Racial and Ethnic Differences in Deaths by Suicide, Drug Overdose, and Opioid-Related Overdose in a National Sample of Military Members With Mild Traumatic Brain Injury, 1999-2019

Abstract

Objective: 

To examine racial and ethnic differences in suicide and drug and opioid-related overdose deaths among a population-based cohort of military service members who were diagnosed with a mild traumatic brain injury (mTBI) during military service.

Design: 

Retrospective cohort.

Setting: 

Military personnel receiving care within the Military Health System between 1999 and 2019.

Participants: 

In total, 356 514 military members aged 18 to 64 years, who received an mTBI diagnosis as their index TBI between 1999 and 2019, while on active duty or activated.

Main Measures: 

Death by suicide, death by drug overdose, and death by opioid overdose were identified using International Classification of Diseases, Tenth Revision (ICD-10) codes within the National Death Index. Race and ethnicity were captured from the Military Health System Data Repository.

Results: 

Overall crude rates were 38.67 per 100 000 person-years for suicide; 31.01 per 100 000 person-years for drug overdose death; and 20.82 per 100 000 person-years for opioid overdose death. Crude and age-specific rates for military members who self-identified as Other were higher than all other racial/ethnic groups for all 3 mortality outcomes. Adjusting for age, suicide rates for those classified as Other were up to 5 times that of other racial/ethnic groups for suicide, and up to 11 and 3.5 times that of other race/ethnicity groups for drug and opioid overdose death, respectively.

Conclusion: 

Findings extend previous knowledge regarding risk for suicide and deaths by drug overdose among those with mTBI and highlight new important areas for understanding the impact of race and ethnicity on mortality. Methodological limitations regarding classification of race and ethnicity must be addressed to ensure that future research provides a better understanding of racial and ethnic disparities in suicide and drug overdose mortality among military members with TBI.

IN THE UNITED STATES, mortality due to external causes, such as suicide, accidents, and drug overdose, are among the leading causes of death in adults younger than 45 years.¹ Since 2001, unprecedented increases in suicide rates have given rise to concerns that the United States is facing a “suicide crisis.” From 2001 to 2019, the unadjusted suicide rate in the US general population rose 33%.² Moreover, certain groups of Americans, particularly military veterans, have been disproportionately impacted by this crisis. The overall suicide rate for veterans is 1.5 times the rate for civilians.³ Since 2005, the suicide rate has also risen faster among veterans than it has for members of the general US population.⁴ These differences are even more pronounced for post-9/11 veterans, with rates up to 2.5 times the rate of adults in the general US population.³

The post-9/11 conflicts, characterized by a rise in improvised explosive devices, counterinsurgency operations, multiple deployments over a protracted length of time, and advances in medicine and technology, have contributed to a high prevalence of deployment-related injuries, including mild traumatic brain injury (mTBI).^5–7 It is well-established that mTBI increases the risk of suicide in the general and veteran populations,^8–11 and a recent study showed that post-9/11 veterans with a history of TBI had higher mortality rates across multiple causes of death, including suicide, compared with post-9/11 veterans with no TBI and members of the general US population.¹² Several factors may contribute to increased risk of suicide among post-9/11 military personnel with mTBI, including increased prevalence of comorbid conditions, such as depression, posttraumatic stress disorder, and chronic pain, as well as difficulties reintegrating into civilian life.^11,12

As the suicide crisis was unfolding in the United States, the opioid epidemic was also developing. Initially, the epidemic was fueled by staggering increases in opioid prescriptions for pain (1990s). This was followed by second (2010) and third (2013) waves of the epidemic, characterized by illicit and synthetic opioids, respectively.¹³ Drug overdose deaths in the United States have quadrupled since 1999, reaching the highest ever recorded in 2020.¹⁴ More than 70% of US drug overdose deaths in 2019 involved an opioid.¹⁵ Trends in veteran overdose mortality have paralleled rising rates observed in the United States more broadly.^16,17 From 2010 to 2019, the age-adjusted rate of drug overdose mortality among veterans increased by 53.2% overall and by 93.4% for opioid-related overdoses specifically.¹⁷

