Making Sense of Poor Adherence in PTSD Treatment from the Perspectives of Veterans and Their Therapists

Abstract

Objective:

Dropout rates from trauma-focused PTSD treatments (TFTs) in VA clinics are particularly high. We conducted in-depth qualitative interviews with 29 veterans and their therapists to better understand this phenomenon.

Method:

Participants part of a multi-site, mixed-methods study of TFT adherence in VA clinics. Veterans were eligible for interviews if they exhibited poor TFT adherence and screened positive for PTSD in follow-up surveys. Interviews were analyzed using qualitative dyadic analysis approaches.

Results:

Therapists relied on stereotypes of poor adherence to understand veterans’ experiences and were missing information critical to helping veterans succeed. Veterans misunderstood aspects of the therapy and struggled in ways they inadequately expressed to therapists. Therapist attempts at course corrections were poorly matched to veterans’ needs. Many dyads reported difficulties in their therapeutic relationships. Veterans reported invalidating experiences that were not prominent in therapists’ interviews.

Conclusions:

Future work is needed to test hypotheses generated and find effective ways to help veterans fully engage in TFTs.

Adherence is defined as the degree to which an individual’s behavior corresponds to the agreed-upon recommendations from a healthcare provider (Sabaté, 2003). Within cognitive-behavioral therapy (CBT), this is typically measured through session attendance and homework compliance (Taylor et al, 2012). Treatment adherence may be especially problematic for veterans receiving trauma-focused CBTs for PTSD within VA facilities (i.e., Prolonged Exposure and Cognitive Processing Therapy). Hale and colleagues (2019) found that among nearly 17,000 veterans who began either CPT or PE in a VA facility, only 31.1% completed an adequate dose of treatment (i.e., eight or more sessions). While we are aware of no published data on the frequency of homework completion among veterans receiving TFTs, suboptimal homework compliance is commonplace in CBT (Helbig & Fehm, 2004). Both treatment completion and homework compliance are robust predictors of CBT response, generally (Glenn et al., 2013), including from TFTs (e.g., Cooper et al., 2017).

Research explaining poor adherence to TFTs among veterans is limited and inconsistent. For treatment dropout, except for age, few predictors have been replicated in more than one study. Demographic, non-mutable personal characteristics (e.g., trauma type, comorbidities), and facility information (e.g. percentage of clients using PE/CPT at the facility) are examined the most frequently (e.g., Hale et al., 2019; Wiltsey Stirman et al., 2018). With a few notable exceptions (i.e., Hundt, Ecker, et al., 2018; Hundt, Helm, et al., 2018; Meis et al., 2019), this work is largely a-theoretical, reliant on administrative records data, focused on variables that cannot be altered, or dependent upon secondary analyses from efficacy treatment trials with a limited scope of predictors. This greatly restricts the clinical utility of the conclusions which can be drawn.

We are aware of only one study examining predictors of homework compliance with TFTs (Wiltsey Stirman et al., 2018). Wiltsey Stirman and colleagues (2018) examined associations with homework compliance among female sexual or physical violence survivors. They found that greater PTSD symptom severity predicted session 1 homework compliance but not homework compliance at later sessions. Other client characteristics were not significantly associated with homework completion. Relatedly, in other client populations, therapist characteristics are better predictors of CBT homework compliance than client characteristics (Holdsworth et al., 2014). Exceptions include greater social support (Hebert et al., 2010) and client capacities (Holdsworth et al., 2014), such as stronger intentions to complete treatment (Hebert et al., 2010) and treatment rationale agreement (Addis & Jacobson, 2000).

Qualitative methods can provide an important avenue for advancing our understanding of poor TFT adherence among veterans. Three published qualitative studies have examined the perspectives of veterans with lived experiences struggling with TFT adherence. While not an explicit study of treatment adherence, Doran and colleagues (Doran, O’Shea, & Harpaz-Rotem, 2019; Doran, O’Shea, & Harpaz-Rotem, 2020), interviewed veterans who participated in TFTs for PTSD. Eight had dropped out of treatment and were analyzed separately. Their explanations for dropout included external reasons (e.g., practical barriers, other life stressors, health problems), internal reasons (avoidance, negative emotions, coping limitations, symptom exacerbation), and therapy-related reasons (e.g., treatment was exhausting).

Hundt and colleagues (Hundt, Ecker, Thompson, Smith, Stanley, & Cully, 2018; Hundt, Helm, Smith, Lamkin, Cully & Stanley, 2018) conducted the remaining two qualitative studies, interviewing veterans who presented for a TFT within a VA PTSD outpatient clinic and either failed to initiate treatment or ultimately dropped out of treatment. Veterans provided descriptions of several barriers to adherence, including practical barriers (e.g., work, school, or caregiving demands), emotional barriers (e.g., treatment was stressful, feeling not ‘ready’), VHA system-related barriers (e.g., scheduling, negative prior VA care experiences), and therapy or therapist-related barriers. Therapy or therapist-related barriers included trouble connecting with therapists, feeling pushed too fast in the therapy, lack of veteran buy-in to treatment, feeling the therapy was not working, or deciding jointly with the therapist to quit (e.g., because treatment was too hard).

