The American Indian and Alaska Native (AIAN) population has considerably higher death rates for nearly all causes of death relative to the White population and the lowest life expectancy of all US racialized groups.1,2 In 2007–2009, the AIAN–White life expectancy gap was 7.3 years.3 Over a decade later, the gap in life expectancy is still about the same, specifically 7.0 years in 2019.4 The main explanation for this life expectancy gap is worse individual health over the life course for AIANs, greatly influenced by the lasting legacy of US settler colonialism and systemic racism contributing to the health outcomes of AIAN persons.5 AIAN persons have high stress levels and a high prevalence of chronic conditions which are factors linked to greater mortality.1,2 Disproportionate exposure to disadvantageous social determinants from social and economic marginalization, a physiological weathering process,6 and intergenerational transmission of historical trauma7 in the context of ongoing violations of Indian treaty rights8 are a few of the interconnected mechanisms of US settler colonialism and racism that have generated AIAN health inequities.3,5,9 Yet, an underexplored explanation for the AIAN–White life expectancy gap is ongoing systemic violence against Indigenous peoples and their disproportionately high victimization rates.

Any critical inquiry about the sources of the AIAN–White life expectancy gap must consider the chronic and disproportionately high victimization of Indigenous peoples in the United States and Canada, particularly for, but not limited to, Indigenous women and girls.10 The AIAN population has the highest violent victimization rate in the United States, with a 2.5 times greater likelihood of experiencing a violent death than all other racialized groups.11 Approximately 84% of AIAN women have experienced violence: 56% have experienced sexual violence and 55% have experienced physical violence.12 Furthermore, in 2016 alone, 5712 AIAN women and girls were reported missing to the National Crime Information Center.13 Together, violent victimization and the number of reported missing represent a human rights crisis, often referred to as Missing and Murdered Indigenous Women and Girls (MMIWG). The MMIWG crisis highlights the disproportionately high victimization of Indigenous women and girls—a result of colonization, which introduced a rigid system of gender that imposes disproportionate gendered violence on Indigenous women and gender nonbinary people within the interlocking systems of settler colonialism, racism, and heteropatriarchy.14 However, the inclusion of all genders in the MMIWG discussion is critical. During the Missing and Murdered Diné Relatives Forum in 2019, the Navajo Nation community urged researchers to include all genders in their investigations.15 Given this community input, our study includes both women and men, and we will describe the crisis as the Missing and Murdered Indigenous Relatives (MMIR) for MMIR to indicate gender inclusivity.

To date, only one nationwide study has attempted to document the prevalence of MMIR.16 Although a critical first step, that study do not fully capture the prevalence of MMIR, as it only focused on selected urban areas and confronted poor data collection from numerous cities.16 Because of rampant race misclassification3,17 and the systemic erasure of Indigenous peoples in data,18 AIAN mortality assessments are likely to be misattributed and underestimated.17 In 2019, US Executive Order 13898 was signed to establish a national task force to address MMIR concerns.19 Despite the important momentum that the executive order generated, an important challenge remains: the lack of data.16,20 To visibilize the impacts of MMIR on Indigenous peoples, we use the national detailed mortality data to tease out the overlooked nuances.

THE CURRENT STUDY

In the context of the MMIR human rights crisis, this study provides the first national overview of the prevalence of MMIR-relevant causes of death in 2010–2019 and their impacts on the AIAN–White life expectancy gap. We use multidecrement life tables and the age–incidence demographic decomposition method to assess the role of the MMIR crisis on the AIAN–White life expectancy gap.21

METHODS Data

We used data from the National Center for Health Statistics (NCHS) Detailed Mortality files, which contain information from death certificates.22 The data cover more than 99% of all deaths in 50 states, the District of Columbia, and the US territories. The NCHS uses an algorithm to bridge multiple race responses on death certificates to a single race that is compatible with the US Census Bureau’s standard practices.23 The data include information on the underlying cause of death, the disease, or circumstance that initiated the process leading to death. Consistent with previous research,2,24 we focused on non-Hispanic AIAN (hereafter AIAN) in our analysis because the updated bridged intercensal population estimates substantially overestimated the Hispanic AIAN population, which can lead to the underestimation of AIAN death rates.25 Similarly, we used non-Hispanic Whites as our reference group. The 2010–2019 population denominators were taken from the NCHS Vintage 2019 bridged-race postcensal population estimates.26

