Anxiety sensitivity and social support in veterans with emotional disorders

Abstract

Objective:

An understanding of the incremental value of social support in predicting psychopathology above transdiagnostic risk factors could speak to the benefit of leveraging social factors into existing, evidence-based interventions in veterans with emotional disorders. This cross-sectional study aimed to expand our understanding of associations between domains of anxiety sensitivity and facets of psychopathology in veterans with emotional disorders. We also determined whether social support predicted psychopathology above anxiety sensitivity domains and combat exposure and explored these relationships with a path model.

Methods:

One hundred and fifty-six treatment-seeking veterans with emotional disorders completed diagnostic interviews and assessments of demographics, social support, symptom measures (e.g., PTSD, depression, anxiety, and stress), and transdiagnostic risk factors (i.e., anxiety sensitivity). After data screening, 150 were included in regressions.

Results:

Using regression analyses with cross-sectional data, cognitive anxiety sensitivity concerns predicted PTSD and depression above combat exposure. Cognitive and physical concerns predicted anxiety, and cognitive and social concerns predicted stress. Above combat exposure and anxiety sensitivity, social support predicted PTSD and depression.

Conclusion:

Focusing on social support in tandem with transdiagnostic mechanisms in clinical samples is critical. These findings inform transdiagnostic interventions and recommendations related to incorporation of assessment of transdiagnostic factors in clinical contexts.

1 |. INTRODUCTION

Veterans are at elevated risk for experiencing mental health problems (Hoge et al., 2004; Milliken et al., 2007), and diagnostic comorbidity is the rule rather than the exception (Knowles et al., 2019). Approximately 23% of veterans meet diagnostic criteria for posttraumatic stress disorder (PTSD; Fulton et al., 2015), 12% for generalized anxiety disorder (Milanak et al., 2013), 12% for depressive disorders (Liu et al., 2019), and 8% for panic disorder (Gros et al., 2011). Prevalence of comorbid emotional disorders is associated with more elevated psychopathology compared to those with single diagnoses (Gros et al., 2022). Further, rates of the triple comorbidity of PTSD, clinical anxiety, and clinical depression among veterans (26.7%–30.1%) are higher than PTSD alone or PTSD comorbid with anxiety or depression (Ginzburg et al., 2010). Identifying mechanisms that maintain psychopathology and underlie comorbidity, such as transdiagnostic protective and risk factors, is paramount to effectively alleviating psychological distress (Barlow et al., 2011) in veterans presenting to psychotherapy.

Anxiety sensitivity (AS) is a well-established transdiagnostic risk factor (Vujanovic et al., 2018) implicated in the etiology and maintenance of both psychopathology (Bernstein & Zvolensky, 2007; Scher & Stein, 2003) and chronic medical conditions (Horenstein et al., 2018). AS is the tendency to interpret innocuous anxiety symptoms (e.g., increased heart rate, sweating) as having disastrous physical (e.g., heart attack), cognitive (e.g., losing control), or social (e.g., rejection) consequences, subsequently amplifying arousal and distress (Reiss & McNally, 1985). AS is hypothesized to exacerbate risk for conditioning of fear responses and is positively associated with emotional and behavioral difficulties commonly experienced among veterans, including PTSD and pain (Armstrong et al., 2021; Naragon-Gainey, 2010; Overstreet et al., 2018; Raines et al., 2022). Though AS is thought to be centrally related to the phenomenology of PSTD and panic disorder (Olatunji & Wolitzky-Taylor, 2009), it still shows meaningful, strong associations with depression symptoms (Mitchell et al., 2014; Naragon-Gainey, 2010).

The association of AS and PTSD or depression symptoms is particularly relevant and well-established in veterans (Overstreet et al., 2018), with cognitive concerns appearing to be particularly salient and corresponding to trauma-related and depressogenic cognitive appraisals (Mitchell et al., 2014; Raines et al., 2017). AS is highly modifiable, and previous research has demonstrated that reducing AS through brief intervention can yield improvements in a range of clinical outcomes (Fitzgerald et al., 2021). Accordingly, AS is one potential mechanism that may contribute to comorbidity and symptom severity among veterans with emotional disorders across psychological outcomes, and that can be directly modified to improve treatment outcomes. As such, understanding associations between facets of AS and domains of psychopathology among a highly comorbid veteran sample may provide further insight regarding modifiable psychological risk factors.