There have also been concerted efforts to understand the impact of the opioid epidemic on persons with TBI. These efforts have been fueled by growing recognition that individuals with TBI may be at greater risk for receiving prescription opioids due to comorbid conditions (eg, chronic pain), and persons with TBI may experience unique vulnerabilities (eg, cognitive and neurobehavioral changes, barriers to accessing care), which create a “perfect storm” of risk factors for opioid misuse and related consequences (eg, overdose, suicide).^18,19 Although evidence is still emerging, studies with civilians and veterans have shown that TBI increases risk for drug and opioid-related overdose.^10,20,21 Among post-9/11 veterans, all levels of TBI severity were associated with increased risk of suicide and drug overdose mortality.¹⁰ In a study of post-9/11 veterans receiving long-term opioid therapy for chronic pain, those with TBI had a 3-fold increase in nonfatal opioid overdose compared with those with no TBI.²²

Trends in suicide and overdose rates also highlight disproportionate increases for minoritized Americans. Although suicide deaths dropped overall in the United States, from 2019 to 2020, and declined by 5% for White Americans, there were increases among Black, Hispanic, and American Indian and Alaska Native (AIAN) Americans.²³ Similar disparities in suicide rate trends have been identified for minoritized veterans.³ Drug overdose death rates also increased across all racial and ethnic groups between 2018 and 2020, but, as seen for suicide, increases were larger for AIAN, Black, and Hispanic people than for White individuals.²⁴ Of particular concern are disparate increases in opioid-related deaths among minoritized groups. Since 2012, increases in opioid overdoses driven by heroin and the introduction of fentanyl have disproportionally impacted non-Hispanic Black and Hispanic individuals.^25,26 Although data on racial and ethnic differences in drug overdose deaths in the military population are limited, a recent study revealed that veterans who identified as multiracial had the highest crude rate of overdose mortality in 2019 (69.5 deaths per 100 000).¹⁶

Understanding health disparities in the military population is critical to ensuring equitable care for all who have served. Among military service members, disparities in external causes of death, such as suicide and drug overdose associated with TBI, may be further compounded for members of minoritized groups. However, to date, no studies have examined racial and ethnic differences in suicide and drug overdose deaths, 2 external causes of death among post-9/11 military members with mTBI. To address this knowledge gap, we examined racial and ethnic differences in specific external causes of death, namely, suicide and drug and opioid-related mortality (1999-2019), among a population-based cohort of military members diagnosed with mTBI during military service.

METHODS

Study cohort

This work was conducted as part of operational (program evaluation and quality improvement) efforts within the Department of Defense (DoD) and the Veterans Health Administration (VHA); the Colorado Multiple Institutional Review Board also determined that efforts were not human subjects research. Reporting follows the STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) guideline. Our analytic sample included all military members aged between 18 and 64 years who received an mTBI diagnosis as their index TBI (ie, first diagnosis) between 1999 and 2019, while on active duty or activated. Those whose index TBI diagnosis was a moderate, severe, penetrating, or unclassified injury were excluded (n = 122 278). Our final sample included 356 514 military members.

Data sources

Traumatic brain injury diagnoses in the Military Health System (MHS) were drawn from the MHS Data Repository, which includes information on all outpatient and inpatient encounters occurring at military treatment facilities and provided by civilians but paid for by TRICARE. Traumatic brain injury diagnoses were also captured from theater-based encounters from the Theater Medical Data Store. Demographic characteristics were obtained using the Defense Enrollment Eligibility Records System and the Defense Manpower Data Center. We linked our sample to the VA-DoD Mortality Data Repository, which compiles National Death Index records, the gold standard source for identifying cause and date of death for all military members and veterans.

mTBI diagnosis

We identified mTBI based on the DoD Standard Surveillance Case Definition for TBI classification,²⁷ enhanced with select codes included in the Rocky Mountain Mental Illness Research Education and Clinical Center TBI classification.²⁸

Race and ethnicity

Race/ethnicity is captured in the MHS Data Repository based on self-report by military members and was collapsed by DoD to create the following racial/ethnic groups: AIAN; Asian or Pacific Islander (AAPI); Black/non-Hispanic; White/non-Hispanic; Hispanic; Other (ie, military member-reported Other for race and Other or Unknown for ethnicity); or Unknown (ie, missing data for both race and ethnicity).

Outcomes

Suicide

Death by suicide occurring any time after the index mTBI was identified using International Classification of Diseases, Tenth Revision (ICD-10) underlying cause of death codes (X60-X84 and Y87.0) within the National Death Index.

