Mental Health Outcomes Among American Indian and Alaska Native U.S. Army Soldiers: A Serial Cross-Sectional Analysis

ABSTRACT

Introduction

American Indian and Alaska Native (AI/AN) individuals in the USA experience higher rates of mental illness and preventable death than the general population. Published research demonstrates that AI/AN veterans experience similar disparities to other minorities compared to non-minority veterans; few studies, however, have assessed mental health outcomes in AI/AN active duty military members. The objective of this study was to determine differences in depression, anxiety, hazardous alcohol consumption, and suicidal ideation among AI/AN soldiers compared to soldiers of other races during the Coronavirus Disease 2019 (COVID-19) pandemic.

Materials and Methods

We conducted repeated cross-sectional electronic surveys to assess the mental health of active duty and activated reserve U.S. Army soldiers within three commands in the Northwestern Continental United States , Republic of Korea, and Germany during May-June 2020 (T1) and December 2020-January 2021 (T2). The primary exposure of interest in the present analysis was race and ethnicity, and the primary outcomes were probable depression with functional impairment (subsequently “depression”), probable anxiety with functional impairment (subsequently “anxiety”), hazardous alcohol use, and suicidal ideation. Multivariable logistic regression models were used to determine the association between demographics and COVID-19 concerns on mental health outcomes for each time point.

Results

A total of 21,293 participants responded to the survey at T1 (participation rate = 28.0%), and 10,861 participants responded to the survey at T2 (participation rate = 14.7%). In the multivariable model, AI/AN participants had 1.36 higher adjusted odds of suicidal ideation (95% CI: 1.02-1.82) at T1 and 1.50 greater adjusted odds of suicidal ideation at T2 (95% CI: 1.00-2.24), when compared to non-Hispanic White participants. During T1, there was no significant difference detected between AI/AN and non-Hispanic White participants for anxiety (adjusted odds ratio: 1.21; 95% CI: 0.91-1.60) (Table IV). However, AI/AN participants had 1.82 greater adjusted odds of anxiety when compared to non-Hispanic White participants at T2 (adjusted odds ratio: 1.82; 95% CI: 1.29-2.57). There were no significant differences detected between AI/AN participants and non-Hispanic White participants in multivariable models for either depression or hazardous alcohol use at both time points.

Conclusions

Although we hypothesized that all adverse mental health outcomes would be higher for AI/AN service members at both time points, there were no significant differences at each of the time points analyzed for most of the outcomes analyzed. However, differences in suicidal ideation were found at both time points. Analyses and proposed interventions should account for diversity and heterogeneity of AI/AN populations.

INTRODUCTION

Approximately 5.7 million individuals in the U.S. population identify as American Indian and Alaska Native (AI/AN) either alone or in combination with another race.¹ Relative to Whites, AI/AN individuals in the USA are at greater risk for psychiatric diagnoses,² including postpartum depression,³ PTSD,⁴ and suicide.⁵ The U.S. Indian Health Service (IHS), a federal agency responsible for providing comprehensive health services to federally recognized AI/AN people, reports considerably higher death rates in the AI/AN population for suicide (60% greater), poisoning (118% greater), and chronic liver disease and cirrhosis (368% greater) when compared to the total U.S. population.⁶ Similar disparities have been documented specifically for AI/AN military veterans and veterans of other races.^7–11 However, it is unclear if the increased risks observed in AI/AN civilians and military veterans are also observed for AI/AN active duty military service members.

The U.S. Military has a universal medical system for active duty members and their beneficiaries. Previous studies have demonstrated disparities for AI/AN groups in military populations for prenatal care utilization,¹² insomnia,¹³ stress fractures,¹⁴ gallbladder disease,¹⁵ diabetes readmissions,¹⁶ and nicotine dependence.¹⁷ However, few studies have examined disparities in adverse mental health outcomes between AI/AN service members and other groups. Mental health disparities are especially important to quantify in a universal medical system since equal access does not always result in lower adverse mental health outcomes.

