Substance Use in U.S. Vietnam War Era Veterans and Nonveterans: Results from the Vietnam Era Health Retrospective Observational Study

Abstract

Background: Substance use (SU) is associated with physical injury and mental health disorders in older persons, but recent research has scarcely examined SU in U.S. Vietnam-era veterans who are mostly in or near their eighth decade of life. Objectives: We compared the prevalence of self-reported lifetime and current SU and modeled current usage patterns in a nationally representative sample of veterans versus a matched nonveteran cohort. Methods: Cross-sectional, self-reported survey data were analyzed from the 2016–2017 Vietnam Era Health Retrospective Observational Study (VE-HEROeS) (n = 18,866 veterans, n = 4,530 nonveterans). We assessed lifetime and current alcohol and drug use disorders; lifetime and current use of cannabis, opioids, stimulants, sedatives, “other drugs” (psychedelics, prescription or over-the-counter drugs not prescribed/used as intended); and current SU patterns (alcohol-use-only, drug-use-only, dual-SU, no SU). Weighted descriptive, bivariable, and multivariable statistics were calculated. Covariates in multinomial modeling included sociodemographic characteristics, lifetime cigarette smoking, depression, potentially traumatic events (PTEs), and current pain (SF-8^TM). Results: Prevalence of lifetime opioid and sedative use (p ≤ .01), drug and alcohol use disorders (p < .001), and current “other drug” use (p < .001) were higher in veterans versus nonveterans. Current use of alcohol and cannabis was high in both cohorts. In veterans, very severe/severe pain, depression, and PTEs were highly associated with drug-use-only (p < .001) and dual-SU (p < .01), but these associations were fewer for nonveterans. Conclusion: This research confirmed existing concerns over substance misuse in older individuals. Vietnam-era veterans may be at particular risk due to service-related experiences and later-life tribulations. Era veterans’ unique perceptions toward healthcare assistance for SU may need greater provider focus to maximize self-efficacy and treatment.

Keywords:

• Drugs
• veterans
• Vietnam War
• alcohol
• substance use
• opioids
• sedatives
• cannabis
• marijuana

Introduction

According to the U.S. Department of Veterans Affairs, over 8.7 million Americans served during the Vietnam War between 1964 and 1973, approximately 3.4 million of whom were deployed to Southeast Asia while nearly 3.0 million of those served in the Vietnam theater of operations (Republic of Vietnam, Laos, Cambodia) (VAntage Point, 2021). A recent, unpublished analysis found that approximately two-thirds of an estimated 9.5 million Americans who served in theater or elsewhere (i.e., Vietnam-era veterans) between 1961 and 1975, were still living across our nation (T. Bullman, personal communication, March 2, 2022).

A major health concern for veterans of this war was the use of substances such as alcohol, cannabis, and heroin (Goodwin et al., 1975; Helzer et al., 1976; Mintz et al., 1979; O’Brien et al., 1980; Robins, 1974; Robins et al., 1974; 1975; Stanton, 1976; Wish et al., 1979). Reasons for this may have been associated with the availability of drugs and alcohol (Goodwin et al., 1975; Helzer et al., 1976; Horowitz & Solomon, 1975; Robins, 1974), the need to escape the stresses of combat and service overall (Helzer et al., 1976; Mintz et al., 1979; Robins, 1993; Stanton, 1976), the lack of support for the war and disillusionment with its objectives (Horowitz & Solomon, 1975; Mintz et al., 1979; Stanton, 1976; Walker & Cavenar, 1982; Wright et al., 2005), and changing sentiments at home about the acceptability of drug use (Stanton, 1976).

Certain perceptions held by Vietnam-era veterans during the war, however, still remained long after the war (Desai et al., 2016). These views include ongoing feelings of social isolation and abandonment (Desai et al., 2016; Olson, 2020), and society’s poor acknowledgement and support of their military service (Boscarino et al., 2018; Desai et al., 2016; Kaiser et al., 2017). These longstanding beliefs may underlie the development of substance use and other mental health disorders in this population (Desai et al., 2016; Wright et al., 2005) largely to manage mental health symptoms (Back et al., 2014; Beckham et al., 1997b; Bremner et al., 1996; McFall et al., 1992). Other factors may include life’s stresses after the war (i.e., difficulties maintaining relationships or employment) (Bookwala et al., 1994; Boscarino, 1981; Nace et al., 1977) and the reemergence of memories of past trauma and wartime experiences (Back et al., 2014; Beckham et al., 1997b; Davison et al., 2016; Kulka et al., 1990; Marini et al., 2020; McFall et al., 1992).

The continued study of substance use in Vietnam-era veterans is relevant because use among older persons has been increasing over time (Chhatre et al., 2017; Saxon, 2021; Upah et al., 2015; Wang & Andrade, 2013) and this trend has been associated with concomitant increased risk of overdose, physical and cognitive impairments (Azubike et al., 2021; Lynch et al., 2021; Saxon, 2021), and mortality (Lin et al., 2022; Saxon, 2021). It is also still the case that Vietnam-era veterans have not recovered from the psychological effects from the war (Corry et al., 2016; Cypel et al., 2022; Desai et al., 2016), and may be less likely than veterans from other conflicts to seek care for those issues (Desai et al., 2016; O’Malley et al., 2020).

Many of the foundational studies on substance use by Vietnam-era veterans were conducted toward or soon after the end of the war and generally based on samples that were not necessarily representative of this population (Boscarino, 1979; Goodwin et al., 1975; Helzer et al., 1976; Mintz et al., 1979; O’Brien et al., 1980; Robins, 1974; Wish et al., 1979). Later nationwide epidemiological studies examined only a limited range of substances or substance-related disorders (e.g., Centers for Disease Control Vietnam Experience Study, 1988a,b; Vietnam Era Twin Registry Study (VETR), Goldberg et al., 1990). The 1980-1985 Epidemiologic Catchment Area (ECA) study on mental health disorders (Eaton et al., 1984) only examined lifetime alcohol and drug use disorders (Norquist et al., 1990). The 1984-1988 National Vietnam Veterans Readjustment Study (NVVRS) (Kulka et al., 1990) and the 2012-2013 National Vietnam Veterans Longitudinal Study (NVVLS) (Schlenger et al., 2015a) collected data on substance use, but only the NVVRS published findings on both veterans and civilian controls, and this was just for alcohol and drug use disorders (Jordan et al., 1991; Kulka et al., 1990)

Veterans may be at greater risk of poor mental health than nonveterans because of their service-related experiences (Boscarino et al., 2018). However, past studies of substance use have not consistently shown that it has been more problematic for veterans. Thus, further research about differences between these groups is highly relevant. For example, lifetime alcohol use disorder was found to be greater in Vietnam war veterans relative to civilians (NVVRS, Jordan et al., 1991), while no significant differences were found for lifetime drug use disorder or current substance use disorders (SUDs) in that study, or for lifetime and current SUDs between Era veterans and civilians in other research (ECA Study, Norquist et al., 1990). For U.S. veterans, little or no difference was demonstrated between veterans and civilians for lifetime SUDs of various types of drugs such as stimulants (1.3% versus 1.3%, respectively), hallucinogens (.2% versus .5%), inhalants (.0% versus .1%), sedatives (.6% versus .4%), and cannabis (5.3% versus 8.0%) (National Epidemiologic Survey on Alcohol and Related Conditions-III (NESARC-III), Rhee & Rosenheck, 2019). But data from the same survey showed past-year and lifetime SUD was significantly higher (p < .001) in veterans than nonveterans (NESARC-III, Boden & Hogatt, 2018). Additionally, 2015–2017 data from the National Survey on Drug Use and Health showed mostly no significant differences between U.S. veterans and nonveterans on any illicit drug use except for cannabis (Agaku et al., 2020), and earlier 2000–2003 data from the same survey similarly showed no significant differences between cohorts for most substances except past-month marijuana use (Wagner et al., 2007).