In addition to psychological risk factors, numerous sociocultural factors are identified as critical to etiology and maintenance of psychopathology in veterans, including social determinants of health, trauma-specific factors, and social support. Social support may be of particular interest, as it functions as both a risk and protective factor for adverse psychological outcomes in veterans (McGuire et al., 2018; Zalta et al., 2021). It is posited that there is a reciprocal association between PTSD and social support, which is informed by both social causation and social erosion models suggesting that social support may both impact and be impacted by PTSD (Johnson et al., 1999; Shallcross et al., 2016). Additionally, social support predicts domains of psychopathology other than PTSD in veterans, such as anxiety and depression symptoms (Romero et al., 2015). Social support can be modified at the individual level, unlike other sociocultural (e.g., minoritized race/ethnicity) or trauma-specific factors (e.g., trauma history). Among veterans, social support moderates mental health treatment outcomes (Campbell et al., 2020; Price et al., 2018). Interestingly, social support more strongly predicts PTSD in studies of combat trauma (Ozer et al., 2003), suggesting that it may also be necessary to account for effects of combat exposure in predicting PTSD, as well as in predicting depression symptoms (Wells et al., 2010) and psychopathology more broadly (Campbell et al., 2021) while simultaneously examining the role of social support.

The literature clearly identifies both AS and social support as having an important role in the onset, course, and treatment of psychopathology and especially affective comorbidities among veterans. Yet, evidence-based interventions in clinical contexts may focus solely on targeting of transdiagnostic mechanisms or symptoms, without leveraging social support. However, social support is a pertinent protective factor for psychopathology, buffers the impact of stressful events on psychopathology, and may play a role in treatment entry (Kelly et al., 2010). Therefore, we would expect that social support may predict psychopathology above transdiagnostic mechanisms. Understanding the unique contribution of social support to different forms of psychopathology (Shallcross et al., 2016), and in particular above AS, could aid in adaptation of evidence-based interventions for veterans with emotional disorders and elucidate how AS and environmental factors uniquely relate to psychopathology. In particular, understanding unique contributions of social support could inform transdiagnostic interventions that have the ability to leverage social support (Coyne & Gros, 2022) to enhance treatment outcomes for veterans with comorbidities.

Accordingly, the current cross-sectional study aimed to examine associations between AS and psychopathology in a sample of veterans with high rates of psychiatric comorbidity presenting for psychotherapy. We also examined whether social support predicted PTSD, depression symptoms, anxiety, and stress above both AS (Cox et al., 2001) and combat exposure (Hoge et al., 2004). Specifically, we hypothesized that each domain of AS would positively predict PTSD, depression symptoms, anxiety, and stress, controlling for combat exposure (Campbell et al., 2021; Ozer et al., 2003; Wells et al., 2010) given the role of combat exposure in contributing to symptomatology and second, that greater social support would negatively predict PTSD, depression symptoms, anxiety, and stress, controlling for effects of AS and combat exposure. We chose to examine whether AS and social support predicted different domains of psychopathology, given differential associations between AS, social support, and psychological outcomes in the extant literature (e.g., Kaniasty, 2012; Olatunji & Wolitzky-Taylor, 2009; Wang et al., 2018) and the potential utility these results may provide within the context of clinical interventions. Finally, we conducted an exploratory path analysis integrating results from regression models to test a parsimonious model of the regression findings.

