Experiential Avoidance, Pain, and Suicide Risk in a National Sample of Gulf War Veterans

Abstract

Objective:

Pain confers risk for suicidal thoughts and behaviors. Experiential avoidance (EA), which is relevant to both pain and suicide risk, has not been studied as a potential mechanism for this relationship. The present study tested the hypothesis that pain indirectly impacts suicide risk through EA in a national sample of Gulf War veterans.

Methods:

Participants included a stratified random sample of United States veterans (N = 1012, 78% male) who had served in the Gulf War region between August 1990 and July 1991. Validated scales were used to quantify levels of pain, EA, and suicide risk.

Results:

Regression analyses indicated independent associations between pain, EA, and suicide risk; moreover, the association between pain and suicide risk was no longer significant once EA was included in model. Bootstrapping analyses confirmed that EA partially accounted for the cross-sectional association between pain and suicide risk, independent of common co-occurring problems, such as depression, PTSD, and alcohol use disorder symptoms.

Conclusions:

EA could be a key modifiable risk factor to target in people experiencing pain.

Physical pain has generally been associated with suicidal thoughts and behaviors (Franklin et al., 2017; Jacob, Haro, & Koyanagi, 2018). Although some research has demonstrated a link between acute pain and risk for suicide, much of the literature to date has largely focused on chronic pain. Indeed, research suggests that individuals who regularly experience pain are significantly more likely to die by suicide relative to people without pain conditions, even after adjusting for lifetime and past-year mental health conditions, such as mood, anxiety, and traumatic stress disorders (Campbell, Drake, & Bruno, 2015). This is particularly critical, given that chronic pain affects as many as 20% of U.S. citizens (Dahlhamer et al., 2018). However, while the association between suicidal thoughts and behaviors and pain is clear, we presently lack insight into the underlying factors contributing to this association, which could be informative for the assessment, prevention, and treatment efforts.

Research to date suggests that certain factors may be particularly prominent in the link between pain and suicide. According to initial studies, aspects of physical pain, such as type of pain (e.g., headache pain; Ratcliffe et al., 2008), and stronger intensity and longer duration of pain (Smith et al., 2004; Tang & Crane, 2006; Carson et al., 2000), confer the greatest risk for suicide among individuals reporting pain. However, results from these earlier studies have not been fully supported by more recent research. For instance, whereas Ilgen et al. (2008) found an independent association between head pain and suicidal ideation and attempt, these findings were not statistically significant in a more recent study which instead found a more significant association between neck pain and suicidal ideation (Campbell et al., 2015). Moreover, most research has failed to demonstrate a significant association between pain duration or intensity and suicide risk variables (see Racine 2018 for review). Hence, it is worth exploring additional underlying factors for this relationship.

In recent years, there has been increasing attention given to psychological variables, specifically those that may influence one’s perception and experience of pain, as potential explaining factors for the pain-suicide link. For example, greater pain catastrophizing (i.e., a tendency to imagine the worst possible outcome associated with pain) is associated with severity of suicidal thoughts (Racine et al., 2014; Racine et al., 2017; Tang et al., 2016) and attempted suicide (Sansone et al., 2014). Pain-related helplessness, a component of pain catastrophizing, is associated with suicidal ideation, whereas optimism about one’s pain prognosis is considered protective (Racine et al., 2017). Additionally, among chronic pain patients receiving specialized treatment within a hospital setting, those who feel socially withdrawn and perceive themselves to be a burden may be especially more likely to report suicidal ideation (Cheatle et al., 2014; Shim et al., 2017). However, similar to studies on physical pain variables on suicide risk, these findings are not consistent across all studies (e.g., Tang et al., 2016). As such, more research is necessary to elucidate the role of psychological variables linking pain and risk for suicide. One such variable that has received less attention in this regard is experiential avoidance (EA).

EA is defined as unwillingness to experience, or remain in contact with, an aversive private (i.e., internal) experience, such as physical pain or difficult emotions (Hayes et al., 1996). As a result, individuals high in EA often attempt to alter the form and frequency of their painful experiences as well as the contexts in which these painful experiences occur, which can be harmful in the experience of physical pain. Laboratory studies have demonstrated that people reporting higher levels of EA have lower acute pain tolerance and endurance (Zettle et al., 2005; Zettle et al., 2012), and recover more slowly in response to acute pain (Feldner et al., 2006). EA also appears to influence adaptation to chronic pain. For instance, EA has been associated with depression and anxiety in patients with chronic pain in primary care settings (Costa & Pinto-Gouveia, 2013), greater pain-related impairment, catastrophizing and hypervigilance among patients with chronic spinal pain (Ramirez-Maestre et al., 2014), and greater dysfunction and lower quality of life in adolescents and adults with various pain conditions (Feinstein et al., 2011; McCracken & Velleman, 2010). For these reasons, emerging research has identified EA as a key modifiable therapeutic target in pain management (Hughes et al., 2017).