Operationalizing MMIR-relevant Causes

Because MMIR-specific information is unavailable in official vital statistics, we created MMIR-relevant categories drawing from the Sovereign Bodies Institute’s cause-of-death categorization27 and Indigenous epistemology.28 Most previous studies disaggregated external causes into three categories (e.g., homicide, suicide, and accidents), plus a heterogeneous residual category [1,2,29]. Potentially MMIR-relevant deaths due to other external causes (e.g., accidental drowning, struck by object, etc.) would be invisible if lumped into “accidents” or assigned to the residual category. To distinguish MMIR-relevant causes, we disaggregated “external causes of morbidity and mortality” into eight categories: (1) homicide, (2) undetermined intent, (3) causes potentially related to victimization and violence, (4) transport accidents, (5) suicide, (6) drug overdose, (7) institutional and/or medical racism, and (8) other accidents (residual category). We conceptualized the first three categories as the MMIR-relevant causes. The drug overdose category includes all deaths attributable to drug overdose regardless of intent (i.e., suicide, homicide, or unintentional), following the typical CDC classification.30 Institutional and/or medical racism includes the causes that are potentially related to institutional racism within the health care system.

After pulling out causes related to homicide, undetermined intent, transport accidents, suicide, drug overdose, and institutional and/or medical racism from “external causes of morbidity and mortality,” we categorized the remaining causes into two categories: “causes potentially related to victimization and violence” and “other accidents” (not related to victimization and violence). For this process, we utilized consensual qualitative research methodology, which is an integrative methodology to construct consensus among multiple researchers.31,32 All three authors independently reviewed the causes and assigned them to the domains and compared results across to identify the consensus. We also consulted with Indigenous elders with our categorization. We assigned the causes that can potentially involve other person(s) as “causes potentially related to victimization and violence.” Causes for which involvement of other person(s) is unlikely (e.g., Y98. Lifestyle-related condition) or beyond human control (e.g., Y97. Environmental-pollution-related condition) were assigned as “other accidents.”

The Age-incident Decomposition Method

We first created multidecrement life tables for both the AIAN and White populations. This demographic technique standardizes two populations to eliminate differences in population structure and estimates cause-specific mortality. We categorized age as younger than 12 months, 1–4 years, and subsequent ages in 5-year groups (i.e., 5–9, 10–14, and so on) up to age 100+. During 2010–2019, the total number of deaths was 174,105 for the AIAN population and 20,927,002 for the White population. We used the International Classification of Diseases (ICD-10) to collapse the more than 10,000 specific causes of death into 21 categories (eTable; https://links.lww.com/EDE/B934).

We then used the age–incidence decomposition method to identify the mechanisms contributing to group differences in life expectancy for multiple causes of death.21 Traditional demographic decomposition techniques can be used to quantify the contribution of a cause of death to the life expectancy gap,33 whereas the age–incidence method can also disentangle the mechanisms through which each cause contributes to the life expectancy gap, by distinguishing, within each cause, between age and incidence.21,34,35

Life expectancy at birth (e0) can be calculated as the sum of the cause-specific mean age at death weighted by the cause-specific probability of death: e0=∑cpcx¯c, where c indicates the cause of death, x¯cis the mean age at death for cause c, and pc is the probability of dying due to cause c. Then, the life expectancy gap between population A and population B is e0A−e0B=∑cpcAx¯cA−∑cpcBx¯cB. A given cause’s contribution to the gap equalspcA(x¯cA−e¯)−pcB(x¯cB−e¯), where e¯ is the average life expectancy for the two populations. If the probability of dying of cause c were the same for both populations (pcA=pcB), then cause c would contribute to the gap only through differences in mean age at death across the two populations. If the mean age at death for cause c were the same for both populations (x¯cA=x¯cB), then cause c would contribute to the gap only through differences in the probability of dying. This method enables identifying (1) how much of the AIAN–White life expectancy gap is due to differences in the prevalence of a specific cause (i.e., incidence component) and (2) how much is due to age-at-death differences between AIAN and White persons who die of the same cause (i.e., age component).21 For example, homicide can contribute to the comparatively lower AIAN life expectancy if (1) the AIAN population has a higher probability of dying from homicide (incidence component), given that the average age at death for homicide is lower than the overall life expectancy; (2) AIAN victims of homicide die, on average, at a younger age (age component); or both. More details about the age–incidence decomposition method are described elsewhere.21