We were unable to locate any published qualitative studies interviewing therapists about their perspectives and experiences with poor TFT adherence. However, Eftekhari and colleagues (2020) surveyed therapists participating in the VA’s PE training program about why their veterans dropped out of treatment when they discontinued prematurely. Forty-five percent reported that their veterans could not tolerate PE, due to increases in distress (35.9%) or avoidance of treatment (9.2%). Thirty-seven percent reported dropout was due to reasons outside of the therapy itself, including external stressors (e.g., family deaths, medical issues) and scheduling problems. Nearly 18% of therapists reported the reason for dropout was “unknown.”

While not about treatment adherence, several in-depth qualitative studies exist exploring therapists’ perceptions of the challenges to implementing TFTs in real-world care more broadly. A common therapist-reported barrier to TFT use is client ‘readiness’ for trauma-focused treatment (e.g., Zubkoff et al., 2016), defined as psychiatric stability, broad readiness to change, and established coping skills for managing distress and/or regulating affect (Cook et al., 2017). Other therapist-reported barriers to TFT delivery include clients’ motivation level, openness to TFTs, support network quality, connection to their TFT therapist, uptake of local resources, competing priorities to address other psychosocial stressors, limited literacy, and cognitive limitations (Cook et al., 2014; Marques et al., 2016). Therapists express beliefs that TFTs only work for some clients, may make clients worse (Marques et al., 2016), lack flexibility, are insufficient to address symptoms for many veterans and are a poor fit to some veterans’ needs (Doran et al., 2019). Given these reports, one can anticipate many of these barriers may also limit adherence for TFTs among veterans. This remains unexplored.

All of the work highlighted above focuses on one member of the veteran-therapist dyad at a time. Psychotherapy is, by nature, interpersonal. This single perspective approach is inadequate for developing a multi-dimensional understanding of a given treatment episode. Without considering both perspectives simultaneously, we miss the opportunity to make meaning from the overlap and contrasts in the experiences of each dyad member. A qualitative dyadic approach allows one to capture each reporter’s description of how the treatment episode unfolded and how dyad members worked together or at cross-purposes. This information is crucial to understanding when and how disconnects occur between veterans and therapists and what can be done about it. Quantitative research has shown that agreement between veterans and therapists on measures of the quality of their therapeutic relationship can predict psychotherapy outcomes (e.g., Rubel et al., 2018). Understanding more about the experience of the therapy, considering both dyad members’ perspectives, may help to identify processes contributing to veteran adherence.

The purpose of the current study was to conduct a qualitative dyadic analysis of veterans’ and therapists’ experiences when veterans struggled with TFT adherence. Dyadic qualitative approaches, such as when each dyad member is interviewed alone but analyzed together (Eisikovits & Koren, 2010), allows for drawing meaning from each dyad member’s personal experience as well as from the two interviews side-by-side (i.e., dyad as the unit of analysis), deepening and broadening our knowledge of a dyad’s experience. We were interested in examining veteran’s and therapist’s descriptions of their therapy experience, how experiences compare, and how to characterize and understand congruencies and incongruences.

Method

Participants and Methods

Veterans were interviewed for a larger, mixed-method study of TFT adherence, using a prospective embedded design (Creswell & Plano Clark, 2018; see Meis et al., 2019). For the parent study, veterans received mailed surveys (Time 1) as they initiated TFTs across four VA hospitals (Time 1; N = 598). Veterans were surveyed again four months later (Time 2). Veterans were eligible for qualitative interviews if they completed Time 2 surveys, either completed or dropped out of treatment before interview recruitment, had poor CPT or PE adherence (defined below), and significant Time 2 symptoms of PTSD (i.e., PTSD Checklist score of 50 or greater, PCL-S; Weathers et al., 1994). Poor adherence was defined as dropping out of TFT before successful completion, as indicated in therapists’ electronic medical record notes (EMR) and/or completion of less than half of assigned homework tasks, on average. Homework completion was assessed through therapists’ responses to an item embedded within EMR notes: The veteran completed __ of the homework assigned: none (0%), some (25%), about half (50%), most (75%), all or nearly all (100%). Veterans eligible for interviews were recruited by mail and phone for a 90-minute interview about their experiences with CPT or PE and were compensated $50 (see Meis et al., 2019). We used purposive sampling to ensure at least 25% of the interview sample were women and 25% identified as from racial or ethnic minority groups. We recruited eligible veterans until thematic saturation was achieved (i.e., no new information was obtained from interviews; Creswell & Plano Clark, 2017). Fifty-eight veterans were contacted for recruitment, and 53.4% of these veterans completed study interviews (n = 31). Veterans were asked if the study team could interview their CPT or PE therapist and a loved one (not incorporated in the present study). All therapists approached agreed to be interviewed. Therapists were consented by telephone and completed a 60-minute interview, with no compensation due to VA policy. Four therapists had treated multiple veterans in the sample. These therapists completed separate interviews for each veteran.

While 31 veteran interviews were conducted, two interviews were excluded from this analysis. One veteran declined to provide permission to interview his/her therapist. The second was excluded because the audio recorder failed during the therapist’s interview, yielding a final n of 29 veteran-therapist dyads (29 veterans across 23 therapists). The study was conducted in compliance with the Institutional Review Boards (IRBs) at each of the participating facilities. Veterans in the final sample averaged 50.9 years old and were 59% male, 17% African American/Black, and 20.7% Hispanic. Many had received CPT (75.9%); 24.1% received group CPT; 9.5% received CPT without a trauma account (CPT-Cognitive; CPT). Most of the veterans interviewed had dropped out of TFT (89.7%). The average time from veterans’ last session to participant interview was 6.3 months (SD = 1.6) for veterans and 7.8 months for therapists (SD = 1.9). See the online supplement for additional characteristics (Tables 1 and 2).

