Article Text
PDF
Abstract
Background Missing and Murdered Indigenous People is a historic and contemporary issue that has gained national attention. In 2021, homicide was the eighth leading cause of death among American Indian/Alaska Native (AIAN) persons aged between 1 and 54 years old, and homicide is the sixth leading cause of death among all AIAN males aged 1–54 years old.
Aim These data will build knowledge around AIAN homicides and to identify circumstances that can aid in comprehensive Missing and Murdered Indigenous People prevention efforts.
Methods AIAN homicide data came from Centers for Disease Control and Prevention’s National Violent Death Reporting System, a state/jurisdiction-based surveillance system that collects detailed information about characteristics and circumstances of violent deaths. We examined data from 2003 to 2020 (all available years) from participating states/jurisdictions. We also assessed sociodemographic characteristics of victims and suspects, incident characteristics and differences across dichotomised urban/rural status. The study was conducted in 2022.
Results The National Violent Death Reporting System provided data on 2959 AIAN homicides from 2003 to 2020 (54.2% urban and 45.8% rural). Significant differences based on the two locations included type of weapon used, the location of the injury, race of the primary suspect, the victim’s relationship to the suspect and select circumstances precipitating the homicide including crimes precipitating the homicide and homicides stemming from intimate partner violence.
Outcomes These findings provide crucial information to strengthen public health efforts for prevention.
- Rural
- Urban
- Indigenous
- Epidemiology
- Violence
Data availability statement
Data are available upon reasonable request. NVDRS data used are available via approved application to use the NVDRS Restricted Access Database (RAD). More details are available online: https://www.cdc.gov/nvdrs/about/nvdrs-data-access.html.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Missing or Murdered Indigenous People is a preventable and historic public health issue that has gained national attention. Comprehensive efforts to prevent violence can incorporate contextual understanding while strengthening existing protective factors such as cultural and community assets that include Native traditions and language while decreasing risk factors.
WHAT THIS STUDY ADDS
This study provides characteristics of American Indian/Alaska Native homicides in urban versus rural settings and helps to provide historical and contemporary contextualisation.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Findings regarding urban and rural American Indian/Alaska Native homicides provide critical information to tailor and strengthen public health endeavours to prevent Missing and Murdered Indigenous People. Comprehensive solutions to prevent violence may include an integrated primary prevention response that involves tribes, the federal government, the public health sector, the criminal justice system and victim services.
Introduction
Missing or Murdered Indigenous People (MMIP) (MMIP includes Missing or Murdered Indigenous Women; Missing or Murdered Indigenous Women and Girls; Missing or Murdered Indigenous Women, Girls, and Two Spirit Individuals; Missing or Murdered Indigenous Relatives; and other acronyms used) is a preventable and historic national public health issue. American Indian and Alaska Native (AIAN) persons are at increased risk of violence compared with the general US population. The Department of Justice reported that violent crime against AIAN people was more than twice the rate of the general population between 1992 and 2002.1 More recent violent crime reports have aggregated rates of violent crime against AIAN people into ‘other’ racial categories. For example, from 2008 to 2021, rates of violent crime against AIAN people were aggregated with rates of multiracial people. This aggregated rate remained more than twice the rate of the general population.2 In 2021, homicide was the eighth leading cause of death among AIAN persons aged between 1 and 54 years old.3 Understanding the circumstances surrounding AIAN homicides is crucial so that data may inform prevention approaches following the public health model.4
Preventing violence among AIAN persons will require indigenous, cultural and traditional knowledge.5 Comprehensive efforts can incorporate contextual understanding while strengthening existing protective factors such as cultural and community assets (eg, Native traditions and language) while decreasing risk factors.6 AIAN people can experience violence risk factors (eg, social determinants of health factors) that are compounded by historic trauma (eg, federal government’s ethnocidal and genocidal policies (see Sand Creek Massacre National Historic Site and other resources on ethnocidal and genocidal policies and military actions7–11) and intergenerational and ongoing trauma.6 More research is needed around the circumstances and characteristics of AIAN homicides to understand impacts.6
Prior work by Petrosky et al examined differences between homicides of AIAN females and males using data from the Centers for Disease Control and Prevention’s National Violent Death Reporting System (NVDRS).12 Findings showed that more AIAN females were victims of intimate partner violence (IPV)-related homicide compared with AIAN males, and that more AIAN male homicides were precipitated by an argument or conflict compared with female homicides. Disaggregating data and assessing differences across other demographics or circumstances can help elucidate factors that drive high rates of AIAN homicides.