We examined each cause’s contribution to the AIAN–White life expectancy gap first for the AIAN and White populations overall and then for women and men separately. Although the MMIWG and MMIR movements honor nonbinary genders and focuses on the disproportionate victimization of two-spirit, lesbian, gay, bisexual, transgender, queer, questioning, intersex, and asexual (2SLGBTQQIA) Indigenous peoples,16 the gender binary classifications on death certificate prevents us from following suit.

Race Misclassification Adjustments

Rampant race misclassification3,17,36,37 leads to important concerns about data validity—specifically, that AIAN mortality may be underestimated.18 In this study, we made adjustments for both AIAN mortality underestimation and AIAN population denominator overestimation.1,2,24

For mortality, race misclassification occurs primarily because race is attributed by others on the death certificate which reflects the complexity of defining an AIAN identity.38 When misclassification occurs, AIAN persons are typically misclassified as White,39 causing an underestimation of AIAN deaths. A recent review by the CDC revealed that this misclassification occurs for about 33.5% of deaths and that it varies by sex and age.4 To adjust for the underestimation in the death counts, we multiplied the deaths by the sex- and age-specific adjustment factors provided by Arias and colleagues (Technical notes, table 2).4 Additionally, previous research showed a 12% overestimation of the AIAN population counts.36 To adjust for race misclassification in the population estimates, we deflated the AIAN population denominators by dividing them by a fixed 1.12 adjustment factor.

RESULTS

Life expectancy in 2010–2019 was 70.6 and 78.8 for the AIAN and White populations, respectively, corresponding to a gap of 8.2 years. Table 1 summarizes the life table-adjusted cause-specific prevalence and average age at death for all causes of deaths in 2010–2019. Approximately 4.0% of all deaths among the AIAN population were due to the three MMIR-relevant causes (i.e., homicide, undetermined intent, and causes potentially related to victimization and violent experience), compared with 2.6% among the White population. The three MMIR-relevant causes accounted for 2.4% of all deaths among AIAN women, compared with 2.9% among White women. These three causes accounted for 5.0% of all deaths among AIAN men, compared with 2.9% among White men. Both AIAN women and men had a considerably lower mean age at death for all three causes than their White counterparts.

