Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 May 1.
Published in final edited form as: Psychol Serv. 2021 Apr 12;19(2):386–395. doi: 10.1037/ser0000538

The Role of PTSD Symptom Clusters and Criterion in Predicting Future High-Risk Drug and Alcohol Use Among Returning Veteran Men and Women

Nicholas A Livingston 1,2,3, Stacey L Farmer 4, Colin T Mahoney 5, Brian P Marx 1,2,3, Terence M Keane 1,2,3
PMCID: PMC9048192  NIHMSID: NIHMS1796375  PMID: 33844563

Abstract

The prevalence of co-occurring posttraumatic stress disorder (PTSD) and substance use disorder (SUD) remains exceptionally high among returning veterans, with numerous studies linking PTSD, but not specific PTSD symptoms, to future SUD risk. Further explication of PTSD symptom effects on future SUD risk will likely promote intervention development and refinement while offsetting SUD risk. Accordingly, In this study we explored the prospective associations between PTSD symptom clusters, symptoms, and future SUD risk and use of specific drug classes. Returning veterans (N = 1,295; Mage = 42.3, SD = 9.89; 51% female; 66.8% White) completed structured diagnostic interviews to assess PTSD symptoms and self-report measures of substance use 14–36 months later (M = 24.59, SD = 2.97). Hyperarousal and reckless/self-destructive symptoms specifically predicted future high-risk drug use and binge drinking behavior, and avoidance of internal stimuli (i.e., of trauma memories, thoughts, and feelings) differentiated individuals classified as high-risk for alcohol use based on their AUDIT total score. Further, negative alterations in cognition and mood predicted future opioid (i.e., nightmares) and stimulant use (i.e., flashbacks), whereas concentration difficulties were inversely associated with future binge drinking. This longitudinal study identified prospective and enduring associations between specific PTSD symptom clusters, symptoms, and future high-risk substance use patterns among returning veterans. Accordingly, careful assessment of specific PTSD criteria and differential motivations for substance use is warranted, along with tailored interventions to offset risk for opioid, stimulant, and alcohol use among returning veterans.

Keywords: PTSD, drug use, alcohol use, veteran, longitudinal

The prevalence of alcohol use, drug use, and co-occurring posttraumatic stress disorder (PTSD) symptoms are exceptionally high among veterans (Hoge, 2015; Petrakis et al., 2011; Stein et al., 2017; Teeters et al., 2017) and remain especially high in veterans of the Iraq and Afghanistan conflicts (Lan et al., 2016; Seal et al., 2011). Data from the National Health and Resilience in Veterans Study (NHRVS), a nationally representative and population-based veteran sample, suggest that veterans with PTSD are 3.9–4.5 times more likely to also meet the criteria for substance use disorder (SUD; Wisco et al., 2016). Further, as many as 63–75% of returning veterans with a SUD also meet the criteria for PTSD (Debell et al., 2014; Hoge et al., 2004; Seal et al., 2011), in contrast to rates of 15–30% in the general population (Debell et al., 2014; Hoge et al., 2004). Although longitudinal data are limited, PTSD symptom severity significantly predicts future substance use, higher frequency use, and negative consequences (Berke et al., 2019), as well as poorer treatment adherence and outcomes related to long-term substance use and PTSD recovery (Norman et al., 2007; Ouimette et al., 1998, 2006; Possemato et al., 2017; Schäfer & Najavits, 2007; Tate et al., 2007).

The self-medication hypothesis, which is the theoretical model invoked most often to account for PTSD-substance use co-occurrence, postulates that substance use becomes a negatively reinforced behavior to the extent that it provides relief from PTSD symptoms (Berke et al., 2019; Khantzian, 2012, 2003; Lane et al., 2019; Mahoney et al., 2020). However, PTSD represents a complex and heterogeneous disorder, comprised of both internalizing and externalizing symptoms with unknown associations with substance use risk, or use of particular substances. An interesting yet still understudied possibility is that substance use, and the use of certain classes of substances, might be characteristic of elevations within specific PTSD symptom clusters. For instance, hyperarousal symptoms (Criteria E) might distinguish the use of disinhibitory substances or use patterns, such as binge drinking or stimulant use, whereas the use of opioids or cannabis might be especially appealing to individuals motivated to suppress or alleviate intrusive symptoms (i.e., Criteria B). In a recent study, Dworkin et al. (2018) found among treatment-seeking individuals that hyperarousal symptoms did in fact distinguish patients with drug use disorders, whereas avoidance symptoms were more characteristics of patients with alcohol use disorders. This sample was not veteran specific and further research is needed.

Unfortunately, most research among veterans to date only supports the link between global PTSD symptom severity and substance use. In a recent study conducted among veterans, Tripp et al. (2020) found that greater overall PTSD symptom severity was associated with greater future alcohol use. Similar findings were found in a large sample of active-duty service members, where researchers found that PTSD symptom severity was associated with increased problem alcohol use prospectively (Berke et al., 2019). Further, research has shown that PTSD, and experiencing a combat-related TBI, were a significant risk factor for future binge drinking (Adams et al., 2016), and ecological momentary assessment research supports positive associations between PTSD symptom activation and daily alcohol consumption (Possemato et al., 2015). Cross-sectional studies support these associations, but also highlight associations between PTSD symptom clusters and substance use (Jakupcak et al., 2010). In one prospective study that analyzed self-report of DSM-IV-TR PTSD symptoms data from Gulf War veterans, Shipherd et al. (2005) found that hyperarousal symptoms predicted drug use (but not alcohol) among veterans 4–4.5 years later after adjusting for other PTSD clusters. While laying the groundwork for future research, this study did not include assessment of specific substances and, as such, no known reports exist documenting both the role of DSM-5 PTSD clusters, or symptoms, in predicting future high-risk substance use, or particular patterns of alcohol and illicit drug use among veterans. Replication of research by Shipherd and colleagues, using a latent difference score modeling approach, has produced different results and highlights the possible impact of intrusion and numbing symptoms, but not avoidance or hyperarousal, in predicting future alcohol misuse among combat veterans (Langdon et al., 2016).

In the current study, we sought to extend works by Shipherd et al. (2005) and Langdon et al. (2016) by examining PTSD symptoms at criterion level, and their associations with particular patterns of high-risk substance use among returning veterans over time. We also sought to extend previous literature by using PTSD symptom data gleaned using clinician-administered structured diagnostic interviews, rather than common-use self-report PTSD measures to predict future high-risk substance use. Consistent with the above-mentioned research documenting the effects of PTSD on future risk of substance use, we hypothesized positive prospective associations between PTSD symptom clusters and substance use outcomes. Despite robust literature supporting the association between global PTSD and substance use, and because very little research has examined prospective effects of PTSD symptom clusters on future substance use risk, we regarded these analyses as largely exploratory. As such, our primary objective was to examine whether intrusions (Criteria B), avoidance (Criteria C), negative alterations in cognitions and mood (Criteria D), and hyperarousal (Criteria E) symptoms were associated with future high-risk alcohol and drug use, with follow-up post hoc analyses at the symptom level to generate hypotheses for future research. As a follow-up and exploratory aim, we examined whether these criteria were differentially related to future binge drinking, and the use of cannabis, stimulant drugs, and opioids. We regarded these novels and follow-up post hoc analyses as exploratory and, as such, expected to observe positive associations but made no a priori hypotheses regarding the particular patterns of effects we would observe.

