Behavioral Activation as a Treatment for Posttraumatic Stress Disorder Among Returning Veterans: A Randomized Trial

Despite the established efficacy of trauma-focused psychotherapies for posttraumatic stress disorder (PTSD), including prolonged exposure and cognitive processing therapy, many civilians and veterans with PTSD are not receiving or completing these treatments (1–7). This finding is noteworthy because unprecedented training and implementation policies ensure that all U.S. Department of Veterans Affairs (VA) medical centers provide access to prolonged exposure and cognitive processing therapy.

Avoidance of trauma reminders is a central feature of PTSD, and research has indicated that PTSD-related avoidance is associated with lower utilization of psychiatric care (8–10). Among veterans, commonly reported treatment-seeking barriers include perceived stigma and competing responsibilities (work, childcare; 11–13). Implementation barriers also exist regarding use of trauma-focused treatments in many settings, including limited provider hours and variability in training (14, 15).

To address the need for additional PTSD treatments, Wagner et al. (16) and Jakupcak et al. (17) developed a brief adaptation of behavioral activation for treating PTSD (and comorbid depression). Behavioral activation is a present-focused, well-established treatment for depression (18–21) that targets patterns of avoidance through identification and enactment of reinforcing activities aligned with patients’ long-term goals (22). Case formulations and treatment planning are ideographic and patient centered. The rationale for applying behavioral activation to PTSD was based on the role of avoidance in the development and maintenance of PTSD impairments (23) and contemporary approaches identifying unified intervention targets (avoidance) across many psychiatric disorders (24).

Preliminary studies have supported adaptations of behavioral activation for PTSD among both civilians (16) and veterans (17, 25–27). Two pilot studies conducted by independent researchers found that the brief behavioral activation for PTSD protocol developed by Wagner et al. (16) and Jakupcak et al. (17) was clinically effective for reducing PTSD symptoms in military and veteran samples (28, 29). Furthermore, behavioral activation for PTSD has been effectively delivered in a range of settings, including primary care (17, 29), rehabilitation clinics (27), and home-based care (16, 28).

Two additional studies have included behavioral activation in phased treatments for comorbid depression and PTSD; in these protocols, several sessions of behavioral activation preceded trauma processing (30, 31). In both studies of behavioral activation plus trauma processing, reductions in PTSD were evident during the behavioral activation phase; Gros and colleagues (30) found treatment-response trajectories unchanged following the initiation of imaginal exposure. Together, these initial studies suggest that behavioral activation holds promise as a stand-alone treatment for PTSD for individuals who do not want or cannot access trauma-processing therapies.

This study builds on previous work by evaluating behavioral activation for PTSD in a randomized trial, comparing behavioral activation with specialty PTSD treatment (standard care) in a sample of veterans with PTSD. The behavioral activation protocol tested is the eight-session version developed by Wagner et al. (16) and Jakupcak et al. (17). We predicted veterans receiving behavioral activation would demonstrate significant and greater reductions in PTSD relative to those receiving standard care.

We further predicted that behavioral activation would be associated with significant reductions in depression severity and increases in overall functioning compared with standard care. We chose a standard care comparison group (rather than another specific treatment) on the basis of our interest in comparing behavioral activation with the existing standard of care within the VA health care system, which promotes referral to specialty PTSD care for trauma-focused treatments but also allows for patient choice and provider recommendations.

Methods

Participants

Participants were 80 veterans of the wars in Iraq and Afghanistan, enrolled at either the VA Portland Health Care System or the VA Puget Sound Health Care System (Seattle) and diagnosed as having PTSD related to military trauma(s). Study referrals were primarily from providers in primary care, reflective of recruitment efforts to enroll veterans who might not typically seek psychiatric care. Eligibility criteria included current diagnosis of PTSD; willingness to refrain from additional PTSD care during the treatment phase; psychiatric medication stability (no medication changes within 4 weeks before enrollment); and absence of current psychosis, bipolar disorder, substance dependence (as defined by the DSM-IV; substance abuse was not a rule out), and high current suicidality or homicidality. Data were collected between June 2009 and March 2013.

Primary Outcome Measures

PTSD.

The Clinician-Administered PTSD Scale (CAPS; 32, 33) was used to assess initial PTSD diagnosis (DSM-IV; 34) and PTSD severity across assessment time points. The CAPS is considered a gold standard for assessing PTSD. Items are ranked on Likert scales according to both frequency (0–4) and intensity of symptoms (0–4), yielding an overall severity score based on the sum of frequency and intensity ratings across 17 items (range 0–136). For this study, a diagnosis of PTSD was based on the scoring rule in which a symptom with a “1” on frequency and “2” on intensity is counted as present, and the overall CAPS severity score is at least 45 (35). The CAPS has strong psychometric properties (32, 33).