EA has also been linked with suicidal thoughts and behaviors (Roush et al., 2019) as well as various proximal risk factors for suicide, such as depression, hopelessness, and nonsuicidal self-injury (Anderson & Crowther, 2012; Berking et al., 2009). Moreover, reduction of EA in treatment is associated with reductions in depression and suicidal ideation (Ellis & Rufino, 2016; Zeifman et al., 2020). Suicidal behavior also frequently functions as a mechanism to escape aversive private experiences (Brown et al., 2002; Bryan et al., 2013). Therefore, EA may be an underlying factor driving the association between pain and suicide risk. For instance, individuals experiencing severe and persistent pain may have difficulty tolerating these pain-related aversive internal experiences and have a desire to escape them, possibly through suicidal thoughts and behaviors. Yet, despite evidence for EA as a factor independently linked with both pain and suicide risk, to our knowledge, no research has investigated if EA contributes to this relationship.

The Present Study

To address this gap in literature, a secondary analysis was conducted on data gathered from a national sample of 1012 Gulf War veterans, who completed surveys assessing pain, EA, history of suicidal thoughts and behaviors, and other relevant variables. Notably, suicide rates among military veterans have steadily increased over the last two decades, increasing by 35.9% between the year 2001 and 2019 (VA Office of Mental Health and Suicide Prevention, 2021). Moreover, 25–72% of military veterans report regularly experiencing some form of physical pain (Van Den Kerkhof et al., 2014), which is likely higher than the estimated 18–34 % of civilians reporting such pain (Dahlhammer et al., 2018; Johannes et al., 2010; Yong et al., 2022). Hence, this line of research may be especially relevant to veterans’ health and could lead to modifiable therapeutic targets to reduce suicide risk in this population.

The aims of the present study were to: (1) provide further evidence of the hypothesized associations between pain, EA, and suicide risk among Gulf War veterans; and (2) examine if EA would account for the cross-sectional association that we anticipated observing between pain and suicide risk score determined by the revised version of the Suicide Behaviors Questionnaire (Osman et al., 2001). It was hypothesized that pain would be positively associated with suicide risk and that EA would at least partially account for this association independent of common co-occurring factors, such as depression, PTSD, and alcohol use disorder (AUD) symptoms.

Methods

Participants and Procedure

The current study involved secondary analysis of data from a study aimed at evaluating the healthcare needs of Gulf War-era veterans, entitled Project GRIT (grant #1I01HX001682). For this study, a survey was administered to a national sample of Gulf War veterans. Eligible participants included U.S. military veterans who were on active duty between August 1990 and July 1991. All study procedures were approved by the Institutional Review Boards at the Durham Veterans Affairs (VA) Health Care System and the VA Salt Lake City Health Care System.

For a more detailed description of the recruitment and data collection procedures, see Blakey et al. (2021). Briefly, administrative data were used to identify Gulf War veterans who had sought care at the VA at least one time. In order to provide nationwide representation, stratified sampling procedures were used to sample veterans across eight geographic regions. Women veterans were oversampled to approximately 25% to ensure adequate representation.

A total of 3,272 veterans were mailed surveys that included questionnaires pertaining to healthcare utilization, certain psychiatric and substance use disorders, history of suicidal thoughts and behaviors and other factors relevant to the healthcare needs of veterans. To incentivize timely survey completion, all veterans who returned surveys in a timely manner were entered into a lottery raffle in which cash prizes were awarded, and the earlier surveys were returned the greater the likelihood of being selected in the lottery. Ultimately, a total of 1,153 surveys were returned (35.2% return rate). The sample for the present study includes the 1,012 veterans who completed questionnaires for all of the present study’s variables of interest.

Materials

Demographic Information.

Participants provided age, sex, gender identity, race, ethnicity, sexual orientation, marital status, and Department of Veterans Affairs (VA) service-connection status (see Table 1).

Table 1.