Method

The study sample included returning veterans from the VALOR Project, a national and longitudinal registry of post-9/11 combat-deployed veterans of Iraq and Afghanistan conflicts. All participants in this study completed a psychiatric at a Veteran Affairs (VA) medical center prior to participation. Eligible patients were identified using the VHA national claims database. Eligible patients included post-9/11 combat-deployed veterans who had received a mental health evaluation at a VA medical center (indexed by receipt of a diagnostic interview or psychotherapy procedure code) between July 2008 and December 2009. Eligible patients were then randomly selected from VHA claims data and called to participate in the study. The average date of enrollment was 2011 (range: 2009–2012), with an average length of time since the deployment of just over 6 years (M = 6.14, SD = 2.02). We oversampled veterans with a documented history of PTSD (3:1 PTSD: no PTSD ratio) and women veterans (50% of the original sample). We oversampled patients with a PTSD diagnosis to ensure sufficient representation of PTSD symptomology, given our primary interest in PTSD symptoms and course over time. We oversampled women to overcome issues related to the underrepresentation of women in veteran health research, a rising proportion of women joining the military and veteran population (Cucciare et al., 2013; Kamarck, 2019), and for planned analyses by collaborating members of the research team. Since this was a national study and study enrollment occurred remotely, all participants were informed of study risks and benefits over the phone, then asked to provide verbal consent to participate. Once they provided verbal consent, participants were sent a hard-copy informed consent for their records, outlining identical information about the study to which they consented verbally. All study procedures were approved by the local IRB and Human Research Protection Office, U.S. Army Medical Research and Material Command.

The current study includes 1,295 veterans who completed the Structured Clinical Interview for DSM-5 (SCID-5) for PTSD at time one (T1), as well as self-report questionnaires regarding past-year substance use at T2 (no structured interviews took place during T2), approximately 2 years later (Mmonths = 24.59, SD = 2.97; range: 14–36 months), thereby ensuring at least a 14-month gap between clinical interview and self-report of past-year substance use. The procedures for this longitudinal registry included a combination of self-report measurement batteries, designed to be administered at least a year apart, and structured diagnostic interviews. For self-report questionnaires, participants were mailed packets to complete and return. For the overall study (i.e., including all assessment waves), self-report measures were completed at five time points. On average, questionnaires at time two, three, four, and five were completed 2.42, 3.38, 4.52, 6.94 years after participants’ enrollment date, respectively. Structured diagnostic interviews occurred via phone and were conducted by trained, doctoral-level assessors working under licensed supervision. Participants were called for their diagnostic assessments separately and between waves of self-report measurement. Data used for the current study include structured diagnostic PTSD interview data from the first wave of diagnostic interviews (T1), during which no substance use data were collected. We also used self-report substance use data collected from the subsequent self-report assessment wave (T2), during which no diagnostic PTSD interview data were collected. Thus, data for the current study include SCID-5 PTSD interview data from T1 and self-report substance use data from T2. While we do not have PTSD data for T2, other studies published from these data have demonstrated chronic and persistently high severity of PTSD symptoms over time among participants (e.g., Lee et al., 2020; Livingston, Lee, et al., 2021), indicating that veterans in this study likely experienced significant distress from PTSD symptoms at T2. Further, given the fact that we performed a structured and diagnostic assessment of PTSD at T1, and the time lag between T1 and T2, the current analysis stands to contribute insight into the prospective correlates of PTSD symptoms and criterion on future substance use behavior in the absence of baseline substance use assessment. Demographic variables were collected at baseline (see Table 1).

Table 1.

Demographic Characteristics of the Sample (N = 1,295)

Demographics M SD
Age 42.36 9.89
n %
Gender
 Male 634 48.9
 Female 661 51.0
Sexual orientation
 Heterosexual 1,018 78.6
 Gay/Lesbian 41 3.2
 Bisexual 22 1.7
 “Other” 8 .6
 Unspecified 206 15.9
Ethnicity
 White/Caucasian 731 56.4
 African American 171 13.2
 Asian 9 .7
 Native American/Alaskan Native 7 .5
 Native Hawaiian/Pacific Islander 2 .2
 Hispanic 8 .6
 Mixed race 153 11.8
 Unspecified 214 16.5
Relationship status
 Married 623 48.1
 Living with a domestic partner 64 4.9
 Divorced/separated 226 17.5
 Widowed 14 1.1
 Single 140 10.8
 “Other” 23 1.8
 Unspecified 205 15.8
Level of education
 High school Diploma/GED 67 5.2
 Vocational training post-High School 34 2.6
 “Some college” 242 18.7
 Associate’s degree 190 14.7
 Bachelor’s degree 320 24.7
 Master’s degree 210 16.2
 Doctoral degree 20 1.5
 Unspecified 212 16.4
Employment
 Full time 518 40
 Part time 54 4.2
 Unemployed and looking for work 39 3
 Temporarily laid off, sick/other leave 4 .3
 Full-time homemaker 22 1.7
 Full-time student 38 2.9
 Disabled 217 16.8
 Retired 106 2.9
 Other 93 7.2
 Unspecified 212 16.4
Open in a new tab

Note. GED = high school graduate equivalency degree.

Measures

Posttraumatic Stress Disorder Symptoms

The Structured Clinical Interview for DSM-5 (SCID-5; First et al., 2015) is a structured diagnostic interview that was used by study staff to assess the presence of PTSD symptoms at T1. Only the SCID-5 for PTSD module was used, which has demonstrated strong psychometric properties with evidence of good diagnostic reliability (Regier et al., 2013) and good test–retest reliability for symptom severity (r = .73; Shankman et al., 2018). Interviews were conducted by doctoral-level clinicians via phone. Each interview was digitally recorded for later review by an independent assessor who was not involved in the original assessment. To calculate interrater agreement, an independent rater randomly selected, reviewed, and coded 100 interviews using the same SCID-5 criteria as the original assessor. We then compared the assessment results of the original and independent assessors, which demonstrated good interrater agreement (κ = .82; Green et al., 2017; Mitchell et al., 2017).