Table 1.

Characteristics of Veteran Participants

Characteristic		Characteristics Continued
Race, N (%)		Time 2 PTSD Checklist, M (SD)	64.8 (9.5)
White	18 (62.1)
Black	5 (17.2)	Deployment Era (N, %)
Native American/American	1 (3.4)	Post-911	9 (31.0)
Indian, N (%)		Vietnam	11 (37.9)
Latinx/Hispanic ethnicity, N (yes, %)	6 (20.7)	Gulf War	2 (6.9)
		Peacetime	5 (17.2)
Income, N (%)		Combat Exposed, N (%)	15 (51.7)
Less than $20,000	7 (26.9)	MST Exposed, N (%)	5 (17.2)
$20–40,000	9 (34.6)	Therapy Info, N (%)
$40,000–60,000	7 (26.9)	Dropped out	26 (89.7)
$60,000–80,000	2 (7.7)	PE	7 (24.1)
Greater than $80,000	1 (3.8)	Group delivery (CPT)	7 (24.1)
Age, M (SD)	50.9 (14.3)	Telemedicine Delivery	4 (13.8)
Education, N (%)		Sessions among dropouts, M (SD)	4.07 (3.41)
Less than a high school diploma	1 (4.0)
High school graduate or equivalent	4 (13.8)
Some college	13 (44.8)
Bachelor’s degree	3 (10.3)
Post-graduate degree	4 (13.8)
Gender, N, (%)  Men	17 (58.6)
Women	9 (31.0)

CPT = Cognitive Processing Therapy. PE = Prolonged Exposure. PTSD = Posttraumatic stress disorder. Percentages represent valid percentages that exclude missing cases from calculations. Categories that do not add up to the total number of veterans in the sample (n = 29) indicate missing data.

Table 2.

Characteristics of Therapist Participants

Characteristic	N	%
Therapist Discipline
Social worker	10	45.5
Psychologist	8	36.4
Psychology Trainees	2	9.1
RN-level Nurse	1	4.6
Nurse Practitioner	1	4.6
Therapist gender
Male	9	39.1
Female	14	60.9

Note. Data was missing on therapist discipline for one therapist.

Interview Procedures

Open-ended, semi-structured interview guides were used for data collection. The interview guide began with broad opening questions and included follow-up probes. Questions covered (1) the course of and experiences in TFT, including the therapeutic relationship, homework, and dropout. For example, “Will you walk me through your decision to quit.” (veteran interview), (2) perceived treatment response: “What, if anything, do you think he/she got out of CPT/PE?” (therapist interview), (3) factors influencing sticking with or dropping out of treatment, including treatment barriers and facilitators: “Why do you think he/she stuck it out as long as he/she did?” (therapist interview). The interview concluded with soliciting opinions about family involvement in TFTs for the parent study.

We used a flexible “conversational partnership” (Rubin & Rubin, 2005) format that allowed interviewers to follow the interviewee’s train of thought and ask additional questions to maximize the depth of response. Over the first four interviews, we expanded on and refined the interview guide (Crabtree & Miller, 1999). Interviews were administered by four individuals (MO, EE, RO, LM). The majority were conducted by a master’s level psychologist (74.1%, EE). The study employed several methods to ensure quality and consistency in data gathering. Each interviewer conducted at least two initial interviews with another team investigator. Two interviewers were present for 13 interviews (22.4%). Interviewers completed written summaries, outlining highlights/key themes. The first author directly observed or reviewed recordings for 14% of all interviews.

Data Analyses

Interviews were professionally transcribed verbatim and de-identified for analyses. An ‘audit trail’ was maintained to ensure accounting of the coding process. We employed a social constructionist framework where the goal was to understand how individuals within a relationship build and interpret their social reality (Reczek, 2014). Consistent with dyadic qualitative analysis approaches, we analyzed each dyad from three perspectives: 1) the veteran, 2) the therapist, and 3) the dyad, at both the descriptive and interpretive level (Eisikovits & Koren, 2010; Hockman et al., 2020). This approach allows for identifying themes that would not have emerged through individual interviews alone (Eisikovits & Koren, 2010). Analysis at the individual level (veteran or therapist) aimed to understand the individual participant’s experience (Hockman et al., 2020). The individual-level analysis was purely foundational and not an analysis end-point, as one of the assumptions made with dyadic analysis is that any participant’s perspective alone is incomplete without actually considering his or her partner’s perspective too (Eisikovits & Koren, 2010). The individual analysis included memo-writing, review of individual interviews multiple times, highlighting significant statements and quotes, and developing clusters of meaning (e.g., horizontalization; Creswell, 2007; Patton, 2015).

