ANs in urban versus rural settings are needed to reduce violence and homicide in these communities.
Interpersonal conflicts, such as arguments, IPV, and problems with family members and other nonintimate partners, were prominent circumstances of AI/AN homicides, most of which were perpetrated by male suspects and suspects the victim knew. AI/AN homicide victims are commonly perceived to be killed by non-AI/AN suspects, particularly White males (44). Information about perpetrators of violence against AI/ANs are frequently based on a 1999 U.S. Department of Justice report in which 60% of American Indian victims of violent crime described a White perpetrator (44). However, the findings in this report indicate that nearly one third of AI/AN victims were killed by AI/AN suspects and approximately 40% were killed by a racially diverse group of non-AI/AN suspects. Further research is needed to understand the interpersonal conflicts and racial/ethnic dynamics between victims and suspects that contribute to AI/AN homicides.

IPV-Related Homicide
The results of this study provide further evidence that violence against AI/AN women is an issue of urgent concern (43). IPV was a contributing factor in nearly half of the homicides of AI/AN females. Rape or sexual assault occurred in nearly one third of IPV-related homicides precipitated by another serious crime. Data from the National Intimate Partner and Sexual Violence Survey and National Crime Victimization Survey indicate that AI/AN women experience higher rates of rape and sexual assault, physical assault, and stalking than women of other racial/ethnic groups (13,43). However, similar to women of other racial/ethnic groups, AI/AN women are likely to be victimized by someone they know (43). The findings in this report indicate that approximately 40% of female victims were killed by a current or former intimate partner and 12% by an acquaintance or friend. Results also indicate that leaving an intimate partner relationship remains dangerous for some victims (45). Ten percent of AI/AN female victims and 4.2% of male victims were killed by a former intimate partner.
The findings also indicate that some AI/ANs were corollary victims of IPV-related homicide, particularly among male victims (46). Such victims are those who were affected by IPV but were not the intimate partners themselves. These can include bystanders to an IPV incident, children of the IPV victim, family members or friends assisting the IPV victim, or persons involved in a true or perceived romantic relationship with the IPV victim (46). A large proportion of female victims of IPV-related homicide were killed by their current or former intimate partners. In contrast, approximately half of the male victims killed during an IPV-related incident were corollary victims. This finding indicates that many AI/AN male victims of IPV-related homicide might reflect a specific category of corollary victims, referred to as "other intimate partner involvement" (46). These are victims who were connected to the suspect through a mutual intimate partner, either currently or in the past, such as new boyfriends and partners killed by a former partner (46). This is also supported by findings related to jealousy over an actual or perceived relationship with a romantic rival. Nearly one third of IPV-related homicides involving male victims were precipitated by jealousy compared with approximately 14% of female victims. These findings indicate that the impact of IPV can extend beyond the couple involved, and that IPV prevention and intervention strategies that consider family, friends, and others who might be exposed to IPV can be most helpful in preventing IPV-related homicide.
The IPV-related homicides identified in this report, particularly among women, have several implications for the criminal justice system. However, AI/AN women experiencing IPV face unique challenges when seeking resources. Complicated jurisdictional issues might lead to inadequate or delayed responses by investigative authorities, and legal constraints on tribal sovereignty might limit tribal authority to prosecute offenders (43). Further, indigenous views of law differ from the U.S. legal system (43). Traditional AI/AN criminal justice systems emphasize communal values that restore harmony within the community (47,48), whereas the U.S. legal system typically emphasizes deterrence in the form of punishment and focuses on individual cases (43). This dissonance can create a cultural barrier that impedes healing and resolution. AI/AN women living on tribal lands might experience additional challenges of social isolation, inadequate access to resources, and cultural barriers when seeking assistance outside their community (43).
