Training with tarantulas: A randomized feasibility and acceptability study using experiential learning to enhance exposure therapy training

Abstract

Background:

Although exposure is a key evidence-based intervention for anxiety, it is infrequently used in clinical settings. This study employed a novel training strategy, experiential learning, to improve exposure implementation. This study aimed to assess the feasibility and acceptability of experiential training and preliminary training effectiveness.

Methods:

Participants were 28 therapists who were randomized to (a) training-as-usual or (b) experiential training (training-as-usual plus a one-session treatment for fear of spiders). Workshops lasted one day and were followed by three months of weekly consultation.

Results:

Experiential training was viewed as feasible and acceptable. Participants, including those who were fearful of spiders, had a positive response to the training and reported it to be useful. There was a significant increase in the number of exposures used by therapists receiving experiential training compared to training-as-usual at 1-month follow-up.

Conclusions:

A one-day training resulted in significant improvements in knowledge, attitudes toward exposure, and self-efficacy in using exposure. Preliminary findings suggest that experiential training resulted in greater use of exposure post-training compared to training-as-usual. Results provide evidence for the feasibility and acceptability of experiential training as a strategy to increase the use of evidence-based interventions.

Trial Registration:

Clinicaltrials.gov: NCT0335497

Although efficacious treatments have been developed for a wide variety of mental health disorders, most practicing therapists do not use evidence-based interventions (EBIs; Garland, Bickman, & Chorpita, 2010; Gyani, Shafran, Myles, & Rose, 2014; Shiner et al., 2013). The field of implementation science emerged out of a recognition of this research-practice gap and focuses on the systematic study of how to integrate research findings into routine clinical practice settings (Eccles & Mittman, 2006). Enhancing strategies to train therapists is a critical aspect behind efforts to improve access to EBIs. In particular, insufficient and ineffective therapist training appears to be a primary driver of the research-practice gap; many approaches to training are ineffective in increasing therapists’ procedural knowledge of and self-efficacy in EBI use (Frank, Becker-Haimes, & Kendall, 2020). Evidence suggests that even following “gold-standard” training, including workshops and follow-up consultation, implementation of EBIs remains limited (Beidas et al., 2019).

The research-practice gap is particularly striking for exposure therapy, which is an EBI for treating anxiety disorders with robust support for its efficacy and effectiveness in treating anxiety (Abramowitz, 2013; Davis, Ollendick, & Öst, 2019; Higa-McMillan, Francis, Rith-Najarian, & Chorpita, 2016). Above and beyond other elements of cognitive-behavioral therapy (CBT), exposure has emerged as a key active ingredient in the treatment of anxiety disorders and is included in nearly all evidence-based treatments for anxiety in adults (Chorpita & Daleiden, 2009; Chorpita, Daleiden, & Weisz, 2005) and youth (Kendall et al., 2005; Peris et al., 2015). Furthermore, a meta-analysis of CBT for childhood anxiety disorders found that greater use of in-session exposure was associated with larger treatment effects (Whiteside et al., 2020), and the greatest improvement in anxiety symptoms occurs after exposures are introduced (Peris et al., 2015).

Despite its strong evidence base, exposure remains one of the least used EBIs in community settings (Chu et al., 2015). This may be attributed in part to the fact that exposure therapy is a complex intervention to deliver (Garcia, 2017). In addition, exposure therapy requires intentionally provoking short-term distress for clients, which elicits ethical concerns and negative beliefs for many therapists (Deacon et al., 2013; Gola, Beidas, Antinoro-Burke, Kratz, & Fingerhut, 2015). Negative beliefs about exposure are in turn associated with suboptimal delivery of exposures (Farrell, Deacon, Kemp, Dixon, & Sy, 2013). Estimates suggest that less than a third of therapists regularly use exposure therapy with their anxious clients (Becker, Zayfert, & Anderson, 2004; Becker-Haimes, Okamura, et al., 2017; Cook, Biyanova, Elhai, Schnurr, & Coyne, 2010). Many therapists report inadequate training in exposure (Wolitzky-Taylor et al., 2019), and those who do receive training report using exposure with only a minority of their anxious clients (Becker-Haimes, Okamura, et al., 2017; Higa-McMillan, Kotte, Jackson, & Daleiden, 2017; Thomassin, Marchette, & Weisz, 2019; Wolitzky-Taylor, Zimmermann, Arch, De Guzman, & Lagomasino, 2015; Wolk et al., 2019). These findings suggest that current training efforts focused on exposure are ineffective. Although enhancing therapist exposure delivery will ultimately require a comprehensive set of implementation strategies, optimizing educational implementation strategies may be a necessary precursor to support clinician behavior change (Wensing, Fluit, & Grol, 2013). For exposure therapy in particular, efforts to train therapists may require a tailored training approach (Becker-Haimes, Franklin, Bodie, & Beidas, 2017).

Research on effective therapist training has emphasized the importance of follow-up consultation (Edmunds, Beidas, & Kendall, 2013) and active learning strategies to improve the adoption of EBIs (Beidas & Kendall, 2010). Active learning strategies refer to an interactive, reflective process of acquiring knowledge and may include strategies such as feedback and role plays (Beidas, Edmunds, Marcus, & Kendall, 2012; El-Tannir, 2002). Training approaches that address therapists’ negative beliefs about exposure therapy have shown promise in improving attitudes (Farrell, Kemp, Blakey, Meyer, & Deacon, 2016; Waller, D’Souza Walsh, & Wright, 2016), but there is limited evidence for such approaches leading to sustainable changes in therapist behavior. One study provided evidence for the effectiveness of online training to increase self-reported exposure use (Harned et al., 2014). However, the key ingredients of this online training approach are unknown given its inclusion of several components, including self-paced didactic materials, e-Books, and simulated clinical scenarios. For interventions that are particularly difficult to implement, such as exposures, efforts should be made to identify specific educational strategies that may enhance learning: Experiential learning (Kolb, 1984) may be one such strategy.