Alcohol Use

At T2, no structured diagnostic interview was performed, just self-report questionnaires. We measured alcohol use using the 10-item Alcohol Use Disorders Identification Test (AUDIT; Saunders et al., 1993). Respondents answered items using a 5-item Likert-type scale, with total scores ranging from 0 to 40 and higher scores indicating higher consumption, greater risk for negative consequences, and/or meeting criteria for alcohol use disorder. Psychometric research suggests that a score of eight or higher maximally differentiates individuals with low versus high risk for alcohol use disorder (i.e., higher consumption and alcohol-related consequences; Conigrave et al., 1995; Maisto et al., 2000). The internal consistency of the AUDIT in this study was high (α = .86).

Drug Use

We indexed drug use using the 10-item Drug Abuse Screening Test (DAST-10) at T2 (Skinner & Goldberg, 1986), which is a self-report screener of past-year consumption and negative consequences associated with the use of illicit drugs and prescription medication. Total scores range from 0 to 10 with higher scores indicating more frequent drug use and more severe negative drug use-related consequences. Internal consistency for the DAST-10 in the current study was also high (α = .85). We collapsed total DAST-10 scores into a binary variable using the established cut score of two which has been shown to differentiate individuals with versus without a SUD diagnosis (Maisto et al., 2000). Although we do not identify individuals as above and below this cut score as having or not having a SUD, we use this instead as an index of high-risk drug use, as recommended (Maisto et al., 2000). We also included single items to assess specific substances used “recreationally” in the past year: Cannabis (i.e., “cannabis (non-prescribed marijuana, pot)” and “cannabis (medicinal marijuana)”), stimulants (i.e., “cocaine” and “methamphetamine (speed, crystal)”), and opioids (e.g., “narcotics (heroin, oxycodone, methadone)” and “opioids (Vicodin, OxyContin, Percocet)”), using a binary yes/no scale in this study.

Data Handling and Analytic Strategy

Independent variables included the number of diagnostic criteria met within each PTSD symptom cluster at T1. We created overall PTSD symptom cluster scores by summing the count of symptom criteria endorsed within each cluster. Since each cluster contains a different number of symptoms, we then standardized participants’ summed PTSD cluster scores to express them as standard deviation units.

The two primary dependent variables included high-risk drug use (DAST-10 scores 0–2 = 0, 3 or higher = 1; Maisto et al., 2000) and high-risk alcohol use (0–7 = 0, 8 or higher = 1; Conigrave et al., 1995; Maisto et al., 2000) as measured at T2. Due to the design of the study, we could not covary substance use at T1 as we did not collect substance use data at T1, nor change in PTSD as we did not assess PTSD at T2. For exploratory and follow-up post hoc tests, we considered secondary dependent variables of binge drinking, cannabis use, stimulant use, and opioid use in the past year, all measured at T2. Cannabis use variable included endorsement of recreational use in the past year (no = 0, yes = 1). Stimulant use included an endorsement of cocaine and methamphetamine in the past year (no = 0, yes = 1). Recreational opioid use in the past year included the endorsement of narcotics and opioid medications. We computed “binge drinking” using item three from the AUDIT (“How often do you have six or more drinks on one occasion?” [in the past year]), coding “never” as zero and everything else as one.

We tested hypotheses using hierarchical logistic regression (Knapp, 2018) and all analyses were carried out using SPSS, version 26 (IBM Corp, 2019). We used a data-driven approach to model building and to identify relevant demographic covariates for inclusion (i.e., age, race, gender), excluding from the final model’s covariates that were not statistically significant. Our final models include each PTSD symptom cluster, with age included as a covariate in models predicting drug use and gender included in models predicting alcohol use outcomes. For our primary aim, we evaluated PTSD clusters as predictors of high-risk drug use and alcohol in two separate models. The remaining analyses were considered post hoc or exploratory. Once we established the statistical significance of a PTSD symptom cluster in our primary analyses, we performed post hoc analyses to examine the specific symptoms within the identified clusters that might account for the overall cluster associations observed. Next, and for our exploratory analyses, we used the same model from our primary analyses to predict future use of specific substances (i.e., cannabis, opioids, stimulants, and binge drinking)—that is, putative substances that might be driving scores on the parent high-risk drug and alcohol use measures used. As with our primary analyses, if we observed statistically significant associations between PTSD clusters and specific substance use outcomes, we performed post hoc analyses to identify the specific symptoms within the cluster that might best account for the overall cluster association observed. For example, after observing hyperarousal symptoms predicted binge drinking, we performed post hoc analysis of each hyperarousal symptom (i.e., irritability/anger outbursts, reckless behavior, hypervigilance, exaggerated startle, concentration problems, and sleep issues) in the same model, covarying for gender, to explicate precise symptoms driving the overall hyperarousal effect observed. We approached these analyses in an exploratory fashion to zero in on precise symptoms of note for future prospective studies.

Our primary analyses included two models to examine predictors of high-risk drug use and alcohol use separately. However, we also examined four separate substances as outcomes in our exploratory analyses. Given the fact that our primary and secondary/exploratory outcomes were related, we used a Bonferroni correction and considered statistically significant any model with an omnibus test that was significant at p < .008. Finally, we performed sensitivity analyses to evaluate whether our predictors were associated with dropout by T2 among participants who completed our T1 diagnostic interview.

Results

Descriptive Results

The average length of time between T1 and T2 was 24.59 months (SD = 2.97; range: 14–36). We analyzed predictors of missing data at T2 among participants who completed our T1 diagnostic interview. PTSD symptom clusters, age, gender, and race were unrelated to missing data at T2 ( p = .431–.992). A total of 276 (21.3%) scored “high-risk” on the AUDIT and 94 (7.3%) scored “high-risk” on the DAST-10. Further, 492 veterans (38.0%) reported past-year binge drinking; specific substances reported included cannabis (n = 197, 15.2%), opioids (n = 78, 6.0%), and stimulants (n = 35, 2.7%). The majority of veterans met criteria for PTSD (n = 966; 74.6%); PTSD criteria counts ranged from 0 to 5/5 for intrusions (Mcount = 3.71, SD = 1.35), 0–2/2 for avoidance (Mcount = 1.62, SD = .65), 0–6/6 for negative alterations in cognitions and mood (Mcount = 4.00, SD = 2.05), and 0–6/6 for hyperarousal (Mcount = 4.40, SD = 1.51). Additional demographics are presented in Table 1. We applied a Bonferroni correction to the omnibus tests of statistical significance and considered statistically significant tests below p < .008 (p < .05 divided by six primary and secondary outcomes). Each omnibus test of statistical significance was significant at p < .001.

