Trauma Center Trauma-Sensitive Yoga Versus Cognitive Processing Therapy for Women Veterans With PTSD Who Experienced Military Sexual Trauma: A Feasibility Study

Abstract

BACKGROUND:

Posttraumatic stress disorder (PTSD) is a common sequela to military sexual trauma (MST) among women veterans. Yoga has shown promise in research examining its benefit for symptoms and sequela of PTSD.

AIMS:

The objective of this study was to evaluate the feasibility of a randomized controlled trial (RCT) of Trauma Center Trauma-Sensitive Yoga (TCTSY) for women veterans with PTSD related to MST.

METHOD:

In this feasibility study, the final sample included women veterans (n = 41) with PTSD related to MST accessing health care in a Veterans Affairs Health Care System in the southeast United States; the majority were African American (n = 33; 80.5%). Interventions used established protocols of 10 weekly sessions of group TCTSY versus 12 weekly sessions of group Cognitive Processing Therapy (CPT). PTSD was assessed via clinical interview and participant report. Additional data collection included multiple participant-reported outcomes commonly associated with PTSD and psychophysiological measures. We also collected data regarding participant satisfaction and feasibility-related feedback from participants and providers.

RESULTS:

Feasibility and acceptability were evaluated via demand, practicality, fidelity, and acceptability. This was measured by expressed interest, attendance, program completion, barriers to care and satisfaction with treatment, and satisfaction with interventions and data collection.

CONCLUSIONS:

Results indicate the RCT design and TCTSY implementation were feasible; a full-scale RCT was subsequently conducted to determine efficacy of the experimental intervention. Recommendations for successful research strategies are provided.

Background and Introduction

Military sexual trauma (MST) is a term used by the Veterans Administration (VA) to describe sexual assault and on-going threatening sexual harassment experienced during one’s military service. MST has widespread negative health and wellness consequences. Approximately 38% of women service members have experienced MST, though the rate of MST varies across samples and study designs, with reports of MST ranging from 20% to 68% (Barth et al., 2016; Boehler, 2019; Wilson, 2018). Posttraumatic stress disorder (PTSD) is the most common mental health disorder following MST; more than 50% of those reporting MST are diagnosed with PTSD (Gibson et al., 2020; Kimerling et al., 2010; Scott et al., 2014; Sexton et al., 2017; Yalch et al., 2018). MST is a risk factor for a host of stress-related physical and mental health conditions in addition to PTSD, including substance use, anxiety, depression, chronic pain, and sleep disturbances (Afari et al., 2014; Carlson et al., 2020; Moeller-Bertram et al., 2012).

Current evidence-based psychotherapies (EBPs) for PTSD have limitations. Among research participants treated with cognitive processing therapy (CPT) and prolonged-exposure (PE), 49% to 70% experience clinically significant symptom improvement; however, mean post-treatment symptom presentations often remain at or above the diagnostic threshold for PTSD (Békés et al., 2016; Larsen et al., 2019; Steenkamp et al., 2015; Surís et al., 2013). PTSD treatment-seeking women veterans with MST have identified numerous barriers to care, including institutional betrayal, which interfere with recovery from PTSD leaving them at risk of developing various conditions related to chronic stress (Kelly, 2021). As such, intervention research for women veterans with MST-related PTSD will benefit from examination of treatments from innovative theoretical perspectives (Boehler, 2019; Kelly et al., 2011). One such innovative approach that is growing in popularity and use is yoga. Yoga reduces stress and improves mood, pain, sleep, and functioning, which correspond to the aforementioned risk factors of MST (Chopin et al., 2020; Cramer et al., 2013; Taylor et al., 2020; Zaccari et al., 2020).

