Szabo et al., 2021_Resilience predicts posttraumatic cognitions after a trauma reminder task and subsequent positive emotion induction among veterans with PTSD

Header: RESILIENCE AND POSTTRAUMATIC COGNITIONS Resilience Predicts Posttraumatic Cognitions after a Trauma Reminder Task and Subsequent Positive Emotion Induction among Veterans with PTSD Yvette Z Szabo1*,2, Sheila B Frankfurt1,2,3, A Solomon Kurz1,2, Austen R Anderson1,2, Adam P McGuire1*,2,4 VA VISN 17 Center of Excellence for Research on Returning War Veterans, Waco, TX, USA Central Texas Veterans Health Care System, Temple, TX, USA Texas A&M University, College of Medicine, College Station, TX, USA Department of Psychology and Counseling, The University of Texas at Tyler, Tyler, TX, USA Yvette Szabo https://orcid.org/0000-0002-5436-7081 Sheila Frankfurt https://orcid.org/0000-0002-8156-1488 Austen Anderson https://orcid.org/0000-0003-2585-8030 A Solomon Kurz https://orcid.org/0000-0001-7597-3745 Adam McGuire https://orcid.org/0000-0002-4512-1207 Author Note This manuscript has been accepted for publication in Psychological Trauma: Theory, Research, Practice, and Policy This postprint version is nearly identical to the accepted paper Corresponding: Yvette Szabo, yvette.szabo@va.gov or Adam McGuire adam.mcguire@va.gov; 4800 Memorial Drive (151C), Waco, TX 76711 RESILIENCE AND POSTTRAUMATIC COGNITIONS Acknowledgments This material is the result of work with resources and the use of facilities at the VISN 17 Center of Excellence for Research on Returning War Veterans and the Central Texas Veterans Health Care System Dr Szabo is supported by Career Development Award IK1-RX003122, Dr Frankfurt is supported by Career Development Award IK1-RX002427, and Dr McGuire is supported by a Small Projects in Rehabilitation Research Award I21-RX003035- from the United States (U.S.) Department of Veterans Affairs, Rehabilitation Research and Development Service The views expressed herein are those of the authors and not necessarily reflect the official policy or position of the Department of Veterans Affairs or the United States Government The authors not have any financial conflicts of interest to disclose Author Contributions (using CRediT categories) Conceptualization / Investigation: YS, AM Methodology / Data curation: AM, supporting YS Formal analysis: YS, supporting SK Validation: YS, supporting SK, AA Visualization: YS, SK Writing – Initial Draft – YS, with SF (Introduction) and AM (methods) drafting sections Writing – Editing and Revising – YS, SF, AK, AA, AM RESILIENCE AND POSTTRAUMATIC COGNITIONS Abstract Objective: Posttraumatic stress disorder (PTSD) is a common problem for veterans Resilience, the tendency to bounce back from difficult circumstances, is negatively associated with posttraumatic cognitions (PTCs) among individuals with a history of trauma, and thus it may be important to understand responses to trauma reminders Method: Using a quasi-experimental design, we examined the association between trait resilience and state PTCs in veterans with PTSD (n = 47, Mage = 48.60, 91.8% male) at two points: following a written trauma narrative exposure (Time [T1]), and following a subsequent positive distraction task (i.e., brief, positive video) (T2) Results: After controlling for PTSD symptom severity and combat exposure, resilience was negatively associated with PTCs at T1 (ΔR2 = 19) and T2 (ΔR2 = 13) However, resilience was a poor predictor of change in PTCs from T1 to T2 We also examined the relationship between resilience and subtypes of PTCs: resilience was associated with negative views of the self (T1, ΔR2 = 24) but not negative views of the world or self-blame (T1, ΔR2s < 07); these results were consistent at T2 Conclusions: Thus, resilience may attenuate negative trauma-related cognitions after trauma recall; however, this study was not designed to test causal pathways Future research could examine whether resilience-building exercises reduce negative PTCs after trauma reminders among veterans Additional research is needed to generalize to other trauma-exposed populations Keywords: resilience, trauma, PTSD, veteran, posttraumatic cognition RESILIENCE AND POSTTRAUMATIC COGNITIONS Clinical Impact Statement: PTSD is a highly prevalent and debilitating disorder for war veterans This study uses a quasiexperimental design to help further understanding about how resilience is related to one proposed mechanism of PTSD – posttraumatic cognitions – in response to