Although most AIAN people reside in urban areas, 40% live in non-metropolitan areas13 compared with 14% of the general public.14 Existing research has not explored urban and rural homicide characteristics of AIAN people or the complexities and impacts related to historical events and trauma such as forced relocation.6 Researching urban and rural homicides of AIAN persons may help to inform prevention interventions in different geographies. We can use the public health model4 as a critical first step, such as using surveillance data to define the problem, so that public health practitioners and partners may use information more effectively.
Nearly half (48.4%) of all AIAN homicides from 2003 to 2018 involved the use of a firearm (51.5% of male victims, 39.1% of female victims).12 Non-Hispanic AIAN firearm homicide rates increased by 27% (6.4 to 8.1 per 100 000) from 2019 to 2020.15 National Vital Statistics System data from 2021 show that age-adjusted rates of firearm homicides of AIAN persons were more than twice as high in non-metro areas compared with metro areas.3 While there is little research on the factors driving this difference, one study used 2002–2019 data from the National Survey on Drug Use and Health to examine adolescent firearm possession which showed that adolescents in rural areas are more likely to report carrying a handgun than their urban counterparts.16 Existing research has not examined potential differences in the characteristics of firearms used in urban versus rural AIAN homicides, such as firearm type. This information may shed light on firearm access in AIAN homicides that occur in urban versus rural areas.
This report extends prior work12 by adding data from 2019 and 2020 and compares characteristics of homicides of AIAN persons that occur in urban and rural areas, including detailed circumstances and weapons used. It seeks to build knowledge around AIAN homicides and to identify important characteristics that can aid in comprehensive MMIP prevention.
Methods
National Violent Death Reporting System
NVDRS is an active state/jurisdiction-based surveillance system that collects information from death certificates and reports from coroners/medical examiners (C/ME) and law enforcement (LE) on the characteristics and circumstances of violent deaths, including homicides.17 It combines information from the three sources for each death and links those that are related (eg, multiple homicides, homicide followed by suicide) into a single incident. Along with death certificate data, trained state-level data abstractors review LE and C/ME reports and code up to 600 variables in NVDRS using standardised guidance from CDC and enter the data into the NVDRS Web-based system. These variables include characteristics of victims and suspected perpetrators (suspects), incidents (eg, when and where the incident occurred), manner of death (eg, suicide, homicide), weapons that inflicted fatal injuries and circumstances that were reported or perceived in the investigative reports (ie, C/ME or LE) as being related to the violent death. Data on these precipitating circumstances often originate from investigator interviews with informants who knew the victim, witnessed the incident, or both.
State participation in NVDRS has expanded over time.18 NVDRS now collects data in all 50 US states, the District of Columbia and Puerto Rico. Florida is the only state without reported data as of 2020 (NVDRS data have been collected in Maryland, Massachusetts, New Jersey, Oregon, South Carolina and Virginia since 2003; Alaska, Colorado, Georgia, North Carolina, Oklahoma, Rhode Island and Wisconsin since 2004; Kentucky, New Mexico and Utah since 2005; Ohio since 2011; and Michigan since 2014. Data have also been collected in Arizona, Connecticut, Hawaii, Illinois, Indiana, Iowa, Kansas, Maine, Minnesota, New Hampshire, New York, Pennsylvania, Vermont and Washington since 2015; Alabama, California, Delaware, District of Columbia, Louisiana, Missouri, Nebraska, Nevada and West Virginia since 2017, Montana, North Dakota and Wyoming since 2019, and Arkansas, Idaho, Mississippi, South Dakota, Tennessee and Texas since 2020. Three states reported data on a subset of counties that represented at least 80% of violent deaths in their state during 2016–2018 (Illinois and Pennsylvania) and 2016–2017 (Washington; statewide since 2018). California reported 2017 data from 4 counties, 2018 data from 21 counties, 2019 data from 30 counties and 2020 data from 35 counties. Texas reported 2020 data from four counties. Data from Illinois, Pennsylvania, Washington, California and Texas are not representative of all violent deaths that occurred because <100% of violent deaths statewide were reported. Hawaii data are included for years 2015–2016 and 2019. New York data are included for 2015–2018 and 2020. Years not reported for Hawaii, New York and Florida are omitted from the data because the threshold for inclusion in the national dataset (>50% of cases must have circumstance information from the C/ME, or LE report) was not met).