While several published studies add support to the feasibility and acceptability of yoga as a clinical intervention for veterans, studies examining the benefits of yoga have been limited by methodological problems including small sample sizes, inadequate power, lack of randomization and/or rigorous control conditions, and inconsistency among the yoga methods tested and the measures used to evaluate the effects of yoga (Chopin et al., 2020; Davis et al., 2020; Gopal et al., 2011; Kelly et al., 2018; Khattab et al., 2007; Kiecolt-Glaser et al., 2010; Raghuraj & Telles, 2008; Streeter et al., 2012; Taylor et al., 2014; Zaccari et al., 2020). In published yoga studies with veterans with combat-related PTSD, women were disproportionately underrepresented, data specific to women were not reported, and sexual trauma was not considered. Furthermore, while a small number of yoga studies for PTSD have included psychophysiological variables as outcomes, none have focused on women with histories of sexual trauma and PTSD, nor evaluated the validity or clinical significance of their measures (Gopal et al., 2011; Kelly et al., 2018; Khattab et al., 2007; Kiecolt-Glaser et al., 2010; Raghuraj & Telles, 2008; Streeter et al., 2012; Taylor et al., 2014).

This study was designed to address these gaps by establishing the feasibility of implementing Trauma Center Trauma-Sensitive Yoga (TCTSY) as a clinical intervention and conducting a randomized controlled trial (RCT) of the yoga intervention for women veterans with PTSD related to MST. To address the limitations of single-arm and wait-list control studies, we used a gold-standard PTSD treatment, CPT, as the control condition. To evaluate feasibility, we assessed criteria based on the feasibility framework proposed by Bowen and colleagues (2009). Specifically, we examined the demand via expressed interest, practicality via actual program attendance and intervention and data collection completion, and the acceptability via barriers to and satisfaction with treatment, and data collection procedures. In addition to assessment of symptoms via clinician interview and participant self-report, we measured psychophysiological variables to determine the feasibility and utility of measuring these variables in future studies which seek to determine yoga’s mechanism of action for reducing symptoms of PTSD.

Method

Participants

Data for this feasibility study were collected from women veterans who sought out-patient PTSD treatment in a VA located in the Southeastern United States. The research was approved by the appropriate Institutional Review Board and the VA research oversight committees. The study used a partial Health Insurance Portability and Accountability Act (HIPAA) waiver to conduct pre-screening and recruitment which facilitated identification of eligible participants. Forty-two women veterans were enrolled and one withdrew prior to baseline data collection, yielding a final baseline sample size of 41 women (age ranged from 30 to 72, Md = 45; SD = 9.9), the median highest education level obtained was 15 years, and most (70.7%) had a household monthly income of $2,000 or more. The majority were African American—non-Hispanic (80.5%), nearly one-third (29.3%) had served in a warzone, and one-quarter had been hospitalized for psychiatric or mental health treatment (24.4%) and/or had attempted suicide (26.8%) in their lifetime. The TCTSY and CPT groups did not differ significantly on age, race, or education. See Table 1 for sample demographics.

Table 1.

Sample Demographics.

Characteristic	All (n = 41)			TCTSY (n = 17)			CPT-C (n = 24)
	n	Md *	SD **	n	M **	SD	n	M	SD
Age (range, 30–72)	41	45.0	9.9	17	46.1	12.4	24	44.2	7.9
	n	Md	IQR 〸	n	Md	IQR	n	Md	IQR
Education (range, 12–20 years)	40	15	[14, 16]	16	16	[14.5, 16.8]	24	14	[13, 16]
Attendance (range, 1–13 sessions)	36	7.5	[1.3, 10.8]	14	8.5	[1, 10]	22	5	[1.8, 11]
Race	n	%		n	%		n	%
AAǂ; non-Hispanic	33	80.5%		14	82.4%		19	79.2%
White; non-Hispanic	5	12.2%		2	11.8%		3	12.5%
Other/mixed Marital status	3	7.3%		1	5.9%		2	8.3%
Single, never married	7	17.1%		3	17.6%		4	16.7%
Married/partnered	17	41.5%		8	47.1%		9	37.5%
Divorced/separated	15	36.6%		5	29.4%		10	41.7%
Widowed	2	4.9%		1	5.9%		1	4.2%
Past suicide attempt	11	26.8%		2	11.8%		9	37.5%
Past psychiatric hospitalization	10	24.4%		4	23.5%		6	25.0%

^*= Median

^**= Standard Deviation

^***= Mean

^〸= Interquartile range

^ǂ= African American

TCTSY = Trauma Center Trauma-Sensitive Yoga; CPT = cognitive processing therapy.

Procedure

Recruitment.