reminders of the trauma Our findings suggest that resilience may buffer against negative trauma-related cognitions after trauma recall, though future research is needed to understand the long-term clinical utility RESILIENCE AND POSTTRAUMATIC COGNITIONS Resilience Predicts Posttraumatic Cognitions after a Trauma Reminder Task and Subsequent Positive Emotion Induction among Veterans with PTSD Approximately 23% of post-9/11 veterans are diagnosed with posttraumatic stress disorder (PTSD) (Fulton et al., 2015) Trauma-focused cognitive behavioral therapies (CBTs) such as Cognitive Processing Therapy (CPT; Resick et al., 2017) propose that negative or distorted beliefs about oneself and the world generated in the aftermath of a trauma give rise to PTSD (Resick, 2001) Some prototypical negative beliefs, called posttraumatic cognitions (PTCs), include ‘I am a bad person,’ ‘The world is a completely dangerous place,’ and ‘I am to blame for my trauma.’ Individuals with a history of trauma may try to avoid or distract themselves from their posttraumatic cognitions and thus miss potentially disconfirming experiences that could counter their negative beliefs Posttraumatic cognitions are believed to maintain PTSD and so are the focus of PTSD therapies such as CPT (LoSavio et al., 2017) However, little is known about predictors of individual differences in PTCs It is possible that treatment could target factors that influence PTCs and thus enhance treatment through individualized tailoring Resilience, the tendency to bounce back from difficult circumstances (Smith et al., 2008), may be one such modifiable factor that accounts for variance in PTCs In some studies, resilience was negatively associated with PTSD symptoms and positively associated with posttraumatic growth (e.g., Bensimon, 2012) In terms of evidence of an association with PTCs, greater resilience was associated with fewer PTCs among active-duty United States (U.S.) soldiers and veterans (Sexton et al., 2018; Zang et al., 2017) Although this suggests that resilience buffers trauma-related distress, a recent study reported that self-reported trait resilience did not predict resilience in PTSD-related impairment above the contribution of other personality factors, self- RESILIENCE AND POSTTRAUMATIC COGNITIONS reported psychological flexibility, and mental health symptoms among veterans (Meyer et al., 2019a) However, the latter study used a unique technique to derive their definition of resilience based on residuals in a PTSD-related factor score and thus, it is difficult to directly compare it to the larger body of research Because the specific mechanisms by which resilience might reduce PTSD are unknown, additional work is needed to further clarify how resilience is associated with PTCs As highlighted by previous research, resilience may play an important role in understanding individual differences in PTCs and PTSD symptom severity among traumaexposed people However, most previous research on the role of resilience in PTCs and PTSD has relied on cross-sectional or observational assessment (e.g., Infurna & Jayawickreme, 2019; Zang et al., 2017) Naturalistic research designs limit our ability to draw conclusions about the associations between trait resilience and state-level experiences of trauma-related cognitions For example, cross-sectional studies typically assess associations between global ratings of resilience and outcomes such as PTSD, failing to account for short-term fluctuations in the strength of belief in PTCs One exception to this trend was an experience sampling study that measured PTCs several times a day over days and revealed both between-person and within-person variability in PTCs (Carlson et al., 2016) Further, the strength of belief in PTCs are expected to change within individual sessions of CPT After the patient identifies a trauma-related belief and rates the strength of that belief on a scale of 0-100, they complete a worksheet that promotes cognitive restructuring, followed up re-rating the strength of their belief in that thought (Resick et al., 2017) Empirically, the degree of change in PTCs is associated with PTSD improvement during a course of treatment and often precede symptom improvement (Brown et al., 2019) Among patients receiving a trauma-focused CBT treatment, nearly a third experience sudden RESILIENCE AND POSTTRAUMATIC COGNITIONS gains in both trauma-related appraisals and PTSD symptoms, with the former predicting the latter (Wiedemann