Emily Jones and Katharina Manassis
Outcomes for cognitive behavioral therapy (CBT) and other evidence-based psychotherapies have been evaluated extensively, but the research literature on methods of training in these therapies and on disseminating them is relatively sparse. Recent changes in the mandates of governments, mental health agencies, and research associations have, however, sharpened the focus on knowledge translation and evidence-based practices. Therefore research has started to examine what constitutes effective training for CBT and how to best help trainees implement CBT in real-world conditions of practice when working with clients of various ages who suffer from internalizing disorders. This chapter examines training methods generally, then more specifically in relation to the treatment of anxiety disorders, depression, children and adolescents. It concludes with a review of obstacles to effective training.
Effective strategies for presenting training material have been derived from learning models. Three different but complementary models are reviewed and then linked to specific training methods. Beidas and Kendall (2010) seek to understand how training models affect the translation of evidence-based practices (EBP) through a systems-contextual perspective. Theirs is considered a broad, holistic approach that emphasizes that an individual works within a system and thus quality of training, organizational supports, therapist variables, and client variables interact to effectively translate EBP. Provisional evidence from Beidas and Kendall’s (2010) literature review suggests that, when all levels of the systems-contextual model are addressed, therapists reach proficiency levels in competence, adherence, and skill that facilitate effective client change.
Bennett-Levy (2006) presents a cognitive model for training therapists that emphasizes changes in therapists’ information processing and reflection. In Bennett-Levy’s model there are three principal systems at work: declarative, procedural, and reflective (DPR). The declarative system consists of factual information typically learned through didactic teaching, observed learning, supervision, and reading assignments. The procedural system consists of applied skills; and the reflective system consists of metacognitive skills that include observations, interpretations, and evaluations of one’s self in the past, present, and future. Bennett-Levy demonstrates how this model accounts for a new therapist’s progressive learning until he or she becomes an expert. The new therapist begins by relying on the declarative system using didactic learning, modeling, practice, and feedback but further develops clinical expertise through the procedural system. Finally, he or she moves to the stage of clinician, as a teacher or expert, through the reflective system. The transition to becoming an expert happens when the therapist goes from reflection-on-action (retrospective review of therapeutic interventions) to reflection-in-action (reflection that occurs during therapy sessions).
Other theoretical reviews – such as Milne, Aylott, Fitzpatrick, and Ellis (2008) and Rakovshik and McManus (2010) – emphasize the relationship between trainee changes and methods of instruction. Here it is found that trainees in CBT need supervision that provides not only didactic teaching but also experiential learning methods such as role-play, cotherapy, and modeling. Further, Rakovshik and McManus (2010) provide evidence to suggest that, if supervision is removed prior to consolidation, then the learned skills can deteriorate over time.
On the basis of these learning models, many types of training have emerged. Bennett-Levy, McManus, Westling, and Fennell (2009) examine which training methods are most effective, testing the didactic versus experiential distinction described above. Although their study relies on therapists’ perceptions, it does allow for predictions to be made about the effect of training methods on CBT skills and competence. In this study, 120 therapists were surveyed regarding the effectiveness of 6 popular training methods and 11 different CBT skills. Bennett-Levy and colleagues found that different methods of learning were perceived to be effective across different CBT skills. Passive training strategies such as reading manuals and books, attending lectures and workshops, and completing non-interactive online training increased declarative knowledge and some conceptual and technical skills. Active training strategies such as role-play, reflective practice, self-experiential work, graded training and supervision were found to increase the therapist–client exchange and the level of technical, interpersonal, procedural, and reflective skills.
Herschell, Kolko, Baumann, and Davis (2010) reviewed 55 different studies on training therapists in psychosocial treatments. Their review found that passive training – such as reading and workshops – develops knowledge but is not sufficient for achieving competence, as skills learnt through these methods do not reliably persist over time. The addition of skill feedback and short-term consultation or supervision after a workshop was shown to increase both knowledge and skills. In particular, clinical supervision and consultation with field experts consistently produced superior training outcomes, despite difficulties that may arise with cost and availability. Overall, Herschell and colleagues concluded that multi-component training programs are superior to other (single-component) training methods.