Overdose deaths

We identified drug overdose and opioid overdose deaths within the National Death Index similarly to suicides. Drug overdose deaths codes included “suicide” codes X60-X64, “unintentional” codes X40-X44, and “undetermined” codes Y10-Y14.²⁹Opioid overdose deaths were identified as a subgroup of drug overdose deaths using ICD-10 multiple cause-of-death codes, which implicated opioids (T40.0, T40.1, T40.2, T40.3, T40.4, or T40.6), natural/semisynthetic opioids (T40.2), methadone (T40.3), synthetic opioids other than methadone (T40.4), or heroin (T40.1) as the underlying cause of death.²⁹

Analyses

Demographics captured at the time of index mTBI were summarized for the overall mTBI cohort and among suicide and overdose deaths, using frequencies. Crude average annual suicide rates and drug and opioid death rates were calculated per 100 000 person-years for 1999-2019, overall and by racial/ethnic group. Person-years at risk for 1999-2019 were calculated as the sum, across military members, of the time from mTBI to date of death or December 31, 2019, whichever came first. Rates with cell sizes less than 16 are denoted as unreliable (relative standard error of the rate is ≥25%) and less than 10 are suppressed, given these rates are highly unreliable and to protect confidentiality.³⁰

Age-adjusted rates were used to calculate age-adjusted rate ratios (RRs) to compare rates across racial/ethnic groups. Age-adjusted rates were standardized on the basis of the 2000 US population,³¹ using age categories 18 to 29 and 30+. Crude RRs are presented for all nonsuppressed rates as it was not feasible in all cases to calculate the corresponding age-adjusted RRs (ie, if either age group had fewer than 5 events). Consistent with best practice recommendations, no single racial or ethnic group was used as a reference group for comparisons. Specifically, the exploratory nature of this study and lack of a specific hypothesis with respect to racial/ethnic differences precluded selection of a single reference group.^32,33 Although all pairwise comparisons were made where feasible, results are presented only for crude RRs where the reference group had the lower rate for ease of interpretation. Corresponding age-adjusted RRs are then presented, when reasonably reliable, to account for differences in age distributions across study groups, given that risk can vary considerably by age. All rates and RRs are presented with 95% confidence intervals (CIs). As all outcomes are rare events, crude rates are presented with exact CIs, and RRs are presented with CIs based on the inverse of the F distribution.³⁴

RESULTS

Numbers and percentages of suicide and drug overdose and opioid overdose deaths by race/ethnicity and age are listed in Table 1. Data on drug overdose deaths classified as suicide, unintentional and undetermined, are included in Supplemental Digital Content Table 1, available at: https://links.lww.com/JHTR/A630. Distribution of rank by race/ethnicity groups is included in Supplemental Digital Content Table 2, available at: https://links.lww.com/JHTR/A631.

TABLE 1 - Sample characteristics of military-acquired mTBI population based on DoD and MIRECC case definitions

	Total, N = 356 514		Suicides, N = 1105 (0.31%)		Drug overdose deaths, N = 886 (0.25%)		Opioid overdose deaths, N = 595 (0.17%)
	N	%	N	%	N	%	N	%
Gender
Male	303 145	85.03	1043	94.39	839	94.70	567	95.29
Female	53 369	14.97	62	5.61	47	5.30	28	4.71
Race/ethnicity
American Indian/Alaskan Native	5292	1.48	20	1.81	10	1.13	5	0.84
Asian American/Pacific Islander	19 629	5.51	48	4.34	33	3.72	20	3.36
Black/non-Hispanic	47 946	13.45	90	8.14	35	3.95	26	4.37
White/non-Hispanic	191 805	53.80	587	53.12	512	57.79	350	58.82
Hispanic	38 091	10.68	80	7.24	49	5.53	33	5.55
Other	26 300	7.38	197	17.83	190	21.44	125	21.01
Unknown	27 451	7.70	83	7.51	57	6.43	36	6.05
Age, y
18-24	149 682	41.98	589	53.30	437	49.32	309	51.93
25-29	74 616	20.93	239	21.63	250	28.22	175	29.41
30-34	40 658	11.40	127	11.49	78	8.80	38	6.39
35-44	54 108	15.18	132	11.95	108	12.19	68	11.43
45-54	25 186	7.06	17	1.54	13	1.47	5	0.84
≥55	12 259	3.44	1	0.09	0	0	0	0.00

Crude and age-specific suicide rates and adjusted rate ratios, 1999-2019

The overall crude average annual rate of suicide (all ages and racial/ethnic groups) was 38.67 per 100 000 person-years (see Table 2). Across both age groups, suicide rates for military members in the Other group were higher than in all other racial/ethnic groups. Suicide rates were higher in military members aged 18 to 29 years than in those 30 years of age or older for all racial and ethnic groups (age-specific suicide rate: 44.18 [18-29] vs 28.16 [age 30+]).