Women Veterans with PTSD related to MST were recruited by clinician referral, chart review, review of PTSD clinic schedules, and via flyers and brochures placed in waiting areas and bulletin boards in various locations within the VA. After reviewing diagnosis and treatment records, clinical judgment was used to determine potential eligibility, excluding women with predominant symptoms of personality disorders or a documented history of disruptions in group treatment. Potentially eligible participants were informed of the study by the Women’s Trauma Program (WTP) Director or the veteran’s WTP clinical treatment coordinator, and permission to be contacted was requested.

Eligibility and Evaluation.

Those who agreed to be contacted were screened by telephone or in-person visits. Women veterans who were diagnosed with PTSD (based on Diagnostic and Statistical Manual of Mental Disorders—Fourth Edition, Text Revision [DSM-IV-TR]; American Psychiatric Association, 2000) related to MST and who also reported chronic pain and sleep disturbance were included if they were able to give informed consent and willing to participate in either intervention.

Exclusion criteria included psychotic symptoms that precluded ability to participate in TCTSY and CPT, active suicidal ideation with intent or plan, current moderate or severe alcohol or substance use disorder, dementia, moderate or severe traumatic brain injury, pain due to acute injury, restless leg syndrome, and/or current engagement in trauma-focused therapy or any non-VA mental health treatment. Medical clearance was conducted by a physician co-investigator or the study PI, a dual certified Psychiatric/Mental Health Nurse Practitioner and Adult Nurse Practitioner.

Interested and eligible participants provided informed consent and were randomized. The randomization sequence was generated using the block randomization algorithm “random sorting using maximum allowable percent deviation” using the PASS v.15 power analysis software package (NCSS, 2020). Following randomization, historical and baseline data were collected. Participants were provided remuneration for the data collection sessions ($60/session), focus group participation ($10), and intervention sessions attended ($15/session).

Overview of Interventions and Fidelity Procedures.

Group interventions were co-facilitated by two interventionists. The target group size was 12 per intervention per cohort, with an expected 20% attrition. Two cohorts were completed. To assess feasibility relative to session attendance, intervention sessions were held during normal business hours in the first cohort and in late afternoon in the second cohort. A WTP clinician and the study PI were available during sessions if participants had urgent mental health needs.

Trauma Center Trauma-Sensitive Yoga.

TCTSY is a standardized Hatha yoga (a gentle, slower-paced approach to the physical forms of yoga) for use by women with treatment-resistant, complex PTSD (Price et al., 2017). The intervention protocol was developed by David Emerson and Jennifer Turner at the Justice Resource Institute and consisted of 10 weekly 60-min group sessions (Emerson, 2009). The protocol integrates themes related to establishing safety, individual choice, interoception, being in the present moment, and taking effective action. Group facilitators were experienced in facilitating trauma-sensitive yoga and had received 20 hr of in-person TCTSY training and 20 hours of telephone supervision with David Emerson, co-developer of TCTSY. The TCTSY facilitators completed the TCTSY Certification Program during the course of this study. Having two teachers present at each session facilitated the protocol implementation, allowed for demonstrating modifications and options in various yoga forms, and increased support that was needed at times during sessions (e.g., when a participant had an emotional reaction while doing yoga). There were no adverse events related to TCTSY in this study.

The TCTSY facilitators completed Intervention Fidelity Self-Reflection Field Notes developed by David Emerson immediately following each session, in which they recorded examples of providing choice and emphasizing interoception, as well as participants’ responses (implicit and explicit) to these options. The facilitators co-led the sessions, allowing for peer monitoring and feedback. Supervision calls were held with David Emerson, the PI, and the TCTSY facilitators within 48 hr of each session to review intervention fidelity, debrief, and provide on-going supervision. Meeting Minutes were recorded and reviewed to determine intervention fidelity. Treatment fidelity challenges and deviations from the protocol were categorized, quantified, and tracked over time to identify potential causes and solutions. Fidelity to the protocol was maintained despite individual teacher preferences for sequencing yoga components differently based on their own teaching training philosophies and prior experiences.

Comparative Intervention: Cognitive Processing Therapy.