The Posttraumatic Stress Disorder Checklist-Military version (PCL-M; 36, 37) was used to assess participants’ subjective PTSD-related distress. The PCL-M is a 17-item, self-report scale that assesses DSM-IV PTSD symptoms. Items are rated on a 5-point Likert scale according to how much symptoms bothered the respondent over the past month. Psychometric properties of this measure are excellent (37, 38).

Depression.

The Beck Depression Inventory-II (39) is a 21-item measure of subjective depression severity. This widely used measure is commonly included in outcome studies to determine treatment effects on severity of depressive symptoms and has excellent psychometric properties (40).

Secondary Outcome Measures

Functioning.

The Sheehan Disability Scale (41) is a widely-used three-item self-rated scale of functional impairment. Items ask respondents to rate (on a Likert scale ranging from 0 to 10) to what extent their symptoms interfere with functioning in the areas of work, social, and family life. Scores range from 0 to 30, higher scores representing greater functional impairment. The scale’s reliability and concurrent validity have been demonstrated in anxiety disorder samples (42).

Behavioral activation.

The Behavioral Activation for Depression Scale (BADS; 43) was used to assess changes in behavioral activation. The BADS is a 25-item, self-report measure that assesses overall degree of behavioral activation. Items are rated on a 7-point Likert scale ranging from 0 (not at all) to 6 (completely). Total scores range from 0 to 150, with higher scores indicating more behavioral activation. The BADS has demonstrated strong reliability and validity (44).

Treatment Acceptability Measures

The Credibility-Expectation Scales (45) were administered only after behavioral activation, session 1, to assess perceived credibility of treatment and expectancy for change. This six-item measure is widely used and has demonstrated strong reliability and validity (46). Items 1–3 assess credibility (the degree to which the treatment is viewed as believable, convincing, and logical); items 4–6 assess expectancy (the degree to which the respondent believes improvements will occur).

The Client Satisfaction Questionnaire (47) is an eight-item, self-report questionnaire used to assess satisfaction with care. Items are rated on a 4-point Likert scale, with higher scores reflecting greater satisfaction (range 0–32). The Client Satisfaction Questionnaire has demonstrated strong reliability and validity (48).

Procedures

Assessors contacted patients referred to the study for a preliminary phone screen to confirm interest in PTSD treatment and psychiatric medication stability. Those eligible were scheduled for a pretreatment session that included informed consent and a clinical interview to assess remaining eligibility criteria and baseline symptom severity and functioning levels [see online supplement].

Participants completed in-person posttreatment and 3-month follow-up assessments. Assessors were master’s- or doctoral-level clinicians who were blinded to treatment assignment. Because of the nonspecific nature of standard care (described later), follow-up assessments occurred at predetermined time points (vs. after a specific number of sessions). A scheduling “clock” was started at the time of participants’ first therapy appointment or 1 month after pretreatment assessment, whichever came first. The posttreatment assessment was scheduled for 12 weeks after clock initiation, and the 3-month follow-up was scheduled for 24 weeks after clock initiation.

Study assessors were trained to reliability on the CAPS. To determine interrater reliability on the CAPS, approximately 10% of completed CAPS (N=21) audio recordings were randomly selected (across all assessment points) and rated by an expert CAPS assessor unaffiliated with this study. Interrater reliability was high (intraclass correlation coefficient=0.97). This study was approved and monitored by the institutional review boards of the VA Portland and Puget Sound health care systems as well as a centralized VA institutional review board.

Interventions

Behavioral activation for PTSD.

The eight-session protocol is based on behavioral activation for depression as described by Martell and colleagues (22). The protocol maintains the central structure and components of this approach, including behavioral analyses for assessing avoidance patterns and tracking difficulties and benefits of engaging in activities; activity scheduling and monitoring of reactions based on patient values, physical abilities, and life circumstances; and present-centered strategies to decrease rumination and enhance emotional engagement and awareness of subjective experiences, referred to as “attention to experience” (similar to mindfulness training). Psychoeducational materials for patients included information on PTSD and its association with avoidance patterns. The manual allows for the inclusion of a significant other in an early session with encouragement to involve significant others in goal setting and goal-directed activities.