TABLE 2 - Age-specific suicide rates per 100 000 person-years, 1999-2019

	Age-specific rates (95% CI)
	All ages, y	18-29 y	30+ y
Overall	38.67 (36.42-41.02)	44.18 (41.23- 47.30)	28.16 (24.95- 31.68)
Race/ethnicity
American Indian/Alaskan Native	46.96 (28.69-72.53)	41.94a (21.67-73.27)	b
Asian American/Pacific Islander	28.37 (20.92-37.61)	32.64 (22.74-45.40)	20.97a (11.17-35.87)
Black/non-Hispanic	25.39 (20.42-31.21)	39.39 (31.08-49.23)	8.18a (4.36-13.99)
White/non-Hispanic	39.56 (36.42-42.89)	44.11 (39.93-48.61)	32.07 (27.56-37.12)
Hispanic	29.40 (23.31-36.59)	29.58 (22.28-38.50)	29.02 (18.78-42.84)
Other	82.05 (70.99-94.35)	82.28 (70.31-95.70)	80.81 (54.12-116.1)
Unknown	28.13 (22.40-34.87)	32.21 (25.29-40.44)	b

Abbreviation: CI, confidence interval.

^aUnreliable, N < 16.

^bSuppressed, N < 10.

Crude and adjusted suicide RRs are presented in Table 3. Adjusting for age, significant differences were found for suicide rates in the Other group compared with all other racial/ethnic groups, except for AIAN. Suicide rates for those classified as Other were 2.34 times that of White/non-Hispanic and up to 5.44 times that of Black/non-Hispanic. Age-adjusted rates for AIAN military members were 2.30 times that of AAPI and 3.62 times that of Black, non-Hispanic military members. White/non-Hispanic and Hispanic military members had rates 2.32 and 1.95 times that of Black/non-Hispanic military members, respectively.

TABLE 3 - Crude and age-adjusted suicide RRs, 1999-2019^a

	Suicide RRs (95% CIs)
	Crude	Age adjusted
Reference: American Indian/Alaskan Native
Other	1.75 (1.10-2.92)	1.50 (0.76-3.21)
Reference: Asian American/Pacific Islander
American Indian/Alaskan Native	1.66 (0.93-2.84)	2.30 (1.01-4.89)
White/non-Hispanic	1.39 (1.04-1.91)	1.48 (0.97-2.32)
Hispanic	1.04 (0.72-1.51)	1.24 (0.72-2.15)
Other	2.89 (2.10-4.05)	3.45 (2.05-5.90)
Unknown	See inverse comparison	See inverse comparison
Reference: Black/non-Hispanic
American Indian/Alaskan Native	1.85 (1.08-3.03)	3.62 (1.70-7.05)
Asian American/Pacific Islander	1.12 (0.77-1.60)	1.58 (0.92-2.61)
White/non-Hispanic	1.56 (1.25-1.97)	2.32 (1.71-3.16)
Hispanic	1.16 (0.85-1.58)	1.95 (1.24-3.02)
Other	3.23 (2.51-4.19)	5.44 (3.55-8.24)
Unknown	1.11 (0.81-1.51)	1.19 (0.69-2.02)
Reference: White/non-Hispanic
American Indian/Alaskan Native	1.19 (0.72-1.85)	1.56 (0.78-2.84)
Other	2.07 (1.76-2.44)	2.34 (1.68-3.20)
Reference: Hispanic
American Indian/Alaskan Native	1.60 (0.93-2.63)	1.85 (0.86-3.71)
Asian American/Pacific Islander	See inverse comparison	See inverse comparison
White/non-Hispanic	1.35 (1.06-1.72)	1.19 (0.84-1.70)
Other	2.79 (2.14-3.67)	2.78 (1.77-4.39)
Unknown	See inverse comparison	See inverse comparison
Reference: Unknown
American Indian/Alaskan Native	1.67 (0.97-2.75)	3.04 (1.32-6.54)
Asian American/Pacific Islander	1.01 (0.69-1.46)	1.32 (0.70-2.47)
White/non-Hispanic	1.41 (1.12-1.79)	1.95 (1.25-3.12)
Hispanic	1.05 (0.76-1.44)	1.64 (0.94-2.89)
Other	2.92 (2.25-3.82)	4.57 (2.66-7.92)

Abbreviations: CIs, confidence intervals; RRs, rate ratios.