CPT is a cognitively-based, trauma-focused treatment and gold standard therapy for PTSD treatment (Resick et al., 2008). The CPT manualized intervention was delivered in 12 weekly 90-min group sessions (Resick et al., 2008). CPT was led by two licensed clinical social workers certified in this treatment. The group sessions focused on identifying how thoughts change as a result of trauma exposure and teaching participants ways to realistically evaluate these maladaptive, rigid thoughts and come up with more accurate, flexible alternative thoughts. It is important to note that subsequent to this study, Resick et al. (2017) reported group format for CPT to be inferior to individual CPT, although approximately 50% of those receiving individual treatment still had a PTSD diagnosis and clinically significant symptoms post-treatment.

Certification in CPT includes training in maintaining fidelity to the model, and facilitators are trained on and observed maintaining fidelity as a condition of completing certification. The CPT clinicians maintained fidelity by adhering to the treatment manual as it is designed with attention paid to theoretically important treatment components (e.g., use of Socratic questioning, prioritizing assimilation, attending to practice assignments, and encouraging affect; Farmer et al., 2017). They used resources available to CPT-certified providers within VA to maintain adherence and fidelity, including accessing a VA intranet website with treatment materials and resources, case consultation resources, and a community of practice call. Our CPT providers adhered to the 2014 protocolized treatment manual which provides session-by-session guides consisting of assessment materials, a session outline, transcribed samples of interventions, and a participant workbook. CPT facilitators debriefed following sessions and consulted as needed with a study Co-Investigator, who is a certified CPT trainer and clinical supervisor. They also consulted with the PI when questions arose related to prioritizing research versus clinical protocols (i.e., CPT participants who missed several intervention sessions were removed from the CPT group, per WTP clinical protocols).

Measures and Data Collection.

Medical history, trauma history, self-report instruments, diagnostic instruments, and physiological data were collected at multiple time points (baseline, mid-intervention, 2 weeks post-intervention, and 3 months post-intervention). We collected focus group data after the first intervention session, psychological self-report questionnaires and clinical interviews at mid-intervention, TCTSY intervention fidelity self-reflection field notes after each session, weekly TCTSY supervision notes, and study staff field notes after each data collection session.

Demographics, Medical History, Trauma History, Psychological and Functioning Measures.

Minimal demographic data were collected at screening followed by in-depth baseline demographic assessment via participant self-report and review of medical records. Psychiatric comorbidities were assessed using the Structured Clinical Interview for DSM-IV (SCID-IV; First et al., 1995). Historical data collection included extensive childhood, adult, and military-specific trauma histories. The primary psychological outcome was PTSD, assessed using the Clinician-Administered PTSD Scale (CAPS) and the PTSD Checklist-Civilian (PCL-C; Blanchard et al., 1996; Hovens et al., 1994; Weathers et al., 2001). Additional self-report measures included the Post-traumatic Cognitions Inventory, the Posttraumatic Growth Inventory, the Dissociative Experiences Scale, the Difficulties in Emotion Regulation Scale, the Beck Depression Inventory-II, multiple PROMIS measures, the Pain Outcomes Questionnaire—Short Form, the Patient Health Questionnaire-15, the Pittsburgh Sleep Quality Index, PSQI Addendum for PTSD, the Kentucky Inventory of Mindfulness Skills, and the Inventory of Psychosocial Functioning, Quality of Life Inventory, and the Centers for Disease Control Behavioral Risk Factor Surveillance System (BRFSS) Health-Related QOL items (Baum et al., 2010; Beck et al., 1996; Berstein & Putnam, 1986; Bovin et al., 2018; Buysse et al., 1989; Centers for Disease Control and Prevention [CDC], 2018; Clark et al., 2003; Foa et al., 1999; Frisch et al., 1992; Germain et al., 2005; Hallion et al., 2018; Northwestern University, 2020; Tedeschi & Calhoun, 1996; van Ravesteijn et al., 2009).

Physiological Measures.