2 |. MATERIALS AND METHODS

2.1 |. Participants

The present data were collected via baseline assessments for two ongoing randomized clinical trials. The sample consisted of 156 veterans requesting evidence-based psychotherapy for symptoms of depression and anxiety at primary care and mental health clinics within a large Southeastern Veterans Affairs Health Care System (VAHCS). Study inclusion criteria involved: (1) competence to complete study consent and procedures, (2) DSM-5 diagnostic criteria for a principal diagnosis of panic disorder and/or agoraphobia, PTSD, or major depressive disorder, and (3) participant age of 18 to 80 years old. Study exclusion criteria involved: (1) recent history (≤2 months) of psychiatric hospitalization or a suicide attempt, (2) acute, severe illness or medical condition that likely will require hospitalization and/or otherwise interfere with study procedures, (3) recent start of new psychiatric medication (≤4 weeks), or (4) diagnosis of a significant psychotic disorder, personality disorder, and/or bipolar disorder based on medical chart review.

2.2 |. Measures

2.2.1 |. Structured assessment of emotional disorders

The Anxiety and Related Disorders Interview Schedule for DSM-5 (ADIS-5; Brown & Barlow, 2014) was utilized to formally assess current emotional disorders. The ADIS-5 is a semi–structured, clinician-administered diagnostic interview and assesses anxiety disorders and common comorbidities (e.g., depressive disorders). Examiners assess the severity of each emotional disorder on a 0–8 scale (e.g., 0 = none, 8 = very severely disturbing/disabling), and scores of 4 or higher are generally indicative of clinically significant scores for each respective disorder. The ADIS-5 has strong psychometric properties (Brown & Barlow, 2014). While ADIS reliability was not calculated for the present study, extant studies from the current authorship team replicating the methodology used in the current study demonstrate excellent inter-rater reliability for PTSD (100%), major depressive disorder (85.0%), and panic disorder/agoraphobia (100%) in 20% of interviews scored by an independent rater (Gros & Allan, 2019).

2.2.2 |. Depression, anxiety, and stress

The Depression Anxiety Stress Scales-21 (DASS-21; Lovibond & Lovibond, 1995) is a 21-item self-report measure assessing depression, anxiety, and stress symptoms over the past week. The current study used each of the three subscales, which measure dysphoric mood, as well as physiological and cognitive symptoms. The DASS-21 employs a 4-point Likert-type scale (e.g., 0 = Did not apply to me at all, 3 = Applied to me very much or most of the time), where participants rate the extent to which they have experienced depression, anxiety, and stress symptoms over the past week. Summed scores for each subscale range from 0 to 21, and higher scores indicate more severe symptoms. The DASS-21 has strong psychometric properties (Lovibond & Lovibond, 1995). Internal consistency α found in the current study was 0.91 for depression, 0.87 for anxiety, and 0.88 for stress, respectively.

2.2.3 |. Anxiety sensitivity

The Anxiety Sensitivity Index-3 (ASI-3; Taylor et al., 2007) is an 18-item self-report measure and was utilized to assess physical, social, and cognitive domains of AS. No timeframe is defined for the ASI-3; participants answered based on whether they agree with each item. The ASI-3 employs a 5-point Likert-type scale (e.g., 0 = Very Little, 4 = Very Much). Scores are summed, with higher scores indicating greater AS. Psychometric properties for the ASI-3 are strong (Taylor et al., 2007). Internal consistency found in the current study was 0.88 for the physical subscale, 0.89 for the cognitive subscale, and 0.81 for the social subscale.

2.2.4 |. Posttraumatic stress

The PTSD Checklist for DSM-5 (PCL-5; Blevins et al., 2015) is a 20-item self-report measure utilized to assess PTSD symptoms over the past month. The PCL-5 employs a 5-point Likert-type scale (e.g., 0 = Not at all, 4 = Extremely) and measures the extent to which participants have experienced PTSD symptoms. Scores are summed and range from 0 to 80, and higher scores indicate more severe PTSD symptoms. The PCL-5 has excellent psychometric properties. Internal consistency found in the current study was 0.93 for the PCL-5.

2.2.5 |. Social support

The Medical Outcomes Study Social Support Survey (MOS; Sherbourne & Stewart, 1991) is a 19-item self-report measure utilized to assess social support. No timeframe is defined for the MOS. The MOS employs a 6-point Likert-type scale (e.g., 1 = All of the time, 6 = None of the time) and assesses perceived social support across different domains (i.e., emotional/informational support; tangible support; affectionate support; positive social interaction). A final item measures whether the participant has “Someone to do things with to help you get your mind off things.” For purposes of the current study, a mean score was calculated (Sherbourne and Stewart, 1991), with mean scores ranging from 1 to 6. In the case of the MOS, higher scores are indicative of lower levels of social support. The MOS exhibits adequate reliability and validity (Sherbourne and Stewart, 1991). Internal consistency found in the current study was 0.96.