et al., 2020) Thus, observational research of daily life, the theory underlying effective evidence-based treatments, and clinical research findings all suggest that belief in PTCs fluctuate in ways that are clinically meaningful However, controlled study designs that provide a higher degree of sensitivity to changes in outcomes and internal validity are needed In line with the research demonstrating within-person variation in PTCs and the literature indicating that affect or mood impacts our thinking (Dolan, 2002; Forgas, 2017), we designed a controlled study to measure cognitive responses to specific stimuli to understand whether resilience can act as a personal resource in situations that might elicit PTCs The present study used data from a quasi-experimental study to examine the association between trait resilience and state PTCs at two points: following a written trauma narrative exposure (Time [T1]), and following a positive mood-inducing video (T2) We had three hypotheses: resilience would be negatively associated with PTCs following a trauma reminder (T1; Hypothesis [H1]); resilience would be negatively associated with PTCs following positive stimuli (T2; H2); and, resilience would be positively associated with decrease in PTCs from T1 to T2 assessments (H3) We assessed whether the association between resilience and PTCs at both timepoints differed by PTCs subscale and considered these analyses exploratory Method This study is part of a larger study on the role of positive psychology constructs in the development and maintenance of PTSD; detailed information about the parent study is reported elsewhere (McGuire & Mignogna, 2021) Briefly, the parent study recruited veterans between the ages of 18–90 with subclinical or clinical PTSD symptoms (defined as a positive screen on a symptom checklist, as described below) The sample was recruited from a southern Veterans RESILIENCE AND POSTTRAUMATIC COGNITIONS Affairs Hospital via mail, flyer, and provider referrals Exclusion criteria were active psychosis, active mania/hypomania, and cognitive disorders (e.g., dementia, traumatic brain injury) Eligibility was assessed during a phone screen Veterans were asked to self-report on the domains that were theoretically important for completing this study Additionally, mailed recruitment letters were only sent to Veterans without documented traumatic brain injury (TBI) or dementia in the medical record Of the 59 veterans who were screened, 48 veterans met criteria and participated in the study One participant was excluded because of repeated disruptions to their session that compromised the fidelity of session tasks Thus, the final sample comprised 47 veterans All participants provided written informed consent and this study was approved by the local Institutional Review Board Participants were financially compensated for their participation Procedures The present study was completed in a single session that lasted approximately 90 minutes First, participants completed baseline measures and wrote a trauma narrative describing their worst traumatic event for 15 minutes Next, participants completed the measure of PTCs (Time 1; T1) Then, participants were randomized to watch one of two positive emotion-inducing videos (one condition targeted amusement or humor and the other targeted moral elevation (i.e., inspiration by virtuous acts) Following the videos, all participants completed a brief reflection exercise where they wrote about their reactions to the videos for approximately 10 minutes Lastly, participants completed the PTCs measure for a second time (T2) Participants completed T1 and T2 PTCs approximately 50 minutes apart For the purposes of the current study, the two positive emotion-inducing video conditions were collapsed into one group because video condition did not predict PTCs at T2 RESILIENCE AND POSTTRAUMATIC COGNITIONS Measures The 20-item Posttraumatic Stress Disorder Checklist-5 (PCL-5; Weathers et al., 2013) assesses the respondent’s severity of PTSD symptoms over the past month Items are rated on a Likert-style scale from (not at all) to (extremely) Symptom severity was calculated by summing all items PCL-5 scores > 32 indicated subclinical or clinical PTSD symptom severity (Bovin et al., 2016) Previous research has demonstrated convergent and discriminant validity of the PCL-5 in veteran samples (Bovin et al., 2016) Mean scores on the PCL were comparable to a sample of veterans seeking treatment for PTSD (Zalta