Dark-enhanced startle (DES) response, heart rate variability (HRV), sleep actigraphy, and blood serum cytokines (interleukin [IL]-6, IL-10, and C-reactive protein) comprised the psychophysiological data collection. DES response was assessed using a paradigm in which acoustic startle (blink) response is assessed in darkness versus light conditions. HRV data were acquired using the ECG module of the BIOPAC system MP150 for Windows (BIOPAC Systems, n.d.). Sleep quality was objectively measured using wrist actigraphy, via an Actiware device (Phillips North America Corporation, n.d.). Blood was collected via venipuncture using EDTA-containing tubes, stored in a −80° freezer, and analyzed at the study site lab.

Participant Satisfaction.

Short-answer, open-ended questions were asked of participants at each standardized data collection visit to elicit opinions regarding intervention and study procedures/data collection acceptability and intervention utility. Research assistants transcribed participant responses verbatim and entered the text into the data management software, REDCap (Harris et al., 2009). Focus groups were conducted by the PI using nominal group process following the first intervention session; participants completed a feedback form with questions related to study procedures and intervention (Allen et al., 2004). Nominal group process ensures equal input from all members and produces a set of prioritized recommendations that represent the group’s preferences and opinions using a four-step process: (1) generating ideas, (2) recording ideas, (3) discussing ideas, and (4) voting on ideas (U.S. Department of Health and Human Services, 2018).

Treatment Adherence.

Reminder calls were made prior to each session with follow-up calls to those who did not attend. Reasons for missed sessions were tracked. Session attendance and attrition/drop out over time were tracked to evaluate treatment adherence. Study staff field notes captured participant reasons for non-adherence. The number of intervention sessions and determination of intervention completion was defined based on published standards for the specific interventions at the time of the study. Emerson and colleagues, who designed the TCTSY intervention, defined intervention completion as attendance at seven of the 10 sessions (Emerson, 2009). Completion definitions for CPT in the literature vary, with eight or more of 12 sessions being common, and more practical than a higher threshhold, particularly in the VA setting (Khan et al., 2020; Maguen et al., 2020). For CPT, we defined treatment completion as attendance at eight or more of the 12 sessions.

Data Analysis

Quantitative treatment adherence data (session attendance, homework completion, and attrition/drop out over time) were analyzed using descriptive prevalence, Fisher’s exact test, and chi-square analyses (p < .05; two-tailed). All data were examined for completeness and descriptive assessment of missing data was made to see if any biases occurred. The small sample size precluded analysis of demographic and clinical factors related to subject attrition and adherence. SPSS Version 27 was used for all statistical analyses (IBM Corp., 2020).

Qualitative data were aggregated into Excel documents separated by data source (participant text responses, field notes, and fidelity meeting minutes) and were individually coded by two coders using inductive descriptive codes. Discrepancies were discussed and resolved, codes were consolidated, and descriptive topics and excerpts to represent experiences shared by participants were identified through discussion among the two coders. All qualitative data were analyzed using descriptive qualitative content analysis (Weber, 1990). Descriptive qualitative content analysis is a low-inference and easily replicable interpretation of qualitative data, which involves describing the quality and nature of the data instead of developing themes (as seen in thematic content analysis; Weber, 1990).

Results

Interest/Demand

Two periods of enrollment were conducted. Recruitment goals were to establish each cohort with 24 participants (12 per TCTSY and CPT groups), totaling 48 participants overall. Once a participant was screened and enrolled in the study, they were informed that the intervention sessions would begin as soon as the cohort was filled. Recruitment was slow and wait times proved to be too long for many participants (nearing 2 months from enrollment to the study start date), impacting attrition. The recruitment strategy and cohort size were modified to reduce participant time burden—cohort size was changed to 10 to 12 participants per group, and study staff broadened recruitment efforts to include in-person recruitment at clinical sites.

In total, 250 women were screened for inclusion and 42 were enrolled in the study (see Figure 1 for participant enrollment and randomization flow). Once enrolled, participants were randomized to either TCTSY or CPT prior to baseline data collection. Of the participants assigned to CPT (n = 25), one refused CPT and withdrew from the study prior to the completion of the baseline assessment (yielding a baseline sample size of n = 41), seven completed the baseline assessment but did not attend a single CPT session (29.2%; 7 out of 24). Study staff reported, via field notes, that some participants’ affect worsened when they found out they were randomized to CPT and not TCTSY. One participant shared, “[I would have chosen] Yoga. I felt that because I always feel tense that would have been a good way to re-learn how to relax and release pain.” The difference in attrition after baseline assessment and prior to the first intervention session between those assigned to CPT (29.2%) versus TCTSY (17.6%), while not statistically significant (Fisher’s exact test p = .480) due to the small sample size, is clinically noteworthy.