Co-occurring alcohol and drug use (i.e., dual-substance use) have largely been unaddressed (Wachsmuth et al., 2016) with only limited findings available for the Vietnam-era veteran population (Upah et al., 2015; Wachsmuth et al., 2016). Notably, multiple substance use in older individuals is prevalent and on the rise (Saxon, 2021). One study found that admissions for alcohol and other drug use in adults at least 55 years of age increased from 16% to nearly 30% in approximately one decade (Chhatre et al., 2017). Also, NESARC-IIII 2012–2013 data revealed that US civilians 50 years or older who currently used cannabis had a higher risk of current alcohol, other drug, and tobacco/nicotine use disorders (Choi et al., 2016). Polysubstance use (i.e., the use of more than one drug including when two or more are taken together or within a short period of time intentionally or unintentionally) (Centers for Disease Control, 2022) in older persons has been associated with increased risk of pulmonary disease and drug interactions (Choi et al., 2016), and possibly worse symptoms and treatment outcomes (Teeters et al., 2017).

Other characteristics have been associated with substance use. Early-life substance use and psychosocial issues (e.g., psychiatric illness such as depression or posttraumatic stress disorder (PTSD), poor social integration, delinquency, and employment and educational status) have been examined in prior studies of these veterans (Boscarino, 1995; Bremner et al., 1996; Fischer, 1991; Helzer et al., 1976; Kulka et al., 1990; Nace et al., 1980). Mental health disorders and substance use have been frequently evidenced together in civilian (Jeffirs et al., 2019) and veteran populations (Jeffirs et al., 2019; Petrakis et al., 2011; Teeters et al., 2017). This comorbidity may become more complicated and more difficult to treat when the persons affected have less social support (Gros et al., 2016). Vietnam-era veterans may be particularly vulnerable because poor homecoming support was found to be significantly greater (p < .001) for Vietnam veterans (44.3%) relative to either Gulf (2.0%), Iraq/Afghanistan (2.1%), or other veterans (post-Vietnam conflict) (9.5%) (Boscarino et al., 2018).

Pain has also been associated with substance misuse (Azubike et al., 2021; Bennett et al., 2022; Dobscha et al., 2013; Enkema et al., 2022; Gros et al., 2015; Hasin et al., 2022; Vowles et al., 2022) and other mental health issues (Bair et al., 2003; Beckham et al., 1997a; Griffin et al., 2022; Umucu et al., 2021) in veterans of various conflicts. In addition, trauma has been positively associated with a greater likelihood of substance use (Solomon et al., 2023) and addiction (Thege et al., 2017). Veterans may have more childhood traumas (i.e., physical, sexual abuse) than nonveterans (Blosnich et al., 2014; Schultz et al., 2006). Notably, lifetime traumatic events (i.e., potentially traumatic events) were reported at relatively high frequency by Vietnam war veterans (Blosnich et al., 2022; Schnurr et al., 2002) and other older veterans (Moye et al., 2022). Furthermore, cigarette smoking/nicotine dependence has been studied in association with mental health (Beckham et al., 1997b; Lyons et al., 2008) and substance use (Alderman et al., 2000; Scherrer et al., 2008) in Vietnam-era veterans.

The scarcity of recent data on a range of substances and the extent to which they co-occur represent important research gaps in our understanding of these veterans’ health and well-being. The current analysis used data from the nationwide Vietnam Era Health Retrospective Observational Study (VE-HEROeS), the first comprehensive and large comparative survey of Vietnam-era veterans’ health in nearly 30 years (Fales et al., 2019), to: (1) obtain estimates of lifetime and current use for U.S. Vietnam-era veterans and nonveterans (i.e., by veteran status) for varied drug types and alcohol, (2) and examine the single and combined use of substances by veteran status using bivariable and multivariable models. We examined veteran status overall rather than just service in theater because all those who served were confronted with the aftermath of the war’s unique sociopolitical climate (Wright et al., 2005), as well as by exposure to military culture (e.g., stresses related to training, peer/social influences, sexual trauma).

Materials and methods

Participants and procedures

The 2016-2017 VE-HEROeS was a mail survey of the health of U.S. Vietnam-era veterans who served in the U.S. military between February 28, 1961, and May 7, 1975, and an age- and sex-matched control sample of nonveterans without U.S. military service. The veteran sampling frame (n = 9.87 million) was derived from the Department of Veterans Affairs (VA) U.S. Veterans Eligibility Trends and Statistics (USVETS) database consisting of 38 million-plus data records compiled from over 34 VA and other federal data sources (VA Information Resource Center, 2019).

A household screener was designed to develop the nonveteran sampling frame (n = 21,101) from which a control group of eligible nonveterans (i.e., born prior to 1958, no military service) was identified. Approximately 300,000 households received the screener based on address-based sampling of all U.S. postal service delivery addresses. Survey administration followed Dillman’s Tailored Design Method (Dillman et al., 2014). Veteran and nonveteran versions of the mail questionnaires underwent cognitive testing for usability and comprehension (Fales et al., 2019). Probability-based samples were created and persons who were deceased or had missing or invalid information (e.g., addresses, veteran status, age) were excluded. Questionnaires were sent to 42,393 veterans and 6,885 nonveterans. Further details about VE-HEROeS methods have been reported elsewhere (Cypel et al., 2022; Davey et al., 2023; Fales et al., 2019).

Response rates were 45.0% (n = 18,866: 6,735 theater, 12,131 nontheater) for veterans and 67.0% (n = 4,530) for nonveterans (total n = 23,396; analytical sample). Data were self-reported. All procedures were approved by the VA Central Institutional Review Board; the latter allowed for study information to be enclosed within survey mailings and a returned survey in lieu of informed consent (Davey et al., 2023).