The Violence Against Women Act (VAWA) and its reauthorizations provided legislation for protecting female victims of violence and dedicated funding to combat and respond to violence against women (49,50). VAWA included initiatives to strengthen law enforcement in preventing violence against women; allocated funds to expedite the processing of rape kits and other victim services; and authorized the administering of grants to several programs designed to reduce domestic violence, dating violence, sexual assault, and stalking (49,50). VAWA's reauthorization in 2013 addressed tribal sovereignty and jurisdictional issues related to violence against AI/ANs occurring on tribal lands. In particular, VAWA's reauthorization in 2013 recognized the inherent ability of tribes to prosecute non-AI/AN domestic violence and stalking perpetrators and violators of orders of protection after certain criteria were met by tribes and the alleged perpetrators (50). In 2014, the Pascua Yaqui Tribe became the first tribe to prosecute a non-AI/AN domestic violence perpetrator in Pascua Yaqui Tribal Court (51). Tribal sovereignty and jurisdictional issues related to violence against AI/ANs and cross-sectoral collaborations (e.g., among public health professionals, law enforcement, and direct service providers) remain key focus areas for those working to address AI/AN homicide.

Implications for Prevention
Violence prevention efforts might have greater impact if they include trauma-informed strategies, provide accessible behavioral health services, and use strengths-based approaches to promote resilience and other protective factors that are inherent in AI/AN communities (21). AI/AN populations possess community and cultural assets that protect against violence, such as community mindedness, connection to tribal leaders and elders, participation in tribal ceremonies, and spirituality (21,52,53). Native teachings and traditions help persons develop a sense of identity, and generational sharing of knowledge through songs and storytelling contribute to a sense of connectedness and community resilience (52,53). Such cultural practices are protective against violence among AI/AN youths (52) and can be integrated into violence prevention strategies in AI/AN communities.
CDC developed technical packages that summarize the best available evidence for preventing different types of violence (54). One critical strategy for preventing interpersonal violence is to teach safe and healthy relationship skills through social-emotional learning programs for youth (55). For example, "Safe Dates," a school-based dating violence prevention program, has helped reduce teen dating violence among racial/ethnic minority adolescents (56) and could be adapted for use with AI/AN youths. "Fourth R: Strategies for Healthy Teen Relationships" (https://youthrelationships.org) is another evidence-based, social-emotional learning program that has been successfully adapted and implemented in AI/AN communities (57). Tribal communities also have developed similar programs that leverage their unique cultural and community assets. "Discovery Dating" is a strengths-based, healthy relationships curriculum rooted in Native American values that increased feelings of personal agency among Native American middle school youths (58) and reduced rates of teen pregnancy (59). Further research is needed to evaluate specific impacts on violence outcomes; however, programs like "Discovery Dating" represent culturally relevant interventions that can be tailored for tribal communities and are consistent with strategies identified in CDC's technical packages.
Community-led action and attention on missing and murdered indigenous women and girls have led discussions and policy actions around AI/AN homicide (1). However, homicides of AI/AN males, particularly among youths and young adults, contribute to many AI/AN homicides. AI/AN communities can strengthen, evaluate, and adapt existing programs created specifically to address violence among AI/AN males. For example, "Boys Run I toowú klatseen" (https:// boysrun.org) aims to build healthy relationships by promoting traditional tribal values and equitable gender norms among third-and fifth-grade Alaska Native boys (60). Multnomah County in Oregon partners with the Native American Youth Association to implement a culturally adapted version of "Coaching Boys Into Men" (https://www.coachescorner.org), an evidence-based program engaging male youths that builds knowledge and prosocial attitudes about healthy masculinity, healthy relationships, nonviolent problem solving, and being an active bystander (61).
This report identified circumstances of IPV-related AI/AN homicides that could guide other evidence-based approaches, such as bystander programs and IPV lethality risk assessments, that can be adapted for use in AI/AN communities. Approximately 15% of AI/AN female victims and 4.5% of AI/AN male victims of IPV-related homicide experienced some form of violence in the preceding month, which could have provided opportunities to intervene and prevented escalating violence that resulted in the homicide. Bystander programs such as "Green Dot" teach participants how to recognize situations or behaviors that might become violent and how to intervene safely and effectively in IPV and other violence (62). IPV lethality risk assessments are used by law enforcement officers responding to the scene of a domestic violence incident to facilitate immediate safety planning and to connect IPV victims with other services, such as crisis intervention and counseling, housing, medical and legal advocacy, and other community resources (63). These assessments help identify victims at risk for future violence and might have promise in preventing intimate partner homicide (63).