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.18 Homicide is classified by the International Classification of Diseases, 10th Revision cause-of-death codes X85–X99, Y00–Y09, Y87.1 and U01–U0219 provided on death certificates and on data found in LE and C/ME reports. Victims and suspects were classified as AIAN if they had ancestries of the original inhabitants of North America who maintained their cultural identification through tribal affiliation or community recognition (information on race and ethnicity are recorded as separate items in NVDRS consistent with the U.S. Department of Health and Human Services (HHS) and Office of Management and Budget standards for race/ethnicity categorisation. HHS guidance on race/ethnicity is available at https://aspe.hhs.gov/basic-report/hhs-implementation-guidance-data-collection-standards-race-ethnicity-sex-primary-language-and-disability-status).18
Data abstractors select from a list of potential circumstances20 and are required to code all that are known to relate to each victim in an incident; therefore, circumstances are not mutually exclusive. Certain circumstances are coded for only a specific manner of death (eg, ‘gang-related’ is collected for homicides); other circumstances are coded across all manners of death (eg, ‘argument or conflict led to the victim’s death’). The data abstractor does not endorse any unknown circumstances (eg, a body was found in the woods with no other details reported); deaths with no known circumstances are excluded from the denominator for circumstance values. Detailed descriptions of all variables collected by NVDRS are available in the NVDRS Coding Manual.18
Urban or rural injury location was defined using rural–urban commuting area (RUCA) codes which classify geographical areas into metropolitan and non-metropolitan areas based on the proportion of commuters travelling to urban centres.21 Injury location zip codes were matched to their corresponding RUCA code. A location was classified as ‘urban’ if the RUCA code was between 1 and 3 (metropolitan area core, high commuting and low commuting) and was classified as ‘rural’ if the RUCA code was between 4 and 10 (micropolitan, small town and rural areas).
Analysis
This report summarises AIAN homicide data from 49 states and the District of Columbia that participated in NVDRS during 2003–2020 (all available years; no Florida data available at the time of report). Of the 49 states and Washington, DC, five states (Arkansas, Connecticut, Delaware, New Hampshire and West Virginia) had no reported AIAN homicides during 2003–2020 and thus were not part of this analysis, resulting in inclusion of 44 states and Washington, DC. This report includes reported AIAN homicide data from 39 states that collected statewide data, 5 states that collected data from a subset of counties (California, Illinois, Pennsylvania, Washington and Texas) and the District of Columbia.
We analysed all AIAN homicides and variables relevant to homicides from all data captured by NVDRS. We conducted descriptive analyses of sociodemographic characteristics of victims and suspects (ie, age, sex, ethnicity, education and victim’s relationship to the suspect), and incident characteristics (ie, the mechanism or method used to inflict fatal injuries, location of injury and whether the victim was injured at their home). Categories of precipitating circumstances included victim’s mental health and substance use, interpersonal problems and conflict (eg, IPV, family relationship problems), life stressors (eg, crisis during previous or upcoming 2 weeks), crime and criminal activity (eg, drug involvement), and other events of the homicide that were relevant to the death (eg, victim used a weapon). We used χ2 tests of independence and Fisher’s exact tests to determine differences between demographics and circumstances by urban/rural status. All analyses were conducted using R.
