The efficacy of cognitive behavior therapy (CBT) in children with internalizing disorders has been demonstrated internationally in numerous randomized controlled trials (1). Unfortunately, the availability of child CBT in Ontario is often limited to specialized academic centers and private mental health practitioners, typically psychologists and social workers. Community access to child CBT is limited, especially for families living distant from an academic center or unable to afford private practitioners. Training in CBT received by psychiatrists currently in practice is highly variable (2). Increasing the number of CBT practitioners in publicly funded, community settings is therefore essential. Community mental health professionals (social workers, child and youth counselors, pediatricians, and family doctors with mental health interests) often have little formal training in CBT, but are highly motivated to learn, use, and adapt evidence-based interventions with children. Unfortunately, few studies have examined optimal training models to disseminate CBT skills to these professionals. Moreover, CBT may require adaptation to the diverse clientele of community mental health settings and to the parameters of practice in these settings. In this study, we (three child psychiatrists and one child psychologist from an academic center) sought to disseminate CBT skills by training community mental health professionals and evaluating the effects of that training. We considered several training models in light of existing literature before evaluating a 20-session group supervision seminar with a didactic component. We used a mixed-method design including quantitative and qualitative methods to ensure a comprehensive evaluation.

Models for delivering CBT training to professionals outside academic centers have received limited study, and direct comparisons of models are rare. Workshops/didactic teaching models are less time-consuming than case-based training and therefore often preferred. Unfortunately, their benefits have not been consistently demonstrated. Jensen-Doss and colleagues (3), for example, found that workshop-based training in trauma-focused CBT did not result in changes to the use of this modality after training, although therapists indicated that they knew it was highly effective. By contrast, Hides and colleagues (4) reported that a 2-day training workshop in CBT for youth with concurrent mental health and substance use problems had a positive impact on knowledge, skills, and confidence of health professionals using this modality. We did a 6-month follow-up of community practitioners attending a 3-day intensive training workshop in child CBT (5) and found that most participants felt more knowledgeable about child CBT, but few were confident they could practice it. Case supervision as a form of CBT training has shown consistently positive effects on trainee knowledge and confidence, regardless of trainee discipline or patient population treated (6–9). Most studies are open trials, and length of supervision is variable. In a rare randomized study, Mannix and colleagues (6) discontinued supervision in half of a group of 20 CBT trainees and continued it for 6 months in the other half. CBT skills (evaluated using session audiotapes) and self-reported confidence were higher in the group receiving continuing supervision. Supervision allows for greater dialogue between trainer and trainee than didactic teaching, facilitating the adaptation of CBT techniques to meet the needs of diverse populations. Because it is time consuming (up to 40 sessions) (8), community practitioners may have difficulty obtaining coverage of other professional duties while attending weekly supervision. Agencies may have to weigh the benefits of CBT training against those of having staff see additional clients during the training time. Supervisor unavailability may be an obstacle to training in this model, even in psychiatric residency programs (10). Group supervision appears promising and requires fewer supervisors (7, 8), but has received limited study. Use of new technologies such as web- or video-based training provides additional CBT training options and may be uniquely suited to remote areas. Sholomskas and colleagues (9) compared community-based clinicians providing either review of a CBT manual, the manual plus access to a CBT training web site, or the manual plus a didactic seminar and case-based supervision. The supervised clinicians showed greater CBT skills in structured role-playing than those receiving only the manual. Scores for those receiving web-based training were intermediate, suggesting some promise for this type of training. Rees and Gillam (11) provided a 20-session videoconference CBT course to 12 rural community mental health professionals in Australia and demonstrated increased knowledge of CBT on a test and increased confidence using it. They did not measure CBT skills. It is unclear how to best address barriers to implementation of CBT, though they have been identified in several studies. Brooker et al. (12) identified caseload size, lack of an implementation plan or training strategy, and lack of a supportive team as potential barriers in community mental health settings. Additional barriers have been identified in family practice settings (13).