Other strategies to prevent violence include reducing exposure to community-level risks and strengthening economic supports. Community-and societal-level programs, such as "Crime Prevention Through Environmental Design" (64), business improvement districts (65,66), and alcohol-related policies, promote protective community environments and reduce the risk for violence (33). The earned income tax credit enhances household financial security by raising family income while incentivizing work and counterbalancing the costs of child rearing, which in turn improve home environments and encourage healthy development (67,68). A quasi-experimental study of the effects of cash supplements from casino revenue on an Eastern Cherokee reservation found that an additional $4,000 per year for the poorest American Indian households increased educational attainment by 1 year at age 21 years and reduced the incidence of criminality by 22% at ages 16 and 17 years (69). The strategies described indicate that a multidisciplinary approach that cuts across different sectors and addresses the unique values and cultural needs of AI/AN communities might be warranted to reduce violence and AI/AN homicide.

Limitations
The findings in this report are subject to at least six limitations. First, racial misclassification of AI/AN decedents on death certificates and in other investigative reports is a known concern (22,70) and might have contributed to an underestimate of AI/AN homicides. Racial misclassification is often caused by inaccurate information submitted by health care personnel or by death certifiers basing classification on appearance or surname alone (22). This is compounded by jurisdictional issues in death investigations that occur on tribal lands because information from these investigative reports might not be captured in federal systems, such as NVDRS, because of a lack of routine data sharing (43). Each tribe has inherent legal and political authority to govern itself (i.e., tribal sovereignty) (19) and might not interact with the investigative and public health entities that typically provide data for NVDRS.
Second, NVDRS does not collect information on tribal affiliation, residence on reservations, or federally recognized tribally governed lands or whether homicides occurred on tribal lands. Linking with tribal registries or other tribal data sources would enable more reliable collection of data. Third, the rates and characteristics of homicides of different tribes could not be assessed; therefore, the findings in this report might not be generalizable across all AI/AN communities. Future studies might consider examining differences across tribes, given the diverse cultural practices and beliefs that might affect health outcomes in AI/AN communities (36). Fourth, NVDRS data are not nationally representative, and not all states joined the system at the same time (25). Therefore, it was not possible to use NVDRS data to report regional disparities in homicide rates, which have been demonstrated by other studies of AI/AN homicide (22).
Fifth, the availability and completeness of the data are dependent on successful partnerships among local VDRS programs and their partners in vital records, medical examiner and coroner offices, and law enforcement (24). NVDRS data might be limited or incomplete for areas in which these datasharing relations are not fully developed. Further, some state VDRS programs do not receive detailed law enforcement reports until cases are adjudicated (24), which might result in an underestimate of circumstance information. Child abuse and neglect (8,11,71) and human trafficking (72)(73)(74) are known issues of concern in AI/AN populations but were likely underreported in NVDRS. In addition, medical, mental health, and substance use information are not captured directly from medical records and might not be systematically collected for victims unless they were directly related to the homicide. Therefore, the completeness and accuracy of this information is limited.
Finally, NVDRS collects limited information about suspects. Data on certain suspect characteristics, such as alcohol and other substance use, were recently added to the system but were not available for this report (26).

Conclusion
This report presents a detailed examination of the circumstances surrounding AI/AN homicides and highlights the role of IPV in these deaths. NVDRS data can be used to characterize and monitor AI/AN homicides and identify effective and early intervention strategies for preventing such deaths. Future studies of NVDRS data could show unique risk factors for homicide in the AI/AN population, changing trends, and regional variations in AI/AN homicide. An integrated primary prevention and health promotion response that coordinates across tribes, the federal government, public health and health services sectors, the criminal justice system, and victim services is important in addressing violence among AI/AN populations. To support prevention efforts and to fully address the complexities of MMIP and, particularly, AI/AN homicide, AI/AN communities might need collective healing of historic and intergenerational traumas and resolution of structural inequities (20). AI/AN communities are rich in native knowledge, teachings, and protective factors, which can be used to further violence prevention efforts (20). When possible, violence prevention efforts should include community-developed, culturally relevant, and evidencebased strategies; incorporate traditional native knowledge and solutions; and consider the influence of historical and larger societal factors that might increase the likelihood of violence in AI/AN communities.