A team of five research clinical psychologists with expertise in the treatment of PTSD then summarized each dyad member’s interviews side-by-side. Two researchers were randomly assigned to each dyad. Each researcher independently completed a summary table for the dyad, including direct quotations (see Table 3 for example excerpt). The two researchers then met together to review their summary tables. When researchers’ summaries diverged, they discussed the divergence until they achieved agreement on a shared interpretation (Patton, 2015). A final consensus table was created for each dyad. These consensus tables were then reviewed by the first author for common patterns and themes. Patterns were brought back to the team for discussion and clarified, merged, and separated until reaching a consensus on the final themes.

Table 3.

Example Excerpt of Dyadic Analysis from One Veteran-Therapist Dyad

#	Veteran	Therapist	Dyadic Analysis
1	It (CPT) was not as effective as it could have been. It made me think, but I’m not any different than I was before. It went okay.	It (CPT for this veteran) was good.	Dyad members have clearly discrepant views of how effective the treatment was for the veteran.
2	Trauma-focused therapy was not a good fit to my real problems: My problem was not my experience in Vietnam. It was my reaction to the praise returning veterans are getting. I’ve already spent a lot of time writing and talking about the trauma.	Therapist was aware of veteran’s reaction to the praise returning veterans were receiving but didn’t target this in treatment.	They agree that veteran’s reaction to returning veterans coming home caused distress, but therapist did not see this as the veteran’s central problem. The veteran saw this as his primary problem, so he found the treatment unhelpful.
3	Veteran did not have a good connection with his therapist. Described her as “unskilled,” “uncaring,” and “she drives me crazy.”	Therapist reported they had a good therapeutic alliance.	Veteran had a strong dislike of his therapy and his therapist of which the therapist appeared unaware.

Results

Convergence between dyad members was common around how the episode of therapy went overall. Typically, dyads agreed that therapy went poorly or was not helpful. Less often the dyad agreed it was helpful in some ways but not sufficient to address the veteran’s needs. Dyads also agreed on the basic details of what happened, such as if the veteran finished treatment and basic details on how therapy ended (e.g., termination was directly discussed in session). Other points of convergence included mutual expressions of positive regard for each another, awareness of the veteran’s motivation and that the treatment was challenging for the veteran, shared understanding of the veterans’ reports of symptom exacerbation or fears about the treatment (e.g., anticipatory anxiety regarding exposure exercises), shared understanding around barriers and facilitators to participation (practical barriers, level of the support in the veteran’s social environment, relationships with therapists before the TFT). However, discrepancies in dyad members’ perceptions of the treatment experience occurred considerably more often than these convergences. In comparison, convergences were also less substantive than discrepancies in their depth and significance to the interviewees. The content of the divergences is captured in the themes below. See Table 4 for a summary of all themes and subthemes.

Table 4.

Summary of Primary Themes and Subthemes

Theme	Description and Subthemes
1. Therapists relied on stereotypes while veterans’ explanations were nuanced	Therapists’ explanations of their veterans’ adherence problems relied on simple clinical short hands that did not fit with veterans’ experiences; veterans’ explanations were rich, unique, and multi-causal.
2. Therapists were in the dark	Therapists lacked important details about their veterans’ experiences, challenging therapists’ efforts to accurately understand their veterans’ circumstances.
3. What’s therapy supposed to do?	Veterans held large misunderstandings about the nature of trauma-focused therapy of which therapists appeared unaware.
4. Signals, misses, and misfires	Therapists were aware of signs and signals that veterans were likely to dropout. In response, therapists proceeded with minimal adjustment or made changes that were mismatched with veterans’ perception of the problem.
5. Problematic veteran-therapist relationships and interactions	2A. Lack of a bond between veterans and therapists (Disconnection). 2B. Veterans described invalidating interactions with therapists; therapists were either unaware of the salience of these interactions or experienced them differently (Invalidating Experiences).

Theme 1: Therapists Relied on Stereotypes While Veterans Explanations Were Nuanced

The language therapists used to describe their experiences relied on clinical shorthand or stereotypes commonly used to explain poor adherence. Examples include clients’ who were “not ready,” “highly avoidant,” “not motivated,” “stuck,” could not “tolerate” the treatment, could not “stay on task,” had scheduling problems, and/or comorbid personality features that interfered with treatment (e.g., borderline features). Therapists often attributed root causes of non-adherence to characteristics of the veterans and under their control:

That’s all I could do at that point. He had to do the rest. I can’t see what else I could do as a therapist or the CPT model itself. I think it came to a point where he [veteran] had to really grab the ball and start running with it, but he just couldn’t or didn’t want to, or maybe a combination of both. (Therapist 13230)

It is not surprising that veterans’ accounts of why they quit or did not fully participate were more rich, complex, and multifaceted. A veteran’s understanding of their behavior should, by nature, be more comprehensive than an outsider’s. However, the simplicity of therapists’ explanations in combination with the use of stereotypes and the poor match in content between therapists’ and veterans’ explanations was striking. Importantly, these differences were present even when veterans expressed warm or positive regard for their therapists.

Veterans attributed noncompliance to their fears of the therapy, doubts that the treatment could work for them, emotional reactions to the therapy (e.g., feeling worse), failings of the therapist, life events, and their limitations. These causal explanations were woven together in complex ways. One therapist (14120) reported that the veteran she worked with had practical barriers to participation because she was a single mom with young children and that, “She may not have wanted to allow herself to be that vulnerable [in therapy].” However, the veteran reported a much more complex set of circumstances that could have been addressed in treatment in several different ways:

I felt very overwhelmed, and I felt very lost and very frantic in what I had to do [for therapy]. Especially having the homework to do, but then knowing I couldn’t do it. Because I knew my reaction was going to upset my daughters. So, then trying to find time later at night, but then writing about all of that would disturb my sleep -- what little bit I get. And the disturbed sleep would mean I was up all night … so it just kind of snowballed.