A previous report of the cross-sectional survey data used for this study found mental health disparities by sex, rank group, and race/ethnicity.¹⁸ However, only four categories of race/ethnicity were used (non-Hispanic White, non-Hispanic Black, Hispanic, and Other), and therefore, the AI/AN population was subsumed within the “Other” group. The objective of the current study is to determine if there are differences in probable depression, probable anxiety, hazardous alcohol consumption, and suicidal ideation among AI/AN soldiers compared to non-Hispanic White, non-Hispanic Black, and Hispanic soldiers using an expanded set of race/ethnicity categories than those originally reported in the previous study.¹⁸

METHODS

The Walter Reed Army Institute of Research and the U.S. Army Public Health Center formed the Behavioral Health Advisory Team (BHAT) to collaboratively assess the health and well-being of U.S. Army soldiers during the Coronavirus Disease 2019 (COVID-19) pandemic. The BHAT administered repeated cross-sectional surveys that assessed numerous domains, including soldiers’ demographic characteristics, behavioral health, concerns about COVID-19, and experiences during the pandemic. All soldiers within three commands headquartered in the Northwestern Continental United States, Republic of Korea, and Germany were invited to participate during May-June 2020 (T1) and December 2020-January 2021 (T2). Both surveys were administered electronically, and soldiers were instructed to complete the survey via their smart phone, computer, or other web-enabled device. All responses were anonymous, and therefore not linked to personnel records, to encourage honest reporting. However, the focus on anonymity meant that no unique identifiers were collected to link the surveys over time. Prospective participants were only allowed to take the survey if they were active duty or activated Army Reserve at the time of the survey. No participants were deployed at the time of the survey. No incentive was provided for survey participation.

The goal of this study was to determine potential mental health disparities by race/ethnicity, and therefore, race/ethnicity was the primary predictor. The outcomes of interest included probable depression with functional impairment, probable anxiety with functional impairment, hazardous alcohol use, and suicidal ideation as measured by thoughts about being better off dead or hurting oneself.

Measures

Demographics

Participants were asked to identify their sex as male or female. Participants were also asked to identify their rank/pay grade as E1-E4 (Junior Enlisted), E5-E9 (Senior Enlisted), W1-W5 (Warrant Officer), O1-O3 (Company Grade Officer), O4-O6 (Field Grade Officer), or O7 and above (Flag Officer). For the purposes of the analysis, Warrant Officer and Officer groups were collapsed into one group. Participants were asked one multiple-choice question to report their race/ethnicity. The question included seven options, including AI/AN, Asian, Black or African American, Hispanic or Latino, Native Hawaiian or Other Pacific Islander, White, and/or Other (please specify). Participants who answered “American Indian/Alaska Native” either alone or in combination with another race/ethnicity were coded as “AI/AN” for the purposes of the analysis. The other categories coded for this analysis were non-Hispanic Black, non-Hispanic White, Hispanic, and Other (including Asian, Native Hawaiian or Other Pacific Islander, and Other). All demographic questions also included a “Prefer not to respond” option if a participant elected to not report their demographic characteristics.

Depression and anxiety

Depression was assessed with the two-item Patient Health Questionnaire (PHQ-2) where the summed score of 3 or higher was used to determine probable depression.¹⁹ Anxiety was assessed using the two-item Generalized Anxiety Disorder (GAD-2) where the summed score of 3 or higher was used to determine probable anxiety.²⁰ In addition, these measures incorporated functional impairment, which was measured with the question, “How difficult have these problems made it for you to do your work, take care of things at home, or get along with other people?” Participants who responded “Somewhat Difficult,” “Very Difficult,” or “Extremely Difficult” were classified as having functional impairment. Individuals who responded “Not Difficult at All” were classified as not having functional impairment.

Hazardous alcohol consumption

Hazardous alcohol consumption was assessed using the three-item Alcohol Use Disorder Identification Test-Consumption.²¹ A summed cutoff score of 5 for men and 4 for women was used to indicate a positive screen for potentially hazardous alcohol consumption.²²

Suicidal ideation

Suicidal ideation was measured as thoughts of being better off dead or self-harm. The BHAT used the ninth item from the PHQ-9: “Over the last 2 weeks, how often have you been bothered by the following problems?—Thoughts that you would be better off dead or of hurting yourself in some way.” Participants responded to the item on a four-point scale: “Not at all,” “Few or several days,” “More than half the days,” and “Nearly every day.”²³ Participants’ responses were coded as a binary outcome where “Not at all” was coded as “No” and “Few or several days,” “More than half the days,” and “Nearly every day” were coded as “Yes” to thoughts of being better off dead or self-harm. Respondents who indicated suicidal thoughts were provided with suicide hotline and military mental health resources at the end of the survey.