Although active learning strategies are undoubtedly important in therapist training (Beidas & Kendall, 2010), they have not been sufficient to lead to meaningful change in therapist behavior. Experiential learning extends beyond active learning strategies by requiring participants themselves to engage in the activity that is the focus of training. To date, experiential training has been only been incorporated into one study of therapist training (Farrell et al., 2016). In this study, therapists in an enhanced training condition completed a series of five brief interoceptive exposures (hyperventilation) and then discussed the safety and tolerability of the experience of completing those exposures. However, this experiential practice was one of several activities included in the enhanced training condition focused on addressing clinician concerns about exposures and was not the isolated focus of study. Research in other domains suggests that experiential learning is an effective method to increase critical thinking and engagement and to enhance theory-informed practice (Clem, Mennicke, & Beasley, 2014; Coker, 2010; Hawtrey, 2007; Knecht-Sabres, 2013). Although experiential training is aligned with behavioral principles that are applied to change client behavior for a wide array of mental health concerns, these same principles have not been applied to changing therapist behavior. Much like a client would not practice new skills in a role play with a CBT therapist and expect improvement from that practice alone, meaningful change comes from applying skills in real-life scenarios. Experiential learning expands upon active learning by providing an opportunity for therapists to participate in the intervention (exposure) themselves.

This study evaluated the feasibility, acceptability, and preliminary effectiveness of an experiential approach for training therapists in exposure therapy by using behavior change principles that have been shown to be effective with clients (i.e., exposure) to train therapists. Specifically, therapists were randomized to training-as-usual (TAU) or to experiential training (ET), which used a modified version of the one-session phobia treatment (Davis et al., 2019) to allow participants to undergo a treatment protocol for exposure to spiders. The first aim of this study was to examine feasibility by monitoring rates of recruitment and retention, completion of assessments, and consultation call attendance. The second aim was to assess acceptability using mixed methods to merge quantitative and qualitative findings. The third aim used exploratory inferential analyses to compare the two conditions on (a) knowledge, (b) attitudes toward exposure, (c) intentions, (d) self-efficacy, (e) self-reported exposure use, and (f) role play competence. Both conditions were expected to demonstrate improvements in some domains (e.g., knowledge, competence, self-efficacy), but the ET condition was hypothesized to have fewer negative beliefs, higher intentions, and higher exposure use than the TAU condition.

Methods

Participants

Therapist participants (N=28) were mental health therapists who signed up to receive training in exposure therapy. Participants were primarily female (75%) and worked in community mental health (CMH) clinics (57.1%). Demographic comparisons between conditions (Table 1) indicated that TAU participants were more likely than ET participants to have a doctoral than a master’s degree. The CONSORT diagram illustrates participant study flow and randomization (Eldridge et al., 2016; Figure 1).

Table 1.

Characteristics of study participants

	Full Sample Mean (SD) or N (%)	Experiential Mean (SD) or N (%)	TAU Mean (SD) or N (%)	Tor χ2 value
Therapists	n=28	n=17	n =11
Age	41.00 (12.42)	38.67 (13.32)	44.50 (10.80)	1.03
Sex (Female)	21 (75)	14 (82.4)	7 (63.6)	1.25
Race				1.08
White	19 (67.9)	12 (70.6)	7 (63.6)
Black/ African American	3 (10.7)	2 (11.8)	1 (9.1)
Asian	3 (10.7)	1 (5.9)	2 (18.18)
Multiracial/Other	3 (10.7)	2 (11.8)	1 (9.1)
Hispanic/Latinx	1 (3.6)	1 (5.9)	0 (0)	0.71
Licensed	19 (67.9)	10 (58.8)	9 (81.8)	1.62
Degree				6.21*
Master’s	22 (78.6)	16 (94.1)	6 (54.5)
Doctorate	6 (21.4)	1 (5.9)	5 (45.5)
Theoretical Orientation				4.24
Cognitive-Behavioral	10 (35.7)	5 (29.4)	5 (45.5)
Eclectic	6 (21.4)	3 (17.6)	3 (27.3)
Acceptance-Based	3 (10.7)	2 (11.8)	1 (9.1)
Humanistic/Client-Centered	3 (10.7)	3 (17.6)	0 (0)
Psychodynamic	2 (7.1)	1 (5.9)	1 (9.1)
Systems/Family Systems	2 (7.1)	1 (5.9)	1 (9.1)
No response	2 (7.1)	2 (11.8)	0 (0)
Practice Setting				2.93
Community Mental Health	16 (57.1)	11 (39.3)	5 (45.5)
Private Practice	8 (28.6)	5 (29.4)	3 (27.3)
Children’s Hospital	3 (10.7)	1 (5.9)	2 (18.18)
Residential Treatment	1 (3.6)	0 (0)	1 (9.1)
Years in Practice	9.74 (9.80)	9.44 (10.56)	10.18 (9.06)	0.19
Level in Training in Exposurea	1.89 (0.96)	1.82 (1.07)	2.00 (0.78)	0.47
Study Status (Drop Out)b	3 (10.7)	3 (17.6)	0 (0)	2.17

^*p < .05

^aMeasured with a 1–5 point Likert scale, with 1 being no training and 5 being extensive training

^bEither requested to drop or did not complete any additional study measures/assessments after training

Figure 1.

CONSORT Flow Diagram

Inclusion criteria.

Participants were required to have an advanced degree in a mental health field, be working in a clinical care setting, and be able to read and speak English. They were required to be treating at least one client with clinically significant anxiety and planning to continue providing therapy to a client(s) with anxiety for the duration of the study. Recruitment targeted CMH therapists; however, therapists working in other settings were also eligible.

Exclusion criteria.