TABLE 1. - Distribution of Deaths and Average Age at Death, by Specific Cause: Total and Gender-Specific, 2010–2019 (Life Table Results) TOTAL WOMEN MEN Proportions Mean age Proportions Mean age Proportions Mean age AIAN (%) White (%) AIAN White AIAN (%) White (%) AIAN White AIAN (%) White (%) AIAN White MMIR-relevant causes  Homicide 0.9 0.2 37.0 41.1 0.1 0.4 36.7 43.3 1.3 0.3 37.1 40.0  Undetermined intent 0.9 0.5 49.6 64.2 0.4 0.7 53.6 71.1 1.1 0.5 47.1 58.2  Potentially related to victimization 2.2 2.0 64.2 77.2 1.8 1.8 71.6 82.3 2.6 2.1 59.5 72.4  MMIR total proportions 4.0 2.6 2.4 2.9 5.0 2.9 Other external causes  Transport accidents 2.7 1.0 42.3 48.6 0.6 1.8 41.5 50.5 3.5 1.5 42.7 47.8  Suicide 1.6 1.2 37.7 49.7 0.4 0.7 33.0 46.7 2.5 2.0 38.9 50.3  Drug overdose 2.6 1.6 43.8 42.9 1.1 2.1 44.3 45.3 3.1 2.1 43.4 41.5  Institutional/medical racism 0.2 0.2 61.9 71.0 0.1 0.2 61.3 74.1 0.3 0.2 62.3 68.5  Other accidents 0.6 0.3 62.9 79.7 0.3 0.4 69.8 84.3 0.7 0.3 59.4 74.1  Other external causes total proportion 7.8 4.3 2.6 5.2 10.2 6.1 All other causes  Infectious and parasitic diseases 3.0 2.2 68.5 76.9 3.2 2.2 71.1 79.4 2.8 2.2 65.7 74.2  Neoplasms 18.2 20.9 72.3 75.6 17.5 19.0 72.8 76.1 18.9 22.9 71.8 75.1  Immune, endocrine, and metabolic  diseases 7.6 4.1 71.8 77.4 7.9 4.0 74.3 80.0 7.4 4.3 69.3 74.8  Mental and behavioral disorders 4.7 6.5 79.3 87.3 5.5 8.3 84.1 89.1 4.0 4.5 73.0 83.6  Nervous system diseases 4.9 8.0 80.6 84.7 6.1 9.5 83.3 86.5 3.8 6.5 76.7 81.7  Diseases of the eye or ear 0.0 0.0 67.0 76.6 0.0 0.0 73.3 80.0 0.0 0.0 61.8 72.4  Diseases of the circulatory system 27.1 32.5 76.6 82.7 27.2 32.8 80.2 85.7 27.0 32.1 73.2 79.3  Diseases of the respiratory system 10.1 10.8 76.9 81.0 10.7 10.9 78.1 82.1 9.6 10.6 75.7 79.8  Diseases of the digestive system 7.3 3.4 62.5 74.4 7.6 3.3 64.7 77.8 6.9 3.5 60.2 70.9  Diseases of the skin and  musculoskeletal system 1.0 0.7 72.0 79.9 1.4 0.8 73.3 81.4 0.7 0.5 69.8 77.3  Diseases of the genitourinary system 3.0 2.5 76.7 82.7 3.4 2.5 77.9 83.9 2.5 2.5 75.1 81.4  Prenatal and postnatal causes 0.7 0.5 14.5 16.4 0.7 0.5 16.6 18.8 0.7 0.5 12.4 14.2  Ill-defined causes 0.6 1.0 71.5 84.0 0.8 1.3 78.7 87.2 0.5 0.7 60.4 77.9 All other causes total proportion 88.2 93.1 91.9 95.1 84.8 91.0 Total 100 100 70.6 78.8 100 100 74.1 81.8 100 100 67.3 75.6

Totals may not sum exactly because of rounding.

Table 2 shows the contribution of each cause of death on the AIAN–White life expectancy gap (decomposition results) for the total population and separately by gender. The incidence component accounted for 2.33 years (28.4%) of the total AIAN–White life expectancy gap, while the age component accounted for 5.88 years (71.6%). That is, the majority of the gap is explained by a younger age at death among AIANs, within the same causes. The remaining 29.4% of the gap was explained by the higher prevalence among AIANs of causes of death that tend to strike at younger ages. Remarkably, the three MMIR-relevant causes account for nearly 0.74 years (9.6%) of the life expectancy gap, even though they accounted for 4.0% of AIANs deaths. Within the incidence component, transport accidents (5.8%), drug overdose (4.0%), and homicide (7.7%) were the first, second, and fourth largest contributors. Transport accidents, drug overdose and homicide contributed to the gap mostly through the greater prevalence of these causes among AIAN persons than among White persons (incidence component). On the other hand, causes potentially related to victimization and violence, suicide, and undetermined intent contributed to the gap primarily through the younger ages at death among AIAN persons than among White persons (age component).

TABLE 2. - Components of the AIAN–White Life Expectancy Gap, by Total Population and by Gender, 2010–2019 (Decomposition Results) Total Women Men Incidence Age Total Incidence Age Total Incidence Age Total MMIR-relevant causes