Measures

Substance use outcomes

Lifetime and current use of five types of drugs were examined as outcomes: cannabis, opioids (e.g., heroin and other opiates), stimulants (i.e., amphetamines, cocaine), sedatives (i.e., barbiturates, tranquilizers), and “other drugs” (i.e., psychedelics and prescription or over-the-counter drugs not prescribed or used as intended). A 1988 NVVRS (Kulka et al., 1988a) question was modified and two items, one on lifetime drug use and a follow-up item on current use, were administered: (1) “The next questions ask about your experience with drugs. Please mark “Yes” or “No” to indicate if you have ever used any of the following drugs - marijuana, amphetamines, barbiturates, tranquilizers, cocaine, heroin, opiates, psychedelics, other prescription or over-the-counter drugs not prescribed for you or not used as they were intended. For these questions, drug use refers to the use of prescribed or over-the-counter drugs in excess of the directions or the use of any nonmedicinal drugs.” Respondents were asked to mark either yes or no for each drug type. The current drug use question was: “Now, think about the past 12 months. How often, if at all, have you used these types of drugs in the past 12 months?” For each drug type, these options were used: “more than twice a week”, “once or twice a week”, “two to three times a month”, “about once a month”, “less than once a month”, and “do not currently use”. Responses were recoded to “currently used” and “do not use currently.” For “currently used”, all responses except “do not currently use” were combined to create a binary variable (“use” versus “no use”).

Current alcohol use was based on the AUDIT-C (Bush et al., 1998), the three-item abbreviated version of the 10-item Alcohol Use Disorders Identification Test (Saunders et al., 1993). The AUDIT-C asked about alcohol use over the past year, has been shown to perform as well as the full 10-item test for identifying heavy drinking and alcohol use disorder (Bush et al., 1998), and has demonstrated validity (Bradley et al., 2003; 2016; Bush et al., 1998). The three items are: (1) “How often did you have a drink containing alcohol?” (“never”, “monthly or less”, “2 to 4 times a month”, “2 to 3 times a week”, and “4 or more times a week”), (2) “How many drinks containing alcohol do you have on a typical day when you are drinking?” (“1 or 2”, “3 or 4”, “5 or 6”, “7, 8, or 9”, and “10 or more”), and (3) “How often do you have six or more drinks on one occasion?” (“never”, “less than monthly”, “monthly”, “weekly”, and “daily or almost daily”). Current alcohol use was defined dichotomously as either reporting use of alcohol in the past year or not. Lifetime alcohol use was not asked in VE-HEROeS.

Alcohol use disorder, per the AUDIT-C, was examined as both continuous and dichotomous variables. The AUDIT-C score was calculated by summing the scores on three questions, with each of their respective response options scored from 0–4. Total scores ranged from 0 to 12, with higher scores suggesting greater risk of alcohol misuse. For the dichotomous variable, males having a total score of 4 or more (cutpoint) were assigned a “1” (i.e., alcohol use disorder) (Bradley et al., 2016), while those who scored lower than the cutpoint were assigned a “0”. For females, a cutpoint of 3 or more signified alcohol use disorder (Bradley et al., 2016). Lifetime “alcohol and drug dependence” (per the VE-HEROeS questionnaire) was derived from the National Health Interview Survey’s (NHIS) dichotomous item on whether a health professional ever told the veteran that he/she had an alcohol/drug dependency (Centers for Disease Control, 2013).

Current substance use patterns included four mutually exclusive patterns: (1) current alcohol-use-only, (2) current drug-use-only (i.e., current use of one or more types of drugs), (3) current drug use and current alcohol use combined (i.e., dual-substance-use—whether alcohol and drugs were both reported in the past 12 months and not whether they were used at the same time), and (4) no current substance use.

Key explanatory variable

Veteran status was the key independent variable and was defined dichotomously as Vietnam War Era service in the U.S. military (”veteran”) or no military service (“nonveteran”). Veteran status was ascertained using USVETS data and defined with the VE-HEROeS’ question: “During your service, did you serve on the ground, in the air, or in the inland waters in any of the following areas between 1961 and 1975? (1) South Vietnam (Republic of Vietnam), Laos, or Cambodia; 2) North Vietnam (Democratic Republic of Vietnam); 3) Southeast Asia other than Vietnam, Laos, or Cambodia; 4) Asia other than Vietnam or Southeast Asia; 5) Europe; 6) U.S.; 7) Other.” Nonveterans were U.S. civilians with no military service whose status was determined via the VE-HEROeS’ U.S. household screener.

Other explanatory variables

Lifetime depression and PTSD questions came from the NHIS yes/no items that asked whether a health professional ever told the individual that he/she had the condition (Centers for Disease Control, 2013). The Brief Trauma Questionnaire, shown to have good interrater reliability (Kappa = 0.89) and criterion validity, screened for ten types of lifetime, potentially traumatic events (PTEs) (Schnurr et al., 2002) and were summed to obtain a count of the number of types of PTEs/individual. The pain variable was derived from one of the Short-Form Health Survey’s (SF-8^TM) eight subscales that assess quality-of-life (e.g., “How much bodily pain have you had during the past 4 weeks?”) (Ware et al., 2001). The SF-8^TM’s eight subscales were shown to have good test/retest reliability, ranging from 0.59–0.70 (Ware et al., 2001). Pain responses were classified as very severe/severe, mild/moderate, and very mild/none. Lifetime cigarette use (i.e., ever smoker versus never smoker) was based on 2013 NHIS questions (Centers for Disease Control, 2013).

Sociodemographic covariates were age (in years) at time of survey, sex, race/ethnicity, marital status, education, annual income, and employment status. Age was assessed as both continuous and categorical variables. Sex was based on sex assigned at birth (binary variable: male, female). Race/­ethnicity was categorized as White, Black, other race, and Hispanic. White and Black races, respectively, comprised non-Hispanic single race persons while “other race” represented multiracial non-Hispanic persons. “Hispanic” included persons who only reported being of Hispanic, Latino, or Spanish origin, or multiracial persons who also reported being of Hispanic, Latino, or Spanish origin. Current marital status, originally a six-response option question, was collapsed into a dichotomous variable (“Married/with partner” versus “Not married or partnered”). The education (lifetime) question, which had 10 response options, was evaluated as a three-level categorical variable: “high school or less/General Educational Diploma (GED)”, “Some college/vocational”, and “Bachelor’s or more.” Current annual income was analyzed as a four-level categorical variable (continuous data were not collected): “<$15,000”, “$15,000-$49,999”, “$50,000-$99,999”, and $100,000 or more”. Current employment status was assessed as either “paid work” or “not working.”

Statistical analysis

Responses were weighted to account for the complex sampling design and nonresponse. Jackknife repeated replication estimated the sampling variability of statistics generated from survey data. Descriptive statistics were computed. For current drug use and use of different drug types, we used the count of those who ever used drugs and those who ever used a specific type of drug (e.g., opioids) respectively as denominators. P-values were based on the design–adjusted, second-order Rao-Scott chi-square test in contingency table analyses (Heeringa et al., 2017). Tukey-Kramer adjustments were made to p-values to reduce multiplicity (the inflation of Type I error with repeated comparisons) (SAS Institute, 2017; Streiner & Norman, 2011). P-values less than .05 were deemed statistically significant, and two-tailed tests were employed. Analyses were performed using SAS Enterprise Guide (Version 8.2, SAS Institute, Inc., Cary, NC, USA) for Linux.