Figure 1.

Study Enrollment, Attendance, Completion, and Data Collection.

Intervention Practicality: Treatment Attendance and Program Completion

Of the 42 veterans who were allocated to a treatment group, 31 attended at least one session and 20 completed treatment (i.e., they attended ≥7 of 10 TCTSY sessions or ≥8 of 12 CPT sessions). Throughout the course of the study, participants reported several reasons for missing intervention sessions including, pain (n = 5), child care (n = 4), transportation and schedule conflicts (n = 4), and severe psychological symptoms or adverse health (n = 4). The treatment completion rate in the TCTSY group (58.8%) did not differ significantly from the CPT group, 41.7%; Figure 1; χ²₍₁₎ = 1.172, p = .279. In the second cohort, attrition in the CPT group was >50% within the first two sessions. Therefore, that group was discontinued and a new CPT group was enrolled.

Intervention Acceptability: Barriers to Care and Satisfaction With Treatment

Barriers to care and satisfaction with treatment were assessed using qualitative interviews to help determine intervention acceptability. Seventeen participants completed pre-intervention qualitative interviews and 12 completed post-intervention qualitative interviews.

TCTSY.

Prior to the start of TCTSY sessions, many participants were concerned about the effect of TCTSY on their current pain (n = 2) and physical limitations (n = 7). However, in the post-intervention qualitative interviews, participants reported the intervention to be calming/relaxing and helpful for stress reduction (n = 9). For example, one participant shared, “I’ve enjoyed it. It’s relaxing. When I’m in yoga my mind is clear. I enjoy the group. It’s calm.” Several participants reported an increased awareness of their bodies as a result of yoga (n = 3); one participant stated, “[Yoga] relaxes my muscles, [I have] less pain (noticed changes in carpal tunnel) . . . Afterwards, I’m not a rubber band wound too tight.”

Participants were generally receptive and engaged in TCTSY sessions, as evidenced by instructor field notes and qualitative interviews (n = 12). Participants found the TCTSY sessions to be “relaxing” and “a nice break from life” or their daily routine (n = 10). However, the main dislikes among TCTSY participants were the lack of physical space in the session room (n = 5) and the time of day and commute to the sessions (n = 5). One noted the “. . . cramped space, being on top of everybody, can’t be completely relaxed because you might run into somebody,” while another participant noted, “By the time I get here, I’m stressed [after] 1.5 hours of traffic.” Finally, participants desired longer or more frequent yoga sessions (n = 3). One participant shared,

I wish they offered it 4 times a day so we could do it after one-on-one (therapy sessions). Lots of women there can’t afford a gym so this would be a way to offer them yoga in the VA setting.

CPT.

Prior to the start of the CPT sessions, some participants expressed feeling “nervous” or unsure about participating in group sessions (n = 3) and some were concerned about being “judged” by their peers (n = 3). As seen among the TCTSY participants, some CPT participants found getting to the sessions difficult due to scheduling, traffic, and parking (n = 3). Participants also shared that they would prefer to have the group sessions somewhere other than the VA, as this location triggered PTSD symptoms. After participating in several sessions, some CPT participants stated that they appreciated the process of learning how to challenge their thoughts (n = 2), though many found the process difficult (n = 3). One participant shared, “[It’s] very hard for me to discuss feelings with a group of strangers but also good for me to hear that others share my experience and issues.”

Data Collection Practicality and Acceptability

Participants were randomized to TCTSY (Cohort 1, n = 9; Cohort 2, n = 8) or CPT (Cohort 1, n = 11; Cohort 2 n = 14); one withdrew immediately following randomization to CPT prior to completing baseline data collection. The remaining 41 women completed baseline data collection. The percent of TCTSY participants retained in data collection was higher at each data collection time point than in the CPT group; however, these differences were not statistically significant, Figure 1; T2 χ²₍₁₎ = 0.874, p = .350; T3 χ²₍₁₎ = 2.476, p = .116; T4 χ²₍₁₎ = 0.672, p = .412. A flowchart detailing recruitment, enrollment, and retention throughout the course of the study, including intervention and data collection sessions, is provided in Figure 1.