2.2.6 |. Combat exposure

To assess combat exposure, a single self-report item was used that asked participants the following: “Have you ever participated in combat (e.g., fired your weapon or were fired upon)?” Participants were able to select either “Yes” or “No” for this item.

2.3 |. Procedure

Participants were recruited from October 2020 to December 2022. Project staff advertised the study to local providers in the primary care and outpatient mental health clinics within the VAHCS. Interested participants were referred to project staff and scheduled for an intake appointment to complete consent documents, evaluate study inclusion and exclusion criteria, and complete diagnostic and self-report measures. Additional study procedures were completed for the larger study, such as randomization to therapy condition, but omitted here as they were not relevant to the present study.

2.4 |. Data-analytic plan

2.4.1 |. Data screening

Before conducting main analyses, data were screened for accuracy errors, missing data, multivariate outliers (i.e., using Mahalanobis distance), and multivariate assumptions. Fully Conditional Specification was used to replace missing data. Analyses were conducted in R and JASP.

2.4.2 |. Pearson correlations

Pearson correlations were calculated between the following variables: PTSD, depression symptoms, anxiety, stress, each domain of AS, and social support.

2.4.3 |. Hierarchical regression

Four hierarchical multiple regression analyses assessed whether each predictor variable (i.e., combat exposure, AS domains, social support) predicted PTSD (PCL-5), depression (DASS-21), anxiety (DASS-21), and stress (DASS-21). A three-step multiple regression procedure was followed. First, the dichotomous combat exposure variable was entered into step one. Then, the three domains of AS (ASI-3) were entered at step two. Step three included social support scores (MOS). R² and F change statistics were calculated for each model.

3 |. RESULTS

3.1 |. Data screening

We screened the data for accuracy errors, missing data, outliers, and multivariate assumptions. Of the 156 veterans, 6 participants were missing data for the combat exposure question and were excluded from regression analyses. Twenty-six participants missing data on measures used in the main analysis had their data imputed; the remainder (n = 130) had complete data. Specifically, data were imputed using the mice package in R (Van Buuren & Groothuis-Oudshoorn, 2011). The mice package utilizes Fully Conditional Specification (Van Buuren & Groothuis-Oudshoorn, 2011) to facilitate replacement of missing data. There were no identified outliers using Mahalanobis distance. Again, the final sample for regression analyses included 150 participants after data screening.

3.2 |. Sample demographic and diagnostic characteristics

Participants were middle-aged adults (M_age = 46.55, SD_age = 14.25) and mostly White (55.1%) or Black/African American (34.6%). Participants were predominantly male (69.9%). Regarding employment, many participants indicated that they were employed full-time (38.5%). Most participants indicated that they were Army veterans (48.1%) and served in Afghanistan/Iraq (33.3%). Finally, 39.1% of veterans reported engaging in combat.

In total, 46.2% of the sample met diagnostic criteria for a principal diagnosis of major depressive disorder, 42.3% met diagnostic criteria for a principal diagnosis of PTSD, and 9.6% met diagnostic criteria for a principal diagnosis of panic disorder. Comorbidity was common among the sample, with an overall comorbidity rate (i.e., at least two emotional disorders) of 75.6%.

3.3 |. Pearson correlations

Physical AS (r = 0.277, p < 0.001, Lower 95% CI = 0.126, Upper 95% CI = 0.416), cognitive AS (r = 0.461, p < 0.001, Lower 95% CI = 0.327, Upper 95% CI = 0.576), and social AS (r = 0.278, p < 0.001, Lower 95% CI = 0.127, Upper 95% CI = 0.417) were positively correlated with depression symptoms. Using effect size interpretation guidelines from Cohen (1988), associations were small to medium for physical AS, medium to large for cognitive AS, and small to medium for social AS using confidence intervals.