Primary Analyses

High-Risk Drug and Alcohol Use

To develop our final models, we first examined age, race, and gender as covariates in the first block. Gender was a statistically significant covariate regarding high-risk alcohol use and was retained, adjOR = 1.79, 95% CI [1.32, 2.42], p < .001, but not age, adjOR = .99, p = .080, or race, adjOR = .94, p = .745. Age was the only significant covariate regarding high-risk drug use, adjOR = .96, 95% CI [.93, .99], p = .010, but not gender, adjOR = 1.41, p = .191, or race, adjOR = 1.31, p = .395, which were removed from final drug use models

In our final models, individuals meeting the criteria for a higher number of hyperarousal symptoms were more likely to be classified into the high-risk drug use group by the next assessment wave, adjOR = 1.72, 95% CI [1.11, 2.67], p = .015. No other PTSD symptom clusters predicted high-risk drug use after accounting for age and other PTSD symptom clusters. On the other hand, and after controlling for gender, avoidance was positively associated with future high-risk alcohol use, adjOR = 1.24, 95% CI [1.01, 1.53], p = .044 (see Table 2).

Table 2.

Primary Logistic Regression Output for High-Risk Alcohol and Drug Use

High-risk drug use
 High-risk drug use
Variables β SE β adj OR [95% CI] β SE β adj OR [95% CI]
Intercept −1.25* 0.61 .29 — −1.60*** 0.12 .20 —
Age −.04* .02 .96 [.94, .99] — — — —
Gender — — — — .58*** .15 1.78 [1.32, 2.40]
Intrusions .39† .21 1.48 [.98, 2.25] −.02 .10 .98 [.80, 1.20]
Avoidance .09 .20 1.09 [.74, 1.62] .22* .11 1.24 [1.01, 1.53]
NACM −.17 .18 .85 [.60, 1.20] .00 .11 1.00 [.82, 1.23]
Hyperarousal .54* .22 1.72 [1.11, 2.67] .20† .11 1.22 [.98, 1.52]
Open in a new tab

Note. β = Unstandardized regression coefficients; adjOR = adjusted odds ratio; 95% CI = 95% confidence interval; NACM = Negative alterations in cognition and mood.

†

p < .10.

*

p < .05.

**

p < .01.

***

p < .001.

Post Hoc and Exploratory Analyses

High-Risk Drug and Alcohol Use.

As a follow-up to these analyses, we evaluated specific hyperarousal symptoms as predictors of the future high-risk drug, and avoidance symptoms as predictors of future high-risk alcohol use. As with primary analyses, age was included as a covariate for drug use and gender was retained as a predictor of alcohol use. We found that reckless or self-destructive behavior was the only significant hyperarousal symptom associated with high-risk drug use, adjOR = 4.19, 95% CI [2.37, 7.42], p < .001, and avoidance of internal trauma reminders was the only predictor of high-risk alcohol use, adjOR = 1.95, 95% CI [1.26, 3.01], p = .003.

Exploration of Substance Types.

Using the same model, controlling for gender, we found that hyperarousal symptoms predicted past-year binge drinking, adjOR = 1.24, 95% CI [1.04, 1.48], p = .017. After accounting for age, intrusion symptoms but no other symptom clusters predicted future opioid use, adjOR = 2.19, 95% CI [1.40, 3.43], p = .001, and stimulant use, adjOR = 2.58, 95% CI [1.17, 5.69], p = .018. PTSD symptoms did not predict future cannabis use (Table 3).

Table 3.

Exploratory logistic Regression Output for Past Year Opioid, Stimulant, and Cannabis Use, and Binge Drinking

Opioid use
 Stimulant use
 Cannabis use
 Binge drinking
Variables β SE β adj OR [95% CI] β SE β adj OR [95% CI] β SE β adj OR [95% CI] β SE β adj OR [95% CI]
Intercept −.87 .62 .42 — −1.07 1.05 .31 — .17 .42 .68 — −.74*** .09 .48 —
Age .05** .02 .95 [.92, .98] −.08** .03 .93 [.88, .98] −.05*** .01 .95 [.94, .97] — — — —
Gender — — — — — — — — — — — — .62*** .13 1.87 [1.46, 2.40]
Intrusions .79*** .23 2.19 [1.40, 3.43] .95* .40 2.58 [1.17, 5.69] .13 .13 1.14 [.89, 1.46] .02 .09 1.02 [.86, 1.20]
Avoidance .14 .20 1.15 [.78, 1.70] .15 .34 1.16 [.60, 2.25] .07 .13 1.07 [.84, 1.37] −.02 .08 .99 [.84, 1.16]
NACM −.17 .18 .84 [.60, 1.20] −.34 .27 .72 [.42, 1.21] .09 .13 1.10 [.86, 1.40] −.15 .09 .87 [.73, 1.03]
Hyperarousal −.09 .19 .92 [.62, 1.35] .58 .37 1.78 [.86, 3.66] .10 .13 1.10 [.85, 1.43] .21* .09 1.24 [1.04,1.48]
Open in a new tab

Note. β = Unstandardized regression coefficients; adjOR = adjusted odds ratio; 95% CI = 95% confidence interval; NACM = negative alterations in cognition and mood.

†

p < .10.

*

p < .05.

**

p < .01.

***

p < .001.

In the final step, we examined individual criterion-level associations to isolate specific symptom effects potentially driving the overall PTSD symptom cluster associations observed. These results further implicated reckless/self-destructive behavior regarding future binge drinking, adjOR = 1.67, 95% CI [1.28, 2.19], p < .001, but also trauma-related nightmares in predicting opioid use, adjOR = 5.42, 95% CI [1.26, 23.32], p = .023, and dissociative flashbacks in predicting future stimulant use, adjOR = 2.65, 95% CI [1.14, 6.18], p = .024. Interestingly, we observed inverse effects of concentration difficulties on future binge drinking, adjOR = .69, 95% CI [.50, .97], p = .03.

Discussion

In the current study, we explored the prospective associations between PTSD symptom clusters, specific symptoms, and future high-risk drug and alcohol use among returning veteran men and women. Using PTSD symptom data gathered via structured diagnostic interview and reports of drug and alcohol use collected 14–36 months later, we found that the number of endorsed PTSD hyperarousal symptoms predicted future high-risk drug use, and PTSD avoidance symptoms predicted future high-risk alcohol use as indexed by the AUDIT total score cut-off. These results are consistent with those of Dworkin et al. (2018), who found in a sample of treatment-seeking individuals that hyperarousal symptoms differentiated patients diagnosed with drug use disorders, and elevated avoidance symptomology differentiated patients with alcohol use disorder.

We also found that distinct PTSD symptoms predicted some substance use outcomes but not others. For instance, intrusions predicted opioid and stimulant use. Upon further examination, we found that recurrent trauma-related nightmares predicted future opioid use, which may be motivated by a desire to slow, suppress, or seek comfort from intrusive nightmares (Garland et al., 2016). The fact that trauma-related nightmares are a common and debilitating symptom endorsed by many returning veterans with PTSD makes it a chief consideration in general (Levin & Nielson, 2007; Neylan et al., 1998), but especially to the degree that it predicts future opioid use risk in the context of an ongoing opioid epidemic (Manchikanti et al., 2012). In contrast, dissociative flashbacks predicted future stimulant use, which is notable given the rising incidence of stimulant use in the U.S. (Chan et al., 2019).