Data collection sessions took 3 to 6 hr and several participants noted that data collection presented a significant time burden (n = 6). One participant shared that the data collection process was “too long” and another participant requested “. . . less questionnaires during data collection visits [because] they cause [me] to feel down about stuff.” A third participant stated that the data collection was “like interrogation.” Sleep actigraphy and sleep diaries were the most problematic component of data collection for the participants and yielded minimal usable data. The sleep actigraphy data alone indicated minimal sleep among most participants, at levels not documented in the literature, for example, less than 4 hr per 24 hr repeatedly. Actigraphy data were eliminated from data analysis due to the lack of participant acceptability and reliable data.

Participants’ willingness and tolerance of blood collection and dark-enhanced startle/heart rate variability assessments were greater than anticipated. However, two participants expressed being skeptical of treatment and data collection procedures and subsequently withdrew from the study. One participant noted that data collectors should provide a “. . . better rationale for the blood draws” because she believed they were a diagnostic test for PTSD. In addition, two participants identified their main dislike of the study to be the “startle lab.” It was hypothesized that the hyperarousal associated with PTSD might pose potential data collection and interpretation challenges; however, very few participants declined or discontinued DES data collection sessions.

Discussion

This feasibility study provided robust information to develop a successful proposal for a large-scale RCT of the TCTSY intervention. The data related to the feasibility of study implementation were most informative and guided changes in recruitment, retention, and data collection procedures when developing the full-scale RCT proposal and protocol. Overall, some findings were expected (e.g., traffic, parking, and the male-dominated VA environment) and acted as barriers to study engagement. Participants expressed a preference for the yoga intervention; however, other findings provided new information and lessons learned. A discussion of the unexpected findings and recommendations follows.

Feasibility of Recruitment and Data Collection

Recruitment.

We encountered more challenges with recruitment than we anticipated, given the volume of women veterans served at the study site and the number of monthly referrals to the WTP. One significant factor was the passage of H.R.3230, the Veterans Access, Choice, and Accountability Act of 2014 (Choice Act) in response to public perception that the VA health care system was unable to see patients in a timely manner (Chang & Brannen, 2015). Passage of the Choice Act resulted in decreased numbers of veterans seeking care at the study site during the recruitment phase of the study.

The second significant challenge to recruitment was participants’ unwillingness to be randomized to CPT. Internally, this study became known as “the yoga study.” Thus, the majority of interested participants expected a yoga intervention, not the current standard of trauma-focused psychotherapy. While this phenomenon supported the rationale for the study (that EBPs for PTSD have limited desirability and acceptability in this population), it created challenges for recruitment. For this reason, a non-yoga specific study name, Project Stress-Less, was used for the full-scale RCT.

Logistics were an additional challenge to recruitment, particularly given the group format of intervention delivery. This format and study design necessitated enrollment of a sufficient cohort to populate two groups to start simultaneously. The challenges in recruitment described above resulted in delays of several weeks to months for early enrollees in a cohort, leading to some attrition. Also, the group format and randomization meant that potential participants had to be available at two specific times each week, to be eligible due to the scheduling requirements of the four interventionists.

Recommendations.

First, ensure that recruitment information to referral sources (e.g., clinicians) and potential participants do not imply receipt of a popular alternative to a difficult to endure standard of care treatment in an RCT. In short, avoid references to “the yoga study.” In studies with similar designs and disparate intervention arms, we recommend that potential participants be informed that they can receive the intervention they are not assigned to after completion of their participation in the study at no cost (cross-over option) to improve recruitment. Cross-over design was offered in the full-scale RCT.

Second, whenever possible, streamline scheduling so that both intervention groups are held at the same time to minimize potential participants’ schedule conflicts. Finally, exploration of alternative intervention delivery models, such as virtual intervention and data collection sessions may reduce participant burden and increase engagement.

Data Collection.

Participants were notified of their group assignment after consenting to participate in the study but prior to baseline data collection. While only one woman withdrew from the study explicitly because of her assignment to CPT, several indicated that they were interested in the study primarily because they wanted an alternative to CPT.