Additionally, physical AS (r = 0.583, p < 0.001, Lower 95% CI = 0.469, Upper 95% CI = 0.678), cognitive AS (r = 0.604, p < 0.001, Lower 95% CI = 0.494, Upper 95% CI = 0.695), and social AS (r = 0.500, p < 0.001, Lower 95% CI = 0.372, Upper 95% CI = 0.609) were positively correlated with anxiety. Associations were medium to large for all relationships. Physical AS (r = 0.269, p < 0.001, Lower 95% CI = 0.117, Upper 95% CI = 0.409), cognitive AS (r = 0.401, p < 0.001, Lower 95% CI = 0.260, Upper 95% CI = 0.525), and social AS (r = 0.407, p < 0.001, Lower 95% CI = 0.267, Upper 95% CI = 0.530) were related to increased stress. These associations were small to medium for physical AS and small to large for cognitive AS and social AS.

Physical (r = 0.025, p > 0.05, Lower 95% CI = −0.133, Upper 95% CI = 0.182), cognitive (r = 0.074, p > 0.05, Lower 95% CI = −0.084, Upper 95% CI = 0.228), and social AS (r = 0.006, p > 0.05, Lower 95% CI = −0.152, Upper 95% CI = 0.163) were not related to social support and were all indicative of small associations. Physical (r = 0.377, p < 0.001, Lower 95% CI = 0.234, Upper 95% CI = 0.504), cognitive (r = 0.517, p < 0.001, Lower 95% CI = 0.391, Upper 95% CI = 0.623), and social AS (r = 0.356, p < 0.001, Lower 95% CI = 0.211, Upper 95% CI = 0.486) were positively associated with greater PTSD. These associations were small to large for physical AS, medium to large for cognitive AS, and small to medium for social AS. Separate Pearson correlations for principal diagnoses of both PTSD and major depressive disorder are presented at the following open science framework (OSF) link: https://osf.io/vsf4j/?view_only=806a90b84dbd46a4b85ee386adf7adc6.

3.4 |. Hierarchical regression analyses

Using a Bonferroni correction based on four regression analyses (0.05/4), we used a significance level of 0.0125 to guide interpretation of regression results for each outcome (PTSD, depression, anxiety, stress). See Table 1 for regression statistics for coefficients for the models discussed below. Step 1 of the first hierarchical regression was computed to examine the association between combat exposure and PTSD. Combat exposure accounted for minimal variance in PTSD and was not a significant predictor, F(1, 148) = 3.36, p = 0.069, Adjusted R² = 0.02. In the second step of the model, AS domains were entered next, accounting for an additional 28% of the variance in PTSD, ΔF(3, 145) = 18.87, p < 0.001, ΔR² = 0.28. In the second step with combat exposure and all three AS domains, only combat exposure and cognitive AS predicted PTSD. In particular, those who reported not having experienced combat reported lower PTSD, while higher cognitive AS was related to greater PTSD. Finally, social support was entered in the third and final step of the model, accounting for an additional 4% of variance in PTSD, ΔF(1, 144) = 8.60, p = 0.004, ΔR² = 0.04. Lower social support predicted greater PTSD.

TABLE 1.

Hierarchical regression analyses predicting posttraumatic stress, depression, anxiety, and stress.