Descriptive statistics for demographic variables

Variable	% / M (SD)
Age	58.60 (7.19)
Sex Assigned at Birth	78% Male
Gender
Man	78%
Woman	21%
Other/Not reported	<1%
Sexual Orientation
Heterosexual	96%
Gay or Lesbian	2%
Bisexual	1%
Not sure or Other	<1%
Race
White/Caucasian	65%
Black/African American	21%
American Indian	1%
Asian or Pacific Islander	2%
More than one race	5%
Other/Not reported	6%
Ethnicity	9% Hispanic
Marital Status
Married or Living Together	74%
Separated or Divorced	18%
Widowed	2%
Single/Never Married	6%
Reported at least one physical disability	17%
Service Connected	78%

Pain.

Pain was assessed using a well-established 3-item scale quantifying Pain, interference with Enjoyment of life, and interference with General activity (PEG), through which participants provided “past week” average pain levels using a 10-point Likert-based rating scale of pain intensity (0 = “no pain,” 10 = “the most intense pain you could possibly imagine”). The PEG has been validated to assess past week pain intensity and related consequences in chronic pain patients (Krebs et al., 2009). It has been found to be reliable, have good construct validity, and be sensitive to change (Kean et al., 2016; Krebs et al., 2009). It demonstrated excellent internal consistency in the present study (α = .96).

Suicide risk.

Suicide risk was assessed with the 4-item Suicide Behaviors Questionnaire-Revised (SBQ-R; Osman et al., 2001), a reliable and validated self-report measure that quantifies overall risk for suicide as a combination of past ideation and planning, prior suicide attempts, and likelihood of future attempts. Scores for the four items were summed for a composite score of risk, where higher scores indicate greater risk for suicide. Prior studies have demonstrated a moderately high coefficient alpha estimate, along with excellent sensitivity (93%) and specificity (95%; Osman et al., 2001). Internal consistency for the SBQ-R in the present sample was adequate (α = .76).

Experiential avoidance.

EA was assessed using the Brief Experiential Avoidance Questionnaire (BEAQ; Gámez et al., 2014), a validated 15-item measure based on the more comprehensive Multidimensional Experiential Avoidance Questionnaire (MEAQ; Gámez et al., 2011). Participants used a 7-point Likert scale ranging from 0 (Strongly disagree) to 6 (Strongly agree) to rate the degree to which they agreed with each item (e.g., “I go out of my way to avoid uncomfortable situations”). Higher BEAQ scores indicate greater (i.e., worse) EA. Previous work has shown good convergent and discriminant validity of the BEAQ (Gámez et al., 2014), as well as good internal consistency. Internal consistency for the BEAQ in the present study was excellent (α = .90).

Depression symptoms.

This was assessed using an abbreviated version of the Patient Health Questionnaire (PHQ-2), a validated 2-item measure of depression symptoms (Kroenke et al., 2003). The measure asks participants to rate the frequency of anhedonia (i.e., “Little interest of pleasure in doing things”) and depressed mood over the past two weeks using a Likert scale ranging from 0 (not at all) to 3 (nearly every day). The PHQ-2 has shown good convergent and discriminant validity, and an overall valid alternative to the longer version, the PHQ-9. Internal consistency for the PHQ-2 in the present study was excellent (α = .90). Although this measure is not meant to indicate a clinical diagnosis, for readability purposes we refer to scores on this measure as “depression” hereafter. This is also the case for the PTSD and AUD symptom measures described below.

Posttraumatic stress disorder symptoms.

This was assessed using the Primary Care-PTSD scale (PC-PTSD; Prins et al., 2003), a 6-item scale validated and commonly used to screen for PTSD symptoms in primary care settings. The scale asks participants to answer “Yes” or “No” to indicate if they experienced certain PTSD symptoms over the past month (e.g., “In the past month, have you been constantly on guard, watchful, or easily startled?”). The PC-PTSD has demonstrated good psychometric properties in prior studies (e.g., Ouimette et al., 2008), and its internal consistency for the present study was good (α = .85).

Alcohol use disorder symptoms.

AUD symptoms were assessed using an abbreviated version of the Alcohol Use Disorder Identification Test (AUDIT-3), which is a validated 3-item measure of at-risk and hazardous drinking (Bradley et al., 2003). The AUDIT-3 includes questions pertaining to frequency, intensity, and severity of alcohol use (e.g., “How often do you have a drink containing alcohol?”) with multiple choice answers corresponding to each item. The AUDIT-3 is a widely used screening measure that has demonstrated good psychometric properties. The internal consistency for the AUDIT-3 for the present study was adequate (α = .75).