The data collection sessions took longer than anticipated. This was due in part to challenges with DES and wrist actigraphy equipment malfunctioning and/or user error and the separate building locations for interviews, phlebotomy, and DES. Qualitative feedback from participants suggested a need for streamlining data collection to reduce participant burden and enhance the quality of data collection to uphold the study design. This feedback was integrated into the design and methods of the full-scale RCT. On the other hand, participants were far more willing to undergo data collection, including blood draws and DES, than anticipated. We theorize that this reflects a combination of veterans’ generally higher willingness to participate in research studies than civilians, women veterans appreciating a study designed specifically for them as they are underrepresented in VA studies, as well as the benefits of research staff experienced in interpersonal trauma-related research and able to establish trust and rapport with participants (Campbell et al., 2007; Danan et al., 2019)

Striking a balance between thorough data collection and participant burden in this study was challenging. The planned aims for the full-scale RCT included a comprehensive evaluation of PTSD symptoms and commonly co-occurring symptoms of depression, chronic pain, and sleep disturbance. In addition, given that the state of the science lacks clear mechanism of action of yoga for PTSD, we aimed to establish the feasibility of multiple psychophysiological measures, adding significantly to participant and research staff burden. For the full-scale RCT protocol, we used shorter measures and eliminated redundant measures and non-essential outcomes. We reduced the total number of response items on clinician-administered and self-report questionnaires (baseline and outcome) from 820 in this feasibility study to 364 in the full-scale RCT design.

Recommendations.

Similarly designed studies are recommended to collect baseline data before randomization and participant notification of their intervention assignment. In terms of aims and outcome variables, participant burden can result in less successful collection of essential data. We recommend the use of high-quality, up-to-date equipment to obtain psychophysiological data, though this can be financially prohibitive in small studies. Teach-back methods should be used with participants for psychophysiological data collection instruction with follow-up phone calls to ensure understanding and proper use of equipment.

Feasibility of TCTSY Implementation

Findings related to the TCTSY intervention indicated that the intervention protocol did not require changes or adaptations, nor did processes to assess and maintain intervention fidelity. Efforts such as co-facilitation of TCTSY, 20 hr of trauma-sensitive yoga training for the facilitators, debriefing sessions, and ongoing consultation calls allowed for fidelity to the model to be maintained without deviations from the protocol. Future studies would benefit from use of TCTSY-Certified Facilitators and video-recorded sessions to allow for robust and objective treatment fidelity assessment. Similarly rigorous CPT treatment fidelity utilizing established measures should be employed as well.

Limitations.

This feasibility study has several limitations that are important to note. First, this study compared two interventions delivered in group format: group TCTSY versus the gold-standard treatment at the time, group CPT. Since the time of the study, individual CPT has been shown to be superior to group CPT (Resick et al., 2017). Future studies utilizing CPT might use the individual format, depending upon the setting. In addition, criteria regarding missed CPT sessions were applied, in that participants who missed four sessions but did not initiate withdrawal from the study (n = 3) were removed from the group by the study team (Resick et al., 2015). There was not a parallel criterion for the TCTSY group. Assessment of treatment fidelity in CPT did not include established quantitative measurement.

Conclusion

The findings of this feasibility study supported the design of a subsequent VA-funded full-scale RCT of TCTSY for women veterans with PTSD who experienced MST and are receiving PTSD treatment in the VA health care system (NCT02640690). This feasibility study, the subsequent full-scale multi-site RCT, and published studies in recent years firmly establish that yoga is a feasible and acceptable intervention worthy of exploration for PTSD symptom reduction, functional improvement, and overall quality of life improvement. Additional studies are needed to investigate the effectiveness of yoga, specifically TCTSY as an alternative, precursor, or concurrent adjunct to PTSD treatment and to identify demographic and clinical predictors of potential effectiveness (i.e., for whom is TCTSY most likely to improve PTSD symptoms). Dismantling studies should be conducted to better understand the mechanism of action when applying yoga as a clinical intervention or treatment. These suggested studies are needed to provide evidence to support adoption of TCTSY as a clinical intervention for PTSD in the VA system nationally.