Sessions are designed to be delivered within 60 minutes, typically lasting 45 minutes. Sessions 1 and 2 focus on orientation to the treatment model, assessing and discussing PTSD avoidance in the maintenance of functional impairment, assessing values and goals, and identifying and initiating areas for activation. Sessions 3–7 focus on continuing assignment of behavioral activation tasks, assessing the effects of activation, and promoting problem solving. Sessions are loosely but consistently structured, anchored by agenda setting, homework review, and troubleshooting, with flexibility to address additional concerns. The last session includes a review of activation progress, relapse prevention, and discussion of future behavioral activation targets and additional treatment.

Standard care.

Participants randomly assigned to standard care received standard treatments delivered in the VA PTSD specialty clinics in Portland and Seattle. In both clinics, providers are trained in either prolonged exposure, cognitive processing therapy, or both; VA performance measures promote offering these treatments. Both clinics also offer individual and group-based alternative treatments and coping-skills training (e.g., acceptance and commitment therapy, anger management).

Both PTSD clinics offer pharmacotherapy. Actual treatment received was determined collaboratively between the standard care provider and the veteran. Therefore, standard care reflected “real-world” care per policies, training, and procedures consistent with PTSD specialty clinics across VA medical centers. In an effort to equate access to care between groups as much as possible (without interfering excessively with the spirit of standard care), participants randomly assigned to standard care were offered a minimum of six individual psychotherapy appointments.

Therapists and Treatment Fidelity

Behavioral activation was delivered by three doctoral-level psychologists, including the principal investigators (the first and second authors) and a provider in Seattle who was trained and supervised by the principal investigators. Most cases were seen by the principal investigators. Providers of standard care were not preselected for study purposes and were representative of the providers of the PTSD clinics, composed of psychologists, psychiatrists, nurse practitioners, and social workers.

Recordings of 25 randomly selected sessions (11%) underwent fidelity coding, with roughly equal numbers chosen across behavioral activation sessions and sites. A fidelity checklist consisted of 12 “essential” components that were rated on a 3-point scale (0, absent; 1, partial and more was needed; and 2, present and no more was needed). The fidelity assessor was trained by the first author to reliability by using additional tapes not selected for fidelity assessment. Most (91%, N=273) essential components were listed as present with an average score of 1.8, indicating most were present fully.

Data Analyses

Descriptive statistics were obtained for all variables, and tests of normality and homogeneity of variance were performed. Data were analyzed by using mixed-effects regression modeling to account for missing data across assessments (both groups experienced fairly high rates of study dropout; number of participants lost to 3-month follow-up was 13 for both groups). Primary hypothesis tests were evaluations of changes from baseline through follow-up assessments on measures of PTSD severity (PCL-M and CAPS). Secondary hypothesis tests were evaluations of changes on measures of depression (Beck Depression Inventory-II) and quality of life (Sheehan Disability Scale). All analyses included the intent-to-treat sample, used IBM SPSS Statistics (version 24), and were two-tailed with α=0.05.

Results

Participant characteristics are presented in Table 1. Groups did not differ significantly on any demographic variables. Table 2 provides a summary of the treatment received by both groups, determined by review of electronic medical records by the assessors. Behavioral activation participants received significantly more individual treatment (all behavioral activation therapy) at posttreatment than the standard care participants despite equal access to care (t=6.70, df=73, p<0.001). Similarly, total treatment received, defined as the sum of individual, group, and medication management sessions across time periods, was greater among the behavioral activation group than the standard care group (t=2.67, df=72, p<0.009). Despite all standard care providers being trained in prolonged exposure and cognitive processing therapy, few veterans received these treatments, even by 3-month follow-up. To control for the amount of treatment received on outcomes, the composite variable of total treatment received was computed and used as a covariate in analyses.

Table 3 displays means and standard deviations for primary and secondary outcome variables. Controlling for amount of treatment received, the analyses found that levels of PTSD symptoms decreased for both groups over time on the CAPS total score (F=30.47, df=2, 111, p<0.001) and PCL-M (F=35.07, df=2, 114, p<0.001). Contrary to prediction, the behavioral activation group did not demonstrate a significantly greater reduction in the CAPS total score than the standard care group (group × time interaction), although there was trend toward significance in favor of behavioral activation (F=2.69, df=2, 111, p=0.07). There was a significant group × time interaction on the PCL-M, indicating greater reductions in PTSD symptoms for the behavioral activation group compared with the standard care group (F=5.87, df=2, 114, p<0.01).

