Moderators of PTSD symptom change in group cognitive behavioral therapy and group present centered therapy

https://doi.org/10.1016/j.janxdis.2021.102386Get rights and content

Highlights

  • Moderators of change during 2 group-based interventions were explored for specific symptom clusters of PTSD.

  • Group Cognitive-Behavioral Treatment (trauma-focused) and Group Present Centered Therapy (non-trauma focused) were studied.

  • For GCBT, pre-treatment PTSD moderated avoidance reduction while treatment expectancies moderated reduction in negative cognitions/mood.

  • For GPCT, pre-treatment depression moderated change in intrusion and arousal/reactivity symptoms.

  • Results support the importance of examining trajectories of change and their moderators for specific treatments.

Abstract

To examine moderators of change during group-based intervention for Posttraumatic Stress Disorder (PTSD), multilevel models were used to assess trajectories of symptom clusters in male veterans receiving trauma focused Group Cognitive Behavioral Treatment (gCBT; N = 84) or non-trauma focused Group Present Centered Therapy (gPCT; N = 91; Sloan et al., 2018). Separate models were conducted for symptom clusters in each intervention, examining pre-treatment PTSD symptoms, pre-treatment depression severity, age, index trauma, and outcome expectancies as potential moderators. Unconditioned growth models for both gCBT and gPCT showed reductions in intrusions, avoidance, negative cognitions/mood, and arousal/reactivity (all p < .001). Distinct moderators of recovery emerged for each treatment. Reductions in avoidance during gCBT were strongest at high levels of pre-treatment PTSD symptoms (low PTSD: p =  .964, d = .05; high PTSD: p < .001, d = 1.31) whereas positive outcome expectancies enhanced reductions in cognitions/mood (low Expectancy: p =  .120, d = .50; high Expectancy: p < .001, d = 1.13). For gPCT, high levels of pre-treatment depression symptoms negatively impacted change in both intrusion (low depression: p < .001, d = .96; high depression: p =  .376, d = .22) and arousal/reactivity (low depression: p < .001, d = .95; high depression: p =  .092, d = .39) symptoms. Results support the importance of examining trajectories of change and their moderators for specific treatments, particularly when contrasting trauma focused and non-trauma focused treatments.

Introduction

Substantial advances have been made in the psychological treatment of Posttraumatic Stress Disorder (PTSD) during the past two decades (e.g., American Psychological Association Guideline Development Panel for the Treatment of PTSD, 2017; Department of Veterans Affairs & Department of Defense, 2017). Despite the identification of effective treatments, a significant percentage of individuals do not obtain clinically meaningful reductions in PTSD symptoms. Although many individuals show reductions in PTSD symptoms following treatment, some continue to score above established norms for clinically significant change or retain PTSD diagnostic status. This is especially the case for veterans who have been shown to have a less positive response to evidence-based treatments, relative to non-veterans (e.g., Steenkamp, Litz, Hoge, & Marmar, 2015). In response to calls for increased information about factors that impact treatment outcome (e.g., Resick, Monson, Gutner, & Majlej, 2014), investigators have begun to examine variables that may influence the effects of psychosocial treatments. Although this research has many practical and clinical uses, the existing literature is somewhat limited by a focus on treatment delivered in an individual-format. The current study examined moderators of treatment outcome for veterans with PTSD receiving one of two group-based interventions, Group Cognitive-Behavioral Treatment (gCBT; Beck, Coffey, Foy, Keane, & Blanchard, 2009) and Group Present Centered Treatment (gPCT; Schnurr et al., 2003). In this study, factors that previously have been noted to impact PTSD treatment outcome in veterans were examined, with attention to whether different factors moderate outcome in a trauma focused group treatment (gCBT) versus a non-trauma focused group treatment (gPCT).

The literature on efficacy and effectiveness of group treatments for PTSD is less well-developed, compared to the literature on individual treatments (e.g., Beck & Sloan, in press). Despite this relative lack of empirical foundation, group treatment approaches are used often in clinical settings, particularly within the Department of Veteran Affairs (e.g., Sloan, Unger, & Beck, 2016). To expand this literature, Sloan, Unger, Lee, and Beck (2018) compared gCBT with gPCT. gCBT is trauma focused and incorporates interventions designed to address avoidance, emotional numbing, dysfunctional post-trauma thoughts, and physiological hyperarousal. gCBT has been shown to be efficacious in the treatment of PTSD among non-veteran samples (Beck & Sloan, in press). gPCT is centered on group-facilitated coping, is not trauma focused, and has been shown to be effective in comparison with Prolonged Exposure (Foa et al., 2018) and Cognitive Processing Therapy (e.g. Surís, Link-Malcolm, Chard, Ahn, & North, 2013). Sloan et al. (2018) found that both treatment conditions resulted in significant reductions in symptoms of PTSD, depression, and anxiety and improvements in functioning. Overall effects of gCBT and gPCT were comparable, with gains maintained through 12-month follow-up. No significant differences in outcome were observed between the two group treatments. Given the relative lack of information on factors that influence outcomes of group treatments, the current report examined moderators separately within gCBT and gPCT.