COVID-19 concerns

Participants’ COVID-19 concerns were assessed with 19 items (e.g., access to food, access to household supplies, and finances), and participants rated the extent to which they were worried or concerned with each item on a five-point scale (0 = not at all, 1 = slightly, 2 = moderately, 3 = very, 4 = extremely) developed for the original study (Cronbach’s alpha = .93).^24,25 A composite score was then summed (range: 0-76) to demonstrate each participant’s level of concern during the COVID-19 pandemic. This variable was included as a potential confounder because the original analysis found demographic differences in adverse mental health outcomes based on COVID-19 concerns.¹⁸

Data Analysis

We conducted Chi-square tests to determine if there was a statistically significant difference in the prevalence of each adverse mental health outcome at each time point. For the purposes of the prevalence analysis, AI/AN was compared to all other racial/ethnic groups in a simplified two by two table. Multivariable logistic regressions were used to determine the association between demographics (sex, rank group, and race/ethnicity) and probable depression with functional impairment (subsequently “depression”), probable anxiety with functional impairment (subsequently “anxiety”), hazardous alcohol consumption, and suicidal ideation. Models were run separately for each time point (T1 and T2), and each controlled for COVID-19 concerns, for a total of eight separate multivariable logistic regressions. Because of the increased probability of type I errors from multiple testing, an alpha level of .01 was used to determine statistical significance. To preserve comparisons with the original study,¹⁸ no additional covariates were added to the models. SAS (Version 9.4) was used for all analyses (Cary, NC).

RESULTS

Demographics

A total of 21,293 participants responded to the survey at T1 (participation rate = 28.0%), and 10,861 participants responded to the survey at T2 (participation rate = 14.7%). Soldiers could participate at T1 only, T2 only, or both time points; therefore, we cannot infer any within-subjects differences over time. The sample distribution by sex (87.0% male at T1 and 83.3% male at T2) and rank group (51.4% Junior Enlisted at T1 and 47.5% Junior Enlisted at T2) was representative of the overall military.²⁶ Approximately 2.8% and 3.0% self-identified as AI/AN at T1 and T2, respectively.

Depression

About 1 in 5 AI/AN participants screened positive for depression at T1 and T2 (19.6% at T1 [Table I] and 23.0% at T2 [Table II]). The AI/AN participants had a statistically significant higher prevalence of depression at T1 (P = .0008) but not T2 (P = .016) when compared to other racial/ethnic groups (Supplementary Table).

In the multivariable models, males had lower odds to screen positive for depression when compared to females at both T1 (Table III) and T2 (Table IV). In addition, Junior Enlisted and Senior Enlisted service members had greater odds to screen positive for depression when compared to Officers. There was no significant difference in depression between AI/AN and non-Hispanic White participants at TI (adjusted odds ratio [AOR]: 1.22; 95% CI: 0.93-1.60) or T2 (AOR: 1.34; 95% CI: 0.93-1.92).

Anxiety

About 1 in 5 AI/AN participants screened positive for anxiety at T1 (18.5%) and 1 in 4 screened positive at T2 (27.0%). The AI/AN participants had the highest prevalence of anxiety at both T1 (P = .0006) and T2 (P < .0001) when compared to other racial/ethnic groups.

In the multivariable models for both time points, males had lower odds to screen positive for anxiety when compared to females. At T1, Junior and Senior Enlisted had greater odds for anxiety when compared to Officers at T1 (Table III), but there was no statistically significant difference at T2 (Table IV). At T1, there was no significant difference between AI/AN and non-Hispanic White participants for anxiety (AOR: 1.21; 95% CI: 0.91-1.60). At T2, AI/AN participants had 1.82 greater adjusted odds of anxiety when compared to White participants (AOR: 1.82; 95% CI: 1.29-2.57).

Hazardous Alcohol Consumption

Approximately 1 in 4 AI/AN participants screened positive for hazardous alcohol consumption at T1 and T2 (23.1% at T1 [Table I] and 24.7% at T2 [Table II]). Although AI/AN participants had the highest prevalence at both time points for hazardous alcohol consumption when compared to all other racial/ethnic groups, the differences were not statistically significant (Supplementary Table).

In the multivariable models, males had greater odds of hazardous alcohol consumption when compared to females at T1 (Table III), but there was no significant difference at T2 (Table IV). At both time points, Junior Enlisted had greater odds of hazardous alcohol consumption when compared to Officers, but there was no difference between Senior Enlisted and Officers. There was no significant difference in hazardous alcohol consumption between AI/AN and non-Hispanic White participants at T1 (AOR: 1.05; 95% CI: 0.83-1.34) or T2 (AOR: 1.21; 95% CI: 0.87-1.70).

Suicidal Ideation

About 1 in 6 AI/AN participants were coded as “yes” to thoughts of being better off dead or self-harm at T1 and T2 (16.2% at T1 [Table I] and 17.1% at T2 [Table II]). The AI/AN participants also had a higher prevalence of thoughts of being better off dead or self-harm than all other racial/ethnic groups at T1 (P = .0011), but there was no statistically significant difference at T2 (P = .07) (Supplementary Table).