et al., 2018) Internal consistency in the present sample was α = 89 The PCL-5 was administered during the phone screen The 6-item Brief Resilience Scale (BRS; Smith et al., 2008) assesses one’s tendency to bounce back and recover from stressful situations The initial study demonstrated good internal consistency, factor structure, and convergent and discriminant validity in healthy and physical rehabilitation samples Items are rated on a Likert-style scale from (strongly disagree) to (strongly agree), with items were reverse scored Then, an item average score is calculated that ranges from to (Smith et al., 2008) Scores on the BRS were slightly lower but comparable to an online sample of veterans with mental health problems (Umucu et al., 2021) Internal consistency of the sum score in the present study was α = 74 The BRS was administered at baseline prior to the experimental procedure The 7-item Critical Warzone Experiences (CWE; Kimbrel et al., 2014) assesses negative warzone and military experiences, such as combat exposure The initial study demonstrated good internal consistency, factor structure, test-retest reliability, and validity, evidenced by correlations ranging from 51 to 71 with PTSD diagnosis and symptom severity, global functional impairment, anxiety and depression Each item is rated based on how often it occurred using a RESILIENCE AND POSTTRAUMATIC COGNITIONS 10 Likert-style (not at all) to (five or more times) scale Internal consistency in the present study was α = 81 Average scores on the CWE were higher than a recent sample of OEF/OIF Veterans recruited from the same geographical area (Meyer et al., 2019b) The CWE was administered at baseline The 36-item Posttraumatic Cognitions Inventory (PTCI; Foa et al., 1999) assesses strength of agreement with negative cognitions and beliefs associated with an index trauma Each item is rated from (totally disagree) to (totally agree) The present study modified the measure so each question was anchored to the Veteran’s cognitions “at this moment.” In addition to a total score, three subscales can be calculated: negative views of the self, negative views of the world, and self-blame The PTCI was administered twice, at T1 and T2 Internal consistency for the total measure was α = 94 at T1 and α = 96 at T2; see table S1 in supplementary materials for PTCI subscale reliability Average levels of the PTCI were higher at T1 and more comparable at T2 to levels of the PTCI among a sample of treatment seeking Veterans (Sexton et al., 2018) Data Analysis Plan Primary analyses were run in R (R Core Team, 2021), using the base R stats, psych (Revelle, 2020), and effsize (Torchiano, 2000) packages We also used additional online confidence interval calculators (Soper, 2006) There were no missing data Descriptive data are available in Table All hypotheses were tested using multiple linear regression For each of the three criterion (PCTI at T1, PTCs at T2, and change in PTCs), we fit two models: a baseline model and a theoretical model Each of the baseline models contained PTSD symptom severity and combat exposure as the control covariates The three theoretical models added resilience to the two control covariates, which allowed us to assess the unique variance resilience explained above and beyond the potential effects of baseline PTSD symptoms and history of combat RESILIENCE AND POSTTRAUMATIC COGNITIONS 12 $45,000, on average Scores on the PCL ranged from 33-80 Additional demographic information and military history is provided in Table Table Demographics and Participant Characteristics Characteristic % (n) Gender Male 91.5% (n = 43) Female 8.5% (n = 4) Ethnicity Hispanic / Latino 19.1% (n = 9) Not Hispanic / Latino 80.9% (n = 38) Racial identity* American Indian / Alaska Native 4.3% (n = 2) Black or African American 27.7% (n = 13) White or Caucasian 63.8% (n = 30) Asian or Asian American 2.1% (n = 1) Other 4.3% (n = 2) Branch of service* Air Force 4.3% (n = 2) Army 78.7% (n = 37) Marine Corps 10.6% (n = 5) National Guard 6.4% (n = 3) Navy 8.5% (n = 4) In what capacity did you serve?