Whereas opioid and stimulant use might be motivated by desires to manage/reduce nightmares and dissociative flashback symptoms, respectively, null findings for the remaining intrusive symptoms might suggest that veterans manage these symptoms using other unmeasured adaptive (e.g., cognitive reappraisal) and/or maladaptive (e.g., binge eating) coping strategies. An alternative explanation for the pattern of effects observed in this study might relate to our examination of PTSD cluster and criterion effects, as opposed to global PTSD severity as is more common in research. If so, the unique configuration of significant and nonsignificant PTSD symptom associations might actually reflect the specific, item-level, and previously unreported associations driving the associations between global PTSD symptom severity scores and substance use among veterans. This possibility is one that requires replication, as the scope of the current article was exploratory and largely intended to generate hypotheses for future research.

Veterans who met the criteria for reckless/self-destructive behavior were more likely to report high-risk drug use and future binge drinking. These findings are consistent with previous scholarship highlighting distinctions between internalizing and externalizing dimensions of PTSD expression, with the former being characterized by internalizing symptoms such as anxiety and negative emotionality, and the latter by impulsivity, antisocial behaviors, and substance use (Castillo et al., 2014; Livingston, Brief, et al., 2021; Miller et al., 2003). Given this, it is possible that “self-medication,” which accounts for motives and behaviors aimed at reducing actual or anticipatory distress, applies yet manifests differently across the internalizing–externalizing spectrum. That is, some of our findings signify regulation through gratification of higher risk-taking propensities and may point to a more inclusive pathway to self-regulation through engagement in high-intensity acts intended to evoke desired states of disinhibition and physiological arousal (e.g., high-risk drug use and binge drinking). On the other hand, we found that avoidance differentially predicted future high-risk alcohol use, as measured by the AUDIT total score, and that it was the avoidance of distressing trauma memories, thoughts, and feelings specifically that accounted for the overall effect. This association is highly consistent with the self-medication hypothesis (Berke et al., 2019; Khantzian, 2003, 2012; Lane et al., 2019; Mahoney et al., 2020) and suggests the possibility that specific PTSD symptoms might be primary in predicting some patterns of substance use relative to others.

It is commonly believed that hypervigilance and motives to reduce associated arousal account for the association between PTSD and substance use (Back et al., 2006; Hellmuth et al., 2012); however, hypervigilance was nonsignificant in this study. It is possible that hypervigilance was simply nonsignificant over and above the reckless/self-destructive behavior criterion, or other factors such as threat/risk aversion (e.g., decreased use of illicit drugs or alcohol use at hazardous levels), or reluctance to use substances in a manner that might render them more vulnerable or impair their ability to remain vigilant. We also observed an unexpected inverse association between concentration difficulties and binge drinking. This effect was surprising in the sense that endorsement is suggestive of more severe PTSD and, therefore, elevated substance use risk. Alternatively, to the extent that veterans are not experiencing concentration difficulties, they might actually desire substance-induced cognitive-affective change as a means of regulating distressing thoughts given their intact mental faculties. Other hyperarousal symptoms, such as irritability/anger outbursts, exaggerated startle response, and sleep disturbance, were not prospectively associated with high-risk drug use or binge drinking. This could be attributable to the model, wherein they were simply not significant over and above reckless/self-destructive behavior, or other reasons. For instance, alcohol has been shown to increase anger and irritability (Rehm, 2011), startle reactivity and subjective arousal (Gorka, 2020), and sleep problems (Angarita et al., 2016). For some veterans with PTSD, alcohol-related exacerbation of these symptoms may be undesirable, avoided, thereby accounting for the null associations observed here.

While this was a largely exploratory study, strengths of the current paper include representation of a large and gender-balanced sample of returning veterans and featured results from an examination of clinician-assessed DSM-5 PTSD symptom criterion on future high-risk substance use patterns. Future research may benefit from diagnostic assessment of substance use as well, in addition to a finer-grained assessment of use to enable an analysis of high-risk drug and alcohol use patterns within a month, week, or day among returning veterans. Due to the study design, we were unable to evaluate alternative models of PTSD-substance use co-occurrence, such as the mutual maintenance model (see Berke et al., 2019). While our primary objective in the current study was to provide insight into potential functional associations between particular PTSD symptoms, as assessed using structured diagnostic interview, and future high-risk substance use patterns, the lack of baseline substance use measurement and T2 PTSD assessment remain significant limitations. Previous research supports that combat veterans are characterized by internalizing versus externalizing premilitary characteristics are more likely to manifest posttraumatic symptoms in internalizing and externalizing ways, respectively (Miller et al., 2003). While we would suspect similar consistency among veterans in our sample, we are unable to directly test this and our study design undermines our ability to disaggregate reckless/self-destructive behavior, as measured by E2, from unmeasured substance use that could have been present at T1. On the other hand, other PTSD symptoms assessed and found to be associated with future substance use in this study, such as intrusions (i.e., trauma-related nightmares and dissociative flashbacks and opioid and stimulant use, respectively) and avoidance (i.e., internal avoidance and alcohol use) are trauma-specific and distinguishable from unmeasured substance use at the time of baseline PTSD assessment. Future research would still benefit from extending these findings by examining baseline substance use as a covariate but also, importantly, assessing changes in PTSD over time as we were unable to do here.

Despite being limited by a lack of follow-up PTSD assessment, the fact that we used exact PTSD data from structured diagnostic interviews, and examined correlates of symptoms with high-risk substance use 14–36 months (∼1–3 years) later, provides insight into some of the potentially enduring correlates of specific PTSD criterion on high-risk substance use over time among returning veterans. This pattern of results is further supported by prior publication from this data set highlighting the stability of PTSD severity over 20 years following exposure to their index trauma (Lee et al., 2020). The results of this study should be interpreted in light of the fact that we oversampled returning veteran women and those with documented PTSD diagnoses, and that results might differ if replicated in a sample of veterans whose gender (i.e., <20% women; Kamarck, 2019) and proportion of PTSD diagnoses (11–30%; National Center for PTSD, 2018; Reisman, 2016) more closely align with current proportions of returning veterans (see Kamarck, 2019; National Center for PTSD, 2018; Reisman, 2016). Finally, while we observed significant effects regarding stimulant and opioid use, these findings should be interpreted in light of the relatively small number of veterans who endorsed the use of these substances.