Previously, many studies of treatment moderation in this literature have either relied on data drawn from uncontrolled studies (e.g., via chart review; Schumm, Walter, & Chard, 2013) or collapsed across treatment types if data from a randomized controlled trial were involved (e.g., Litz et al., 2019). Both of these approaches have limitations. Studies using chart review data are unable to account for factors such as treatment fidelity, therapist competence, and whether patients self-selected the specific treatment under investigation. Use of data from a randomized controlled trial is preferable, as these factors are either monitored or controlled through the design. However, collapsing across treatment approaches in this type of investigation reduces the applicability of the results to clinical practice, as findings are not unique to any specific treatment. In the current report, the same set of variables were examined as moderators for gCBT and gPCT, in separate analyses.

Selection of potential moderators to include in the present study was guided by the literature. Across multiple studies, pre-treatment PTSD symptom levels have been identified as important influences of outcome. For example, Elliott, Biddle, Hawthorne, Forbes, and Creamer (2005) found that high pre-treatment levels of PTSD symptoms appear to be associated with larger decreases in symptoms during treatment in a sample of Vietnam veterans treated in an intensive outpatient setting. A similar finding was noted by Felleman, Stewart, Simpson, Heppner, and Kearney (2016) in a study of moderators of response to mindfulness-based stress reduction among veterans with PTSD. Pre-treatment depression severity also has been found to moderate treatment outcomes, although findings are mixed. For example, Elliott et al. (2005) noted that veterans with higher pre-treatment levels of depression showed greater reduction in PTSD following treatment. In contrast, Stein, Dickstein, Schuster, Litz, and Resick (2012) reported that higher levels of pre-treatment depression were associated with less response to PTSD treatment. When considering the role of pre-treatment symptom levels as a moderator of outcome, it is notable that within the larger literature, findings are quite mixed. Some reports have documented that individuals who enter therapy with high levels of anxiety symptoms demonstrate worse outcomes, relative to individuals who enter therapy with lower anxiety symptoms (e.g., Newman, Crits-Christoph, Connolly Gibbons, & Erickson, 2006). Other reports note more improved outcomes following CBT for individuals with high levels of pre-treatment depression, relative to those with low levels (Driessen, Cuijpers, Hollon, & Dekker, 2010). As such, examination of the role of pre-treatment levels of PTSD and depression symptoms as moderators of outcome following group treatments is an important issue to pursue further.

Previous studies also have noted that the nature of the index trauma and select demographic factors may impact treatment outcomes for veterans. Across studies, combat exposure as the index trauma is associated with more modest treatment outcome effects as compared to other index trauma events, coded as sexual or physical assault, and mixed/other (e.g., Bradley, Greene, Russ, Dutra, & Westen, 2005; Elliott et al., 2005; Schumm et al., 2013). Data from these studies suggest that qualifying events may deserve closer attention as a moderator of group treatment outcome. Age has also been noted in several studies to moderate treatment outcome, with younger age associated with more positive outcomes, relative to older age (e.g., Chard, Schumm, Owens, & Cottingham, 2010; Litz et al., 2019; Schumm et al., 2013).

Although the literature on moderation of treatment response has focused primarily on pre-treatment symptom severity, trauma characteristics, and demographic factors, there is increased discussion of the role of treatment process variables in affecting outcome (e.g., Zayfert & Black Becker, 2020). Outcome expectancies, defined as a patient’s pre-treatment beliefs about expected gains from psychotherapy, have been highlighted as important predictors of outcome. Constantino, Arnkoff, Glass, Ametrano, and Smith (2011) noted a significant effect size (d = 0.24, p <  .001) between outcome expectancies and treatment outcomes, in a meta-analysis of 46 independent samples. Within the trauma literature, Price et al. (2015) reported a significant negative association between outcome expectancies and both clinician and self-reported post-treatment PTSD symptoms in a sample of combat veterans. This study suggests that outcome expectancies are an important factor to examine as a potential moderator.