See Table 5 for additional examples of these discrepancies.

Table 5.

Example Summaries of Veteran Versus Therapists Explanations of Poor Adherence (Theme 1).

ID	Therapist	Veteran	Veteran’s regard for therapist
13230	Therapist described the veteran as unmotivated and unwilling to confront his trauma.	Veteran feared he could not manage the strong feelings that would be evoked from discussing his trauma.	Negative
17920	Therapist thought the veteran’s work schedule changed, and he could not participate in therapy.	Veteran sought treatment to get an official PTSD diagnosis and not for trauma focused therapy. The therapy made him more aware of the significance of his mental health problems, making him feel worse.	Negative
12650	Therapist thought the veteran did not want to talk about the trauma because it would be too painful and too hard.	Veteran had done CPT in the past. ‘Redoing’ the trauma-focused parts of therapy would be redundant and was not what he needed.	Positive
17160	Therapist thought the veteran was not ready because she had so much going on in her personal life. Therapist gave “her room to decide” in hopes she would “come back when she was ready.”	Veteran worried about managing her emotional reaction to homework. When she had an appointment, she would “allow” life to get in the way. She “shared more” with the study interviewer than her therapist.	Positive

Theme 2: Therapists Were in the Dark

Therapists were often missing important details about their clients’ experiences that might have helped them better understand their clients and intervene. For example, as veterans began thinking more about their trauma histories in treatment, some veterans would report symptom exacerbations they did not share with their therapists. When these symptom exacerbations (e.g., severity and frequency of nightmares) were shared with therapists, veterans would not disclose just how debilitating the symptoms were to their day-to-day functioning. Veterans described changes in their commitment to therapy that they did not disclose to therapists, preventing therapists from helping veterans through these changes. Critical events happened for veterans between therapy sessions that the veterans linked to their difficulties in treatment but did not discuss with their therapists (e.g., family members discouraging participation, deaths in the family). Veterans’ descriptions of practical barriers to engagement (e.g., work and/or childcare responsibilities) were more problematic than therapists’ descriptions of the same barrier. On occasion, therapists were aware they were missing something: “Was it his work schedule? Is it something else going on in his life? Was it specifically the identified trauma that he didn’t want to address? That’s the part that’s unknown to me,” (therapist 17920).

In one example, a therapist (16670) was broadly aware that the client had practical challenges (housing instability, school responsibilities) and comorbid symptoms of paranoia. She accurately surmised that something stimulated the veteran to quit. Still, she was not aware of the extent to or duration with which he struggled in therapy before quitting, nor did the veteran reach out when a crisis led him to quit:

He just suddenly stopped coming… I’m thinking that he either got really, paranoid of what my motive was… or that it [the therapy] was, in his mind, very ineffective, or dangerous, or too overwhelming for him to do. So, something happened. (Therapist 16670)

The veteran reported privately struggling on his own for weeks and keeping many important experiences from his therapist. “I was starting to have a lot of doubts [about the therapy]. Started to question myself.” He and a military “buddy” regularly discussed how his PE exposure exercises (in vivo exposure) were harmful. He was assigned crowd-related exposure exercises that he approached as “surveillance” rather than new learning opportunities. The veteran’s buddy and wife encouraged him to quit, followed by this event:

I was at school, and one of the kids there, he was playing with his phone. And, his camera accidentally went off. That launched me into a huge rage… I jumped up and, I’ll be honest, in that moment of white, hot, flashing rage, I probably could have killed that kid and not even thought twice about it… I am a person of many, many controls. To lose control like that? That’s not good. (Veteran 16670)

When looking back on their therapy experiences, veterans acknowledged they did not discuss these matters sufficiently or at all with their therapists. When veterans were able to verbalize reasons for not disclosing, veterans reported discomfort with their therapists, shame for struggling with compliance, fears of disappointing their therapists, desires to maintain control over their choice to quit, and fears that discussing their problems may lead the treatment to fail.

Theme 3: What’s Therapy Supposed to Do?

An important group of veterans had basic misunderstandings of the therapy itself, including confusion about why they were asked to face their trauma-related distress or misinformation on how the treatment works, leading to engagement problems. These misunderstandings were largely missed by therapists or misinterpreted as a lack of motivation. For example, this therapist (12670) was working with a veteran with comorbid chronic pain. She explained why she thought the veteran quit treatment in this way:

A lot of it centered around her physically not being able to do as much… I know there had been some medication changes. I know the medication was making her a little more tired... I think the fact that she discontinued her treatment says a lot that maybe it was just too much for her to do the more intensive work. (Therapist 12670)

Alternatively, the veteran reported:

I was just burnt out, I think. Just burnt out with talking about stuff [trauma memories], because I would explain to her [the therapist] that I don’t want to talk about stuff that I blocked out for many, many, many years. And, then, of course, she would explain to me why it was important to talk about it, and I just didn’t feel that way… It wasn’t working for me. I don’t want to talk about bad stuff... I really don’t understand CPT. Like what it’s supposed to do? I wouldn’t really know. (Veteran 12670)

Theme 4: Signals, Misses, and Misfires

On many occasions, despite working with only partial information, therapists would successfully identify signals a veteran struggled with the treatment and might drop out. Some therapists would continue moving forward without adjustment. Others would adjust their approach. They described using motivational interviewing, trying to convince the veteran of the rationale, postponing imaginal exposure exercises, discussing other psychosocial stressors, and recommending veterans end treatment when homework adherence or attendance was chronically problematic. However, therapists deployed these interventions with incomplete information, and the interventions were often incompatible with veterans’ perceptions of the underlying problems.