Clinical Implications

In light of the expressed limitations and largely exploratory nature of this study, we believe replication is needed, and that the unique pattern of effects reported here provides novel direction for future research. If replicated, and particularly in studies that address some of the aforementioned limitations, we believe these results also provide implications for evidence-based assessment and intervention. Rather than focusing on global PTSD severity at the expense of specific PTSD criteria, we might be able to refine assessment strategies and use the information obtained to develop individualized treatment plans for returning veterans with PTSD. This point is further supported by the observed odds ratios which, in some cases (e.g., trauma-related nightmares on future opioid use), signify strong clinically significant correlates of high-risk substance use that persist over time (see Tables 1 and 2). Specific criteria that may be especially important to assess relate to risky or destructive behavior, nightmares, dissociation and flashbacks, and avoidance as they relate to future substance use risk. From a substance use treatment perspective, understanding the precise mechanisms of PTSD-substance use co-occurrence, and possible symptom-level motivations for using certain substances, may help to individualize PTSD and SUD treatment to improve outcomes. For example, if a returning veteran is seeking care for opioid use, the assessment of nightmares and related sleep disturbance may help to better understand the function of the substance use and provide guidance in PTSD treatment planning that also targets opioid use risk reduction.

Administrators of mental health clinics could support these efforts by encouraging PTSD symptom assessment or screening as part of a regular intake process, and by supporting training opportunities for clinicians with less experience with PTSD or substance use. Granted, the above-mentioned assessment recommendations may be lofty for some clinics, given training and time constraints, and the fact that many clinics are not designed to provide specialty PTSD or substance use treatment services. Though we used PTSD symptom data gleaned using a structured diagnostic interview in this study, it may be more feasible for clinics to implement PTSD screening at intake, using abbreviated measures such as the PTSD Symptom Checklist-5 (PCL-5; Weathers et al., 2013). Specific items on this measure correspond to each PTSD symptom and, for screening purposes, could be used similar to the manner in which we examine PTSD cluster and criterion effects here.

Relatedly, we believe that careful screening of alcohol and drug use is warranted. The substance use screeners used in the current study are often used for clinical and research purposes. Although used frequently in research, the AUDIT is widely used as a clinical tool and mandated as part of care within Veteran’s Health Administration (Teeters et al., 2017). Likewise, the DAST was developed for research and clinical use, and both as a screener and for treatment outcome evaluation (Skinner & Goldberg, 1986; Young et al., 2006). That said, other screeners might be more suitable or feasible for some clinics than others and further assessment, potentially using a structured diagnostic interview, is recommended following a positive screen (see Young et al., 2006 for a listing of valid alcohol and drug screeners and assessments).

Conclusion

In this study, we identified potentially enduring associations between specific PTSD symptom clusters and future high-risk substance use. Although largely exploratory, these findings extend prior research by highlighting differential pathways linking particular PTSD symptoms to specific substances and high-risk use patterns, which might also vary as a function of internalizing– externalizing PTSD symptom expression. As research continues to move toward longitudinal modeling of associations between PTSD symptoms and substance use, we encourage investigators to also carefully consider the level of analysis that might lead to the greatest acceleration of clinical insight. For example, a demonstration of the impact of global PTSD symptom severity on future substance use provides some direction for future research and intervention. However, we believe that studies highlighting PTSD symptom cluster or individual criterion effects on substance use risk will lead to an unprecedented degree of precision and specificity that will maximize research and clinical impact. Accordingly, we advocate for careful assessment of PTSD symptom clusters and the individual diagnostic criterion in future research, and for tailored assessment and clinical intervention to lower risk for opioid, stimulant, and alcohol use among returning veterans.

Impact Statement.

We found that PTSD clusters predicted future high-risk substance use among returning veterans, as well as specific patterns of substance use longitudinally (e.g., opioid use). In addition to identifying future research directions, study findings highlight specific PTSD symptoms to consider in assessment and treatment in order to offset risk for hazardous substance use.

Acknowledgments

The first and third author’s work on this project was funded by the Office of Academic Affiliations, U.S. Department of Veteran Affairs. The second author’s work was funded by T32 MH019836. This research was funded by a grant from the Department of Defense; Contract grant numbers: W81XWH-08-2-0100/W81XWH-08-2-0102 and W81XWH-12-2-0117/W81XWH-12-2-0121.