	Step	β	b	SE	t	p Value	Partial	95% lower CI	95% upper CI
Posttraumatic stress
Combat exposure	1	−.15	−4.98	2.72	−1.83	.069	−.15	−10.34	0.39
ASI physical	2	.03	0.08	0.26	0.33	.746	.03	−0.42	0.59
ASI cognitive	2	.43	1.09	0.26	4.20	<.001	.33	0.58	1.61
ASI social	2	.12	0.33	0.26	1.26	.211	.10	−0.19	0.85
Social support	3	.20	2.55	0.87	2.93	.004	.24	0.83	4.27
Depression
Combat exposure	1	.06	0.69	0.96	0.72	.472	.06	−1.21	2.60
ASI physical	2	−.07	−0.06	0.10	−0.61	.543	−.05	−0.25	0.13
ASI cognitive	2	.48	0.44	0.10	4.50	<.001	.35	0.24	0.63
ASI social	2	.03	0.03	0.10	0.31	.760	.03	−0.16	0.22
Social support	3	.23	1.02	0.32	3.18	.002	.26	0.39	1.66
Anxiety
Combat exposure	1	.02	0.25	0.94	0.27	.792	.02	−1.61	2.11
ASI physical	2	.28	0.24	0.08	3.06	.003	.25	0.09	0.39
ASI cognitive	2	.34	0.30	0.08	3.79	<.001	.30	0.15	0.46
ASI social	2	.14	0.14	0.08	1.71	.089	.14	−0.02	0.30
Social support	3	.04	0.17	0.27	0.61	.540	.05	−0.37	0.71
Stress
Combat exposure	1	−.003	−0.04	0.91	−0.04	.968	−0.003	−1.83	1.75
ASI physical	2	−.10	−0.08	0.09	−0.89	.376	−0.07	−0.26	0.10
ASI cognitive	2	.29	0.25	0.09	2.70	.008	.22	0.07	0.43
ASI social	2	.30	0.29	0.09	3.12	.002	.25	0.11	0.47
Social support	3	.10	0.41	0.31	1.32	.190	.11	−0.21	1.02

Note: Regression statistics here reflect statistics when that particular variable(s) was/were added to the model. Beta values are indicative of standardized coefficients, with small b values indicative of unstandardized coefficients. Bold p Values are significant at the p < .05 level.

Abbreviations: ASI, Anxiety Sensitivity Index-3; Partial, partial correlations.

Next, we followed the same hierarchical regression procedure for depression symptoms, anxiety, and stress outcomes; the only difference is the outcome variable used. Step 1 of the hierarchical regression was computed to examine the association between combat exposure and depression. Combat exposure accounted for minimal variance in depression and was not a significant predictor, F(1, 148) = 0.52, p = 0.472, Adjusted R² = −0.003. In the second step of the model, AS domains were entered next, accounting for an additional 21% of the variance in depression, ΔF(3, 145) = 12.77, p < 0.001, ΔR² = 0.21. After entering AS domains to predict depression, only cognitive AS predicted depression symptoms, such that higher cognitive AS was related to greater depression symptoms. Finally, social support was entered in the final step of the model, accounting for an additional 5% of the variance in depression, ΔF(1, 144) = 10.13, p = 0.002, ΔR² = 0.05. Lower social support predicted greater depression symptoms.

Step 1 of the hierarchical regression examined the association between combat exposure and anxiety. Combat exposure accounted for minimal variance in anxiety and was not a significant predictor, F(1, 148) = 0.07, p = 0.792, Adjusted R² = −0.01. In the second step of the model, AS domains were entered, accounting for an additional 44% of the variance in anxiety, ΔF(3, 145) = 38.04, p < 0.001, ΔR² = 0.44. In examining predictors for Step 2, cognitive and physical AS, but not social AS, predicted anxiety, such that higher cognitive and physical AS was related to greater anxiety. Finally, social support was entered in the final step of the model and did not account for additional variance in anxiety, ΔF(1, 144) = 0.38, p = 0.540, ΔR² = 0.001.

Finally, for stress, the final dependent variable, Step 1 of the hierarchical regression was computed to examine the association between combat exposure and stress. Combat exposure accounted for minimal variance in stress and was not a significant predictor in the initial model, F(1, 148) = 0.002, p = 0.968, Adjusted R² = −0.01. In Step 2, AS domains were entered, accounting for an additional 21% of the variance in stress, ΔF(3, 145) = 12.95, p < 0.001, ΔR² = 0.21. In the stress model at Step 2, cognitive and social AS, but not physical AS, predicted stress, such that higher cognitive and social AS were related to greater stress. For Step 3 which added social support, addition of social support to the model was not an improvement and accounted for minimal variance in stress, ΔF(1, 144) = 1.74, p = 0.190, ΔR² = 0.01.¹