Participants who had previously attended a prior full day workshop on exposure therapy or who had received equivalent training (e.g., working as a therapist in an anxiety specialty clinic) were not eligible.

Procedures

Participants were recruited from December 2017 through February 2018 via professional listservs, email outreach to clinic directors, contact with CMH providers who were part of an advisory committee for another study, distribution of flyers to CMH clinics, and word of mouth. Interested participants completed a brief phone screen and those who were eligible were scheduled for one of four training dates (in March 2018: two on weekdays, two on weekends) taking place at Temple University based on their availability. Participants were compensated minimally ($5–20 depending on time point) upon completion of the training and follow-up assessments. Continuing education credits were awarded for workshop attendance.

Training Conditions.

Random assignment to training conditions occurred using equal allocation at the level of training date, such that conditions were randomly assigned to training dates. Training condition assignment for each of the four dates was determined by the first author using a random number generator. Training condition was revealed to participants after they completed informed consent. Both conditions involved eight hours of training in one day plus one hour for completion of questionnaires. Input from a community advisory board comprised of three therapists working in CMH settings was used to tailor the training to therapists working in routine clinical settings. For three months following training, participants attended consultation calls and received monthly emails prompting them to complete additional study measures.

TAU.

The first half of training included lecture-style teaching with three primary goals: (a) Describe the behavioral model underlying exposure-based treatments; (b) Review populations for whom exposure is appropriate, as well as diagnostic and developmental considerations; and (c) Provide an overview of the “nuts and bolts” of implementing exposures, including (1) Psychoeducation/Rationale; (2) Assessment/Hierarchy Building; (3) Implementing Exposures; and (4) Homework/Relapse Prevention. Efforts were made to incorporate strategies to maintain participant engagement, such as showing client videos and providing tickets for a raffle when participants answered questions. The second half of training incorporated active learning components. The instructor provided case examples and demonstrated role plays, and participants worked in small groups to complete: (a) Behavioral rehearsals, in which they each played different roles (therapist, client, observer) to practice delivering the components of exposure therapy using a pre-determined case example; and (b) Interteaching, in which they engaged in small-group discussion with guided questions.

ET.

The first half of the ET was identical to the first half of TAU. The second half of training was experiential; all participants underwent a modified version of a one-session phobia treatment for spiders, which has been shown to be an efficacious treatment for phobias (Davis et al., 2019; Ollendick et al., 2009). This allowed for a unique experience in which participants completed an entire exposure protocol and gained familiarity with the core elements of exposure. Although most participants were not expected to meet diagnostic criteria for a spider phobia, spiders were selected because of the relatively high base rate of arachnophobia (11.4% lifetime prevalence; Oosterink, De Jongh, & Hoogstraten, 2009). In addition, the wide variety of spider-related stimuli available were designed to elicit some anxiety for most participants, including those without extreme fear. The primary goal of ET was to allow participants to experience exposures from a client’s perspective. In addition, participants alternated between the client and therapist roles throughout ET. Prior to beginning the experiential portion of the workshop, therapists rated their fear of spiders on the Fear of Spider Questionnaire (FSQ; Szymanski & O’Donohue, 1995a) and small groups were formed based on level of fear reported. This ensured that participants with high fear levels were not asked to lead a much less fearful partner through an exposure that was too difficult for them. Groups were determined by scores on the FSQ relative to other attendees on that day, with lower scores indicating less fear.

Training proceeded in the four phases introduced during the first half of training, including Psychoeducation/Rationale, Assessment/Hierarchy Building, Implementing Exposures, and Homework/Relapse Prevention. The trainers (HEF and EMBH) provided a review of each phase before asking small groups to practice and apply the skills with each other. First, participants worked in pairs to practice providing a rationale of treatment to each other. Second, participants collaboratively generated a detailed hierarchy of their partner’s fear of spiders. During the third phase, participants implemented spider-related exposures with one another, beginning with easier exposures and working toward more difficult ones. Pre-selected stimuli were available (e.g., photos and videos of spiders, plastic spiders, dead spiders, live tarantulas), and the use of other stimuli (e.g., finding additional videos) was encouraged. Finally, participants discussed relapse prevention and home practice with their partners. Ethical concerns and methods for troubleshooting difficulties during exposure were discussed as a small group.

Role plays.

Role plays to assess competence were conducted within two weeks of the initial training and post-consultation via Webex video. Two alternate versions of standardized case vignettes for a young adult with social anxiety were developed using expert input. Four experts in the treatment of anxiety disorders provided feedback on the difficulty and clarity of both vignettes. Ratings for the two vignettes were compared and found not to be significantly different. The order of case vignettes was counterbalanced across participants. Participants read the vignette and prepared for the 10-minute role play, during which they were asked to prepare the client for and complete an exposure task. An advanced graduate student trained in exposure therapy was trained to portray an anxious client with standardized responses.

Consultation.

For three months following training, participants in both conditions were asked to participate in at least 75% of the 12 weekly consultation phone calls to receive an adequate “dose” of consultation. Consultation calls were conducted by doctoral students with expertise in exposure for anxiety and supervised by a licensed clinical psychologist. Consultation calls were blind to condition and included a mix of participants from both conditions. Participants on each call were determined by availability. Following initial reminders, participants did not reveal their training condition during the calls. Consultation calls included participant case presentations, didactic topics, and assistance in generating exposure ideas, and were audio recorded and coded for participation.

Questionnaires.

Measures were administered in person before and after training. Participants completed follow-up questionnaires online via Qualtrics (see Table 2 for timeline).