Simple logistic regression was used to calculate crude odds ratios (ORs) and 95% confidence intervals (CI) for associations between binary, single substance outcomes and veteran status. ORs were converted to Cohen’s d to assess effect size (Ferguson, 2009; Sanchez-Meca et al., 2003) where d ≤ 0.2, 0.5, and 0.8 were small, medium, and large effect sizes respectively (Cohen, 1988). For “current substance use patterns,” multinomial logistic regression computed ORs and adjusted odds ratios (AORs) with their concomitant 95% CIs, where “no current substance use” served as the referent category. To obtain AORs, current substance use pattern was regressed onto veteran status, age, sex, race/ethnicity, marital status, education, income, employment status, cigarette smoking, bodily pain, depression, and PTEs. Domain (subpopulation) analysis was used to obtain estimates for veterans and nonveterans separately (Heeringa et al., 2017). ORs or AORs exceeding 3.0 (or less than 0.33) were considered strong (Haddock et al., 1998). For continuous covariates, unstandardized regression coefficients (B) and standard errors were reported. Unweighted linear regression tested for multicollinearity (where tolerance <0.40) to identify linear dependencies among the explanatory variables (Allison, 2012). All tolerances fell above 0.70 (not shown in tables).

The covariates included in regression were evaluated in prior substance use research. Sociodemographic characteristics, mental health, pain, and trauma had been evaluated in other substance use studies of Vietnam-era veterans (Boscarino, 1979; 1981; 1995) and older civilians (Salas-Wright et al., 2017). We retained depression and PTEs over PTSD because depression and PTSD co-occur (Ginzburg et al., 2010; Goetter et al., 2020), and the greatest change in AORs during model development happened with the addition of depression (not shown). Also, PTSD was more strongly correlated with PTEs compared to depression in our preliminary work, and both depression (Fu et al., 2007; Helzer et al., 1979; Nace et al., 1977; 1980) and PTEs were found to be relevant to substance use (Blosnich et al., 2014; Boscarino et al., 2011; Panza et al., 2022). Pain was included in our model because of its association with substance use (Azubike et al., 2021; Dobscha et al., 2013; Gros et al., 2015; Hasin et al., 2022; MacLean et al., 2023; Reif et al., 2022; Vowles et al., 2022) and depression (Bair et al., 2003; Dobscha et al., 2013; Griffin et al., 2022). Cigarette smoking was a covariate because of its relationship to other substance use among veterans (Dobscha et al., 2013; Nieh et al., 2021).

Results

Mean (SE) age was 68.5 (0.01) and 66.7 (0.11) years for veterans and nonveterans respectively (p < .001) (Table 1). Respondents were predominantly White NonHispanic (83.2%, veterans; 85.7%, nonveterans), male (96.8%, veterans; 85.3%, nonveterans), and married/partnered (76.6%, veterans; 76.8%, nonveterans). Veterans were less likely (p < .001) to have earned a bachelor’s degree or higher (29.6%, veterans; 46.8%, nonveterans) and more likely to have incomes below $50,000 (51.3%, veterans; 41.5%, nonveterans). Veterans were more likely (p < .001) to report very severe/severe bodily pain (12.3%), depression (27.6%), PTSD (16.3%), and PTEs (Mean = 1.92) than nonveterans (5.9%, 18.7%, 2.7%, Mean = 1.14 for pain, depression, PTSD, and PTEs, respectively).

Table 1. Characteristics of Vietnam-era veteran and nonveteran respondents.

Characteristic	Veterans (n = 18,866)		Nonveterans (n = 4,530)		p-value^b
	n (%)	SE	n (%)	SE
Sex					<.001
Male	18,400 (96.8)	0.05^a	4,246 (85.3)	0.87^a
Female	466 (3.2)		284 (14.7)
Age, years					<.001
58-65	4,402 (29.4)	0.13	2,442 (54.4)	0.75
66-70	8,296 (42.6)	0.21	1,028 (22.4)	0.61
71-75	4,127 (19.4)	0.17	561 (11.3)	0.48
76-80	1,410 (5.4)	0.09	314 (6.4)	0.39
> 80	631 (3.2)	0.07	185 (5.5)	0.36
Age Mean SE^c	68.5	0.01	66.7	0.11
Race/ethnicity					<.001
White NonHispanic	15,730 (83.2)	0.32	3,866 (85.7)	0.51
Black NonHispanic	1,462 (8.0)	0.23	244 (5.8)	0.32
Other race NonHispanic	669 (3.8)	0.16	216 (4.8)	0.31
Hispanic	816 (4.9)	0.18	157 (3.6)	0.25
Education, lifetime					<.001
High school or less/GED	4,874 (26.6)	0.34	1,062 (25.0)	0.72
Some college/vocational	7,960 (43.9)	0.41	1,250 (28.2)	0.77
Bachelor’s or more	5,665 (29.6)	0.32	2,183 (46.8)	0.77
Marital status, current					.788
Married/with partner	14,358 (76.6)	0.29^a	3,500 (76.8)	0.67^a
Not married or partnered	4,226 (23.4)		960 (23.2)
Income, current					<.001
< $15,000	1,291 (7.8)	0.22	392 (10.2)	0.48
$15,000-$49,999	7,763 (43.5)	0.37	1,341 (31.3)	0.70
$50,000-$99,999	6,133 (34.1)	0.36	1,437 (32.2)	0.67
$100,000 or more	2,574 (14.6)	0.21	1,200 (26.3)	0.66
Employment status, current					<.001
Paid work	5,336 (30.8)	0.32^a	2,028 (44.5)	0.71^a
Not working	13,022 (69.2)		2,435 (55.5)
Cigarette smoking, lifetime					<.001
Yes	12,702 (68.8)	0.40^a	2,248 (49.2)	0.77^a
No	5,836 (31.2)		2,246 (50.8)
SF-8^TM bodily pain, current					<.001
Very mild/none	6,970 (36.9)	0.35	2,336 (50.2)	0.66
Moderate/mild	9,490 (50.8)	0.35	1,930 (43.8)	0.69
Very severe/severe	2,229 (12.3)	0.24	252 (5.9)	0.39
Depression, lifetime
Yes	5,076 (27.6)	0.35^a	785 (18.7)	0.70^a
No	13,790 (72.4)		3,745 (81.3)
PTEs, lifetime Mean SE^c	1.92	0.01	1.14	0.02
PTSD, lifetime					<.001
Yes	3,054 (16.3)	0.30^a	102 (2.7)	0.28^a
No	15,812 (83.7)		4,428 (97.3)

Note. GED, General Educational Diploma; PTE, number of potentially traumatic events; PTSD, posttraumatic stress disorder; SE, standard error; SF-8^TM, Short-Form Health Survey (Ware et al., 2001). All statistics except counts were weighted. Missing data ranged from 0–5%.

^aSEs for the levels of binary variables were equivalent and not repeated.

^bp-values per the second-order design-adjusted Rao-Scott chi-square test (Heeringa et al., 2017).