For example, one veteran (15500) reported considerable initial enthusiasm for treatment: “I went in for the first appointment, with gung-ho-let’s-get-with-therapy [attitude]. I’ll do anything.” The prominent reason she ended CPT was that it did not make sense to her: “[My therapist] said… ‘You’re going to talk about it and talk about it and try to get over it.’ I said, ‘… I can’t make my life go away by just talking about it.’ (Veteran 15500)

In contrast, the veteran’s therapist believed the veteran was unwilling to change: “[She] came in pretty resistant to the idea of change… I think the idea of looking at challenging her thinking… [she] was not willing to do that.” In this case, in the dyadic analysis, the therapist was limited to intervening solely from her perspective. Consequently, she tried to increase the veteran’s willingness to change rather than addressing the veteran’s misunderstandings about how change works:

[I] tried some motivational interviewing with her… I explained to her the pros and cons of changing and engaging in this treatment… When she came in and decided that it wasn’t a good fit for her… I encouraged her to reconsider in the future, but I had already done two sessions of trying to convince her. (Therapist 15500)

Theme 5: Problematic Veteran-Therapist Relationships and Interactions

Subtheme 5A: Disconnection.

A less widespread theme was difficulties in the relationships between veterans and therapists. In some cases, the basic bond between veterans and their therapists was problematic, ranging from an absence of connection to a clear dislike. Okay, so I’ll say, admittedly, she wasn’t the most fun veteran to work with… Gosh, I hate to even say it, but I mean it’s true, I think I sort of dreaded-- dread would be a strong word-- But she just had such a heavy, depressed mood. (Therapist 16270)

In this case, both dyad members possessed negative regard for one another: “I would say the person [therapist] was not sympathetic to my issues… She’s looking at another computer and writing things down. No eye contact… You’re [therapist] not interested in what I have to say, and you are already preoccupied.” (Veteran 16270)

On some occasions, these relationships seemed mixed, with veterans expressing respect or warm regards for therapists while also providing examples of interactions that veterans experienced as unempathetic or disingenuous: “I felt like I was with a salesman. I mean, he was really nice and everything, but I just felt like he was too nice — like he wasn’t real ” (Veteran 13990). Therapists frequently appeared unaware of their veteran’s reactions: “I think he felt comfortable talking generally to me, and I think we got along pretty well,” (Therapist 13990).

Subtheme 5B: Invalidating experiences.

Some veterans described specific interactions they found invalidating that contributed to their poor adherence. Therapist interviews either did not discuss these interactions, suggesting they were not as salient to therapists or described the interaction quite differently. For example, some veterans found therapists’ efforts to emphasize completing homework alienating. One therapist (15510) described the following:

He still hadn’t done any of that [homework]. So, when we met for the third session, I asked him to really consider if this really was something that he’s able and willing to do at this time… Because, I don’t want it to give you the impression that treatment doesn’t work, because you didn’t work the treatment… There comes a point of diminishing returns. Where, if I’m working so much harder than they are, it’s just not worth it. I’m not dragging them through treatment.

The veteran described the same exchange in the following manner:

I have three little boys… and it gets busy. And, I forgot all about this homework, so I went back the next time, and he’s like, “Oh, he didn’t do the homework, so you don’t really care.” And, I was really taken back by [that]. I was like, “Who the f*** do you think you are?” [therapist:]“Oh, you just don’t want to try.”… What am I, in the sixth f***ing grade?

Other veterans described therapists’ efforts to normalize discomfort talking about trauma as disrespectful or tone-deaf to the gravity of what they experienced:

[He said] he’d been [doing CPT] for however long, so he was like, “Well, you’re not--There’s nothing new to me.” It was just too--he was too--nonchalant, I guess… [He’s] like, “Let’s just talk about this [trauma] and get it over with.” (Veteran 13230)

In this case, the therapist appeared unaware: “[Our relationship] was very short, because he only completed four sessions. I guess, in this case, I tried to be as empathetic and calm as I could [be] with him.” (Therapist 13230)

On several occasions, veterans complained about therapists’ strict use of the treatment manuals, including reading directly from the manual and denying veteran requests for adjusting the treatment to the veteran’s needs. One therapist reported: “If it appears that I’m reading out of the book, a lot of the reason being is it [CPT] is very standardized and very specific in how I’m supposed to come across and what I’m supposed to say,” (therapist 17450). The veteran found this problematic: “You can have a curriculum, but you’ve got to let loose… It was just so impersonal. Maybe it was her first shot at it. Maybe she’ll get better.” (Veteran 17450).