Table 2:

Means and standard deviations (SDs) for measures at each time point by training condition

Measures	Pre-Training		Post-Training		1-Month Follow-Up		2-Month Follow-Up		3-Month Follow-Up
	Training-as-Usual	Experiential	Training-as-Usual	Experiential	Training-as-Usual	Experiential	Training-as-Usual	Experiential	Training-as-Usual	Experiential
Knowledge	3.91 (1.87)	3.71 (1.79)	6.73 (1.35)	6.24 (1.39)	—	—	—	—	5.75 (1.49)	5.36 (2.25)
ETCUS	1.30 (1.55)	0.75 (1.08)	—	—	2.32 (2.37)	4.18 (5.29)	3.73 (2.99)	4.53 (5.75)	1.82 (0.95)	3.49 (2.71)
% Exposure Use	22.88 (22.88)	17.81 (28.92)	—	—	40.25 (28.94)	56.56 (44.69)	41.55 (26.66)	52.82 (34.43)	39.48 (18.72)	64.03 (31.25)
Total Exposure	1.36 (2.01)	0.59 (0.87)	—	—	1.40 (1.51)	2.94 (3.57)	4.23 (2.69)	4.15 (3.48)	3.88 (2.36)	4.55 (4.74)
TBES	29.04 (7.81)	31.91 (8.89)	20.37 (9.49)	24.35 (10.55)	22.82 (9.63)	23.87 (11.01)	20.50 (10.99)	22.39 (11.54)	20.25 (9.78)	24.18 (11.74)
Exposure Therapy Intentionsa	—	—	3.09 (0.47)	2.89 (0.74)	—	—	—	—	—	—
BAQ	3.12 (0.89)	2.75 (0.85)	3.89 (0.32)	3.80 (0.80)	—	—	—	—	3.83 (0.53)	3.68 (0.79)
FSQ	25.18 (9.71)	34.00 (17.37)	21.18 (5.91)	22.65 (7.94)	—	—	—	—	—	—
Role Play Competence	—	—	3.00 (1.61)	3.75 (1.84)	—	—	—	—	4.38 (1.60)	3.80 (1.23)

^*p <.05

^**p <.01;

^aIntentions questionnaire at post-training asks about likelihood of using exposure; at follow up, instead asks reasons they have not used exposure (if applicable)

Notes: ETCUS is the sum of ETCUS items divided by the number of anxious clients; % Exposure Use is the percentage of anxious clients with whom exposure was used; Total Exposure is the raw sum of number of imaginal, in-vivo, and interoceptive exposures used

Abbreviations: BAQ: Behavioral Anticipation and Confidence Questionnaire; ETCUS: Exposure Therapy Clinical Use Survey; FSQ: Fear of Spiders Questionnaire; TBES: Therapist Beliefs about Exposure Scale

Qualitative interviews

Participants in the ET condition were contacted for qualitative interviews until thematic saturation of interviews was reached (n=12 interviews). Interviews were conducted and audio recorded via Webex after the completion of all other study activities.

Measures

Descriptive Characteristics

Demographics and Background Questionnaire.

The demographics questionnaire assessed participants’ age, gender, race, and education. It also assessed work setting, theoretical orientation, years of experience, and level of familiarity with and understanding of exposure therapy.

Feasibility

Consultation call attendance and involvement.

Consultation call attendance was calculated as a percentage of total calls attended. Each participant’s involvement during consultation was rated on a scale ranging from 0 (Uninvolved) to 6 (Extensively Involved) by an undergraduate research assistant (Edmunds, Kendall, et al., 2013).

Acceptability

Post-Training TCU Workshop Evaluation (WEVAL: Bartholomew, Joe, Rowan-Szal, & Simpson, 2007).

The WEVAL is a 26-item measure that assesses participants’ reactions to training materials and procedures, factors that might influence their likelihood of adopting the procedures taught in training, and intentions to use the materials. The measure has demonstrated internal consistency (Cronbach’s α: .72–89). It includes the following subscales (Cronbach’s α value for each subscale in the current sample is in parentheses): Quality and Utilization (.79), Resources and Skills (.79), Training (.59), Support and Commitment (.76), and Barriers Expected (.82). Responses were rated on a 5-point Likert scale ranging from 1 (Disagree Strongly) to 5 (Agree Strongly) and yielded total scores of 10–50.

Qualitative interviews.

Qualitative interviews assessed the extent to which the training was acceptable and satisfactory in terms of content and training activities. Interviews also included questions about perceptions of exposures and the extent to which participants applied what was learned in the training.

FSQ (Szymanski & O’Donohue, 1995a).

The FSQ is an 18-item self-report questionnaire that assesses spider phobias. Responses are made on a 7-point Likert scale ranging from 1 (Strongly Disagree) to 7 (Strongly Agree). It has demonstrated internal consistency (Cronbach’s α=.92), split-half reliability (.89), and test-retest reliability (.63; O’Donohue & Szymanski, 1993; .97; Szymanski & O’Donohue, 1995b). The FSQ was used in the present study to identify low-, medium-, and high-fear groups for participants in the ET condition. Cronbach’s α in the current sample was .95 at pre and .91 at post.

Inferential Analyses

Knowledge Test (Harned, Dimeff, Woodcock, & Contreras, 2013).

The knowledge test for the present study (10 items) was adapted from Harned and colleagues’ 50-question version to include a subset of questions determined to be most relevant to the current training. Questions were selected through a consensus process conducted by the workshop co-leaders; they independently reviewed the original measure and identified items that related most closely to the workshop content for the current study. Then, they selected the questions that best represented the topics emphasized in the training.

Therapist Beliefs about Exposure Scale (TBES: Deacon et al., 2013).

The TBES is a 21-item self-report measure that assesses therapists’ negative beliefs about exposure therapy. Agreement with each item is rated on a 5-point scale ranging from 0 (Disagree Strongly) to 4 (Agree Strongly). The TBES has evidence for internal consistency (α=.90–.96) and test-retest reliability (r=.89). Higher scores indicate more negative beliefs about exposure, and scores have been shown to decrease after a didactic workshop on exposure, suggesting that it is sensitive to change (Deacon et al., 2013). Cronbach’s α in the current sample ranged from .65–.93 across the four time points.