^aFor each crude RR, groups with lower rates were assigned as the reference for ease of interpretation. Corresponding age-adjusted rate ratios are presented where feasible (ie, each age group cell has at least 5 events). Age adjustment was performed on the basis of age groups of 18 to 29 years and 30+ years using the US 2000 Census. The values of RRs in boldface are statistically significant.

Crude and age-specific drug and opioid mortality rates and adjusted rate ratios, 1999-2019

The overall crude average annual rate was 31.01 per 100 000 person-years for drug overdose death and 20.82 per 100 000 person-years for opioid overdose death (see Table 4). Across age groups, drug and opioid mortality rates were highest for military members in the Other group. For drug overdose deaths, age-specific rates were higher in the 18 to 29 years age group than in those 30 years of age or older in all racial/ethnic groups except for those in the Other group where rates were higher for those aged 30+ years. When rates were calculable for opioid overdose deaths (ie, Other; White/non-Hispanic), age-specific rates were higher in those aged 18 to 29 years than in those aged 30 years and older.

TABLE 4 - Age-specific drug and opioid overdose death rates per 100 000 person-years, 1999 to 2019

	Age-specific rates for drug overdose (95% CI)			Age-specific rates for opioid overdose (95% CI)
	All ages, y	18-29 y	30+ y	All ages	18-29 y	30+ y
Overall	31.01 (29.00-33.12)	36.66 (33.97-39.51)	20.23 (17.52-23.25)	20.82 (19.18-22.56)	25.83 (23.58-28.23)	11.29 (9.28-13.59)
Race/ethnicity
American Indian/Alaskan Native	23.48a (11.26, 43.18)	b	b	b	b	b
Asian American/Pacific Islander	19.50 (13.43-27.39)	22.38 (14.34-33.31)	b	11.82 (7.22-18.26)	13.06a (7.14-21.91)	b
Black/non-Hispanic	9.87 (6.88-13.73)	12.79 (8.28-18.88)	6.29a (3.02-11.57)	7.34 (4.79-10.75)	9.21 (5.46-14.55)	b
White/non-Hispanic	34.50 (31.58-37.62)	41.40 (37.35-45.76)	23.16 (19.35-27.50)	23.59 (21.18-26.19)	29.91 (26.49-33.66)	13.19 (10.35-16.55)
Hispanic	18.01 (13.32-23.81)	19.36 (13.56-26.80)	15.09a (8.03-25.80)	12.13 (8.35-17.03)	14.52 (9.57-21.13)	b
Other	79.14 (68.28-91.23)	77.38 (65.79-90.43)	89.17 (60.99-125.9)	52.06 (43.34-62.03)	53.87 (44.28-64.93)	41.80a (23.39-68.94)
Unknown	19.31 (14.63-25.02)	23.07 (17.28-30.18)	b	12.20 (8.54-16.89)	14.80 (10.25-20.68)	b

Abbreviation: CI, confidence interval.

^aUnreliable, N < 16.

^bSuppressed, N < 10.

Crude and adjusted drug and opioid mortality RRs are presented in Table 5. Adjusting for age, significant differences were found for drug overdose death rates in the Other group compared with all other racial groups. Age-adjusted drug overdose deaths for those classified as Other were more than 3 times the rate of White/non-Hispanic, more than 5 times the rate of AAPI and Hispanic, and more than 11 times the rate of Black/non-Hispanic military members. Age-adjusted drug overdose rates for White/non-Hispanics were more than 1.6 times the rate of AAPI and Hispanic military members and more than 3.5 times that of Black/non-Hispanic military members. Black/non-Hispanic military members had significantly lower age-adjusted risk of drug overdose deaths compared with all other racial groups.