On the CAPS subscales, there was a significant group × time interaction for the avoidance subscale, such that the behavioral activation group had greater avoidance reductions relative to standard care (F=3.63, df=2, 110, p=0.03). Nonetheless, mean scores remained in the clinical range of presumptive PTSD for both groups at both follow-up assessments. Depression scores also significantly decreased over time for both groups (F=8.36, df=2, 112, p<0.001) with a significant group × time interaction, reflecting greater reductions among the behavioral activation group (F=4.08, df=2, 112, p=0.02). However, mean scores remained in the moderate range of depression severity for both groups at both follow-up time periods.

On the Sheehan Disability Scale, participants in both groups showed improvement in functioning over time (F=10.28, df=2, 108, p<0.001), but there was no group × time effect. Likewise, levels of behavioral activation, as measured by the BADS, increased over time for both groups (F=7.1, df=2, 97, p<0.01); however, there was no group × time effect on behavioral activation.

Both groups reported comparable satisfaction with treatment, with overall means suggesting high treatment satisfaction; at the 24-week assessment, the mean score was 25.14±2.35 for behavioral activation (N=28) and 25.39±2.87 for standard care (N=23). Overall, behavioral activation was viewed as credible (items 1–3, 6.7±1.09) with moderate expectation of improvement (items 4–6, 5.3±1.48).

Discussion

Results of this randomized trial suggest that brief behavioral activation may be effective for reducing PTSD and depression symptoms among veterans within VA health care settings. Although both groups showed improvement over time on measures of PTSD (interview and self-report), depression, and overall functioning, the behavioral activation group showed greater reductions on subjective measures of PTSD and depression relative to standard care. However, groups did not differ on overall PTSD severity as measured by the CAPS, suggesting that behavioral activation may not be more effective than referral to specialty PTSD care.

Surprisingly, the behavioral activation group did not show greater increases in behavioral activation than the standard care group, given the specific focus of behavioral activation. Perhaps behavioral activation is an outcome of PTSD improvement regardless of the therapy type.

It is notable that, despite the range of therapies available within the specialty PTSD clinics, veterans randomly assigned to behavioral activation for PTSD attended more psychiatric appointments. Perhaps this outcome reflects veterans’ preferences for present-focused and skill-based therapies over the trauma-focused therapies often associated with specialty PTSD care (49). Indeed, the standard care group’s rates of utilization are comparable with national patterns of psychiatric use observed among VA-enrolled Iraq and Afghanistan veterans diagnosed as having PTSD (6, 49). Alternatively, administrative factors associated with referral processes to PTSD specialty services (scheduling the clinic intake before therapy initiation) may have influenced utilization patterns in the standard care condition (although study procedures also involved extensive pretreatment processes).

Although the group differences in treatment received can be considered a limitation of this study, we intentionally designed the comparison group to reflect real-world PTSD care within the VA health care system. Nonetheless, it cannot be ruled out that administrative, as well as provider, factors (greater motivation within the behavioral activation providers, who were primarily the treatment developers) influenced the results. Future studies should examine the effectiveness of behavioral activation for PTSD relative to research-structured, trauma-focused (evidence-based) therapy protocols with therapists independent of the investigation.

Our study is limited by a modest sample size, a predominantly white and male population, a noncontrolled comparison treatment, and patients dropping out over time. Furthermore, although clinically meaningful reductions in PTSD and depression symptom severity were observed in both groups, follow-up outcomes remained in the diagnostic range across assessments. Meta-analyses have found smaller effect sizes associated with treatment of military-related PTSD relative to PTSD from other types of traumatic exposure (50). Still, PTSD and depression severity scores at follow-up assessments suggest that the majority of study participants would benefit from ongoing therapy.

Finally, our PTSD measures were based on DSM-IV criteria because the study began before the publication of DSM-5. This basis may somewhat limit the generalizability of our findings to persons who have currently been diagnosed as having PTSD (although DSM-IV and DSM-5 diagnoses of PTSD are highly correlated), and we have no knowledge of the effects of behavioral activation on the current “cluster D” criteria, which are related to changes in cognition and mood.

Conclusions

Behavioral activation may be a viable alternative treatment for PTSD for individuals who do not want or who are unable to access trauma processing therapy. With its focus on behavioral avoidance and increasing functionality, behavioral activation is also well-suited for the treatment of conditions commonly comorbid with PTSD, such as depression and chronic pain (31). Elements of behavioral activation are relatively straightforward and easy to implement in a range of health care settings (16, 27, 51) by providers with varying backgrounds and training (52). Taken together, our findings suggest that behavioral activation is a promising therapy that may expand the reach of PTSD treatment (53). Behavioral activation may be strengthened by including additional sessions. Future studies should compare behavioral activation with specific evidenced-based treatments for PTSD; behavioral activation should also be used within implementation designs that include additional populations of individuals with PTSD.