To expand available knowledge about moderators of group treatment outcome for PTSD, the current report examined five variables for their association with trajectories of change during gCBT and gPCT (pre-treatment PTSD, pre-treatment depression, age, whether combat was the index trauma, and outcome expectancies). The current study is also designed to expand on the findings of Sloan et al. (2018) by providing a more fine-grained exploration of person-specific change, using a secondary data analysis. In particular, we examined trajectories of change for each PTSD symptom cluster as defined by DSM-5: intrusion, avoidance, cognitions/mood, and arousal/reactivity (American Psychiatric Association, 2013) as we anticipated moderation differences between the two group treatment conditions. Because gCBT and gPCT differ with respect to their approach, techniques, and trauma focus, we anticipated the potential for moderating effects on avoidance-related outcomes specifically within gCBT given that aspects of the intervention directly target trauma-focused avoidance. A similar hypothesis was made concerning moderation of intrusion-related outcomes; these effects would only be noted in gCBT, owing to the explicit trauma focus in this intervention. By contrast, we hypothesized that significant moderation of change in cognitions/mood and arousal/reactivity outcomes would be observed in both gCBT and gPCT, given that both treatments incorporated a supportive group environment and emphasis on positive coping. As noted by Asnaani, Kaczkurkin, Fitzgerald, Jerud, and Foa (2020), positive appraisal coping appears to moderate change in hyperarousal, independent of treatment condition. Although Asnaani et al. (2020) did not examine negative cognitions/mood symptoms, we speculate a similar trans-treatment relationship between moderators and outcome in this domain given observations by Hurlocker, Vidaurri, Cuccurullo, Maieritsch, and Franklin (2018) that the negative cognitions/mood and arousal/reactivity clusters appear to be non-specific to PTSD among trauma-exposed veterans. No specific hypotheses were made concerning individual moderator variables, as this set of factors has not been examined together previously.

Section snippets

Participants

Participants included 198 male veterans randomized to either gCBT (n = 98) or gPCT (n = 100). Participants were recruited from the VA Boston Healthcare System and the Providence VA Medical Center. As noted in Sloan et al. (2018), all participants were diagnosed with PTSD using the Clinical-Administered PTSD Scale for DSM-5 (CAPS-5; Weathers, Blake et al., 2013) and had stable use of psychotropic medication prior to study entry. Of the 100 veterans randomized to gPCT, eight were excluded from

gCBT

Unconditioned, random-coefficient regression models for gCBT indicated reductions in intrusion, avoidance, cognitions/mood, and arousal/reactivity scores over the course of intervention (all p <  .001; see Table 2). Random effects for intercept and slope parameters at Level 2 provide evidence for differences in initial status and linear change across participants.

gPCT

Unconditioned growth models for gPCT also indicated change in PCL-5 symptom clusters over time (all p < .001; see Table 2). Change in

Discussion

In this study, separate analyses for gCBT and gPCT permitted examination of change processes and moderators of recovery in both trauma-focused and non-trauma focused group-based treatments. Selected moderators included pre-treatment PTSD, pre-treatment depression, age, whether combat was the index trauma, and outcome expectancies. Considering trajectories of change, both gCBT and gPCT showed reductions in intrusions, negative cognitions/mood, and arousal/reactivity. Only gCBT showed evidence

Author note

This study was funded by Department of Veteran Affairs (I01 CX000467). The first author is supported in part by the Lillian and Morrie Moss Chair of Excellence.

References (46)

  • A.T. Beck et al.

    Manual for beck depression inventory-II

    (1996)
  • C.A. Blevins et al.

    The posttraumatic stress disorder checklist for DSM-5 (PCL-5): Development and initial psychometric evaluation

    Journal of Traumatic Stress

    (2015)
  • M.J. Bovin et al.

    Evolving DSM diagnostic criteria for PTSD: Relevance for assessment and treatment

    Current Treatment Options in Psychiatry

    (2015)
  • R. Bradley et al.

    A multi-dimensional meta-analysis of psychotherapy for PTSD

    The American Journal of Psychiatry

    (2005)
  • K.M. Chard et al.

    A comparison of OEF and OIF veterans and Vietnam veterans receiving cognitive processing therapy

    Journal of Traumatic Stress

    (2010)
  • J. Cohen

    Statistical power analysis for the behavioral sciences

    (1988)
  • M.J. Constantino et al.

    Expectations

    Journal of Clinical Psychology: In Session

    (2011)
  • Department of Veterans Affairs & Department of Defense

    VA/DoD clinical practice guideline for the management of post-traumatic stress and acute stress reactions

    (2017)
  • E. Driessen et al.

    Does pretreatment severity moderate the efficacy of psychological treatment of adult outpatient depression? A meta-analysis

    Journal of Consulting and Clinical Psychology

    (2010)
  • P. Elliott et al.

    Patterns of treatment response in chronic Posttraumatic Stress Disorder: An application of latent growth mixture modeling

    Journal of Traumatic Stress

    (2005)
  • B.I. Felleman et al.

    Predictors of depression and PTSD treatment response among veterans participating in mindfulness-based stress reduction

    Mindfulness

    (2016)
  • E.B. Foa et al.

    Prolonged exposure therapy for PTSD

    (2019)
  • E.B. Foa et al.

    Effects of prolonged exposure therapy delivered over 2 weeks vs 8 weeks vs present-centered therapy on PTSD symptom severity in military personnel: A randomized clinical trial

    JAMA Psychiatry

    (2018)
  • View full text