4 |. DISCUSSION

The current cross-sectional study aimed to understand associations between AS, social support, and psychopathology in a clinical sample of treatment-seeking veterans with emotional disorders. We also identified the relative contributions of AS and social support in symptoms of PTSD, depression symptoms, anxiety, and stress, above factors (i.e., combat exposure) that predict veteran symptomatology (Campbell et al., 2021; Ozer et al., 2003; Wells et al., 2010). Overall, 75.6% of patients met criteria for two or more emotional disorders. Domains of AS were universally, yet differentially associated with each outcome, while social support predicted PTSD and depression symptoms. An exploratory path model integrating these results was a good fit to the data. Results underscore the differential impact of AS facets and social support across PTSD, depression symptoms, anxiety, and stress, and the need to consider both individual and environmental factors to adequately address comorbid symptomatology among veterans with emotional disorders.

AS was significantly and positively associated with psychopathology. These correlational findings are congruent with meta-analytic evidence indicating AS is a salient risk factor for emotional disorders (Naragon-Gainey, 2010). Additionally, social support was associated with PTSD and depressive symptoms, but not with anxiety or stress. Social support may be particularly relevant within the context of PTSD and depression symptoms. These associations are somewhat reflective of the literature, with strong associations between social support, depression symptoms, psychological well–being, and PTSD (e.g., Kaniasty, 2012; Wang et al., 2018). Despite mixed results, social support is related to anxiety and stress (Budge et al., 2013; Cohen & McKay, 2020). Additional work may consider use of intensive longitudinal methods, such as experience sampling, to examine temporal associations between social support, anxiety, and other environmental factors, or moderators to understand who would benefit most from leveraging social support.

Beyond relationships discussed above, regression models and the exploratory path model depicted a nuanced relationship of AS and social support to PTSD, depression symptoms, anxiety, and stress symptoms among treatment-seeking veterans with emotional disorders. First, PTSD and depression symptoms were predicted by AS cognitive concerns and social support. The unique contribution of AS cognitive concerns to PTSD and depression symptoms (while controlling for combat exposure) is reflective of the broader literature (Farris et al., 2016; Mitchell et al., 2014; Saulnier et al., 2018). Despite controlling for pertinent factors predicting psychopathology, social support maintained a contribution to PTSD and depression symptoms. Social support, an environmental factor, may act as a protective or risk factor for veterans, even in the context of other robust risk factors (e.g., combat exposure, AS). Indeed, social support buffers the development of suicidal ideation (DeBeer et al., 2014), and lack of social support is associated with greater severity of psychopathology among veterans (Cox et al., 2017). Having the support of loved ones may also promote engagement in treatment (e.g., attendance, exposure, behavioral activation), thereby contributing to greater reductions in PTSD and depressive symptoms (Campbell et al., 2022; Price et al., 2018). Leveraging environmental factors (e.g., social support), in addition to targeting other modifiable psychological risk factors including AS, may optimize treatment outcomes for veterans with symptoms of PTSD and/or depression symptoms.

It is expected that physical concerns of AS would predict physiological anxiety. AS physical concerns correlate with panic disorder and PTSD (Naragon-Gainey, 2010; Olatunji & Wolitzky-Taylor, 2009), with both disorders characterized by fear of physiological sensations. AS is an amplifier of distress; being preoccupied with and fearf of physiological sensations of anxiety (e.g., heart rate increase, sweating) likely results in higher anxiety symptoms. However, the finding that AS social concerns predicted stress is somewhat surprising, given that this AS facet is typically most closely associated with social anxiety (Wheaton et al., 2012) which was likely present to a smaller degree as it was not specifically targeted in the present study. Previous work has suggested that AS social concerns are associated with threat sensitivity (McWilliams et al., 2000), which could account for their relation to stress in our sample of veterans with comorbid emotional disorders. That is, nearly all anxiety disorders are characterized by preoccupation with a perceived threat (e.g., negative evaluation, unforeseeable event occurring, danger), resulting in heightened stress (Connor et al., 2007).