Results
NVDRS collected data on 2959 AIAN homicides between 2003 and 2020. Eighty-eight per cent (87.7%, n=2596) of these homicides had zip codes for injury location that allowed for determining urban/rural status. The remaining 12% (12.3%, n=363) were excluded from analyses due to unknown urban/rural status. Fifty-four per cent (54.2% or 1407) of the 2596 homicides occurred in urban areas and 1189 (45.8%) occurred in rural areas. Most decedents in both areas were aged between 18 and 44 (68.7%, n=966 in urban areas; 62.6%, n=745 in rural areas). (See table 1 for characteristics of AIAN homicide victims with data for the urban/rural status of injury location.)
Number and percentage of American Indian/Alaska Native homicide victims, by urban and rural areas and selected demographic and incident characteristics—National Violent Death Reporting System, 2003–2020 (n=2596)
Eighty-seven per cent (87.4%, n=2270) of all AIAN homicides with urban/rural status for the location of the injury had details regarding the suspected perpetrator of the homicide (87.6%, n=1233 urban; 87.2% rural, n=1037; hereafter ‘suspect’; see table 2). Most known homicide suspects were aged between 18 and 44 (51.0%, n=1144) and 75.5% (n=1708) were male. Eighty-four per cent (84.2%, n=2186) of all AIAN homicides in NVDRS with urban/rural status of injury location data had known circumstances that preceded the homicide (85.4%, n=1202 in urban areas, 82.7%, n=983 in rural areas; table 3).
Number and percentage of American Indian/Alaska Native (AIAN) homicides, by urban and rural areas, selected demographics of homicide suspects, and the victim’s relationship to the suspect—National Violent Death Reporting System, 2003–2020 (n=2270)
Number and percentage of American Indian/Alaska Native (AIAN) homicides, by urban and rural areas and circumstances of the homicide—National Violent Death Reporting System, 2003–2020 (n=2185)
Urban homicides
Most AIAN homicide victims in urban settings were male (76%, n=1069; see table 1) and aged between 18 and 44 (68.7%, n=966). One in nine victims identified as Hispanic/Latino (11.8%, n=166). Most homicides in urban areas involved the use of a firearm (55.7%, n=784). Of these, more than half involved a handgun (57.1%, n=448) as opposed to a rifle or shotgun. More than half (51.2%, n=720) of all homicides occurred in and around a house/apartment (or other rooming house). Just under one-third (31.6%, n=444) of homicides occurred at a victim’s own home. From table 2, primary suspects were racially diverse: of all non-AIAN suspects: 41.1% (n=235) were white, non-Hispanic; 34.8% (n=199) black, non-Hispanic; and 20.8% (n=119) non-AIAN Hispanic/Latino. Additionally, 27.9% (n=307) of primary suspects’ race/ethnic identity was unknown, and 20.2% (n=222) of known primary suspects identified as AIAN. This racial breakdown was significantly different from rural homicide suspects (χ2=136.35, p<0.001). More than half of all urban homicides (59.5%, n=725) were perpetrated by someone the victim knew.
A greater proportion of urban homicides (28.0%, n=337) were precipitated by another crime when compared with rural homicides (20.1%, n=198; χ2=17.801, p<0.001); see table 3. About half (47.0%, n=565) of urban homicides were precipitated by an argument or conflict and more than one in five (21.3%, n=183) were precipitated by a physical fight. The proportion of homicides with drug involvement noted as a circumstance in urban areas (13.3%, n=160) was more than double that in rural areas (6.4%, n=63) (χ2=27.363, p<0.001) and more gang-related homicides occurred in urban areas (8.8%, n=106) than in rural areas (3.1%, n=30) (χ2=29.83, p<0.001). More AIAN victims in urban areas had a diagnosed mental health problem (6.1%, n=63) compared with those in rural areas (2.8%, n=23) (χ2=10.41, p=0.001), and more than double the victims had a history of treatment for a mental health problem (4.6%, n=48) compared with those in rural areas (2.1%, n=17) (χ2=8.1514, p=0.004). Urban areas had more homicide victims who used weapons compared with rural areas (7.5% vs 5.3%, χ2=3.9438, p=0.047).