Training Model in This Study

We developed a 20-week child CBT group supervision seminar with a didactic component, adapted from a program used to train psychiatric residents in CBT since 2001 with consistently positive evaluations (14). Supervisees treated anxious children with “Coping Bear” (15), an extensively evaluated (1, 16, 17) local adaptation of Kendall’s “Coping Cat” (18), an internationally validated child CBT program (1). By involving leaders of community mental health agencies in program development, we hoped to better meet the needs of their staff and ensure ongoing community use of CBT. The group format allowed a small number of supervisors to serve a relatively large group of trainees. The didactic component focused on CBT principles, unique aspects of working with children, selection of appropriate cases, use of CBT manuals in diverse child populations, key aspects of working with parents (based on the book “Keys to Parenting Your Anxious Child,” 19), an introduction to CBT for children with other disorders (particularly depression and obsessive-compulsive disorder), and successful termination of therapy. We discussed adaptations to best serve trainees’ particular clientele and overcome organizational barriers to implementing CBT throughout supervision.

We hypothesized that trainees’ knowledge of child CBT (measured using a multiple-choice test), confidence using child CBT, and desire to do further child CBT would increase with training (key outcomes) and that high training satisfaction would be reported. We also did an exploratory evaluation of trainee and organizational factors predictive of outcome, to allow refinement of our model for future trainees. Qualitative evaluation allowed trainees to comment on the subjective nature of their experience, personal and organizational barriers to participation in training, and ways that training could have been more helpful to them.

After obtaining institutional review board approval and providing study information to children’s mental health agencies, we recruited trainees in partnership with four agencies who were eager to participate: three local and one with multiple sites around the province. We supervised trainees from the latter agency through videoconferencing, using the same model as in face-to-face supervision. Agency leaders nominated trainees, to respect agency autonomy, from among a larger group of volunteers. Trainees had diverse child therapy backgrounds. At least two practitioners had to attend from each site, to ensure ongoing peer support around training and use of child CBT. We offered agencies partial compensation for trainee time spent attending the supervision. All trainees provided informed, written consent. All data were stored in a password-protected computer, with identification of participants by number only. We recruited 24 trainees, and 22 (92%) completed the training and all measures. The trainees were predominantly social workers and child and youth counselors, with a few from other mental health disciplines. There were 19 women and three men, with a mean age of 40.8 years (SD=8.4) and an average of 6.8 years experience doing child therapy (SD=7.1). We supervised trainees in four groups, one per agency, of four to seven participants per group. The trainees met weekly for 1.5 hours per meeting for 20 weeks. Concurrently, each treated one case (child CBT and parent training) for 12 weekly sessions. Because some found suitable cases sooner than others, there was some staggering of treatment start times during the 20 weeks. Participants kept detailed notes of their treatment sessions, but given the need to adapt our CBT protocol to the diverse needs of children in community settings, we did not formally evaluate treatment fidelity.

Evaluation

Based on the learning objectives of the program and the specific, objective CBT content of “Coping Bear” and “Keys to Parenting Your Anxious Child,” we developed a 20-question, multiple-choice test to evaluate practitioners’ gain in knowledge of child CBT from pre- to posttraining. Given the modest sample size, we analyzed results using a paired samples t test. We administered a brief training satisfaction questionnaire using a 5-point Likert scale (1=strongly disagree, 5=strongly agree) after the training. Using one question each, we queried satisfaction with training content, satisfaction with supervisors, perceived knowledge gained, confidence using CBT (all aspects), and desire to do further CBT with clients. We also asked participants about several trainee and organizational factors (e.g., trainee demographics, prior therapy experience, beliefs about therapeutic change, presence of some diagnostic screen at agency intake, agency intake policies) to determine their potential to predict the three key outcomes. We determined bivariate correlations and entered significant correlates in stepwise regression analyses (maximum two variables per analysis, consistent with sample size). To minimize bias, a research technician with no involvement in teaching administered and scored measures. Qualitative evaluation used individual interviews shortly after training for maximal authenticity of response and focus groups 6 months later, when there was a functioning peer support group in place at each agency. Research staff with no involvement in teaching conducted the interviews and focus groups to allow frank discussion and minimize bias. Qualitative data analysis followed procedures consistent with prevailing qualitative approaches (20, 21) to establish key themes.