Discussion

Many veterans who initiate TFTs within VA clinics are unable to effectively complete these treatments. Hale and colleges (2019) estimate that 61.9% of veterans who begin a TFT within a VA clinic fail to reach an adequate dose of treatment (Hale et al., 2019). We interviewed 29 veterans who struggled with TFT adherence and their therapists to gain a richer understanding of this phenomenon from the perspective of its participants. While some convergence of experience was found between dyad members in broad strokes, divergence was prominent and spanned five themes. Therapists’ explanations of veterans’ poor adherence relied on clinical shorthand (e.g., not ready, highly avoidant). Alternatively, veterans’ explanations were complex and multi-causal (Theme 1). Veterans reported withholding information when they struggled with treatment (Theme 2) and held deep and persistent misunderstandings of the purpose of TFTs (Theme 3). For both Themes 2 and 3, therapists were largely unaware of these issues or perceived them quite differently. When therapists identified signals their veterans were struggling or likely to drop out, therapist adjustments to these signals were rooted in therapist explanations and poorly matched to veterans’ experiences (Theme 4). Lastly, while some dyads described positive rapport with one another, other dyad members did not (Theme 5). Dyad members, particularly veterans, described relationships plagued with a lack of connection (5A) and invalidating experiences that were missed or under-emphasized by therapists (5B).

Therapists’ use of clinical short-hands or stereotypes is consistent with the help of heuristics. Heuristics maximize efficiency through a selective focus on a portion of the information (Gigerenzer & Gaissmaier, 2011). Historically, such mental short-cuts have been considered inferior and associated with greater error rates in decision-making. However, in specific contexts, even complex circumstances, heuristics prove more accurate (Gigerenzer & Gaissmaier, 2011). The accuracy of a judgment is dependent upon the correspondence between the rule the individual is attempting to apply (e.g., when veterans repeatedly skip homework this means they are highly avoidant) and the specific environment or context (e.g., the veteran has small children, a history of moderate head injury, and keeps losing the homework forms; Kruglanski & Gigerenzer, 2011). When therapists in our sample intervened on perceived signals that veterans may drop out, their interventions could be poorly matched to the veteran’s needs (Theme 4). Kruglanski and Gigerenzer (2011) argued that, rather than minimizing the use of heuristics, we should question if the circumstances match the heuristic.

CPT and PE manuals explicitly emphasize some heuristics about treatment noncompliance to assist with educating and training therapists. The manuals indicate repeated noncompliance may mean clients are not motivated for treatment (Resick et al., 2017) or may mean clients are avoiding the activities due to trauma-related distress (Foa et al., 2019; Resick et al., 2017). These heuristics are consistent with the theoretical underpinnings of the therapies and are important possibilities for therapists to consider. However, heuristics are cognitive efficiencies, not explanations. They may prove insufficient and problematic in circumstances under which therapists are missing important information, such as clients’ fundamental misunderstandings about the nature and purpose of trauma-focused treatment (Theme 3) or when clients withhold information from therapists regarding their struggles with treatment (Theme 2). The problem may not be the use of such heuristics, but their application under inappropriate circumstances, like when veteran’s circumstances are more complex than therapists realize. Such errors can be observed in Theme 4 findings (signals, misses, and misfires).

Disconnects between what clients are experiencing and what therapists believe they are experiencing can be also be considered from the perspective of an empathic failure (Mordecai, 1991). Empathy is understanding another’s experience or expression (Elliott et al., 2018). It requires being in step with clients’ experiences rather than their words. While not without criticism (Bloom, 2016), the construct of empathy has long been considered key to the change process in psychotherapy. Meta-analysis demonstrates robust associations between therapist empathy and psychotherapy outcome (Elliott et al., 2018). Empathy is considered something that is “co-created” between therapists and clients as therapists work to understand clients, clients must work to be understood (Elliott et al., 2018). However, the veterans’ we interviewed found disclosing unflattering details about their experiences with TFT or with their therapist quite challenging (Theme 2). Similarly, Hill and colleagues (1992) found that when clients and therapists rated videotapes of their therapy sessions, clients’ acknowledged hiding reactions from their therapists, most typically negative reactions. Borges and colleagues (2020) interviewed veterans who had completed a TFT in the past year about their experiences discussing moral injury in treatment. Participants reported withholding moral injury exploration in treatment due to shame, guilt, and concerns about legal implications.

Achieving empathy and appropriate heuristic use may, be particularly problematic when veteran-therapist relationships are characterized by a lack of connection. This was a frequent occurrence for our dyads (Theme 5A). Fundamental problems in the veteran-therapist relationships were also prominent themes among the veterans interviewed by Hundt, Ecker, and colleagues (2018). Bond, trust, and collaboration are all essential to the therapeutic alliance (Bordin, 1979), which is consistently linked to psychotherapy dropout (Sharf et al., 2011). Borges and colleagues (2020) found, not surprisingly, veterans were more willing to discuss their moral injury experiences when they felt connected to their therapists. Importantly, problematic relationships were not universal in our interviews. There were many cases when some combination of Themes 1–4 were present, but dyad members reported broad positive regard for one another. So, positive regard in and of itself may not be enough to safeguard against these problems as positive regard co-existed in our interviews with empathic failure and inappropriate heuristic use.