TABLE 5 - Crude and age-adjusted drug and opioid overdose death RRs, 1999-2019^a

	Drug overdose death RRs (95% CIs)		Opioid overdose death RRs (95% CIs)
	Crude	Age adjusted	Crude	Age adjusted
Reference: American Indian/Alaskan Native
White/non-Hispanic	1.47b (0.79-3.08)	c	c	c
Other	3.37 b (1.79-7.14)	c	c	c
Reference: Asian American/Pacific Islander
American Indian/Alaskan Native	1.20b (0.53-2.50)	c	c	c
White/non-Hispanic	1.77 (1.24-2.60)	1.67 (1.01-2.91)	2.00 (1.27-3.31)	1.61 (0.85-3.34)
Hispanic	See inverse comparison	See inverse comparison	1.03 (0.57-1.89)	0.83 (0.34-2.10)
Other	4.06 (2.79-6.06)	5.34 (2.97-9.93)	4.40 (2.73-7.46)	4.27 (1.98-9.76)
Unknown	See inverse comparison	See inverse comparison	1.03 (0.58-1.88)	c
Reference: Black/non-Hispanic
American Indian/Alaskan Native	2.38 b (1.05-4.91)	c	c	c
Asian American/Pacific Islander	1.98 (1.19-3.27)	2.11 (1.03-4.25)	1.61 (0.85-3.00)	1.75 (0.71-4.15)
White/non-Hispanic	3.49 (2.48-5.07)	3.52 (2.23-5.75)	3.22 (2.16-4.99)	2.83 (1.67-5.05)
Hispanic	1.82 (1.16-2.90)	2.08 (1.08-3.99)	1.65 (0.96-2.88)	1.45 (0.64-3.27)
Other	8.01 (5.66-11.85)	11.26 (6.54-19.74)	7.10 (4.62-11.29)	7.48 (3.82-14.98)
Unknown	1.96 (1.26-3.07)	c	1.66 (0.98-2.87)	c
Reference: White/non-Hispanic
Other	2.29 (1.93-2.71)	3.20 (2.30-4.37)	2.21 (1.79-2.72)	2.64 (1.69-4.03)
Reference: Hispanic
American Indian/Alaskan Native	1.30b (0.59-2.61)	c	c	c
Asian American/Pacific Islander	1.08 (0.67-1.72)	1.01 (0.50-2.02)	See inverse comparison	See inverse comparison
White/non-Hispanic	1.92 (1.43-2.62)	1.69 (1.09-2.73)	1.94 (1.36-2.87)	1.95 (1.09-3.67)
Other	4.39 (3.20-6.15)	5.41 (3.18-9.37)	4.29 (2.91-6.51)	5.17 (2.53-10.82)
Unknown	1.07 (0.72-1.60)	c	1.01 (0.61-1.66)	c
Reference: Unknown
American Indian/Alaskan Native	1.22b (0.55-2.40)	c	c	c
Asian American/Pacific Islander	1.01 (0.64-1.58)	c	See inverse comparison	c
White/non-Hispanic	1.79 (1.36-2.39)	c	1.93 (1.37-2.81)	c
Hispanic	See inverse comparison	c	See inverse comparison	c
Other	4.10 (3.03-5.61)	c	4.27 (2.93-6.37)	c

Abbreviations: CIs, confidence intervals; RRs, rate ratios.

^aFor each crude RR, groups with lower rates were assigned as the reference for ease of interpretation. Corresponding age-adjusted RRs are presented where feasible (ie, each age group cell has at least 5 events). Age adjustment was performed on the basis of age groups of 18 to 29 years and 30+ years using the US 2000 Census. The values of RRs in boldface are statistically significant.

^bUnreliable (ie, at least 1 rate has <16 events).

^cSuppressed (ie, at least 1 rate has <10 events).

Adjusting for age, significant differences were found for opioid-related mortality rates in the Other group compared with AAPI, White/non-Hispanic, Black/non-Hispanic, and Hispanic groups. Rates were suppressed for AIAN due to low numbers. Adjusting for age, opioid-related mortality rates for the members of the Other group were 2.64 to 7.48 times the rates for the other racial/ethnic groups. Age-adjusted opioid-related mortality rate for White/non-Hispanics was 1.95 times the rate for Hispanic and 2.83 times the rate for Black/non-Hispanic military members.