Quantitative Evaluation

Practitioners’ knowledge of CBT increased significantly from pre- to posttraining (t=−4.374, p=0.000). Practitioners also reported a high degree of satisfaction with supervisors, perceived knowledge gained, confidence using CBT, and desire to do further CBT, based on average scores of at least 4.5 out of 5 on a Likert scale for all these queries (Table 1). Satisfaction with training content was slightly lower and was further examined in qualitative analyses. Scores were higher than the sample mean for the videoconference participants, but not significantly so, likely due to small sample size. Significant correlations for the exploratory analysis of trainee and organizational factors in relation to key outcomes are shown in Table 2. Key outcomes were strongly intercorrelated. Also, trainee age correlated positively with confidence using CBT, and agency use of a diagnostic screen at intake correlated positively with confidence using CBT and desire to do further CBT. No additional factors predicted posttraining knowledge. In a stepwise regression analysis using the two factors correlated with confidence, only trainee age remained significantly predictive (adjusted R²=0.28; standardized β=0.59, p=0.04). Agency use of a diagnostic screen at intake predicted desire to do further CBT with clients (adjusted R²=0.21; standardized β=0.50, p=0.03).

Qualitative Evaluation

Reports of participants’ training experience were highly consistent with questionnaire responses, with 80% reporting a very positive experience and 85% feeling competent providing child CBT. Reports remained consistent at 6-month follow-up. All agencies were continuing to provide child CBT at follow-up in both individual and group formats using appropriate manuals to ensure fidelity, even though training had focused only on individual cases. Suggestions for improvement included provision of additional, up-to-date information on CBT theory and practice, observational opportunities, and more leader-directed discussion. Videoconference participants recommended meeting in person once to build initial rapport between supervisors and trainees. Videoconference leaders reported some difficulty interpreting participants’ nonverbal cues, and participants’ interactions were somewhat formal because sites generally took turns speaking, but leaders did not consider these major teaching obstacles.

Although our study is limited by small sample size and lack of direct observation of CBT skills, brief questionnaires and participants’ evaluation of their experience suggest that this 20-week, group supervision seminar with a didactic component increased practitioners’ knowledge, confidence, and desire to practice child CBT. Peer support groups for CBT formed in each participating agency, and all agencies continued to provide CBT at 6-month follow-up. Videoconferencing did not appear to adversely affect training. Exploratory analyses suggest that older, more experienced therapists may gain more confidence with training than less experienced therapists. Given the brevity of training, it may be more effective in teaching child CBT to mature therapists who already have a substantial background in other child therapy skills. The use of an agency-specific diagnostic screen at intake appeared to contribute to therapists’ desire to do further child CBT, perhaps because such screens may allow therapists to obtain suitable cases more readily. Consistent with this idea, therapists identified difficulty finding suitable cases as a significant barrier to training. Organizations may require additional education about case selection for CBT prior to therapists’ enrollment in training to reduce this difficulty. These findings are consistent with those of James and colleagues (22), who found previous experience and careful patient selection was associated with greater therapeutic competence among adult cognitive therapy trainees. Further identified barriers to training included time needed to participate, lack of funding for doing ongoing therapy (compared with briefer interventions), the therapists’ role in the organization, and the client population. Regarding role and clientele, some therapists reported limited influence in selecting cases they were asked to treat and clients that required considerable adaptation of CBT materials, making training in and implementation of child CBT challenging.

Therapists’ recommendations for improving training included a desire for greater structure and a more didactic, theoretically driven approach. It is possible that while emphasizing active learning and supervision, we neglected these elements in the program. These elements may be particularly important when psychotherapy trainees have little familiarity with a subject area (23). CBT adaptations that were particularly valued by trainees focused on children with cognitive limitations, comorbid conditions, and challenging families (14). Making these improvements and addressing the identified barriers may increase the benefits of future training. The lack of observational measures of trainees’ CBT skills and measures of patient change was clearly a limitation of this study. Such measures should be included in future studies. Studies comparing videoconference and face-to-face training and training in other settings (e.g., schools) and with other trainee populations (e.g., different disciplines, rural versus urban practitioners) are also indicated.

The authors received financial support from Centre of Excellence in Children’s Mental Health at the Children’s Hospital of Eastern Ontario and Canadian Imperial Bank of Commerce Children’s Miracles Fund and the community partners Aisling Discoveries Centre, Bloorview MacMillan Centre, George Hull Centre, and Kinark Child and Family Services. The authors thank Dr. Melanie Barwick for her consultation, and Mss. Lisa Fiksenbaum, Karen Sappleton, and Laurie DeOliveira for data management. At the time of submission, the authors declared no competing interests.