Rural homicides
Most AIAN homicide victims in rural settings were male (75.0%, n=891) and aged between 18 and 44 (62.7%, n=745; see table 1). Many rural homicides involved the use of a firearm (42.6%, n=507), but more than one-third of all homicides were committed with a sharp or blunt instrument (34.7%, n=412). Method used differed significantly from urban areas (χ2=52.127, p<0.001). Additionally, more than 25% (n=134) of rural firearm homicides were known to have a rifle or shotgun as the firearm type used. Nearly two-thirds (64.8%, n=770) of rural homicides occurred in and around a house/apartment (or other rooming house).
As seen in table 2, nearly half of all primary suspects with reported race/ethnicity data in rural homicides were AIAN (42.4%, n=385), and 30.4% (n=276) of all primary suspects were non-AIAN. More than two-thirds of non-AIAN suspects (68.8%, n=190) identified as white. More than a quarter (27.1%, n=246) of suspects’ racial/ethnic identity was unknown. More than half of all homicides (70.6%, n=722) were perpetrated by someone the victim knew. Nearly a quarter of all rural homicides (22.4%, n=220) were IPV related (table 3), a significantly higher proportion compared with urban homicides (16.8%, n=202) (χ2=10.431, p=0.001). Around 1 in 10 (10.9%; n=89) involved a family relationship problem, and 11.5% (n=113) involved another relationship problem (non-intimate). Greater than half of homicides were precipitated by an argument or conflict (50.3%, n=494) and about one in four were precipitated by a physical fight (25.8%, n=164), a circumstance that was significantly more common in rural homicides (χ2=3.9682, p<0.05).
Discussion
MMIP remains an urgent issue as indicated by recent executive orders, policies and community action, including Executive Orders 1389822 and 14053,23 Savanna’s Act24 and the Not Invisible Act of 2019.25 Comprehensive violence and MMIP prevention may benefit from incorporating community-developed and culturally relevant, evidence-based approaches.6 Similarities and differences in urban/rural AIAN homicides may inform these efforts. While homicides in urban areas were more likely than rural areas to be precipitated by crime and criminal activity, and a higher proportion of homicides in rural areas were related to IPV, these were leading homicides circumstances across settings. Further, understanding the urban and rural characteristics of AIAN homicides is important to define the problem.4 Historic and contemporary social conditions in these communities may also inform implementation of comprehensive approaches that include honouring and uplifting indigenous values and traditions, supporting collective healing from past and ongoing trauma, and continued comprehensive efforts to prevent MMIP.6
AIAN communities can use data to tailor strategies for those most at risk of homicide. For example, while preventing homicide at all ages is important, prevention efforts may include reaching young AIAN adults and AIAN people in rural areas with local adaptations and implementations focusing on community needs informed by local data. Communities can use data on primary perpetrators to inform prevention. The suspect’s race/ethnicity may be especially relevant as most in urban areas were identified as non-AIAN but less than a third in rural areas were identified as non-AIAN. Data supporting that most homicides involved a perpetrator known to the victim can help to dispel the myth that most homicides are random acts of violence perpetrated by strangers.26 For these AIAN homicides, suspects are generally known to the victim and may be the same or different race and ethnicity. This information may be used to improve and tailor public health and public safety interventions.
Limitations
The findings in this paper are subject to limitations noted previously12 including (1) racial misclassification27 on death certificates and ME or LE reports; (2) NVDRS does not collect data on tribal affiliation; since risk and protective factors may differ for tribes, findings may not be generalisable across AIAN communities; (3) states joined NVDRS at different times and thus data are not nationally representative.18 28 Limitations also include that (4) availability and completeness of data are dependent on successful partnerships among local VDRS programmes and their partners in vital records, C/ME offices and LE17; and (5) NVDRS collects limited suspect data.18 Additionally, (6) data are incomplete for injury location, the primary variable used to determine urban and rural classification. The classification by RUCA code for those that do have injury location zip codes allows for a more robust version of classification based on population density, urbanisation and daily commuting.21 (7) Dichotomous urban/rural geographical areas limit understanding of the 10 types of RUCA codes, thus exploring more localised geographical data analyses to further tailor prevention efforts may be beneficial in the future. Finally, (8) national-level surveillance data limit our ability to understand community-level structural conditions that may underlie urban/rural differences in AIAN homicides.