With the advent of online training, new hybrid training methods that incorporate both passive and active learning methods have become popular (Dimeff et al. 2009; Granpeesheh et al. 2010; Sholomskas and Carroll 2006; Sholomskas et al. 2005; Weingardt 2004; Weingardt, Cucciare, Bellotti, and Lai 2009). For example, Dimeff et al. (2009) investigated the efficacy of three different training methods for dialectical behavior therapy (DBT): (i) a 20-hour interactive online training (OLT) session; (ii) a two-day instructor-led training (ILT) session; and (iii) reading two treatment manuals. These researchers made sure the OLT was interactive and sophisticated, as they felt that variable results in other studies might be due to OLT’s being presented like a book on a screen. Specifically, the OLT included audio and visual presentation as well as a fictional DBT skills group for participants to observe and learn from, expert insights, practical exercises, and knowledge quizzes. Participants were evaluated before and after training, and also in a 90-day follow-up, regarding their knowledge of DBT, satisfaction with training, referencing of materials, and clinical performance in role-play. Dimeff and colleagues found that OLT significantly out-performed both the ILT method and the method of training through manuals.
The duration of training is, potentially, a factor to be considered when investigating effective training methods. However, conclusions on duration are difficult to draw, as the majority of studies do not address this topic. Those studies that do mention training duration only give broad temporal descriptions and vary widely in approach, which makes them difficult to compare. In Rakovshik and McManus (2010), 37 different training programs are compared; but researchers point to a lack of agreement on the terminology that describes the training. For example, they argue that a comparison between a “workshop” that was didactic and interactive and contained role-play and case discussions and a “workshop” that was strictly didactic would be imprecise. Finally, specifying the nature of competence and its measurement presents its own challenges (see below). Therefore further research is needed on measuring competence and on the optimal duration of training. Nevertheless, most training studies reviewed above support multi-component programs that include active learning methods as well as didactic teaching.
Training models in cognitive behavioral therapy (CBT) vary (see previous section), and competencies achieved by trainees are not clearly defined in all studies. Roth and Pilling (2008), however, developed a systematic model of the competencies needed by a therapist in order to treat anxiety and depression in adult populations; to do this, they used a Delphi technique that draws on sharing and refining contributions from individual field experts so as to arrive at a consensus. Roth and Pilling’s model describes five increasingly specific domains of competence: general psychotherapeutic competencies; basic CBT competencies (related to the structure and content of CBT); specific CBT techniques; problem-specific competencies; and competencies that allow adaptation to individual client needs while maintaining treatment fidelity (these are termed “meta-competencies”). The authors highlight that achieving meta-competencies generally requires the highest CBT expertise and the greatest amount of experience, which makes these competencies the most challenging ones to impart to trainees.
Studies of anxious and depressed clients generally emphasize treatment methods and client outcomes (chiefly symptomatic improvement, but also client perceptions of therapy in some studies, as for instance in Hepner, Paddock, Zhou, and Watkins 2011). The therapist’s competence is not consistently measured and evaluated on that basis, and, when measured, this is usually a side issue, subordinated to client outcomes rather than representing a separate goal. Training methods, in the most relevant studies, reflect the multi-component approach recommended more broadly in the training literature – an approach consisting of a blend of didactic teaching, interactive learning, and supervision. Computer-assisted training that incorporates didactic seminars with skills practice, role-play, and videos of “bad” therapy techniques, as well as training via tele-health, are becoming increasingly popular and participants show positive gains in competence and adherence (Reese and Gillam, 2001; Rose et al. 2011). Individual supervision, in particular the kind that focuses on the ongoing refinement of the trainee’s skills on the basis of the supervisor’s input, has shown superiority to other training methods in some studies (e.g., Mannix et al. 2006). Further to these training findings, Karlin et al. (2010), in collaboration with post-traumatic stress disorder (PTSD) therapists, found that incorporating collaborative consultation on actual cases in a training program increased therapists’ competence and helped enhance the adoption of therapy into practice. It is, however, more time-consuming than other learning methods. As an alternative, group supervision models have the potential to optimize the use of a supervisor’s time, but they have received limited evaluation (Manassis et al. 2009; Newton and Yardley 2007).
An approach that diverts from the multi-component recommendation is that of training through participation in cotherapy. For example, Hepner and colleagues (2011) trained addiction counselors in depression-focused CBT group by using a co-therapy model where an inexperienced therapist conducted a treatment program together with an experienced therapist prior to implementing the program independently. The authors were able to demonstrate trainees’ ability to lead depression-focused CBT groups with high fidelity to treatment protocols, without training in individual CBT methods first. This cotherapy option may be a viable addition to other recommended multi-component approaches.