Clinical Implications

Taken together, one important potential implication is the importance of therapists’ challenging their assumptions routinely. In their review of the dangers of therapist assumptions, Lilienfeld and colleagues (2014) suggest therapists work to be “detectives” to identify and intervene upon rival explanations for client behavior. They suggest that clinical supervision should encourage trainees to consider alternative explanations for client behavior, examine the potential for cognitive biases to lead to false inferences, and take safeguards against the potential for erroneous inferences. The safeguards they recommend include building in routine assessments of change and obtaining collateral information from clients’ loved ones on clients’ behavior outside of treatment. Responses to these assessments could be used to prompt therapists to have a richer discussion about what clients are struggling with between sessions.

Within the VA, training for both CPT and PE requires completing therapy cases under consultation, thus providing a natural opportunity for implementing some of these proposed safeguards and training. However, there are no ongoing requirements for consultation or supervision after the successful completion of these initial cases. While more research is needed, ongoing peer consultation may provide an opportunity to perpetuate these safeguards. Therapists may benefit from challenging themselves to view the source of veterans’ struggles as stemming from some “third” party rather than within the veteran to promote greater creativity around how to intervene. Therapists should also consider routinely revisiting their own beliefs about what they think the veteran understands about the therapy. A basic misunderstanding of how CPT and PE work and frank disagreement about the benefits of talking about trauma frequently appeared in veterans’ descriptions of their therapy struggles, findings echoed in Hundt and colleagues’ veteran interviews (Hundt, Ecker, et al., 2018; Hundt, Helm, et al., 2018).

Future Research

Compared to the study of its efficacy, there is a paucity of work on essential therapeutic or potentially counter-therapeutic processes within CBT (Kazantzis et al., 2018). Psychotherapy dropout is a well-recognized problem (Fernandez et al., 2015). Recommendations for how to help clients adhere often assume that clients are rational actors (e.g., motivational interviewing, shared decision making; Corrigan et al., 2014). Many decisions regarding health behavior are implicit, unplanned, and in response to one’s immediate context (Corrigan et al. 2014). Research on decision-making among individuals with trauma histories demonstrates such individuals are more likely to select a smaller, immediate reward than a larger delayed reward (van den Berk-Clark et al., 2018). Thus, a rational argument by therapists that hard work in therapy now will lead to benefits later may be particularly unconvincing to clients with PTSD. Additionally, while commitment and intention to change are important, they are not enough to translate intention into behavior (i.e., the intent-behavior gap, e.g., Chang et al., 2017). In our interviews, noncompliance was often met by therapist efforts to increase veterans’ understanding of the treatment. These efforts may lead to frustration if understanding is not enough to bridge the gap to behavior change. Research could examine methods of adjusting veterans’ environmental contingencies to be more friendly to change and treatment participation, rather than solely appealing to clients’ understanding of the logic behind and long-term benefits of treatment.

Limitations

Qualitative research is, by nature, hypothesis-generating. These observations need further testing to determine if they generalize to the larger population or other settings. Despite the iterative, team-based approach to analysis, it is also possible that additional themes within our data could have been missed. Importantly, we did not compare the experiences of our veteran-therapist dyads to the experiences of dyads with optimal adherence and treatment response. This comparison would allow us to explore if the themes which emerged were unique to dyads with poor adherence. Additionally, due to the time between treatment ending and interview administration, veterans’ narratives may have been more or less coherent than they were during the treatment. Deteriorations in memory may have been particularly problematic for some therapists, given the volume of their caseloads, time since working with the client, and their limited contact with veterans. Both therapists and veterans reported imperfect memories for details of their courses of therapy. This was more common among therapists. A review of therapist interviews administered with the greatest time from the veterans’ last sessions did not reveal clear patterns regarding how to recall problems that may have influenced study findings. However, the degree of incongruence between veterans’ and therapists’ experiences may have been amplified or altered by problems with recall.

Finally, four TFT therapists were interviewed multiple times as they treated multiple veterans in our sample. This is not surprising given TFT delivery is done by the limited set of therapists with this specialized training at each VA facility. It is possible opinions and experiences of these therapists are overrepresented in our data. However, the goal of qualitative analysis is to identify the spectrum of themes relevant to a research question, not their frequency (as this is task is for larger, generalizable samples). We anticipate the repeated interviews were more likely to influence the frequency of themes rather than their nature or content.

Conclusions

At this writing, there is scant research exploring the difficulties some veterans experience when participating in CPT and PE. To our knowledge, none of this work considers both the perspective of the therapist and the client. This study provides a rich, dyadic account of veteran and therapists perspectives on the struggles and challenges of completing trauma-focused treatment. More work is needed to test hypotheses generated, and if supported, to develop interventions that help therapists talk to veterans and veterans talk to therapists.

Clinical Impact Statement:

We interviewed veterans with poor adherence to TFTs and their therapists. Therapists used clinical stereotypes to explain clients’ poor adherence (e.g., not ready, highly avoidant). While veterans’ explanations were complex and multi-causal. Veterans withheld information about their struggles, including persistent misunderstandings about the treatment, leaving therapists challenged to respond effectively. Clinicians may benefit from questioning their assumptions when attempting to understand clients’ problems with treatment, identifying alternative explanations, and soliciting client feedback frequently to encourage greater disclosure around challenges in treatment.