Conclusion
Findings regarding urban and rural AIAN homicides provide critical information to tailor and strengthen comprehensive public health solutions to prevent MMIP. Understanding that urban and rural differences have been driven by historic federal policies that have negatively affected AIAN people is crucial when interpreting findings. Solutions may include an integrated primary prevention response that involves tribes, the federal government, the public health sector, the criminal justice system and victim services.12 Some communities may choose to adopt and tailor prevention strategies29 because all AIAN communities possess numerous cultural and traditional assets that support prevention and may be incorporated into prevention efforts.6 CDC has developed Prevention Resources for Action detailing the best available evidence to prevent violence30 which include approaches and strategies that may be tailored and implemented for AIAN communities. When considering IPV as a precipitating circumstance, AIAN communities, especially those in rural areas, may benefit from the best available evidence compiled in the IPV Prevention Resource for Action.31
AIAN communities may need collective healing and resolution of structural inequities due to historic, intergenerational and ongoing trauma.6 While this study may provide some insight into how violence prevention might be tailored to urban and rural areas specifically, further analysing and contextualising local geographical areas may also guide approaches in regions that may not neatly fit into a binary urban–rural classification. NVDRS provides important and ongoing data related to AIAN homicides that may be used to further understand and guide prevention around violent death characteristics and circumstances. Such efforts may include community-developed, culturally relevant and evidence-based strategies whenever possible to avert the public health issue of MMIP.12
Data availability statement
Data are available upon reasonable request. NVDRS data used are available via approved application to use the NVDRS Restricted Access Database (RAD). More details are available online: https://www.cdc.gov/nvdrs/about/nvdrs-data-access.html.
Ethics statements
Patient consent for publication
Ethics approval
This activity was reviewed by CDC and was conducted consistent with applicable federal law and CDC policy. See, for example, 45 C.F.R. part 46.102(l)(2), 21 C.F.R. part 56; 42 U.S.C. §241(d); 5 U.S.C. §552a; 44 U.S.C. §3501 et seq.
Acknowledgments
The authors acknowledge and honor victims, families and communities affected by Missing or Murdered Indigenous People (MMIP) today; AIAN ancestors who survived; and allies who collaborate on MMIP prevention. We thank supporters and advocates for their decades of work and victim service organisations and criminal justice systems for their contribution. We thank our colleagues across the U.S. federal government, the Health and Human Services’ Intradepartmental Council on Native American Affairs (ICNAA) and the Not Invisible Act Commissioners for their work to improve data and reporting. This work was previously a poster presentation at the Society for the Advancement of Violence and Injury Research 2023 Annual Conference held in Denver, CO, USA.
Footnotes
Contributors All authors took part in conceptualisation, data interpretation and manuscript development. DTC led data analysis and was verified by CJB. Through conceptualisation to early manuscript development, LMMK completed work as part of the Centers for Disease Control and Prevention, National Center for Injury Prevention and Control, Division of Violence Prevention, and DES completed work as part of the Centers for Disease Control and Prevention, Center for State, Tribal, Local, and Territorial Support, Office of Tribal Affairs and Strategic Alliances. No artificial intelligence software was used in the development of this work. LMMK serves as guarantor for this work and accepts full responsibility for the work.
Funding DTC completed coding and a first draft of the manuscript as part of an Injury Control Research Center 2022 summer internship affiliated with the University of Iowa Prevention Research Center (UI IPRC), Grant #R49CE003095 under NOFO RFA-CE-19-001.
Disclaimer The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention and the Agency for Toxic Substances and Disease Registry or the University of Iowa.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.