^cDifference in means between cohorts was statistically significant by Tukey (p < .001).

When comparing veterans to nonveterans, the prevalence of lifetime use was higher for veterans (p < .01) for opioids (16.4% versus 13.3% respectively) and sedatives (18.0% versus 16.5%), while “other drug” use was lower in veterans (14.3% versus 15.7%, p = .02) (Table 2). Veterans, relative to nonveterans, had greater (p < .001) prevalence of lifetime alcohol (11.9% versus 4.3%) and drug use disorders (5.1% versus 1.9%). Of veterans who reported use of one or more drugs during their lifetime, cannabis was the most frequently reported (42.4%) followed by stimulants (20.6%), sedatives (18.0%), opioids (16.4%), and “other drugs” (14.3%).

Table 2. Lifetime and current substance use among Vietnam-era veterans and nonveterans – prevalence, patterns, and crude effects.

	Veterans (n = 18,866)			Nonveterans (n = 4,530)				p-value	OR (95% CI)	Cohen’s d
	n	%	SE	n		%	SE
Any drug use^a, lifetime	8,777	49.1	0.34		2,207	48.7	0.74	.603	1.02 (0.95, 1.09)	0.01
Drug types, lifetime^a
Cannabis^a	7,470	42.4	0.32		1,959	42.7	0.75	.733	0.99 (0.93, 1.06)	0.01
Opioids	2,893	16.4	0.29		592	13.3	0.51	<.001	1.28 (1.16, 1.40)***	0.14
Stimulants	3,533	20.6	0.28		979	21.9	0.67	.065	0.93 (0.85, 1.00)	0.04
Sedatives	3,167	18.0	0.29		735	16.5	0.55	.010	1.12 (1.02, 1.22)*	0.06
Other drugs	2,432	14.3	0.27		724	15.7	0.55	.024	0.90 (0.82, 0.99)*	0.06
Substance dependencies, lifetime
Drug dependence	873	5.1	0.17		89	1.9	0.19	<.001	2.70 (2.20, 3.32)***	0.55
Alcohol dependence	2,162	11.9	0.23		205	4.3	0.30	<.001	3.03 (2.62, 3.52)***	0.61
Any drug use, current^b	2,804	32.4	0.54		665	30.3	1.02	.060	1.10 (1.00, 1.22)	0.05
Alcohol use, current^c	13,314	70.7	0.32		3,333	72.7	0.77	.016	0.90 (0.83, 0.98)*	−0.06
AUDIT-C score Mean SE^d, B (95% CI)	2.50	0.02	─		2.50	0.03	0.04	.082	0.01 (-0.07, 0.09)	─
AUDIT-C, % over male cutpoint	5,307	30.5	0.43		1,418	33.9	0.64	<.001	0.85 (0.80, 0.91)***	−0.09
AUDIT-C, % over female cutpoint	104	23.2	2.15		81	29.2	3.25	.108	0.73 (0.50, 1.07)	−0.17
Current drug use, of those reporting lifetime use of a specified drug
Cannabis	1,530	21.5	0.52		403	21.0	1.01	.647	1.03 (0.90, 1.18)	0.02
Opioids	783	28.0	0.78		157	27.2	1.93	.682	1.04 (0.85, 1.28)	0.02
Stimulants	147	4.2	0.36		51	5.3	0.59	.117	0.79 (0.60, 1.05)	−0.13
Sedatives	976	30.6	1.03		191	28.1	1.83	.254	1.13 (0.91, 1.40)	0.07
Other drugs	239	9.9	0.54		45	6.6	0.87	<.001	1.56 (1.16, 2.10)**	0.25
Substance use patterns, current^e
No current substance use	4,929	25.8	0.29		1,088	24.9	0.75	─	REFERENT	─
Current alcohol-use-only	11,133	58.3	0.34		2,777	60.3	0.85	.123	0.93 (0.86, 1.02)	−0.04
Current drug-use-only	623	3.5	0.14		109	2.4	0.24	.002	1.44 (1.14, 1.81)**	0.20
Dual-substance-use	2,181	12.4	0.25		556	12.4	0.53	.559	0.97 (0.85, 1.09)	−0.02

Note: AUDIT-C, Alcohol Use Disorders Identification Test (3-item); B, unstandardized regression coefficient; CI, confidence interval; d, Cohen’s d = effect size index that converts the odds ratio to d where $d=\mathit{lnOR}\left(\frac{\sqrt{3}}{\pi }\right)$ (Sánchez-Meca et al., 2003). Absolute values of d approximating 0.2, 0.5, 0.8 were considered small, medium, and large effect sizes respectively (Cohen, 1988); OR, crude (unadjusted) odds ratio; SE, standard error; VE-HEROeS, Vietnam Era Health Retrospective Observational Study; ─, statistic not available. Statistics are weighted except for counts. Variance estimated by jackknife repeated replication. 

^a“Any drug use” is a report of ever using 1 or more types of drugs that included opioids (heroin and other opiates), stimulants (amphetamines, coke), sedatives (tranquilizers, barbiturates), cannabis, and other drugs (psychedelics and other prescription drugs or over the counter drugs not prescribed or not used as intended). Drug type categories, such as cannabis (‘marijuana,’ per the VE-HEROeS questionnaire) or opioids, were not mutually exclusive; the sum of the individual counts (or %s) of specific types of drugs will not add to the total (or 100%).

^bCurrent drug use is drug use in the past 12 months (with or without the use of alcohol).

^cCurrent alcohol use is use of alcohol (with or without other drugs) in the past year. No data on lifetime alcohol use was collected by VE-HEROeS.

^dDifference between mean scores was not significant by Tukey (p = .82).

^eSubstance use pattern categories are mutually exclusive. “Current drug-use-only” includes current use of 1 or more drug types without current alcohol use. “Current alcohol-use-only” is current use of alcohol without current drug use. Dual-substance-use includes current drug and alcohol use. “No current use” includes those who report no current drug or alcohol use. The crude association between current substance use pattern and veteran status was significant (p = .002) by the Rao-Scott design-adjusted, second-order chi-square test.

*p < .05; **p < .01; ***p < .001.

Of veterans who reported use of a specific type of drug during their lifetime, current use of sedatives (30.6%) and opioids (28.0%) was the highest followed by cannabis (21.5%), “other drugs” (9.9%), and stimulants (4.2%) (Table 2). Current alcohol use (with or without other substance use) was most reported in veterans and nonveterans, averaging approximately 70%. Only the current use of “other drugs” was significantly (p <.01) higher for veterans (9.9%) than nonveterans (6.6%) (OR = 1.56, 95% CI:1.16-2.10, Cohen’s d = 0.25). For veterans and nonveterans, respectively, the type of “other drug” that was used currently and predominantly reported was other prescription or over-the-counter drugs that were not prescribed or not used as intended (not shown in table).