When direct evaluation of therapist competence has been incorporated into studies of CBT, the Cognitive Therapy Scale (Dobson, Shaw, and Vallis 1985) has been the most frequently reported measure of competence. In this measure, expert raters evaluate videotapes of CBT sessions by using an 11-item scale. Reliability of ratings increases with the number of raters and the number of videotapes rated. Other measures of therapist competence include the Cognitive Formulation Rating Scale, the Cognitive Therapy Awareness Scale, and the CBT Supervision Checklist (Sudak, Beck, and Wright 2003). Nevertheless, there is a paucity of research into standardized measures of therapist competence.
In the depression and anxiety literature, CBT skills and the ability to structure sessions are the aspects of competence that have been regularly linked to therapeutic change (Simons et al. 2010). The literature is not entirely consistent, however, possibly due to difficulty separating the effects of therapist competence from other client- and therapist-related factors. For example, Trepka, Rees, Shapiro, Hardy, and Barkham (2004) found that the relationship between competence and depression outcome was no longer significant when controlling for therapeutic alliance, and Strunk, Brotman, DeRubeis, and Hollon (2010) found that the relationship between therapist competence and change in depressive symptoms was moderated by factors related to the client, as client anxiety, early onset of depression, and chronic course predicted lower levels of change. Kuyken and Tsivrikos (2009), however, found that therapists’ high competence was associated with improved outcomes to depression regardless of clients’ comorbid diagnoses. Factors related to the therapist, such as positive attitudes toward empirically supported treatments, resulted in better implementation of depression-focused CBT in community settings (Lewis and Simons, 2011), and previous cognitive therapy experience and careful case selection were linked to therapist competence in a sample of 20 postgraduate trainees working with depressed clients (James, Blackburn, Milne, and Reichfelt 2001).
Each anxiety and depressive disorder poses unique challenges to training, and this is due to the nature of the disorders themselves. Trainees need to acquire knowledge and experience related to internalizing psychopathology (see Chapter 8). Within the training context some difficulties may arise from the multiple explanations for a single behavior that may occur in anxiety and depression. For example, inconsistent attendance at appointments may occur for different reasons: depressive clients may be fatigued and feel hopeless about the possibility of improvement; anxious clients may wish to avoid the exposure to anxious situations that is integral to therapy. Helping trainees consider multiple alternatives and explanations is likely crucial for an effective training. Similarly, the psychological treatment of mood disorders is more regularly combined with medication than the psychological treatment of anxiety disorders in youth (Treatment for Adolescents with Depression Study Team, 2004). Familiarity with psychotropic medications and regular communication with the prescribing physician are therefore essential for the successful treatment of depressed clients, but not always for the treatment of anxious clients.
The nature of the skills required to treat anxiety and depression may be challenging for novice therapists. For example, encouraging client exposure to feared situations causes discomfort in many training therapists, and is thus a potential barrier to the dissemination of CBT (this is further described below, in the section on obstacles to competent practices). Harned, Dimeff, Woodcock, and Skutch (2011) investigated a solution to this potential barrier by comparing the uses of an interactive online CBT training program in two variants: with and without motivational interviewing. They found that both groups of training therapists gained knowledge, but the group receiving motivational interviewing developed more positive attitudes toward exposure. This suggested that motivational interviewing may aid dissemination through improved attitude on the part of the therapist.
Competence in CBT with anxious or depressed children and adolescents presupposes an understanding of children’s social, cognitive, and emotional development and an appreciation of the impact of environmental factors on their well-being, as these are more salient in child care than in adult care (see Chapter 5). As outlined in Chapter 1, a more sophisticated understanding of competencies needed for child and adolescent CBT has been published by Sburlati, Schniering, Lyneham, and Rapee (2011). Ideally, a therapist will flexibly and skillfully integrate these various competencies to tailor CBT to the needs of each child or adolescent treated (see Chapter 10 for a detailed discussion). In addition, competence in child CBT generally entails the ability to effectively engage and collaborate with the client’s parents. CBT protocols vary in the extent of parental involvement in therapy, but all are predicated on some parental engagement – at least as much of it as would be enough for parents to bring the child to sessions consistently (see Chapter 6). There is a lack of research into the effectiveness of training programs that target those who work with children and adolescents. Logically, however, successful training programs need to incorporate a broad spectrum of examples, which should not only vary in the way they present the problem but also incorporate the developmental and environmental variations that naturally occur within each age group.