Exposure Therapy Intentions (Kaye, 2018).

This measure asks how likely participants are to use each type of exposure (imaginal, in vivo, interoceptive) on a 5-point Likert scale ranging from 0 (Not at All Likely) to 4 (Very Likely). The follow-up version of the measure asks participants to indicate why they have not used exposure if they have not used it during the past month.

Behavioral Anticipation and Confidence Questionnaire (Harned et al., 2013).

This 31-item measure of self-efficacy assesses participants’ confidence in their ability to use exposure therapy after training. This measure has demonstrated internal consistency (Cronbach’s α = .95–.99; (Harned et al., 2014; Harned, Dimeff, Woodcock, & Skutch, 2011)) and predictive validity for self-reported exposure use (Harned et al., 2013). Items begin with “I feel confident in my ability to” and then list specific skills and procedures related to exposures. Responses are rated on a 5-point Likert scale ranging from 1 (Not Confident) to 5 (Very Confident). Scores are represented as the mean score of all items. Cronbach’s α in the current sample was .97.

Exposure Therapy Clinical Use Survey (ETCUS; Harned et al., 2014).

The ETCUS measures therapist-reported use of exposure over the past month, including the frequency of using nine exposure therapy procedures (e.g., hierarchy, rationale). Exposure therapy procedure items were summed and divided by the number of clients treated with exposures, consistent with Harned and colleagues (2014). Frequency of exposure use was assessed in two additional ways: (a) number of times the participant reported using any type of exposure divided by the number of anxious clients treated; and (b) sum of the number of in-vivo, imaginal, and interoceptive exposures used. In Harned and colleagues’ (2014) study, Cronbach’s α was .85–.90. In the current study, Cronbach’s α ranged from .75–.91 across the four time points.

Competence.

Behavioral role plays were examined for competence using a single-item competence observational scoring guide to capture how skillfully and responsively the therapist delivered exposure therapy (consistent with definitions of competence; Perepletchikova, Treat, & Kazdin, 2007). Measure development took place as part of a study of fidelity measurement (Beidas et al., 2016) in collaboration with experts in treatment integrity measurement. Preliminary psychometric analysis of this measure indicated convergent validity with therapist knowledge (Phan, Becker-Haimes, & Beidas, 2019). Item scoring mirrors that of adherence items on the Therapy Process Observational Coding System for Child Psychotherapy-Revised Strategies Scale (McLeod et al., 2015) and includes a 7-point scale ranging from 1 (Very Poor) to 7 (Excellent). Role plays were audiotaped for coding by advanced graduate students (ICCs > .70).

Data Management and Analysis

Qualitative and mixed method analyses

Qualitative analyses were conducted in NVivo. Qualitative interviews were audio-recorded and transcribed by research assistants. Interviews were analyzed in NVivo using an integrated theory approach informed by the Consolidated Framework for Implementation Research (CFIR; Damschroder et al., 2009). Results reported in the current study are limited to themes that related to acceptability, which were generated using an integrated approach (Bradley, Curry, & Devers, 2007). Two advanced doctoral students collaboratively reviewed three transcripts to develop codes and refine code definitions. Then, both coders applied the codes by separately coding three transcripts (kappa=0.79). Disagreements were resolved through discussion, at which point the codebook was refined and finalized (available on request). The revised codes were applied to all interviews, which were coded independently by both coders and then discussed to determine final codes by consensus (across all codes: kappa = 0.78; “substantial” agreement; DeSantis & Ugarriza, 2000; Hill et al., 2005; Viera & Garrett, 2005). Codes were summarized in NVivo and examined for patterns related to acceptability of the workshop training approach. Finally, mixed methods examined how acceptability of the ET varied by fear of spiders and intentions to use exposure. Mixed methods analyses were conducted using a convergent design in which quantitative and qualitative data were collected and analyzed separately (Creswell, 2015). Using NVivo, qualitative data were transformed into counts of the number of times each code appeared in the quantitative categories. This allowed for examination of whether there were differential perceptions of acceptability for therapists with low and high fear of spiders (using a median split) and low and high intentions to use exposure (using a median split).

Quantitative analyses

Quantitative analyses were conducted in SPSS. Mean imputation was used for measures that had subscales missing up to 20% of items. Participants missing more than 20% of items did not receive scores on those measures¹ Paired sample t-tests assessed pre-to-post training changes in knowledge, attitudes, self-efficacy, and fear of spiders, as well as post-training to post-consultation role play competence. Independent samples t-tests compared the two conditions on all measures administered at pre- and/or post-training with the exception of the two measures also administered at 1-month follow-up (Exposure Use, TBES). For these measures, repeated measures mixed-factorial ANOVAs tested the main effects of training condition and time, as well as the interaction between training condition and time, controlling for degree (i.e., the only variable that significantly differed between conditions). Bonferroni corrections were not used for t-tests because of the exploratory nature of this study (Rubin, 2017). Given the reduced sample sizes at 2-month (n=24) and 3-month follow-up (n=19), the measures were not examined with inferential statistics but were reported descriptively.

Results

Aim 1: Feasibility

As shown in the CONSORT diagram (Figure 1), 28 (50%) of participants who initially expressed interest in the study were enrolled. Two participants dropped out (one due to insufficient time to attend calls and one due to lack of anxious clients) and one was lost to follow-up after training. All except one of the remaining participants continued some engagement in study activities through 3-month follow-up. Importantly, no one declined to participate because of lack of willingness to be randomized to the ET condition. Although the three participants who dropped from the study were in the ET condition, the rate of study dropout did not significantly differ between conditions, and the participants who dropped reported reasons unrelated to study condition.