Statistical Approach

The hypothesized associations between pain, EA, and STBs were first assessed using zero order correlations. To establish the path linking pain to EA, we used an ordinary least-squares regression model. The distribution of scores for the dependent variable (suicide risk [i.e., SBQ-R total score]) was positive and right-skewed (skewness = 2.02, kurtosis = 4.04), such that there was a disproportionate number of veterans in the current sample who denied any history of suicidal thoughts and behaviors on the SBQ-R. Therefore, a generalized linear model with gamma distribution and log-link function was used to test the hypothesized associations for the remaining models where the dependent variable was suicide risk score. For these analyses, the association was considered significant if the 95% confidence interval (CI) of the point estimate did not include 0. Finally, to test the hypothesized indirect effect of pain on suicide risk through EA, SPSS PROCESS Macro (Hayes, 2018) was used to conduct bootstrapping analyses for indirect effects. Here, 5000 samples were extracted and bootstrapping confidence intervals (BCIs) were calculated. Indirect, direct, and total effects were considered significant if the 95% BCIs did not contain zero. To determine the relevant contribution of EA to the pain-suicide risk association, two bootstrapping models were conducted. The first model tests EA as an indirect factor linking pain to suicide risk without covariates, whereas the alternative model tests this relationship while also including depression, PTSD, and AUD symptoms as covariates.

Results

Descriptive Statistics, Zero Order Correlations, and Initial Regression Analyses

Descriptive statistics and zero order correlations among all study variables can be found in Table 2. As expected, EA was positively correlated with both pain (r = .40, p < .001) and suicide risk (r = .35, p < .001). Pain was also positively correlated with suicide risk (r = .18, p < .001). Also, as expected, a simple linear regression model with pain as the predictor and EA as the dependent variable was significant (R² = .16, F (1,1010) = 193.01, p < .001), and demonstrated that pain was associated with EA (b = 2.37, t = 13.89, p < .001). In the gamma regression analyses, pain was associated with higher suicide risk (b = .03, 95% CI [.02, .04], $\chi$² = 48.82, p < .001). In a separate regression model, EA alone was also associated with suicide risk (b = .01, 95% CI [.01, .01], $\chi$² = 205.18, p < .001). In a regression model, in which both pain and EA were included as predictors, EA continued to be associated with suicide risk (b = .01, 95% CI [.008, .01], $\chi$² = 154.76 p < .001), whereas pain no longer was (b = .008, 95% CI [−.001, .02], $\chi$² = 2.97, p = .09), suggesting that EA may significantly account for the association between pain and suicide risk. In a final model that included the covariates of depression, PTSD, and AUD symptoms, EA remained significantly associated with suicide risk (b = .002, 95% CI [.001, .004], $\chi$² = 7.57, p = .006), and pain became statistically significant as well (b = −.01, 95% CI [−.02, −.004], $\chi$² = 8.48, p = .004). Depression was also significantly associated with suicide risk (b = .11, 95% CI [.09, .13], $\chi$² = 136.70 p < .001), as was PTSD (b = .02, 95% CI [.006, .04], $\chi$² = 7.08, p = .008). AUD was not significantly associated with suicide risk (b = .008, 95% CI [−.001, .02], $\chi$² = 2.77, p = .09).

Table 2.

Zero-order correlation analyses and descriptive statistics for all study variables

Study Variables	1.	2.	3.	4.	5.	6.
1. Suicide Risk	-
2. Pain	.18***	-
3. EA	.35***	.40***	-
4. Depression	.49***	.47****	.61***	-
5. PTSD	.35***	.45***	.53***	.63***	-
6. AUD	.12***	.01	.11***	.12***	.14***	-
Mean (SD)	4.49 (2.39)	4.25 (2.94)	40.84 (17.41)	1.45 (1.80)	1.42 (1.86)	2.52 (2.56)

Note.

^***= p < .001

Bootstrapping analyses to test for indirect, direct, and total effects

Model 1 (Without covariates).

The full model of pain to suicide risk was significant, [F (1,1012) = 73.20, MSE = 5.01, p < .001, R² = .13]. Consistent with our main hypothesis, pain was positively associated with EA (b = 2.37, 95% CI [2.03, 2.70]), which was, in turn, associated with suicide risk (b = .05, SE = .01, 95% CI [.04, .06]). Pain had a direct effect on suicide risk (b = .15, SE = .03, 95% CI [.10, .19]), but there was also an indirect effect for pain on suicide risk via EA (b =.11, SE = .02, 95% CI [.08, .14]). Notably, there was no direct effect between pain and suicide risk after accounting for the indirect effect (b = .04, SE = .02, 95% CI [−.02, .09]), indicating that EA fully accounted for the association between pain and suicide risk (see Figure 1 [Panel A] and Table 3).