DISCUSSION

This study offers the first comparison of suicide and drug overdose mortality rates, 2 leading causes of external death, for different racial and ethnic groups in a population-based cohort of US military members diagnosed with mTBI, the majority of whom served post-9/11. During 1999-2019, age-specific mortality rates for suicide and drug and opioid overdose were consistently higher among military members who self-identified as Other, a category often selected by minoritized individuals who do not identify with any of the standard categories (eg, multiracial). Increased suicide and opioid overdose death rates were also found for military members aged 18 to 29 years compared with those 30 years of age or older across all racial/ethnic groups.

Findings from this study are consistent with previous work, highlighting increased risk for suicide among those with a history of TBI.^8–12 Among the entire study cohort, the overall crude average annual rate for suicide was 38.67 per 100 000 person-years. The suicide rate among this cohort of military personnel was notably lower than that identified among individuals with mTBI seeking care within the VHA (81 per 100 000 person-years).¹¹ This is in part related to demographic differences among those who do and do not seek VHA care. For example, in comparing risk factors and health characteristics among veterans who did and did not receive VHA care, several studies show that VHA users report lower sociodemographic characteristics and higher comorbidities and disability than non-VA users.^35–37

In terms of risk for suicide among those with TBI by race/ethnicity, research among all cohorts (eg, military, general population) has been extremely limited. Our findings of significantly higher rates among AIAN individuals with mTBI are consistent with work by Mohatt et al,³⁸ who found a suicide rate of 66 per 100 000 person-years among AIAN veterans aged 18 to 39 years (2014-2018). With respect to higher rates found among those who identified as Other, comparison cohorts were not readily identified in the existing literature, highlighting the nascent state of the science.

This study also adds support for the “perfect storm” theory that persons with TBI are at risk for devastating opioid-related consequences including overdose.^18,28 Specifically, among military members with mTBI, over two-thirds of drug-overdose deaths involved an opioid. This finding is similar to the proportion of opioid-related overdose deaths in recent years in the general population.¹⁴ We do not know whether these opioid-related overdose deaths involved prescription opioids only, illicit or synthetic opioids, or a combination of opioids (as well as other drugs), which is important as our study time frame spanned all 3 phases of the opioid crisis. Importantly, military members with mTBI who self-identified as Other were at a highest risk for drug and opioid-overdose deaths. The recent phases of the opioid epidemic characterized by heroin and fentanyl, in combination with other drugs (eg, stimulants), have disproportionally impacted people of color in the United States,²⁵ and these trends have been escalating since the start of the COVID-19 pandemic.³⁹ More research is needed among persons with TBI to investigate disparities in risk for drug and opioid-overdose among racial and ethnic minority groups, in both military and civilian populations. Furthermore, because evidence is mounting that individuals with TBI disproportionally receive prescription opioids due to higher prevalence of pain and other factors,^40,41 more research is needed to examine the role of prescription opioids as a pathway to illicit opioid use in association with risk for drug overdose.¹⁸

This study highlights the importance of examining racial/ethnic disparities in suicide and drug overdose mortality rates in military members with mTBI. Although today's military is more racially and ethnically diverse than previous generations, disparities persist.⁴² Moreover, such disparities are largely shaped by long-standing systemic social, economic, and environmental inequities. In the military, minorities are underrepresented in high-ranking career fields and overrepresented in support and lower-ranking career fields.⁴² This may contribute to a disproportionate number serving in the infantry or in positions that are generally associated with higher rates of injuries and casualties, perhaps increasing risk for TBI. Nonetheless, to date this has not been directly studied. Racial disparities also exist with respect to military justice and discipline processes.⁴³ For example, an independent racial disparity review revealed that Black enlisted members were almost twice as likely as White enlisted members to be involuntarily discharged on the basis of misconduct⁴³ and 72% more likely than White enlisted service members to receive an Article 15.⁴³ Collectively, these disparities may have tangible effects on physical and mental health, retention, reintegration into civilian life, and socioeconomic well-being, which can influence risk for both suicide and drug overdose. In fact, our study revealed that military members in the Other group were more likely to serve in lower-ranking positions (see Supplemental Digital Content Table 2, available at: https://links.lww.com/JHTR/A631), which has been associated with lower socioeconomic status and other social determinants that could in part be contributing to higher suicide and overdose mortality rates observed in this group.⁴⁴