* Active Duty 97.9% (n = 46) Reserves 14.9% (n = 7) What is your gross annual income? $0-$14,999 10.6% (n = 5) $15,000-$29,999 19.1% (n = 9) $30,000-$44,999 31.9% (n = 15) $45,000-$59,999 23.4% (n = 11) $60,000-$74,999 8.5% (n = 4) $75,000-$89,999 6.4% (n = 3) Number of deployments Years of education Note N = 47, *categories were not mutually exclusive M (SD); range 1.47 (1.06); 0-5 13.51 (2.33); 10-20 H1: Resilience is Associated with PTCs After Recalling the Index Trauma The baseline model with the two control covariates explained 19% of the variance in PCTI at T1, R2 = 19, 95% CI [.00, 38] With the addition of resilience, the theoretical model explained 38% of the variance, ΔR2 = 19, [.18, 58] The change in f2 (0.31, [0.05, 0.76]) suggested resilience made a medium- sized contribution to the model See Table for the model RESILIENCE AND POSTTRAUMATIC COGNITIONS 13 comparison statistics of the primary hypotheses In the theoretical model, resilience was negatively associated with PTCs at Time 1, β = -0.51, p < 001 The coefficients for PTSD symptom severity and combat exposure were small and not statistically significant See Table for the coefficient summaries for the models of the primary hypotheses Table Correlation Matrix Variable M (SD) Resilience 2.78 (0.79) Combat exp 11.26 (6.46) PTSD symp 56.45 (11.85) PTCI T1 154.55 (37.59) NVS T1 85.72 (26.18) NVW T1 39.40 (6.15) SB T1 16.17 (7.61) PTCI T2 131.96 (41.69) NVS T2 71.11(27.19) 10 NVW T2 35.47 (9.55) 11 SB T2 14.11 (8.06) 12 Δ PTCI 22.60 (25.72) -.02 -.52*** 25 -.59*** 14 44** -.60*** 15 37* 98*** -.27 10 58*** 50*** 37* -.37* 08 30* 75*** 66*** 20 -.48*** 17 36* 80*** 73*** 37* 74*** -.52*** 19 33* 82*** 80*** 26 70*** 97*** 10 -.18 08 27 25 11 57*** 25 60*** 42** 11 -.37* 06 30* 71*** 65*** 17 90*** 77*** 73*** 20 12 -.08 -.07 05 17 24 13 11 -.46** -.38** -.61*** -.21 Note PTCI = Posttraumatic Cognitions Inventory, T1 = Time (post trauma narrative), NVS = Negative View of the Self, NVW = Negative Views of the World, SB = self-blame, T2 = Time (post positive emotion induction ) Table Model comparison for primary models Baseline model F and R2 Theoretical model F and R2 f2 95% CI Predicting PTCs at T1 F(2, 44) = 5.23, p = 009, R2 = 19 F(3, 43) = 8.84, p < 001, R2 = 38 0.310 [0.05, 0.76] (post-narrative) Predicting PTCs at T2 F(2, 44) = 3.53, p = 038, R2 = 14 F(3, 43) = 5.24, p = 004, R2 = 27 0.180 [0.01, 0.53] (post-positive video) Predicting change in PTCs F(2, 44) = 0.22, p = 801, R2 = 01 F(3, 43) = 0.19, p = 905, R2 = 01 0.003 [0.00, 0.12] Note Baseline model includes PTSD symptoms and combat exposure as covariates Theoretical model includes resilience as an independent variable PTCI = Posttraumatic cognitions inventory, 95% CI presented for f2 H2: Resilience is Associated with PTCs After a Positive Mood-Inducing Video The baseline model with the two control covariates explained 14% of the variance in PCTI, R2 = 14, 95% CI [0, 31] With the addition of resilience, the theoretical model explained 27% of the variance, ΔR2 = 13, [.07, 47] The change in f2 (0.18, [0.01, 0.53]) suggested resilience made a medium contribution to the model In the theoretical model, resilience was RESILIENCE AND POSTTRAUMATIC COGNITIONS 14 negatively associated with PTCs at Time 2, β = -0.42, p = 01 Again, the coefficients for PTSD symptom severity and combat exposure were small and not statistically significant Table Model parameters for primary models Baseline models B 95% CI Theoretical models  B 95% CI  0.00 PTSD symptoms Predicting PTCs at T1 (post-narrative) 75.6 189.5 [25.41, 125.93] 0.00 [112.18, 266.84] 1.36 [0.46, 2.25] 0.43 0.47 [-0.47, 1.40] Combat exposure 0.19 0.10 Intercept [-1.45, 1,84] 0.03 Resilience 0.56 -24.35 [-0.91, 2.03] [-37.88, -10.83] 0.15 -0.51** 0.00 PTSD symptoms Predicting PTCs at T2 (post-positive video) 58.0 162.3 [0.46, 115.61] 0.00 [69.02, 255.69] 1.20 [0.17, 2.23] 0.34 0.38 [-0.75, 1.51] Combat exposure 0.55 0.14 Intercept [-1.33, 2.43] 0.09 Resilience 0.89 -22.32 [-0.89, 2.66] [-38.64, -5.99] 0.11 -0.42** Predicting change in PTCs PTSD symptoms 17.6 0.16 Combat exposure -0.36 Intercept [-20.43, 55.70] 0.00 27.16 [-39.68, 94.00] 0.00 [-0.52, 0.84] 0.07 0.08 [-0.72, 0.89] 0.04 [-1.60, 0.89] -0.09 -0.33 [-1.60, 0.95] -0.08 Resilience -2.04 [-13.73, 9.65] -0.06 Note B = unstandardized regression coefficient, CI = Confidence Interval, β = standardized regression coefficient, PTC = Posttraumatic cognitions, PTSD = Posttraumatic stress disorder, 95% CI presented for R2 H3: Resilience Does Not Account for Change in PTCs PTCs declined from T1 to T2, t (46) = -6.02, p < 001, d = -0.56, [-0.76, -0.36] However, resilience was a poor predictor of change in PTCs The baseline model with the two control covariates explained 1% of the variance in PCTI, R2 = 01, 95% CI [0, 06] With the addition of resilience, the theoretical model explained 1% of the variance, ΔR2 =