Among veterans, significant positive associations were found between drug-related patterns and pain, depression, and PTEs (pain: drug-use-only, AOR = 2.30, 95% CI: 1.56–3.41, p < .001, dual-substance-use, AOR = 1.52, 95% CI: 1.15–2.01, p < .01; depression: drug-use-only, AOR = 2.10, 95% CI: 1.69–2.61, dual-substance-use, AOR = 1.44, 95% CI: 1.27–1.64, p < .001; PTEs: drug-use-only, B(SE)=0.099 (.028), p < .001, dual-substance-use, B(SE)=0.099(.02), p < .001) (Table 3). AORs for current alcohol use (without other substance use) among veterans fell below 1.0 (p < .001) for pain and depression. For nonveterans (Table 4), the number of associations with drug-related patterns for these specific characteristics were fewer (pain: drug-use only, AOR = 3.19, 95% CI: 1.35-7.56, p < .01; depression, drug-use-only, AOR = 2.34, 95% CI: 1.44-3.80, p < .001); PTEs, drug-use-only, B(SE)=0.30(0.06), p < .001, dual-substance-use, B(SE) = 0.13(0.05), p < .01).

Table 3. Association between substance use pattern and veteran characteristic, multinomial logistic regression results.

	Alcohol-use-only		Drug-use-only		Dual-substance-use
Explanatory variable	AOR	95% CI	AOR	95% CI	AOR	95% CI
Age B, SE	−0.025***	0.004	−0.072***	0.013	−0.106***	0.008
Black NonHispanic: Other NonHispanic	1.04	[0.75, 1.45]	0.76	[0.37, 1.57]	1.15	[0.74, 1.77]
Black NonHispanic: Hispanic	0.61***	[0.47, 0.79]	0.41**	[0.22, 0.78]	0.61*	[0.40, 0.93]
Black NonHispanic: White NonHispanic	0.73***	[0.62, 0.86]	0.66	[0.43, 1.01]	0.85	[0.65, 1.12]
Other NonHispanic: Hispanic	0.58**	[0.40, 0.85]	0.54	[0.27, 1.09]	0.53**	[0.33, 0.85]
Other NonHispanic: White NonHispanic	0.70**	[0.52, 0.94]	0.87	[0.44, 1.74]	0.74	[0.51, 1.09]
Hispanic: White NonHispanic	1.20	[0.96, 1.50]	1.60	[0.91, 2.81]	1.41*	[1.01, 1.97]
Male: Female	1.56***	[1.24, 1.96]	1.81	[0.99, 3.31]	2.52***	[1.71, 3.71]
Married-Partner: No	0.95	[0.86, 1.06]	0.76*	[0.62, 0.94]	0.69***	[0.60, 0.79]
$100,000 or more: $15,000-$49,999	2.14***	[1.75, 2.61]	1.30	[0.79, 2.13]	1.65***	[1.22, 2.22]
$100,000 or more: $50,000-$99,999	1.47***	[1.22, 1.77]	1.17	[0.69, 2.00]	1.33	[0.95, 1.85]
$100,000 or more: < $15,000	2.57***	[2.00, 3.31]	1.00	[0.48, 2.09]	1.30	[0.84, 2.01]
$15,000-$49,999: $50,000-$99,999	0.69***	[0.59, 0.80]	0.91	[0.66, 1.24]	0.80*	[0.67, 0.97]
$15,000-$49,999: < $15,000	1.20	[0.96, 1.51]	0.77	[0.50, 1.19]	0.79	[0.59, 1.07]
$50,000-$99,999: < $15,000	1.76***	[1.38, 2.24]	0.86	[0.52, 1.40]	0.98	[0.73, 1.33]
High school or less/GED: Some college/vocational	0.81***	[0.75, 0.89]	1.13	[0.86, 1.50]	0.73***	[0.61, 0.88]
High school or less/GED: Bachelor’s or more	0.67***	[0.60, 0.75]	1.40	[0.94, 2.10]	0.65***	[0.52, 0.82]
Some college: Bachelor’s or more	0.83***	[0.74, 0.92]	1.24	[0.86, 1.80]	0.90	[0.76, 1.06]
Employed: No	0.94	[0.86, 1.03]	0.61***	[0.47, 0.79]	0.81***	[0.72, 0.91]
Ever smoker: No	1.37***	[1.26, 1.48]	1.96***	[1.53, 2.52]	2.12***	[1.90, 2.36]
Moderate/mild pain: Very severe/severe pain	1.51***	[1.28, 1.77]	0.60***	[0.46, 0.78]	0.86	[0.71, 1.04]
Moderate/mild pain: Very mild pain/no pain	0.92	[0.83, 1.01]	1.38	[0.94, 2.01]	1.30**	[1.07, 1.58]
Very severe/severe pain: Very mild pain/no pain	0.61***	[0.52, 0.71]	2.30***	[1.56, 3.41]	1.52**	[1.15, 2.01]
Lifetime depression: No	0.77***	[0.70, 0.85]	2.10***	[1.69, 2.61]	1.44***	[1.27, 1.64]
Lifetime traumatic events (PTEs) B, SE	−0.012	0.014	0.099***	0.028	0.099***	0.019

Notes: AOR, adjusted odds ratio; B, unstandardized regression coefficient; CI, confidence interval; GED, General Educational Diploma; PTEs, potentially traumatic events; SE, standard error. Statistics are weighted. Variance estimated by jackknife repeated replication. The association between substance use pattern and veteran status was statistically significant (p < .01). Domain statement was applied in multinomial logistic regression (Heeringa et al., 2017). Referent = no current substance use.

*p < .05; **p < .01; ***p < .001.

Table 4. Association between substance use pattern and nonveteran characteristic, multinomial logistic regression results.