In the multivariable models, males had higher odds of suicidal ideation when compared to females T1 (Table III), but this association was not detected at T2 (Table IV). In addition, Junior Enlisted and Senior Enlisted service members had greater odds of suicidal ideation when compared to Officers at both time points. AI/AN participants had 1.36 higher adjusted odds of suicidal ideation (AOR: 1.36; 95% CI: 1.02-1.82)) at T1 and 1.50 greater adjusted odds of suicidal ideation at T2 (AOR: 1.50; 95% CI: 1.00-2.24), when compared to non-Hispanic White participants.

COVID-19 Concerns

COVID-19 concerns were significantly associated with depression, anxiety, hazardous alcohol consumption, and suicidal ideation at each time point (Tables III and IV). Specifically, each one-point increase on the COVID-19 concern score was associated with between 1% and 4% increased adjusted odds for each outcome.

DISCUSSION

This study found that AI/AN soldiers had a significantly higher prevalence of anxiety when compared to other racial/ethnic groups at both time points. However, there was no significant difference in the prevalence of hazardous alcohol use at either time point. Depression and suicidal ideation were significantly higher in prevalence at T1, but there was no significant difference in prevalence at T2. When controlling for sex, race/ethnicity, rank group, and COVID-19 concerns, AI/AN soldiers had elevated odds of suicidal ideation when compared to non-Hispanic Whites at both time points. Additionally, AI/AN participants had greater odds of anxiety at T2 in the multivariable model, but this association was not observed at T1. There were no differences between AI/AN and non-Hispanic White participants for either depression or hazardous alcohol use at either time point in multivariable models. Therefore, although we hypothesized that all adverse mental health outcomes would be higher for AI/AN service members at both time points, only suicidal ideation was consistently elevated at both time points for AI/AN soldiers in adjusted models.

The elevated odds of suicidal ideation among AI/AN soldiers are concerning because thoughts of self-harm in AI/AN military personnel and veterans have been linked to increased self-injurious behavior.⁷ Moreover, existing literature suggests that AI/AN soldiers are at elevated risk of non-suicidal self-harm and suicide compared to other racial groups.^7,27 For example, analyses of suicide and accidental death rates of active duty Army soldiers using administrative data found that the suicide rate for AI/AN soldiers was between 1.7 and 2.8 times higher than for other racial groups.²⁷ Furthermore, Veterans Health Administration data analyses indicated that AI/AN veteran suicide rates more than doubled from 19.1 to 47.0 per 100,000 person-years from 2002 through 2018.¹¹ Given high rates of AI/AN participation in the active duty military when compared to other racial/ethnic groups in the USA,^28,29 there is a need to better understand the risk of suicidal ideation of active duty AI/AN populations on an ongoing basis.

This study also extends the limited literature on anxiety and depression in the active duty AI/AN population. A single previous analysis of active duty personnel found that AI/AN service members had higher rates of insomnia before deployment compared to other racial groups (OR 1.86-2.85) and that service members with insomnia were more likely to have a history of depression, anxiety, and alcohol use, but the study did not report diagnoses by race.¹³ A large retrospective study of nearly 800,000 medical records focusing on mental health diagnoses in veterans found that AI/AN male veterans who served in Operation Enduring Freedom, Operations Iraqi Freedom, or Operation New Dawn were diagnosed with most disorders at higher rates than White males, including depression.³⁰ The previous analysis found that anxiety was less common in AI/AN veterans of either sex compared to Whites and that alcohol use disorders were more common.³⁰ A direct comparison of the results is limited, as the previous analysis of veterans used ICD-9/10 diagnoses whereas the current study used self-reported symptoms.³⁰

Disparities in mental health outcomes among civilian AI/AN populations have been attributed to structural racism, income inequality, generational trauma, and physical health inequities.^31–34 When investigating sex differences in use and costs of health care services, previous studies have also consistently reported that men under-utilize health care services. Among the AI/AN population, disparities in physical and mental health outcomes culminate in death rates for AI/AN males 1.5-4.5 times greater than AI/AN females for suicide, poisoning, and chronic liver disease and cirrhosis.⁶ These findings stress the importance of developing resources and programs that improve access to care among AI/AN active duty service members transitioning from military to civilian seeking health care services and treatment. In 2010, IHS and the VA established a memorandum of understanding (renewed in 2021) to improve access to care and the health status of AI/AN veterans.