Figure 1.

Analyses examining the indirect association between pain and suicide risk via EA. Standard regression weights are reported to demonstrate the effect of pain on EA (a), effect of EA on suicide risk (b), the direct effect of pain on suicide risk (c’) and the total effect of the mediation model (c). Panel A above demonstrates the indirect effect of EA without covariates, whereas Panel B demonstrates the indirect effect with depression, PTSD, and AUD included as covariates in the model. Results suggest that EA fully accounts for the link between pain and suicide risk when it is included as the lone factor (Panel A). However, the magnitude of the effect of EA is substantially reduced and only partially accounts for the association when other covariates are included (Panel B). *** = p < .001, ** = p < .01, * = p < .05.

Table 3.

Bootstrapping analyses for both models testing EA as indirect effect for pain predicting suicide risk score

	b	BootLLCI	BootULCI	t	p	R 2	F Value
Model 1 (DV: Suicide risk)
Predictors						.13	73.20***
Pain	.03	−.01	.03	1.36	.17
EA	.05	.04	.06	10.47	<.001
Model 2 (DV: Suicide Risk)
Predictors						.25	69.19***
Pain	−.07	−.12	−.02	−2.75	.006
EA	.01	−.001	.02	2.34	.02
Depression	.57	.47	.68	11.00	<.001
PTSD	.08	−.01	.18	1.75	.08
AUD	.05	−.005	.09	1.67	.08

Note. Some significant values are in bold to aid interpretation. DV = dependent variable. BootLLCI = Bootstrapping lower limit confidence interval, BootULCI = Boostrapping upper limit confidence interval.

^***= p < .001.

Model 2 (With covariates).

The full model of pain to suicide risk was significant [F (1,1012) = 69.19, MSE = 4.28, p < .001, R² = .26]. Pain was positively associated with EA (b = .63, SE = .17, 95% CI [.31, .96]), as was depression (b = 4.14, SE = .31, 95% CI [3.53, 4.76]) and PTSD (b = 1.89, SE = .30, 95% CI [1.30, 2.47]). AUD was not associated with EA (b = .21, SE = .17, 95% CI [−.11, .54]). EA was also associated with suicide risk (b = .01, SE = .004, 95% CI [.001, .02]), as was depression (b = .57, SE = .05, 95% CI [.47, .70]). PTSD was not significantly associated with suicide risk (b = .08, SE = .04, 95% CI [−.01, .18]), and neither was AUD (b = .05, SE = .03, 95% CI [−.005, .09]). In this model, pain again had a direct effect on suicide risk (b = −.07, SE = .03, 95% CI [−.12, −.02]), and there was also a much smaller but significant indirect effect via EA (b = .007, SE = .003, 95% CI [.001, .015]). However, unlike the original model, the direct effect of pain on suicide risk remained significant (b = −.07, SE = .03, 95% CI [−.12, −.02]), indicating that EA only partially accounted for the association between pain and suicide risk (see Figure 1 [Panel B] and Table 3).

Discussion

Pain is a well-known risk factor for suicidal thoughts and behaviors (Jacob et al., 2018), though research has been mixed with regard to the various mechanisms underlying this association (Racine, 2018). Prior research has also demonstrated that EA is independently associated with both pain (Feldner et al., 2016) and suicide (Roush et al., 2019). To our knowledge, however, no prior research has examined whether EA may help to explain the association between pain and suicide risk.

Consistent with our main hypothesis, we found that EA significantly accounted for the cross-sectional association that we observed between pain and suicide risk in this large, national sample of Gulf War veterans. This important new finding provides additional evidence that psychological variables are likely to play a significant role in the well-established association between pain and suicidal thoughts and behaviors, and that EA may be particularly important in this regard; however, the specific reasons for this association remain unclear. It is possible, for example, that failed attempts to control unavoidable pain leads to suicide cognitions that commonly occur in the context of pain, such as defeat and entrapment (Tang et al., 2016). It may also be the case that EA leads to dangerous behaviors to escape from pain, such as opioid misuse and overt suicidal behavior. Accidental and intentional overdoses from opioid misuse in pain patients is a significant contributing factor to the ongoing opioid epidemic in the U.S. (Madadi & Persaud, 2014; Riquino et al., 2018), and there is emerging evidence that low willingness to experience pain is a driving factor contributing to opioid misuse in this population (Rhodes et al., 2020). These are just some of the many plausible explanations for the results demonstrated in the present research. Hence, future work is necessary to clarify the specific mechanisms by which EA confers risk for suicide and other self-destructive behaviors among patients in pain, such as substance misuse.