	Alcohol-use-only		Drug-use-only		Dual-substance-use
Explanatory variable	AOR	95% CI	AOR	95% CI	AOR	95% CI
Age B, SE	−0.027***	0.008	−0.071**	0.021	−0.102***	0.015
Black NonHispanic: Other NonHispanic	1.26	[0.71, 2.22]	1.01	[0.19, 5.41]	2.34*	[1.06, 5.14]
Black NonHispanic: Hispanic	0.40*	[0.20, 0.77]	0.21*	[0.05, 0.92]	0.81	[0.28, 2.33]
Black NonHispanic: White NonHispanic	0.60*	[0.37, 0.95]	0.61	[0.18, 2.09]	0.89	[0.53, 1.48]
Other NonHispanic: Hispanic	0.32***	[0.17, 0.60]	0.21	[0.04, 1.03]	0.35	[0.11, 1.06]
Other NonHispanic: White NonHispanic	0.48***	[0.32, 0.71]	0.60	[0.18, 2.02]	0.38***	[0.21, 0.69]
Hispanic: White NonHispanic	1.51	[0.89, 2.54]	2.90	[0.96, 8.83]	1.10	[0.41, 2.96]
Male: Female	1.33	[0.95, 1.86]	1.71	[0.82, 3.55]	1.47	[0.85, 2.55]
Married-Partner: No	0.97	[0.80, 1.17]	0.75	[0.43, 1.30]	0.62***	[0.48, 0.81]
$100,000 or more : $15,000-$49,999	2.63***	[1.98, 3.50]	0.98	[0.37, 2.60]	2.23***	[1.30, 3.81]
$100,000 or more: $50,000-$99,999	1.74***	[1.31, 2.32]	1.41	[0.46, 4.27]	1.55*	[1.01, 2.39]
$100,000 or more: < $15,000	3.51***	[2.09, 5.91]	0.75	[0.23, 2.40]	2.10*	[1.04, 4.25]
$15,000-$49,999: $50,000-$99,999	0.66**	[0.50, 0.88]	1.44	[0.62, 3.33]	0.70	[0.43, 1.12]
$15,000-$49,999: < $15,000	1.34	[0.84, 2.13]	0.76	[0.38, 1.55]	0.94	[0.54, 1.67]
$50,000-$99,999: < $15,000	2.02***	[1.30, 3.13]	0.53	[0.18, 1.53]	1.35	[0.72, 2.54]
High school or less/GED: Some college/vocational	0.74*	[0.55, 0.98]	0.90	[0.45, 1.79]	0.49***	[0.34, 0.72]
High school or less/GED: Bachelor’s or more	0.63***	[0.48, 0.83]	0.82	[0.39, 1.71]	0.35***	[0.22, 0.55]
Some college: Bachelor’s or more	0.86	[0.66, 1.11]	0.91	[0.49, 1.70}	0.70*	[0.50, 0.99]
Employed: No	0.84*	[0.71, 0.99]	0.72	[0.48, 1.08]	0.74*	[0.55, 0.98]
Ever smoker: No	1.58***	[1.33, 1.88]	2.59***	[1.59, 4.23]	3.37***	[2.59, 4.38]
Moderate/mild pain: Very severe/severe pain	1.79**	[1.13, 2.84]	0.55	[0.28, 1.09]	1.35	[0.77, 2.39]
Moderate/mild pain: Very mild pain/no pain	0.85	[0.68, 1.06]	1.75	[0.88, 3.49]	1.66***	[1.24, 2.22]
Very severe/severe pain: Very mild pain/no pain	0.47**	[0.29, 0.78]	3.19**	[1.35, 7.56]	1.22	[0.69, 2.19]
Lifetime depression: No	0.79	[0.61, 1.02]	2.34***	[1.44, 3.80]	1.34	[0.94, 1.93]
Lifetime traumatic events (PTEs) B, SE	0.04	0.037	0.30***	0.060	0.13**	0.045

Notes: AOR, adjusted odds ratio; B, unstandardized regression coefficient; CI, confidence interval; GED, General Educational Diploma; PTEs, potentially traumatic events; SE, standard error. Statistics are weighted. Variance estimated by jackknife repeated replication. The association between substance use pattern and veteran status was statistically significant (p < .01). Domain statement was applied in multinomial logistic regression (Heeringa et al., 2017). Referent = no current substance use.

*p < .05; **p < .01; ***p < .001.

For veterans and nonveterans, cigarette smoking and age were consistently associated with all three substance use patterns (Tables 3 and 4). For every 1-unit increase in age, there was at least about a 3% concomitant decline (p < .001) in each of the current substance use patterns in both cohorts (veterans: alcohol-use-only, B(SE) = -0.025(0.004); drug-use-only, B(SE) = -0.072(0.013); dual-substance-use, B(SE) = -0.106(0.008); nonveterans: B(SE) = -0.027(0.008); drug-use-only, B(SE) = -0.071(0.021); dual-substance-use, B(SE) = -0.102(0.015)). For cigarette smoking, the estimated odds of current substance use (relative to no substance use) in ever smokers were approximately 1.5 to 3.5 times those for nonsmokers (p < .001) (veterans: alcohol-use-only, AOR = 1.37, 95% CI: 1.26–1.48; drug-use-only, AOR = 1.96, 95% CI: 1.53–2.52; dual-substance-use, AOR = 2.12, 95% CI: 1.90–2.36; nonveterans: alcohol-use-only, AOR = 1.58, 95% CI: 1.33–1.88; drug-use-only, AOR = 2.59, 95% CI: 1.59–4.23; dual-substance-use, AOR = 3.37, 95% CI: 2.59–4.38).

Discussion

Lifetime and current substance use between veterans and nonveterans differed little with some exceptions. Lifetime opioid and sedative use, drug and alcohol use disorders, and current use of “other drugs” were higher for veterans than nonveterans. Pain, depression, and PTE’s generally played more prominently in veterans’ substance use compared to nonveterans. Cannabis and alcohol were still commonly used in older veterans and nonveterans alike, but current trends overall revealed a greater involvement with sedatives and opioids and possibly prescription drug misuse.

This analysis revealed depression’s significant association with drug-related patterns, consistent with much earlier studies of Vietnam-era veterans (Helzer et al., 1979; Nace et al., 1977; 1980; Robins, 1974). Findings on male veteran twins derived from the VETR indicated that associations between mental health conditions and substance use may have genetic underpinnings (Fu et al., 2002; 2007). Some of that evidence relates to the comorbidity of major depression and alcohol dependence (Fu et al., 2002), although we found that among veterans the odds of current alcohol use (without other substance use) were significantly (p < .001) lower for those with lifetime depression than those without depression (Table 3). Genetic links have also been suggested for depression’s association with nicotine dependence in VETR participants (Fu et al., 2007). In other veterans, major depression has been associated with opioid (Dobscha et al., 2013; Rhee & Rosenheck, 2019) and cannabis use disorder (Hill et al., 2021).

Pain has also been known to co-occur with depression, and their joint existence has been referred to as the “depression-pain dyad” (Bair et al., 2003). Overall, complex, and still as of yet undetermined interrelationships exist among pain, depression, and substance use (Maleki & Oscar-Berman, 2020; Morasco et al., 2011). This topic has been addressed to some extent in veterans (Bennett et al., 2019; Boscarino et al., 2011; Dobscha et al., 2013; Irwin et al., 2014; MacLean et al., 2023; Vowles et al., 2022), but further research on modeling relationships with PTSD may be needed (Otis et al., 2003). These interrelationships in our analysis were suggested by results for veterans that showed significant associations for pain and depression with current drug-use-only and dual-use patterns.

Although veterans and nonveterans showed significant positive associations with pain in our analysis, veterans had a greater number of these associations that may relate back to their military service and physical health status. Poor physical health was reported for Vietnam-era veterans (Centers for Disease Control Vietnam Experience Study, 1988b; Eisen et al., 1991); comparisons were not made to nonveterans in those studies. U.S. veterans had more chronic pain than nonveterans even after adjusting for age and other sociodemographic characteristics (24.7% versus 17.4% respectively, p < .001) (Enkema et al., 2022). Physical health quality of life was negatively correlated (r = -0.56, p < .001) with pain level in veterans with mental illness − 10.3% of whom reported SUDs (Umucu et al., 2021).