Together, findings highlight cognitive concerns of AS as a unifying risk factor across symptoms of PTSD, depression symptoms, anxiety, and stress, indicating that AS cognitive concerns may be a particularly salient transdiagnostic target for addressing psychopathology among veterans with comorbidities. Research has demonstrated that AS cognitive concerns are uniquely associated with both PTSD and major depressive disorder (Allan et al., 2014), the two most prevalent diagnoses across our sample. Still, cognitive concerns, such as fear of losing control, span other anxiety disorders, including generalized anxiety disorder, panic disorder, health anxiety, social anxiety, and obsessive-compulsive disorder (see Wheaton et al., 2012, for an overview). Single-session interventions specifically targeting AS cognitive concerns demonstrate promise in reduction of fear of cognitive dyscontrol in high AS individuals (Schmidt et al., 2014). Additionally, Schmidt et al. (2007) showed that individuals reporting high overall levels of AS experienced significant reductions in AS after a brief intervention. Thus, AS concerns may be an optimal transdiagnostic mechanism to target through brief, relatively easy-to-administer interventions to individuals with heterogeneous emotional difficulties. Integration of brief assessments of AS may also be appropriate to incorporate in structured assessments for individuals presenting with PTSD and depression symptoms to (1) identify meaningful connections between AS and symptom domains and (2) facilitate transdiagnostic interventions targeting modifiable factors.

The results of the present study are limited by several factors. Given that this is a cross-sectional study, results are limited in interpretation. It is possible that outcome variables and AS are reciprocally related. Future studies may choose to expand this work by understanding reciprocal associations between AS and different domains of psychological dysfunction in veterans with diagnostic comorbidities. Additionally, this study consisted of primarily males, which limits generalizability to veterans with other gender identifies. The single-item assessment of combat exposure that incorporated a yes/no format may also not be as reliable compared to service record inspection. Regarding constructs assessed, PTSD, anxiety, and stress were strongly and positively correlated (albeit not to the extent that would be considered problematic due to multicollinearity).

Given high levels of comorbidity across treatment-seeking veteran samples as evidenced by this study and others (Ginzburg et al., 2010; Knowles et al., 2019), and the clinical and economic value of addressing comorbidities concurrently rather than sequentially with single-disorder protocols (Gros et al., 2016; Kline et al., 2021), utilizing transdiagnostic mechanisms is essential for maximizing treatment outcomes. Our results suggest that leveraging social support may be particularly impactful when targeting PTSD and depression symptoms among veterans, while reducing AS would crosscut a variety of anxiety-related outcomes. Given symptom overlap and elevated diagnostic comorbidity in veterans, it is essential to examine transdiagnostic mechanisms in conjunction with other environmental factors (e.g., social support) across the spectrum of psychopathology to inform targeted, transdiagnostic interventions and determine when social support may be relevant. To provide one example of such an intervention, Transdiagnostic Behavior Therapy, which is effective in reducing symptoms for veterans with emotional disorders (Gros, 2014; Gros & Allan, 2019), focuses on increasing social engagement using positive emotional exposures (i.e., behavioral activation). Such protocols may also incorporate family members into treatment discussions/approaches to facilitate supports who may help promote consistent treatment engagement (Campbell et al., 2022; Price et al., 2018). Future work may also integrate longitudinal studies within comorbid samples to further understand these relationships and identify optimal treatment approaches.

5 |. CONCLUSION

AS cognitive concerns emerged as a unifying transdiagnostic factor predicting varied psychopathology among treatment-seeking veterans, despite controlling for combat exposure given its connection to symptomatology (Campbell et al., 2021; Ozer et al., 2003; Wells et al., 2010). AS, and as a result psychological distress, may be reduced through brief prevention and intervention (Fitzgerald et al., 2021), and as such, is well-positioned to improve numerous psychological and functional outcomes among veterans with PTSD and/or depression, and potentially other psychiatric comorbidities. Concurrently leveraging social support may also optimize treatment outcomes for individuals with PTSD and depression. Future clinical trials and dismantling studies may seek to understand the additive effect of including social support and AS on psychopathology to improve treatment engagement and clinical outcomes.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.