Although additional research is needed to understand how race and ethnicity shape risk for suicide and drug overdose mortality in military members with TBI, methodological challenges associated with the classification of race/ethnicity must be addressed. In the United States, racial boundaries have shifted considerably, yet race/ethnicity category options are often constrained to a single binary variable. In this study, military members who self-identified as Other had higher rates of all 3 mortality outcomes than all other racial/ethnic groups. However, because choices for race/ethnicity were constrained to a single choice with no category options for multiple races, the racial and ethnic identities of military members in this higher-risk group (7.4%) remain unclear. Such constraints fail to account for within-group heterogeneity, limiting our understanding of how different racial/ethnic identities within the Other group may be associated with risk for adverse outcomes. Addressing these measurement issues both within and beyond the DoD system will be critical to understanding the defining characteristics of those who identify as Other and how these characteristics relate to other factors (eg, discrimination, access to care, connection/belongingness) that may contribute to increased risk of death due to external causes. Furthermore, addressing these measurement issues will ensure that studies on race/ethnicity not only identify disparities but also inform tailored approaches to reducing disparities in those affected.

A strength of our study design is that we identified a population-based sample largely consisting of post-9/11 military members with mTBI diagnosed during their military service and looked prospectively forward from the index mTBI for suicide and overdose mortality outcomes. Although many studies examining suicide or overdose mortality among post-9/11 military members or veterans are limited to those using care in either MHS or VHA,^10,11 or require longer utilization windows in these systems of care,¹² our design allowed us to observe mortality outcomes among the complete population of those exposed (ie, those who remain in MHS, those who leave military service, and either use or do not use VHA care).

This study had some limitations with regard to mTBI capture and analytic methods. Because we did not have a comparison group of military members with no mTBI, our study does not capture risk for suicide or drug overdose death relative to those with no TBI, or for those with an index moderate/severe TBI. Furthermore, while our inclusion criteria required that the index TBI be an mTBI diagnosis, we did not exclude military members with a future or undocumented TBI diagnosis that was more severe. It will be important to replicate these analyses with a comparison group and with military members with more severe TBI to understand whether there are differences in risk for these mortality outcomes specific to racial and ethnic minorities with mTBI that are different from those with moderate/severe TBI or with no TBI. This study was designed to capture mTBI that occurred during military service and did not account for lifetime history of TBI prior to or following military service, which could also impact risk trajectory. Finally, some diagnoses included in the DoD-Case-Definition are nonprecise and capture injuries that may not have been a TBI (eg, head injury, unspecified); future research should examine whether excluding of these codes results in different outcomes.^27,28 Methodologically, we were unable to present rates, especially age-adjusted rates and RR, for some groups due to small cell sizes. Although this was necessary to avoid presentation of unreliable results and preserve confidentiality, it precluded making all comparisons of interest and to fully accounting for differences in age distributions between groups despite knowing that age is strongly related to suicide and drug overdose risk. Similarly, we were unable to calculate sex-adjusted rates for each of the outcomes by race and ethnicity, given small cell sizes.

Based on inherent challenges associated with differentiating deaths by suicide and overdose that may obscure trends in mortality,⁴⁵ additional work is required to highlight factors associated with misclassification. Moreover, excessive alcohol use is consistently identified as a risk factor for suicide.⁴⁶ As such, efforts to evaluate the potential impact of excessive alcohol use on suicide and drug-/opioid-related deaths among military members with mTBI are indicated, especially since military members with deployment-acquired TBI are at an increased risk for frequent binge drinking.^47,48 Finally, collaboration between research teams will be required to combine data in a manner that will allow for exploration of drug/opioid overdoses by race/ethnicity over time, with a specific focus on potential differences between phases of the opioid epidemic.

CONCLUSION

Findings replicate and extend previous work regarding risk for suicide and opioid-related overdose deaths among those with mTBI and highlight new findings regarding the impact of race and ethnicity on mortality. All 3 mortality outcomes were consistently higher among military members who self-identified as Other, a category often selected by minoritized individuals who do not identify with single binary variables frequently provided. Finally, results provide further support for the “perfect storm” theory that persons with mTBI are at risk for negative opioid-related consequences including overdose, and that additional research is needed to increase understanding of risk factors for drug and opioid-related deaths among minoritized military members with TBI across the 3 phases of the opioid epidemic.