Therapists training in child and adolescent CBT often find themselves in a wider variety of settings than those training to treat adults. CBT with youth is not limited to the health care sector, as schools and social service agencies provide psychotherapy as well. School- and community-based CBT training programs are among the most highly evaluated. ACTION is a gender-sensitive CBT program for depressed youth, originally developed on the basis of studies of depression in early teen girls (Stark et al. 2007). It now contains separate workbooks for each gender (20 sessions), as well as gender-specific parent workbooks (8 sessions). Weisz and colleagues (2009) trained community clinic therapists in this program in 6 hours of teaching followed by weekly supervision, and they compared its effectiveness to that of treatment as usual. Both groups of adolescents improved symptomatically, but CBT-treated youth required fewer sessions and fewer additional services than those receiving treatment as usual. The program is currently being adapted for use by school psychologists (Stark, Arora, and Funk 2011).
Manassis and colleagues (2009) also dealt with community mental health providers in their child-focused training program designed to disseminate CBT for children with anxiety disorders. Training consisted of a 20-session group supervision program that included some didactic teaching, either face to face or by videoconference. Trainees showed increased knowledge of child CBT and increased confidence in their ability to practice child CBT after training. Experienced therapists and therapists working in agencies that used a diagnostic screen at intake – which increased their chances of working with children suitable for CBT through thorough assessment – reported the greatest confidence.
The most extensive investigation into training in child CBT has been undertaken by Barrett, Farrell, Ollendick, and Dadds (2006), in their dissemination of the FRIENDS program for child anxiety. Largely a prevention program, FRIENDS is a group CBT intended for use as a universal intervention for all students in a classroom; it can be conducted by teachers or by qualified mental health staff. A single day of group training is required, and implementation occurs via manuals specific to various age groups. In several randomized trials, high-risk students (i.e., those who score above clinical cut-off for anxiety or depression on standardized measures) showed greater symptomatic improvements after intervention, and some maintenance of the gains – up to 36 months after intervention (Barrett et al. 2006). Moreover, benefits were evident regardless of whether psychologists or teachers led the CBT groups.
In summary, although studies that formally evaluate therapists’ competence are lacking in the child CBT training literature, there is accumulating evidence that community- and school-based practitioners can learn and successfully implement this therapeutic modality.
Knowledge translation of CBT is a multifaceted, complex process. Each level of translation (client-directed, therapist-directed, and organizational) needs to be addressed. Adding to this complexity are many barriers that exist at each level. Barriers that relate to knowledge translation of CBT can include the therapists’ attitudes toward CBT and research-based therapies, the lack of access to training and supervisors, the lack of organizational support for implementation, and the therapists’ level of competence.
Attitudes of therapists toward CBT and research-based therapies in general have been identified as a large obstacle to effective knowledge translation (Nelson & Steele, 2007). Training, however, appears to effectively improve these attitudes. For example, Bennett-Levy and Beedie (2007) show increased positive attitudes and self-perception of competence over a one-year cognitive therapy training course. An empirical approach to training that consisted in inviting trainees to observe and comment upon their training experience has also been linked to improved attitudes toward CBT (Sudak et al. 2003).
Due to the inherent structure of CBT, manuals are considered critical to implementation. In these manuals, instructions for the therapist are carefully outlined for each session. For example, one well-known manual for anxiety-focused child CBT is the Coping Cat Manual and Coping Cat Workbook (Kendall and Hedtke 2006a and 2006b). This program and various site-specific, adapted versions of it have been evaluated in multiple studies and have shown consistent efficacy (e.g., Compton et al. 2004; Kendall 1994; Kendall et al. 1997; Khanna and Kendall 2008; Podell, Mychailyszyn, Edmunds, Puleo, and Kendall 2010; Suveg, Sood, Comer, and Kendall 2009). Although CBT manuals have had great success in research and are clearly helpful in standardizing treatment, they have been criticized by therapists in routine clinical settings for limiting the personalization of treatment and the therapist’s ability to respond to the client’s needs. To address this issue, Kendall, Settipani, and Cummings (2012) explain how manuals should be implemented. They advocate employing flexibility as long as the underlying principles of CBT treatment are maintained – namely principles such as having session goals, using a CBT perspective, remaining action-oriented, and using social learning theory. Specific exercises and activities, however, can be modified and individualized to address individual client needs. These researchers believe that the mark of a competent therapist is being able to use good clinical judgment so as to work flexibly within the framework of an evidence-based treatment manual.
