FCS Tribes

According to the 2010 Census, over 500,000 individuals in the FCS identified as single race AIAN [7]. The FCS has numerous tribal lands and reservations including Navajo Nation, which is the largest reservation in the United States [8]. New Mexico has 23 federally-recognized tribes; Arizona has 22, Utah has eight, and Colorado has two [9–12]. In addition to Navajo, the FCS is home to Apache, Goshute, Hopi, Hualapai, Paiute, Pascua Yaqui, Pueblo, Quechua, Shoshone, Tohono O’odham Ute, Yavapai, and Zuni tribes, among others [9–12].

COVID-19 risk factors

Relative to non-Hispanic Whites, Native Americans have higher rates of chronic disease and poverty with lower rates of health insurance coverage and health literacy, which increases their risk for severe COVID-19 and limits their ability to prevent transmission and access medical care [2,3,13–15]. Indigenous people also have higher rates of asthma, COPD, heart and kidney disease, obesity, and diabetes [2,16]. Relative to other racial/ethnic groups, Native Americans are more likely to live below the poverty line, work in a frontline job (i.e., non-remote jobs that require frequent proximity to other people), or be unemployed [2,3,15]. AIAN individuals have the lowest rates of health insurance coverage among all racial/ethnic groups in the US [2,3,13]. For example, in 2021, the uninsured rate for non-Hispanic AIAN was 18.8%, compared to just 5.7% for non-Hispanic Whites [13]. Furthermore, private health insurance coverage was significantly lower among non-Hispanic AIAN individuals (43.1%) than among non-Hispanic Whites (74.2%) [13]. Additionally, historical trauma and discrimination have led to mistrust of the health care system and COVID-19 vaccination hesitancy among indigenous groups, creating additional barriers to care and prevention [17,18]. All these factors have likely contributed to a disproportionate number of COVID-19 cases and deaths in AIAN communities.

Prior research has shown that residents of Navajo Nation experience some of the most acute risks to health and longevity in the US, even relative to other indigenous groups [3,19–21]. Navajo Nation has very high percentages of residents in poverty (40.2%), adults without a high school diploma (26.9%), overcrowded housing conditions (17.4%), households without a vehicle (14.4%), and households with incomplete plumbing (18%) [2,3,21]. Additionally, a relatively large number of Pueblo and Navajo households do not speak English, which can make it difficult to secure resources such as employment and adequate healthcare [21]. Moreover, many reservations in the FCS are highly rural and lack advanced medical facilities capable of treating severe cases of COVID-19 [19,20].

The two primary aims of this study are to (1) investigate how the COVID-19 pandemic impacted life expectancy trends among non-Hispanic AIAN and Whites in the FCS, and (2) understand the extent to which changes in longevity and longevity disparities in the FCS are directly attributable to COVID-19 versus other causes of death. This study will decompose life expectancy changes within (and gaps between) AIAN and White populations in the FCS, allowing us to determine which causes of death contributed most to these changes and gaps during the COVID-19 pandemic. As noted, while some prior studies have examined how COVID-19 impacted longevity among Native Americans, none have employed decomposition methods to disentangle COVID-19 from other causes of death among AIAN and Whites in the FCS region.

Data

To assess longevity changes in the FCS over the period 2018–2022, single-race population data from the US Census Bureau for Arizona, Colorado, New Mexico, and Utah were utilized [22]. In addition, restricted-use mortality data (2018–2022) sourced from the US National Center for Health Statistics (NCHS) were included in our investigation [23].

Measures

Measures in our data include sex, age, race/ethnicity, and cause-of-death. Age at death was organized into 19 categories (less than 1 year old, 1–4 years old, a series of five-year age groups from 5–9 to 80–84, and 85 and above). The measurement for race was based on six revised single-race categories, which includes White, Black or African American, Asian, AIAN, Native Hawaiian or Other Pacific Islander, and more than one race [24]. Analyses included single race non-Hispanic AIAN (hereafter NH AIAN) and non-Hispanic White (hereafter NH White) populations in the FCS [25]. A small number of individuals identified as belonging to “multiple races” was excluded from the analyses. The analyses also excluded the Hispanic population because this group often exhibits socioeconomic, cultural, behavioral, and health profiles that differ from non-Hispanic populations [26].

The 10th revision of the International Classification of Diseases (ICD-10) was utilized to identify the 39 leading causes of death in the US [27]. These categories are mutually exclusive and exhaustive, encompassing a wide range of medical conditions and external causes, including specific types of heart disease and cancer, chronic lower respiratory diseases, cerebrovascular disease, Alzheimer’s disease, diabetes, influenza and pneumonia, and nephritis, among others [27]. Analyses also incorporated alcohol-related causes of death as defined by the Alcohol-Related Disease Impact (ARDI) classification scheme, by isolating different ARDIs and then subtracting them from the 39 leading causes of death [28]. ARDIs were classified into chronic or acute conditions that are either 100% attributable to alcohol (e.g., acute alcohol poisoning) or have alcohol-attributable fractions (AAF) that are directly related to alcohol use (e.g., fatal motor vehicle crashes involving alcohol use) [28]. A detailed description and the complete list of ICD-10 codes used in the ARDI application are available elsewhere [28]. S1 Table provides a list of ICD-10 codes for the 39 causes of death and ARDI.