This study is subject to several limitations. First, anonymous survey data do not exist for these commands before the COVID-19 pandemic (i.e., baseline data). Therefore, we can only compare changes in mental health outcomes during the COVID-19 pandemic, and we cannot determine how these outcomes differ from a baseline period before the pandemic (e.g., adverse mental health outcomes documented in medical records). Second, although data were collected at two time points, the survey was anonymous and we cannot determine if there were within-subjects differences over time. Participants may have participated at either time point or both time points, which further limits our ability to infer any temporal trends. Third, although data were collected from three commands across the U.S. Army, the findings may not be generalizable to the entire U.S. Army. Fourth, the proportion of AI/AN people, either alone or in combination with another race, in our sample (2.8% at T1 and 3.0% at T2) was numerically smaller than the national estimate for T1 (2.9%).¹ Therefore, these data may not be representative of AI/AN service members throughout the Army. Fifth, the small number of AI/AN service members may have led to Type II statistical errors at each time point where an association does exist at the population level but was not detected in the sample, because of lower statistical power than other racial/ethnic groups. Sixth, mental health constructs were assessed with abbreviated screeners to reduce survey fatigue. For example, suicidal ideation was assessed using a single item from the PHQ-9. Although the PHQ-9 is a validated scale, the single question only asks about suicidal ideation. Future studies should use dedicated suicide scales, such as the Columbia-Suicide Severity Rating Scale, which assesses ideation, plan, and attempt.³⁵ Similarly, the PHQ-2 and GAD-2 are abbreviated measures of the more extensive PHQ-9 and GAD-7, respectively, which are more specific. Seventh, we acknowledge that race is a social construct that reduces ancestry, culture, and identity to convenient categories. It is likely that the category of AI/AN, as a construct of race, does not represent the specificity of AI/AN identity, given unique histories of tribal memberships, federal government recognition of 574 tribal entities,³⁶ and individual’s affinities to their families of origin. Eighth, there are likely other unmeasured variables which impact the relationship between race/ethnicity and adverse mental health outcomes including adverse childhood experiences (i.e., poverty, parental substance use, and parental mental health problems) and racial/ethnic discrimination. Future efforts to characterize racial/ethnic health disparities in military populations should collect data on these important domains. Finally, these analyses only compared AI/AN participants to non-Hispanic White participants, and other disparities should be investigated between other racial/ethnic groups.

Given the indicators of racial disparities demonstrated in this sample, further study is needed to more fully investigate mental health outcomes among AI/AN service members in the Army and other branches of the U.S. Military. Future projects could include determining if there are racial differences in perceived barriers to care, utilization of mental health services, and satisfaction with culturally appropriate health care in the military health system. Another potential area of focus is the mitigating effects of variable military training and deployment schedules on AI/AN service members’ levels of resilience. Finally, future studies should elucidate the ways in which structural factors impact AI/AN service members and veterans and explore opportunities to expand objectives within the IHS/VA partnership to improve coordinated, culturally responsive, trauma-informed care.

CONCLUSION

This study demonstrates significant disparities in suicidal ideation among AI/AN service members when compared to non-Hispanic White service members. There are a high number of AI/AN military service members, and these men and women have the potential to return to their communities as transformational leaders. Thus, it is consequential to emphasize the value of identifying and mitigating mental health risks that develop before or during military service, or continue to affect AI/AN service members following military service, after they separate as veterans.

ACKNOWLEDGMENTS

The study team would like to acknowledge the feedback of Dr. Amy Adler and the soldiers who participated in these surveys.

SUPPLEMENTARY MATERIAL

Supplementary Material is available at Military Medicine online.

FUNDING

This study was funded by the U.S. Army Public Health Center and the U.S. Army Medical Research and Development Command.

CONFLICT OF INTEREST STATEMENT

None declared.

DATA AVAILABILITY

Because of the nature of these data, survey participants did not agree for their data to be shared publicly, so supporting data are not available.

CLINICAL TRIAL REGISTRATION

Not applicable.

INSTITUTIONAL REVIEW BOARD (HUMAN SUBJECTS)

The U.S. Army Public Health Center Public Health Review Board no. 22-141 (no. 20-831 BHAT) and WRAIR Institutional Review Board (no. 2766) determined this activity to be public health practice.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC)

Not applicable.

INSTITUTIONAL CLEARANCE

The manuscript was approved by the Director of the U.S. Army Public Health Center for public release on September 22, 2022.

INDIVIDUAL AUTHOR CONTRIBUTION STATEMENT

M.R.B conceived of the study idea, analyzed the data, wrote the manuscript, and critically revised the manuscript. A.A. conceptualized the analysis plan, wrote the manuscript, and critically revised the manuscript. C.M.S. wrote the manuscript and critically revised the manuscript. D.M.P., S.A.Q.G., T.D.J, A.M.B., T.S., and P.J.Q. critically revised the manuscript.

REFERENCES

Author notes