Associations have also been noted between chronic pain and opioid use in other studies of veterans (Dobscha et al., 2013; MacLean et al., 2023) and older persons (Azubike et al., 2021; Galicia-Castillo, 2016; Lynch et al., 2021; Saxon, 2021). In the U.S. there has been a tripling of older persons seeking treatment for opioid-related disorders between 2007 and 2017 (Lynch et al., 2021). This is concerning given the relatively high proportion of current opioid use among drug users in our analysis. Specifically, the percentage who reported currently using opioids of those who ever reported lifetime use of that drug was relatively high in comparison to cannabis, stimulants, and “other drugs” in both veterans and nonveterans. We estimated that approximately 28% of Vietnam-era veterans used opioids currently (and 27% in nonveterans). Past-year opioid use by older veterans (mean age = 69.2 years) who were enrolled in a cannabis patient program was 39.3% (Kang et al., 2021), while it was approximately 40.0% for past-year use of “prescription pain relievers” (details not provided) by U.S. veterans aged 50 years and over (Agaku et al., 2020). Overall, use of opioids by older adults may be reaching epidemic levels (Saxon, 2021). This speaks to the importance of current VA initiatives, namely VA’s Stratification Tool for Opioid Risk Mitigation (STORM) (Oliva et al., 2017), that aim to lessen risks for misuse and overdose in a population for whom these risks may be more likely (Saxon, 2021).

Vietnam-era veterans reported a greater number of PTEs than nonveterans in our analysis, although adjusted findings showed that PTEs were associated with current drug use and dual-substance use for Vietnam-era veterans and nonveterans alike. For veterans, early adverse childhood experiences could exacerbate later-life trauma that may ultimately pose an elevated risk of poor mental health consequences (Blosnich et al., 2014; Panza et al., 2022). Some have posited that childhood traumas may weaken neural function and thereby increase susceptibility to SUD and other psychiatric disorders in adulthood (Khoury et al., 2010). Among Vietnam-era veterans, nearly 30% on average reported being physically abused by a parent, caretaker, or teacher before the age of 18 (Blosnich et al., 2022). Veterans of a similar generation also reported childhood physical abuse and any-age sexual assault, although these traumas were less frequently reported than other traumas (Schnurr et al., 2002). And as adults during their military service, Vietnam-era veterans have likely experienced reemerging, negative war-related memories and traumas (Davison et al., 2016; Marini et al., 2020) that may heighten their vulnerability to substance-related problems. Interestingly, a systematic review of the literature on the relationship between trauma and later-life addictive behaviors found that 64% of the nearly 190 observational studies examined had no significant findings between these events and behaviors, while most of the remaining studies showed positive associations. Only 1% showed a protective effect for trauma exposure (Thege et al., 2017).

Our results indicated that veterans may have a potentially greater risk of current prescription drug misuse than nonveterans. The number and types of medications administered to older individuals, particularly those with dual mental health and substance use diagnoses, may increase health risks due to overmedication, adverse drug reactions, SUDs (Bair et al., 2003), or mortality. Substance-related mortality has minimally been addressed for Vietnam-era veterans with earlier efforts directed generally at mortality from accidental poisoning (Boehmer et al., 2004; Centers for Disease Control Vietnam Experience Study, 1987; Schlenger et al., 2015b) and drugs and/or alcohol specifically (Boehmer et al., 2004; Centers for Disease Control, 1987; Price et al., 2001). One study examined early drug use (i.e., before, during, and soon after Vietnam War service) and its association to overall mortality among U.S. Army veterans (Price et al., 2001). Another analysis of data on Vietnam-era U.S. Army veterans examined alcohol consumption and alcohol use disorder in relation to overall mortality (Lundin & Mortensen, 2015). Alcohol-related mortality has markedly increased among older individuals (Lin et al., 2022; Saxon, 2021), and the high levels of use in our study and in the general US population (National Institute on Alcohol Abuse and Alcoholism, 2022) make this another area of concern.

Vietnam-era veterans’ receptivity to assistance from mental healthcare practitioners may compromise management of SUDs and their treatment. In a qualitative assessment of the perceptions of these veterans to seeking mental health care, these veterans had deep-rooted negative perceptions regarding society’s rejection of their military service and service-related health needs (Desai et al., 2016). Veterans also were concerned about being stigmatized over having “weaknesses” (i.e., “personal/social difficulties”) (Desai et al., 2016). Stigmatization is not uncommon in healthcare and worsens mental health outcomes by diminishing individuals’ sense of social support and beliefs about their abilities to manage their condition (Eagen et al., 2022). Little research has been devoted to examining these types of perceptions as they relate to mental health issues in these veterans (Desai et al., 2016). Therefore, providers must consider these veterans’ unique perceptions toward healthcare assistance to maximize self-efficacy and treatment (Desai et al., 2016).

Limitations require discussion. The cross-sectional design allows no conclusions about causality. The extent of social-desirability bias, the tendency to under-report undesirable behaviors, such as substance use, is unknown in this population. Survey weights were originally constructed to match groups by sex and age, but they did not fully account for differences by these demographics; thus, age and sex were added to regressions to compensate for discrepancies in proportions across cohorts. Although the use of alcohol and drugs was evaluated, we did not have data on whether these substances were actually used at the same time. Confounders potentially relevant to later life usage patterns, such as histories of criminal behavior and family environment (Kulka et al., 1990; Nace et al., 1980; O’Brien et al., 1980; Robins, 1974, 1993; Robins et al., 2010), or geographic considerations (Boscarino, 1979) were not collected and not evaluated.

This study has several strengths. VE-HEROeS provides current data on the health status of older Vietnam War veterans using established survey sampling, administration, and up-to-date analytical guidelines. Its research design includes a matched nonveteran comparison group to better delineate associations between military service and health. A representative veteran sample was systematically selected from a sampling frame derived from a vetted veteran database. Veteran response rate was high (45%) and close to those obtained from other surveys of U.S. veterans (49%, Health and Retirement Study, 2013).

In conclusion, the opportunity to help Vietnam-era veterans by improving their health through directed study is slowly vanishing. Our findings provide insights into a range of lifetime and current substance use, topics that have not been addressed to any great degree in other nationwide studies of these veterans. We hope that these results will prove useful in continuing the study of substance use in this population and support further development of policies and interventions that could improve their well-being.

Disclaimer

This material is based upon work supported by the Epidemiology Program, HOME (12POP5), U.S. Department of Veterans Affairs (VA). The views expressed are those of the authors and do not necessarily represent the views or policy of the VA or the U.S. Government.

Acknowledgements

We would like to thank the veterans who participated in this study. We would also like to acknowledge the assistance of the U.S. Department of Veterans Affairs’ (VA) Central Office library staff who have always responded quickly and adeptly to our requests for information. Other VA staff who we would also like to thank are Ms. Stephanie Eber, Health Science Specialist/Communications Specialist, for her editorial review, and Drs. Aaron Schneiderman and William J. Culpepper, Director, and Deputy Director of the Epidemiology Program (HOME/12POP5), respectively, for their assistance.

Disclosure statement

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to have influenced the work reported in this paper.

Additional information

Funding

This work was supported by the Epidemiology Program, Health Outcomes Military Exposures (12POP5), Office of Patient Care Services.