Study retention was calculated as the percentage of participants who completed at least one measure at each time point. Pre- and post-training measures were completed by 100% of participants (N=28). At follow-up, 96.4% (n=27) of participants completed at least one 1-month follow-up measure, 85.7% (n=24) completed at least one 2-month follow-up measure, and 67.9% (n=19) completed at least one 3-month follow-up measure. There were no significant condition differences in the number of measures completed at each time point.

The feasibility of requiring attendance on at least nine consultation calls was examined. The mean number of calls attended was 8.14 (SD=3.45), and 60.7% (n=17) of participants met the required number of calls. The modal number of calls attended was 10 (n=6 participants). Most participants (92.9%; n=26) attended at least one call. Excluding those who dropped out of the study, there was no significant condition difference in the number of consultation calls attended or in the percentage of people who attended nine calls. For participants who attended calls, involvement was moderate and did not differ significantly between the TAU (M=3.29, SD= 0.65) and ET (M=2.95, SD=0.55) conditions (p=.271 Cohen’s d=0.46).

Aim 2: Acceptability

Acceptability was assessed quantitatively by comparing mean scores on the training satisfaction measure (WEVAL) across conditions. Independent t-tests indicated that there were no condition differences in acceptability of the training. Scores on the WEVAL range from 10–50, with higher scores indicating stronger agreement and lower scores indicating stronger disagreement. The Training subscale measured the quality of training and ability of trainers to train others (mean=39.64, SD=5.47). The Quality and Utilization subscale assessed the extent to which training was well done, useful, and relevant (mean= 44.55, SD=4.41). The Resources and Skills subscale asked about the participants’ organizations’ ability to implement what was learned in the training (mean=37.04, SD=6.52). The Support and Commitment subscale measured the degree to which other staff members in their work setting were likely to support the use of what was learned in the training (mean=40.57, SD=5.30). Finally, the Barriers Expected subscale assessed the reasons why participants may not use what they learned in training (mean=22.28, SD=5.63), and indicated that participants disagreed that there would be substantial barriers to using exposure.

Qualitative interviews.

Qualitative interviews were conducted with a subset (n=12) of participants in the ET condition given that this condition was the focus of the feasibility and acceptability aims. Representative quotes for the qualitative codes are presented in Table 3. A summary of responses for each of the qualitative codes is provided below.

Table 3.

Qualitative Interview Codes and Quotes

Code	Example Quotes
Liked	“I think experiencing it, that it wasn’t as scary, and seeing some examples of it. Seeing that it fits with my style of how I do things, and before it felt very harsh, but that’s not how anyone presented it, and it was presented matter of factly about the anxiety cycle and it made a lot of sense.”
	“I think [the consultation calls were] really helpful with being able to get immediate feedback on any issues I had while I was implementing it, because certainly, it’s one thing to learn it there but it’s another thing to be using it consistently in practice.”
	“I highly enjoyed it, I enjoyed it more than I thought I would. It was very engaging, there was plenty of background material and research to support the theory of exposure therapy for anxiety disorders and debunking the myths that are associated with exposure therapy. I also enjoyed the fact that the participants that were present got to participate in their own level of exposures. For me, it wasn’t a stimulus that was necessarily a fear evoking response, but I understand the value of putting yourself in the client’s shoes and seeing how it would make them feel.”
	“The fact that it was a group was engaging, and I liked how you got a piece of candy if you got a question right because that encouraged participation. Having us split up into the real-life exposures and doing the roleplays with the fear hierarchies, since I do find value in practicing, and I know it’s uncomfortable for a lot of therapists including myself, but it is helpful because I don’t learn by just reading about it, I learn by doing it with trial and error and having one of your team members give feedback. I found more value in the practice.”
Disliked	“During the weekly calls, to be perfectly honest, I was somewhat disengaged during most of them because I found that there were people in the group or the call who were newer to the field. And the questions they were asking or the things they were working on was much more basic than where my skill level is at now.”
	“Also, there was such a wide variety of what people did in your group that sometimes I felt like the people who were there to learn were pulling the topic away from what you were trying to teach to gear it more towards their own particular situation, and I got a little frustrated with that, because I knew that we had only the one day.”
	“I don’t think anybody in my group was afraid of spiders, so I think that it wasn’t valuable, at least that part of it wasn’t valuable to me until I watched another group that had a fear of spiders.”
Suggestions	“So, for me, I would have liked it to have been a two-day training, but I was able to kind of pull it all together and take it away with me and make it usable. I don’t know if somebody who was not as familiar or who was really brand new to it would have felt as comfortable implementing it.”
	“I think it would have been more useful to be with someone who was afraid of spiders. I understood the activity better when we joined another group.”
	“Maybe because [the consultation calls] were a phone call and not a video call. I felt like maybe, a video call or something in person would have been better, because I would have been forced to sit there and focus, but it was generally helpful.”

Liked.

Participants generally had positive perceptions of the training. They reported that the handouts were helpful, as was the focus on explaining the theory underlying exposure and how to relay that rationale to clients. In addition, multiple participants stated that it was helpful that the training directly addressed their resistance to using exposures and discussed how exposures could be integrated into what they are already doing with clients. There was positive feedback about strategies that were employed to make the training engaging, including videos, role plays, structured activities, games, and “hands-on” practice with spiders. Participants reported that they liked hearing about other cases during consultation calls and that this expanded their training beyond the examples provided at the workshop. They also described the calls as being helpful to address questions about their own cases and to apply what was learned in training. Finally, participants said that they appreciated having a weekend option for training, as that made it more feasible to attend.

Disliked.

One frequently endorsed disliked topic was frustration with the logistics of the consultation calls, including scheduling and accessing the technology required for the calls. Some participants noted that they disliked aspects of the content and process of the calls. For example, some people reported feeling frustrated that other participants on their call were at a different level than they were, which affected the pacing and types of questions being asked. Others reported having difficulty focusing and being engaged over the phone. Regarding the training more broadly, there were logistical comments about frustrations with parking and the layout of the room. Many people also said that there was not enough time to cover all of the content they would have liked to review (e.g., interoceptive exposures, theory). Some also said that they wished there was more time for experiential practice. Finally, multiple therapists reported that spiders were not a fearful stimulus for them, so they did not get the full benefit of the experiential portion of the day. Others said that they wished they could have used a stimulus that was more relevant to their clients (e.g., social anxiety exposures).