Another approach to increasing flexibility within treatment manuals is the modular approach to CBT. This approach breaks down CBT into basic modules such as self-calming and cognitive reframing (Chorpita, Taylor, Francis, Moffitt, and Austin 2004). While this approach to treatment can be effective, it does require a high level of expertise and clinical judgment to select, organize, and apply the most appropriate modules in a given case (Weisz et al. 2012); consequently the approach may be more appropriate for experienced clinicians and/or may require specialized training.
As mentioned above, therapists’ attitudes to specific techniques such as exposure tasks can be particularly problematic in CBT. Therapists may fear that exposure would be traumatic to the child and detrimental to the therapeutic alliance, and therefore they may be hesitant to guide their clients through what could be fearful situations. These preconceived ideas about exposure are often due to misinformation. Kendall and colleagues (2009) studied the effect of exposure tasks on the therapist–client alliance. Alliance was reported by both the client and the therapist independently after every session, and it was found that exposure had no harmful effect on the therapeutic relationship. Also, Hedtke, Kendal, and Tiwari (2009) studied the effects of exposure tasks in general and the characteristics of exposure tasks used in CBT for children with anxiety and found exposure tasks to be an integral part of treating anxiety disorders effectively. Despite these findings, some therapists fear clients’ unpredictable reactions to exposure, as they may not have been trained to address these challenging reactions. This problem underscores a further barrier to knowledge translation of CBT: access to effective training and supervision.
Lack of access to effective training as therapist and supervision can impede knowledge translation. CBT training methods have been discussed above; they generally support the inclusion of role-play, modeling, reflective practice, and other forms of experiential learning in addition to didactic teaching (Bennett-Levy et al. 2009). Further, programs that provide ongoing supervision with an expert have been found to result in higher levels of therapists’ satisfaction and competence than those that do not (Bennett-Levy et al. 2009; Herschell et al. 2010; Rakovshik and McManus 2010; Sholomskas et al. 2005). Expert supervision is not available in all jurisdictions, however, and it is often costly and time consuming. Group supervision, including supervision via tele-health, has the potential to increase training access (Manassis et al. 2009). Although sophisticated training models that include case conceptualization and adoption of fidelity measures may further increase knowledge translation, such training models require even greater time commitments and cost.
Karlin and colleagues (2010) show how limiting the eligibility for training programs to therapists who have prior experience and high general competence skills in psychotherapy may be worthwhile, if experienced and competent trainees are available. They found increased completion and post-training success rates among trainees who met predefined competence levels before the training. Consistent with these findings, Manassis and colleagues (2009) found that more experienced therapists reported greater CBT training benefits than less experienced therapists.
Organizational barriers to knowledge translation also exist. Agencies are confronted daily with questions regarding the cost-effectiveness of implementing new treatments and of training therapists in the administration of different treatments. An organization’s ability to flexibly support staff attending training, peer supervision opportunities, funding for training, private therapy space, and other requirements for evidence-based training and practice often determines the success of knowledge translation efforts there. Certain organizational mandates can also play a large role in the implementation of new treatments such as CBT. For example, the mandate to see clients on a first come–first serve basis may result in low availability of clients suitable for CBT when a CBT supervision opportunity arises. In summary, being able to reduce the cost of training and treatment without sacrificing their quality is vital if organizations are to implement CBT effectively.
In order to anticipate obstacles to implementing CBT in an organization, it is important to assess potential knowledge translation challenges. Assessment tools such as the Texas Christian University organizational readiness for change (ORC; Lehman, Greener, and Simpson 2002), the Evidence-Based Practice Attitude Scale (EBPAS), and attitudes toward treatment manuals (Bartholomew, Joe, Rowan-Szal, and Simpson 2007) have been developed to identify possible barriers to implementing new programs. These measures have shown positive results, helping leaders identify areas of need, so that they can be proactive in overcoming barriers. Further research in this area is needed, however, particularly regarding measures that are specifically for use within children’s mental health organizations.
Effective knowledge translation must address all levels of mental health care. Training programs should educate therapists and organizations in the principles and evidence base of CBT, provide active learning styles, promote flexibility in treatment, and provide support to both the therapist and the organization. Further, knowledge translation should aid organizations in developing their own support system for the sustainability of CBT practice, whether through peer supervision, train-the-trainer models, or ongoing consultation with experts. Assessment of obstacles to implementing evidence-based care may be beneficial when one takes a proactive approach to knowledge translation.