References

  1. Adams RS, Larson MJ, Corrigan JD, Ritter GA, Horgan CM, Bray RM, & Williams TV (2016). Combat-acquired traumatic brain injury, posttraumatic stress disorder, and their relative associations with postdeployment binge drinking. The Journal of Head Trauma Rehabilitation, 31(1), 13–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Angarita GA, Emadi N, Hodges S, & Morgan PT (2016). Sleep abnormalities associated with alcohol, cannabis, cocaine, and opiate use: A comprehensive review. Addiction Science & Clinical Practice, 11(1), 1–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Back SE, Brady KT, Sonne SC, & Verduin ML (2006). Symptom improvement in co-occurring PTSD and alcohol dependence. Journal of Nervous and Mental Disease, 194(9), 690–696. [DOI] [PubMed] [Google Scholar]
  4. Berke DS, Yeterian J, Presseau C, Rusowicz-Orazem L, Kline NK, Nash WP, & Litz BT (2019). Dynamic changes in marines’ reports of PTSD symptoms and problem alcohol use across the deployment cycle. Psychology of Addictive Behaviors, 33(2), 162–170. [DOI] [PubMed] [Google Scholar]
  5. Castillo DT, Joseph JS, Tharp AT, C’de Baca J, Torres-Sena LM, Qualls C, & Miller MW (2014). Externalizing and internalizing subtypes of posttraumatic psychopathology and anger expression. Journal of Traumatic Stress, 27(1), 108–111. [DOI] [PubMed] [Google Scholar]
  6. Chan B, Freeman M, Kondo K, Ayers C, Montgomery J, Paynter R, & Kansagara D (2019). Pharmacotherapy for methamphetamine/amphetamine use disorder—A systematic review and meta-analysis. Addiction, (ePub ahead of print). 10.1111/add.14755 [DOI] [PubMed] [Google Scholar]
  7. Conigrave KM, Hall WD, & Saunders JB (1995). The AUDIT questionnaire, choosing a cut-off score. Alcohol use disorder identification test. Addiction, 90(10), 1349–1356. 10.1046/j.1360-0443.1995.901013496.x [DOI] [PubMed] [Google Scholar]
  8. Cucciare MA, Simpson T, Hoggatt KJ, Gifford E, & Timko C (2013). Substance use among women veterans: Epidemiology to evidence-based treatment. Journal of Addictive Diseases, 32(2), 119–139. 10.1080/10550887.2013.795465 [DOI] [PubMed] [Google Scholar]
  9. Debell F, Fear NT, Head M, Batt-Rawden S, Greenberg N, Wessely S, & Goodwin L (2014). A systematic review of the comorbidity between PTSD and alcohol misuse. Social Psychiatry and Psychiatric Epidemiology, 49(9), 1401–1425. [DOI] [PubMed] [Google Scholar]
  10. Dworkin ER, Wanklyn S, Stasiewicz PR, & Coffey SF (2018). PTSD symptom presentation among people with alcohol and drug use disorders: Comparisons by substance of abuse. Addictive Behaviors, 76, 188–194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. First MB, Williams JBW, Karg RS, & Spitzer RL (2015). Structured clinical interview for DSM-5—Research version (SCID-5 for DSM-5, Research Version; SCID-5-RV) (pp. 1–94). American Psychiatric Association. [Google Scholar]
  12. Garland EL, Brown SM,&Howard MO (2016). Thought suppression as a mediator of the association between depressed mood and prescription opioid craving among chronic pain patients. Journal of Behavioral Medicine, 39(1), 128–138. [DOI] [PubMed] [Google Scholar]
  13. Gorka SM (2020). Interpersonal trauma exposure and startle reactivity to uncertain threat in individuals with alcohol use disorder. Drug and Alcohol Dependence, 206, Article 107727. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Green JD, Annunziata A, Kleiman SE, Bovin MJ, Harwell AM, Fox AML, … Marx BP (2017). Examining the diagnostic utility of the DSM-5 PTSD symptoms among male and female returning veterans. Depression and Anxiety, 34(8), 752–760. 10.1002/da.22667 [DOI] [PubMed] [Google Scholar]
  15. Hellmuth JC, Stappenbeck CA, Hoerster KD, & Jakupcak M (2012). Modeling PTSD symptom clusters, alcohol misuse, anger, and depression as they relate to aggression and suicidality in returning US veterans. Journal of Traumatic Stress, 25(5), 527–534. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hoge C (2015). Measuring the long-term impact of war-zone military service across generations and changing posttraumatic stress disorder definitions. JAMA Psychiatry, 72(9), 861–862. [DOI] [PubMed] [Google Scholar]
  17. Hoge CW, Castro CA, Messer SC, McGurk D, Cotting DI, & Koffman RL (2004). Combat duty in Iraq and Afghanistan, mental health problems, and barriers to care. The New England Journal of Medicine, 351(1), 13–22. [DOI] [PubMed] [Google Scholar]
  18. IBM Corp. (2019). IBM SPSS Statistics for Windows, Version 26.0. IBM Corp. [Google Scholar]
  19. Jakupcak M, Tull MT, McDermott MJ, Kaysen D, Hunt S, & Simpson T (2010). PTSD symptom clusters in relationship to alcohol misuse among Iraq and Afghanistan war veterans seeking postdeployment VA health care. Addictive Behaviors, 35(9), 840–843. 10.1016/j.addbeh.2010.03.023 [DOI] [PubMed] [Google Scholar]
  20. Kamarck KN (2019). Diversity, inclusion, and equal opportunity in the armed services: background and issues for congress. Congressional Research Service. https://fas.org/sgp/crs/natsec/R44321.pdf
  21. Khantzian EJ (2003). Understanding addictive vulnerability. Neuro-Psychoanalysis, 5(1), 5–21. [Google Scholar]
  22. Khantzian EJ (2012). Reflections on treating addictive disorders: A psychodynamic perspective. The American Journal on Addictions, 21(3), 274–279. [DOI] [PubMed] [Google Scholar]
  23. Knapp H (2018). Intermediate statistics using SPSS. Sage Publications. [Google Scholar]
  24. Lan CW, Fiellin DA, Barry DT, Bryant KJ, Gordon AJ, Edelman EJ, … Marshall BD (2016). The epidemiology of substance use disorders in US Veterans: A systematic review and analysis of assessment methods. The American Journal on Addictions, 25(1), 7–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lane AR, Waters AJ, & Black AC (2019). Ecological momentary assessment studies of comorbid PTSD and alcohol use: A narrative review. Addictive Behaviors Reports, 10. Article 100205. 10.1016/j.abrep.2019.100205 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Langdon KJ, Fox AB, King LA, King DW, Eisen S, & Vogt D (2016). Examination of the dynamic interplay between posttraumatic stress symptoms and alcohol misuse among combat-exposed Operation Enduring Freedom (OEF)/Operation Iraqi Freedom (OIF) Veterans. Journal of Affective Disorders, 196, 234–242. 10.1016/j.jad.2016.02.048 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Lee DJ, Lee LO, Bovin MJ, Moshier SJ, Dutra SJ, Kleiman SE, Rosen RC, Vasterling JJ, Keane TM, & Marx BP (2020). The 20-year course of posttraumatic stress disorder symptoms among veterans. Journal of Abnormal Psychology, 129(6), 658–669. 10.1037/abn0000571 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Levin R, & Nielson TA (2007). Disturbed dreaming, posttraumatic stress disorder, and affect distress: A review and neurocognitive model. Psychological Bulletin, 133(3), 482–528. 10.1037/0033-2909.133.482 [DOI] [PubMed] [Google Scholar]
  29. Livingston NA, Brief D, Miller M, & Keane T (2021). Assessment of PTSD and its comorbidities in adults. In Friedman MJ, Keane TM, & Schnurr PP (Eds.), Handbook of PTSD: Science and practice. Guilford Press. [Google Scholar]