Suggestions.

Participants had several suggestions to enhance the experiential practice, such as having a wider array of stimuli available and putting participants in “mixed fear-level” groups so that everyone worked with someone during the training who was engaging in a “real” exposure. The most frequently mentioned suggestion was that participants wished the training was two days instead of one. Many therapists suggested booster trainings to serve as a refresher and/or to cover advanced topics in more depth (e.g., working with children, fears other than phobias, and adapting exposure to different practice settings). Participants also suggested the inclusion of more video clips during the training. Finally, suggestions regarding consultation calls included having smaller groups on the calls, spacing the calls out over a longer period of time, and meeting in person or over video rather than by phone.

Mixed method analyses.

Mixed methods were used to merge the quantitative and qualitative findings. Specifically, we examined how two therapist-level characteristics, fear of spiders and intentions to use exposure, corresponded to therapists’ perceptions of exposure and acceptability of the ET. Statements included what therapists disliked about training (N=24 statements), what they liked about training (N=99 statements), and suggestions for training (N=65 statements). Participants with high fear of spiders made fewer statements about what they disliked about the training (n=8 statements) than people with low levels of fear (n=16 statements). People with high and low fear made a similar number of statements about what they liked about the training (low fear: n=51 statements; high fear: n=48 statements). Notably, participants who were more fearful of spiders frequently mentioned the power of better understanding a client’s perspective on exposures and being able to tell clients about their own experience with exposures. Participants who were less fearful of spiders reported that it was still helpful to practice being in the therapist role even if they were not fearful of spiders themselves. Several less fearful participants noted that the most helpful portion of the day was when they had the chance to observe a group of participants who were completing “real” exposures.

Participants with higher intentions to use exposure made more statements on all acceptability codes (liked: n=62 statements; disliked: n=21 statements; suggestions: n=45 statements) than those with low intentions (liked: n=37 statements; disliked: n=3 statements; suggestions: n=20 statements). There were no notable differences in themes related to what participants disliked about the training, but participants with higher intentions tended to provide more detailed and specific comments about what they liked. In terms of suggestions, several participants with lower intentions mentioned a desire to learn more about tailoring exposures to different settings. Participants with higher intentions almost all expressed a desire for more training beyond the one-day workshop.

Aim 3: Inferential analyses

Pre- to post-training comparisons.

Paired samples t-tests examined changes in measures administered at pre- and post-training. Knowledge (d=−1.57), attitudes toward exposure (d=1.04), self-efficacy (d=−1.16), and fear of spiders (d=0.74) significantly improved from pre-to post-training, collapsing across conditions (all ps < .01). Role play competence also was assessed at post-training (M=3.72, SD=1.53) and post-consultation (M=4.06, SD=1.39), and did not significantly improve (p=.430; d=−0.19).

Condition comparisons.

Means and standard deviations for all measures at all time points are displayed in Table 2. Initial analyses compared conditions on pre-training measures as a randomization check; there were no significant differences between conditions on these measures, suggesting that randomization was successful. At post-training, there were also no significant differences between conditions on any of the measures administered, including knowledge (d=0.36), attitudes toward exposure (d=−0.39), self-efficacy (d=0.13), fear of spiders (d=−0.20), intentions (d=0.28), and competence (d=−0.43).

Repeated measures ANOVA examined attitudes toward exposure and exposure use at pre-training, post-training, and 1-month follow up. Specifically, changes in attitudes toward exposure (TBES) were examined using a repeated-measures, mixed factorial ANOVA with a Time (pre-training, post-training, 1-month follow up) × Condition (TAU versus Experiential) design. Neither Condition nor the interaction was significant, but there was a significant main effect of Time, F(1, 23)=12.46, p < .001, Cohen’s f=0.68. These analyses also were conducted using repeated-measures, mixed factorial ANOVAs to assess exposure use with a Time (pre-training, 1-month follow-up) × Condition (TAU versus ET) design. There was a significant main effect of Time for both the ETCUS sum, F(1, 22)=4.62, p=.043, Cohen’s f=0.39, and the percentage of exposure use, F(1,21)=7.72, p=.011, Cohen’s f=0.54. However, neither had significant main effects of Condition or interactions. When examining the raw total of exposures used, there was a significant main effect of Time, F(1, 23)=13.01, p=.001, Cohen’s f=0.69, and a significant Time × Condition interaction, F(1, 23)=7.13, p=.014, Cohen’s f=0.50. This interaction indicates that there was a significantly greater change in exposure use over time for the ET condition than for the TAU condition (Figure 2). The most frequently endorsed reasons for not using exposures in the last month included not having clients with primary treatment goals related to anxiety (n=7), current clients with anxiety disorders not seeming like a good fit for exposure (n=5), and inadequate training/supervision to conduct exposure (n=2).

Figure 2.

Exposure Use over Time

Discussion

Although exposures are recognized as a critical element of EBIs for anxiety, few efforts to train therapists have resulted in improvements in therapists’ delivery of exposure (Farrell et al., 2016; Harned et al., 2014; Harned et al., 2011) and no studies have identified specific educational implementation strategies to enhance clinician learning. This study evaluated the feasibility and acceptability of a novel experiential approach (ET) to train therapists in exposure therapy that was designed to (a) parallel behavior change strategies used with clients and (b) address barriers related to the translation of knowledge into practice. In addition to demonstrating feasibility and acceptability, ET resulted in a significantly greater increase in self-reported exposure use (number of exposures used) at 1-month follow-up compared to the TAU condition. As hypothesized, both conditions demonstrated improvements in most domains, including knowledge, competence, and self-efficacy. Contrary to expectations, there were no significant differences between conditions in beliefs about exposure or intentions to use exposure.