Analysis

The first step was to construct a series of sex-specific period life tables for NH AIAN and NH White populations, spanning the period 2018–2022. Using death records from the NCHS and single-race population data from the U.S. Census Bureau, sex- and age-specific death rates (ASDR) were calculated for both populations. Sex-specific ASDR were used to estimate life expectancies at birth for NH AIAN and NH White populations in the FCS across the study period, following standard procedures for creating abridged life tables [29].

The second step of the analysis involved decomposing life expectancy gaps into age- and cause-of-death components, using the approach proposed by Arriaga [29,30]. Data for decomposition analyses were combined into two- or three-year aggregates of deaths and population data (i.e., pre-pandemic, 2018−19 and peak pandemic, 2020−22), similar to approaches used in previous research [4]. The total life expectancy gap was decomposed into years according to the relative contributions of each age group, followed by the specific causes of death within each age group. Age- and cause-specific decompositions were applied to the life expectancy differences for males and females 1) between NH AIAN and NH White populations at two different time points (i.e., 2018−19 and 2020−22) and 2) within these two populations before and after COVID-19.

The final step was to identify causes of death that were particularly relevant to our selected populations in the FCS. Among the 39 leading causes of death and ARDI we examined, select causes were presented in our final report if they contributed at least 0.20 years to 1) the life expectancy gap between NH AIAN and NH White populations for either males or females in at least one time period, or 2) changes in life expectancy for at least one race/sex group that occurred after the onset of COVID-19. For between-group differences, causes of death that met these criteria included COVID-19, diabetes mellitus, heart disease, malignant neoplasms (i.e., various cancers), symptoms and abnormal clinical/laboratory findings not elsewhere classified (excluding sudden infant death syndrome), influenza and pneumonia, nephritis/nephrotic syndrome/nephrosis, Alzheimer’s disease, chronic lower respiratory diseases, other external causes and unspecified accidents, alcohol abuse and poisoning, alcoholic liver disease, alcohol-related liver cirrhosis, alcohol-related vehicle crashes, alcohol-related homicide, non-alcohol poisoning, and alcohol-related suicide. For within-group changes, the only causes of death that met our inclusion criteria were COVID-19, heart disease, alcoholic liver disease, alcohol-related liver cirrhosis, and non-alcohol poisoning. In both analyses, we included a residual category to account for all other causes of death that did not meet our inclusion criteria.

In line with our research objectives, the results were presented in three parts: first, an overview of life expectancy trends; second, a detailed analysis of the gaps in life expectancy between race/sex groups; and third, an examination of life expectancy changes within each race/sex group. Analyses were conducted using SAS version 9.4 (Cary, NC), R (R Core Team), and MS Excel [30]. Restricted data for the analyses were obtained with permission from the NCHS (Contract number: DVS2024−0259). Our university’s institutional review board considered this study exempt from review (protocol number: 14619).

Public health implications

Despite inherent difficulties in overcoming the social and economic disadvantages faced by AIAN that contribute to reduced health and longevity, research suggests that they can be addressed with targeted interventions. Structural inequalities such as poverty, food insecurity, incomplete plumbing, overcrowded housing, inadequate access to transportation, and barriers to healthcare services (e.g., lower rates of health insurance coverage) all contribute to insufficient preventive care and increased risk for morbidity, including severe cases of COVID-19 [2,3,13,14,16–21,41–43]. Discrimination against indigenous groups and historical events have contributed to mistrust of the health care system and vaccination hesitancy, also placing AIAN individuals at heightened risk for COVID-19 and its complications [17,18]. Cultural sensitivity and improved representation of indigenous people among public health and medical professionals may help overcome distrust and other healthcare barriers, contributing to improved vaccine uptake and better general preventive care.

Drug and alcohol use often coincide with mental illness and trauma [44,45]. Native American adolescents and young adults have high rates of mental illness, particularly as they relate to substance and alcohol use disorders [46,47]. Mental illness, substance use disorders, and alcohol and drug use are all risk factors for accidental deaths (e.g., fatal motor vehicle accidents), homicide, and suicide [48,49]. The mortality age patterns we found suggest that early interventions to support mental health and prevent alcohol use are crucial within the AIAN population. Culturally informed and community-based interventions are crucial to reducing deaths associated with underlying mental illness such as alcohol and drug overdose, suicide, and homicide among indigenous populations [40,50].

Strengths and limitations

Restricted-use NCHS data helped us achieve accurate numerator counts, especially in rural counties where public mortality data tend to be censored due to low numbers. In addition, our use of decomposition methods enabled us to assess how numerous causes of death at different ages contributed to life expectancy changes and gaps during the height of the COVID-19 pandemic. Although we explained most of the life expectancy gaps and changes over this period of observation, 15–30% of these gaps and changes were unaccounted for by the causes of death examined in our analyses. While restricted-use NCHS data provide the most accurate mortality estimates available, they are not linked to economic data or health history information, limiting the ability of analysts to account for these factors. Finally, racial misclassification on death records is common among NH AIAN individuals in the US and may be more prevalent in urban areas and populations with a low percentage of NH AIAN people [51,52]. However, prior research has shown that the Southwest region, which encompasses the FCS, has a much lower misclassification rate than other US regions [51]. Therefore, because our study population lies within the Southwest and is predominantly rural with large NH AIAN majorities in some areas, we expect misclassification bias to be minimal.