  30. Livingston NA, Lee DJ, Mahoney CT, Farmer SL, Cole T, Marx BP, & Keane TM (2021). Longitudinal assessment of PTSD and illicit drug use among male and female OEF-OIF veterans. Addictive Behaviors, 118, Article 106870. 10.1016/j.addbeh.2021.106870 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Mahoney C, Livingston NA, Wong M, Rosen R, Marx B, & Keane T (2020). Parallel process modeling of PTSD symptoms and alcohol use severity in returning veterans. Psychology of Addictive Behaviors. Advance online publication. 10.1037/adb0000569 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Maisto SA, Carey MP, Carey KB, Gordon CM, & Gleason JR (2000). Use of the AUDIT and the DAST-10 to identify alcohol and drug use disorders among adults with a severe and persistent mental illness. Psychological Assessment, 12(2), 186–192. 10.1037//1040-3590.12.2.186 [DOI] [PubMed] [Google Scholar]
  33. Manchikanti L, Fellows B, Janata JW, Pampati V, Grider JS, & Boswell MV (2012). Opioid epidemic in the United States. Pain Physician, 15(3, Suppl), ES9–ES38. [PubMed] [Google Scholar]
  34. Miller MW, Greif JL, & Smith AA (2003). Multidimensional personality questionnaire profiles of veterans with traumatic combat exposure: Externalizing and internalizing subtypes. Psychological Assessment, 15(2), 205–215. 10.1037/1040-3590.15.2.205 [DOI] [PubMed] [Google Scholar]
  35. Mitchell KS, Wolf EJ, Bovin MJ, Lee LO, Green JD, Rosen RC, … Marx BP (2017). Network models of DSM–5 posttraumatic stress disorder: Implications for ICD–11. Journal of Abnormal Psychology, 126(3), 355–366. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. National Center for PTSD. (2018). How common is PTSD in veterans? https://www.ptsd.va.gov/understand/common/common_veterans.asp
  37. Neylan TC, Marmar CR, Metzler TJ, Weiss DS, Zatzick DF, Delucchi KL, … Schoenfeld FB (1998). Sleep disturbances in the Vietnam generation: Findings from a nationally representative sample of male Vietnam veterans. The American Journal of Psychiatry, 155(7), 929–933. [DOI] [PubMed] [Google Scholar]
  38. Norman SB, Tate SR, Anderson KG, & Brown SA (2007). Do trauma history and PTSD symptoms influence addiction relapse context? Drug and Alcohol Dependence, 90(1), 89–96. [DOI] [PubMed] [Google Scholar]
  39. Ouimette P, Goodwin E, & Brown PJ (2006). Health and well-being of substance use disorder patients with and without posttraumatic stress disorder. Addictive Behaviors, 31(8), 1415–1423. 10.1016/j.addbeh.2005.11.010 [DOI] [PubMed] [Google Scholar]
  40. Ouimette PC, Brown PJ, & Najavits LM (1998). Course and treatment of patients with both substance use and posttraumatic stress disorders. Addictive Behaviors, 23(6), 785–795. 10.1016/S0306-4603(98)00064-1 [DOI] [PubMed] [Google Scholar]
  41. Petrakis I, Rosenheck R, & Desai R (2011). Substance use comorbidity among veterans with posttraumatic stress disorder and other psychiatric illness. The American Journal on Addictions, 20(3), 185–189. [DOI] [PubMed] [Google Scholar]
  42. Possemato K, Maisto SA, Wade M, Barrie K, Johnson EM, & Ouimette PC (2017). Natural course of co-occurring PTSD and alcohol use disorder among recent combat veterans. Journal of Traumatic Stress, 30(3), 279–287. [DOI] [PubMed] [Google Scholar]
  43. Possemato K, Maisto SA, Wade M, Barrie K, McKenzie S, Lantinga LJ, & Ouimette P (2015). Ecological momentary assessment of PTSD symptoms and alcohol use in combat veterans. Psychology of Addictive Behaviors, 29(4), 894–905. 10.1037/adb0000129 [DOI] [PubMed] [Google Scholar]
  44. Regier DA, Narrow WE, Clarke DE, Kraemer HC, Kuramoto SJ, Kuhl EA, & Kupfer DJ (2013). DSM-5 field trials in the United States and Canada, Part II: Test-retest reliability of selected categorical diagnoses. The American Journal of Psychiatry, 170(1), 59–70. [DOI] [PubMed] [Google Scholar]
  45. Rehm J (2011). The risks associated with alcohol use and alcoholism. Alcohol Research & Health, 34(2), 135–143. [PMC free article] [PubMed] [Google Scholar]
  46. Reisman M (2016). PTSD Treatment for Veterans: What’s Working, What’s New, and What’s Next. P&T, 41(10), 623–634. [PMC free article] [PubMed] [Google Scholar]
  47. Saunders JB, Aasland OG, Babor TF, de la Fuente JR, & Grant M (1993). Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO collaborative project on early detection of persons with harmful alcohol consumption-II. Addiction, 88(6), 791–804. [DOI] [PubMed] [Google Scholar]
  48. Schäfer I, & Najavits LM (2007). Clinical challenges in the treatment of patients with posttraumatic stress disorder and substance abuse. Current Opinion in Psychiatry, 20(6), 614–618. [DOI] [PubMed] [Google Scholar]
  49. Seal KH, Cohen G, Waldrop A, Cohen BE, Maguen S, & Ren L (2011). Substance use disorders in Iraq and Afghanistan veterans in VA healthcare, 2001–2010: Implications for screening, diagnosis and treatment. Drug and Alcohol Dependence, 116(1–3), 93–101. [DOI] [PubMed] [Google Scholar]
  50. Shankman SA, Funkhouser CJ, Klein DN, Davila J, Lerner D, & Hee D (2018). Reliability and validity of severity dimensions of psychopathology assessed using the Structured Clinical Interview for DSM-5 (SCID). International Journal of Methods in Psychiatric Research, 27(1), Article e1590. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Shipherd JC, Stafford J, & Tanner LR (2005). Predicting alcohol and drug abuse in Persian Gulf War veterans: What role do PTSD symptoms play? Addictive Behaviors, 30(3), 595–599. [DOI] [PubMed] [Google Scholar]
  52. Skinner HA, & Goldberg AE (1986). Evidence for a drug dependence syndrome among narcotic users. British Journal of Addiction, 81(4), 479–484. 10.1111/j.1360-0443.1986.tb00359.x [DOI] [PubMed] [Google Scholar]
  53. Stein M, Campbell-Sills L, Gelernter J, He F, Heeringa S, Nock M, … Ursano R (2017). Alcohol misuse and co-occurring mental disorders among new soldiers in the U.S. army. ACER, 41(1), 139–148. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Tate SR, Norman SB, McQuaid JR, & Brown SA (2007). Health problems of substance dependent veterans with and those without trauma history. Journal of Substance Abuse Treatment, 33(1), 25–32. [DOI] [PubMed] [Google Scholar]
  55. Teeters JB, Lancaster CL, Brown DG, & Back SE (2017). Substance use disorders in military veterans: Prevalence and treatment challenges. Substance Abuse and Rehabilitation, 8, 69–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Tripp JC, Worley MJ, Straus E, Angkaw AC, Trim RS, & Norman SB (2020). Bidirectional relationship of posttraumatic stress disorder (PTSD) symptom severity and alcohol use over the course of integrated treatment. Psychology of Addictive Behaviors. Advance online publication. 10.1037/adb0000564 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Weathers FW, Litz BT, Keane TM, Palmieri PA, Marx BP, & Schnurr PP (2013). The PTSD Checklist for DSM-5 (PCL-5). National Center for PTSD. www.ptsd.va.gov [Google Scholar]
  58. Wisco BE, Marx BP, Miller MW, Wolf EJ, Mota NP, Krystal JH, Southwick SM & Pietrzak RH (2016). Probable posttraumatic stress disorder in the US Veteran Population according to DSM-5: Results from the National Health and Resilience in Veterans Study. The Journal of Clinical Psychiatry, 77(11), 1503–1510. 10.4088/JCP.15m10188 [DOI] [PubMed] [Google Scholar]
  59. Young NK, Nakashian M, Yeh S, & Amatetti S (2006). Screening and assessment for family engagement, retention, and recovery (SAFERR). DHHS Pub. No. 0000. Substance Abuse and Mental Health Services Administration. [Google Scholar]

RESOURCES