Of note, the magnitude of the effect of EA accounting for the pain-suicide risk association was substantially diminished in an alternative model that included depression, PTSD, and AUD symptoms as covariates. In fact, the inclusion of these variables may have contributed to a suppression effect, given that the statistical direction of the association between pain and suicide in the alternative model changed relative to the original model (Smith et al., 1992). However, this is not surprising, as prior research suggests a great deal of shared variance between pain, EA, and psychopathology explaining risk for suicide. Indeed, avoidance is a fundamental process in many forms of high-risk psychopathology (Chawla & Ostafin, 2007; Fledderus, Bohlmeijer, & Pieterse, 2010). In depression, for example, depressive rumination has been shown to be a form of EA in and of itself (Giorgio et al., 2010) and a key factor linking depression to suicide cognitions (e.g., perceived burdensomeness) and active suicidal ideation (Roush et al., 2017). Moreover, people with chronic pain are significantly more likely to report experiences directly associated with risk for suicidal thoughts and behaviors, such as symptoms of depression and anxiety (e.g., McWilliams, Cox, & Enns, 2003), PTSD (Brennstuhl et al., 2015), and alcohol misuse (Egli, Koob, & Edwards, 2012). Nonetheless, our results suggest that EA may still be a unique clinical target to reduce suicide risk among individuals with ongoing physical pain, and additional clinical research should determine the extent and the process by which this is true, particularly in the context of co-occurring factors.

The present research had a number of limitations that should be considered when interpreting the findings. First, this research was cross-sectional in nature. Thus, inferences about mediation should be made with caution until prospective studies have replicated the findings from the present research by utilizing longitudinal designs. Second, the present research relied entirely on self-report. Third, although the pain measure used in the present study is a validated measure of pain and highly correlated with pain chronicity (Chen et al., 2019), it only assesses pain experienced over the past week. As such, the findings outlined above may not necessarily characterize the full experience of those with chronic and persistent pain. There is also some question about the extent to which the BEAQ measures EA as a unitary construct among treatment-seeking veterans (Byllesby et al., 2020). Additionally, the measure of suicide risk used in the present study treats suicidal thoughts and behaviors as a unitary construct, when in reality ideation and behavior are certainly related but otherwise distinct outcomes that do not necessarily share the same risk factors (May & Klonsky, 2016). Finally, although Project GRIT recruited a national sample of veterans representative of sociodemographic characteristics of VHA enrollees at the time of data collection, we did not collect information from participants on other key factors, such as diagnosed psychiatric or medical conditions. Moreover, despite oversampling for women, our sample was still composed of mostly White, heterosexual married men who served during the Gulf War era. As such, the results of the present study may not generalize to people who represent different sociodemographic groups or to veterans of other eras. Future longitudinal research utilizing multiple methods that includes a demographically diverse, clinically diagnosed chronic pain sample and measures distinguishing suicidal ideation and behavior could help to further clarify and provide a more comprehensive view of the role of EA in the pain-suicide association.

The above limitations notwithstanding, if the present study is replicated and extended in future experimental and clinical studies, this line of research could have important implications. Knowledge gleaned from this line of research could improve our ability to screen for and detect risk for suicide among patients who are experiencing pain. Better screening measures in this regard could have implications for prevention efforts in settings where suicide risk may not be routinely or thoroughly assessed (e.g., primary care). From an intervention perspective, it is possible that EA may be a modifiable therapeutic target that could simultaneously reduce dysfunction associated with pain and risk for suicide. For example, EA is a core construct of Acceptance and Commitment Therapy (ACT; Hayes et al., 2009), which in large part focuses on helping patients become more willing to accept and remain in contact with pain. For this reason, ACT has been successfully used to treat acute pain in veterans (Udell et al., 2018) and chronic pain for various medical conditions (Hughes et al., 2017; McCracken & Vellerman, 2010). Moreover, several studies have demonstrated the clinical utility of targeting EA to reduce suicide risk (Ellis & Ruffino, 2016; Zeifman et al., 2020). Future prospective clinical research should examine if leveraging or modifying existing interventions that target EA, such as ACT, is useful for protecting against risk for suicide among military veterans and others with pain conditions.