This study was unique in its application of experiential training. Other studies have referred to incorporating elements of experiential methods into training (e.g., Farrell et al., 2016; Hadjistavropoulos, Thompson, Klein, & Austin, 2012; Webster-Stratton, Reid, & Marsenich, 2014), but these elements were either very brief or required therapists to independently complete them outside of the training setting. In contrast, the experiential approach in the current study incorporated practice that was prolonged and designed to be consistent with a client’s experience of the intervention (exposure). Use of the one-session treatment format (Davis et al., 2019) allowed for experiential practice across all phases of exposure.

Given that this was a novel training approach, one of the primary aims of this study was to assess the feasibility of ET and the study protocol. Study retention and completion of study activities was relatively high, particularly at 1-month and 2-month follow-up when the vast majority of participants completed assessment measures. However, only two thirds of participants completed measures at 3-month follow-up. It is also worth noting that nearly 40% of participants did not attend the required number of consultation calls. On one hand, this suggests that the number of calls required was not feasible for many participants to attend. On the other hand, participants were not paid for attending consultation calls and they still attended a mean of approximately eight calls over a 3-month period. Although this may indicate that nine calls over a 3-month period is not feasible for many therapists, it also highlights the fact that participants were willing and able to attend close to the targeted number of consultation calls without being compensated. This has implications for the design of future studies, especially those that are considering whether providing compensation for consultation is sustainable.

Another study aim was to examine the acceptability of ET using mixed methods, which allowed for a thorough assessment of training acceptability and identification of ways to improve training. Clients who were more fearful of spiders reported that it was valuable to experience what it was like to be a client completing exposures. Participants who were less fearful noted that it was most helpful when they got to observe more fearful participants completing exposures. As noted above, participants were initially paired with others who had similar levels of anxiety about spiders. However, it may be helpful to allow for some practice across fear levels in future work. In addition, less fearful participants requested more fearful stimuli in future trainings. It may be helpful for other types of stimuli to be available (e.g., snakes, social anxiety exposures) to increase the likelihood that all attendees could personally experience an exposure. This notion, combined with comments from participants about the value of observing their peers complete exposures, also raises the question of how important it is for participants to experience exposures themselves, or if practicing using exposure with another person who is actually experiencing anxiety in the moment would provide sufficient opportunity for experiential learning.

Notably, while exploratory, there were significant post-training improvements in many of the targeted domains, including those that have not been successfully addressed by most previous studies of therapist training (e.g., exposure use; Frank et al., 2020). However, exposure use was still relatively low after training, with a mean of approximately three exposures used per month for the ET condition. One reason for this relatively low number might be that therapists have difficulty identifying clients for whom exposure is appropriate; for instance, several potential participants were deemed ineligible and one participant dropped out due to not having any anxious clients, which is unlikely given the high base rate of anxiety. Encouragingly, 83.3% (n=15) of all participants who completed post-consultation role plays achieved competence scores in the “acceptable” range or higher. On the other hand, only 33.3% (n=6) of participants at post-consultation achieved scores in the “good” range or higher. Because therapists in CMH settings are particularly likely to see more severe cases, they may need to achieve higher (than “acceptable”) competence scores in order to feel effective in implementing exposures. Thus, while the percentage of participants achieving acceptable competence scores is encouraging, it is possible that exposure use remains relatively low because higher levels of competence are needed for therapists to consistently implement exposures with severe and complex cases. Evidence suggests that variability in organizational support across settings may make it difficult for some therapists to implement exposures (Becker-Haimes et al., 2020). Therapists working in settings with higher clinical severity (e.g., CMH) may require particularly high levels of implementation support. This highlights the fact that educational implementation strategies need to be optimized and combined with strategies that consider other determinants, such as client and organizational variables (Powell, Proctor, & Glass, 2014). Future efforts to increase implementation will need to target multi-level barriers such as these.

Given its promise as a novel educational implementation strategy, the ET approach expands upon best practices for therapist training; however, these findings should be considered in light of limitations. Inferential analyses were underpowered and may have resulted in Type II errors. This may explain the lack of statistically significant differences between conditions on variables that had substantial differences in means (e.g., post-training fear of spiders). However, because of the exploratory nature of this study, a greater emphasis should be placed on interpreting qualitative analyses and descriptive statistics rather than inferential analyses. In addition, there were limitations in the timing and type of measures. Measures of exposure use were self-report, which may be inflated (Nakamura et al., 2014), and there was no measure of client outcomes. In addition, competence was measured by role plays rather than through client sessions, which may not be representative of competence in actual clinical practice. Finally, some measures had lower internal consistency (Cronbach’s α) than would be expected, which suggests a need to interpret findings using these measures with some caution.

In summary, the experiential approach to training therapists in exposure therapy has support for its feasibility and acceptability. There is preliminary evidence that ET may increase the use of exposure above and beyond TAU, but this requires examination with a larger sample in future work. In addition, ET was described as being most useful for those who experienced fear themselves when practicing with exposure stimuli. This provides additional evidence for acceptability and suggests that it may be useful to provide a wider array of fearful stimuli to ensure that everyone benefits from the experiential practice. Finally, as hypothesized by many implementation science frameworks, it continues to be important that efforts to train therapists address barriers beyond therapist factors, including client- and organization-level factors. Overall, results from this study suggest that experiential approaches can effectively be incorporated into efforts to train therapists in exposure therapy.

Highlights.

• Therapists received training-as-usual (TAU) or experiential training in exposure

• Therapists receiving experiential training underwent exposures themselves

• Experiential training was found to be feasible and acceptable

• Experiential training led to greater use of exposures than TAU