A qualitative examination of trainee perspectives on cognitive behavioural supervision

Abstract

Clinical supervision is the main method by which mental health professionals acquire the competence to deliver safe and effective therapy. The cognitive behavioural supervision (CBS) approach to supervision parallels CBT in structure and form, which may facilitate learning. Although supervision is integral to trainee development, little is known about what CBS interventions trainees consider helpful. Using a qualitative content analysis methodology, we aimed to identify the specific CBS interventions that trainees find most helpful. Eight trainees completing a CBT rotation in an out-patient hospital setting received weekly individual supervision by staff psychiatrists and psychologists. Following each supervision meeting, trainees completed open-ended responses describing what they found most and least helpful. Responses from 127 meetings were coded using a CBS framework. Overall, trainees found many aspects of supervision helpful. The interventions most frequently noted as valuable were teaching, planning, formulating, training/experimenting, and evaluation of their work. When trainees mentioned unhelpful events, insufficient collaboration and a desire for more or less supervision structure were most frequently noted. These results suggest that the perceived helpfulness of supervision may be tied to the use of CBS interventions that provide trainees with concrete skills that facilitate learning. Further suggestions and implications for supervisors are discussed.

Key learning aims

1. (1) To identify the aspects of cognitive behavioural supervision that trainees perceive as most and least helpful for their learning.

2. (2) To integrate trainees’ perspectives with the existing research on supervision satisfaction.

3. (3) To consider limitations, challenges and future directions of cognitive behavioural supervision research.

Introduction

Cognitive behavioural therapy (CBT) is the treatment of choice for many mental health disorders and evidence suggests that it is effective whether treatment is delivered by experienced therapists or trainee therapists under supervision (Forand et al., 2011; Hofmann et al., 2012). Supervision can be defined as ‘the formal provision by senior/qualified health practitioners of an intensive relationship-based education and training that is case-focused, and which supports, directs and guides the supervisees’ (Milne, 2007). Supervision is the main training method by which trainees acquire the competencies to become licensed mental health professionals, and trainees consistently rate clinical supervision as highly influential to their practice (Rakovshik and McManus, 2013; Scott et al., 2011; Watkins and Milne, 2014).

While different models of supervision exist, using cognitive behavioural supervision (CBS) may facilitate skill development for trainees learning about CBT, as it parallels this therapy in structure and form (Beck et al., 2008). In the last decade, the number of guidelines and training manuals related to CBS has grown considerably (e.g. Milne, 2018; Milne and Reiser, 2017; Prasko et al., 2012; Sudak, 2016; Watkins and Milne, 2014), yet the topic remains critically under-studied (Alfonsson et al., 2018; Milne, 2018). Trainee perspectives have been especially neglected, particularly in settings working with pre-licensed trainees (Kelly and Hassett, 2021). Therefore, this paper aims to investigate trainees’ perspectives on the most and least helpful aspects of CBS.

One way to conceptualize the CBS model is Milne and Reiser’s (2017) tandem model, which can be pictured as two people pedalling a two-person bicycle. In the driver’s seat is the supervisor (i.e. ‘supervision cycle’) who directs and leads the trainee (i.e. ‘supervisee cycle’) towards a common goal. CBS emphasizes certain structures and techniques that distinguish it from other supervision models. In particular, CBS is highly structured and agenda-driven, it uses various experiential techniques (e.g. role-play, direct observation), and it involves formal, regular evaluation to promote trainee learning (Reiser, 2014).

Since the 1990s, supervision has become a separate activity from therapy (Milne, 2006). Research is delineating the competencies needed to become supervisors as opposed to therapists and differences have emerged (Prasko et al., 2012). As such, qualification in CBT does not necessarily translate to competency in CBS, and vice versa. Therefore, to establish mechanisms of effectiveness and identify the essential elements that contribute to trainee learning and patient outcomes, CBS merits its own rigorous research. However, conducting high quality research on CBS is not without its challenges. Supervision research is time consuming, and supervision itself is a complex process involving many actors and potential moderators and mediators (Keum and Wang, 2021). As a result, there is paucity of empirical research on supervision, and research in this area is sorely needed.

Prior research on supervision has emphasized the role of the supervisory alliance in promoting trainee satisfaction (Keum and Wang, 2021; Watkins, 2014). The pan-theoretical concept of supervisory alliance, which Watkins (2014) described as the ‘heart and soul’ of supervision is also alluded to in Milne and Reiser’s (2017) tandem model, inasmuch as collaboration and other relational aspects are highlighted through the ‘common factors’. A strong supervisory alliance is crucial as it has been associated with increased trainee openness to disclose difficulties encountered with clients (Falender, 2014; O’Donovan and Kavanagh, 2014; Wilson et al., 2016). Supervisors must create an environment where trainees feel safe or comfortable discussing problems and mistakes, as effective supervision often depends on trainees’ disclosure of their difficulties (Cartwright, 2019; Callahan and Love, 2020; Falender, 2014; Johnston and Milne, 2012; Mehr et al., 2010). This task can be difficult due to the inherent power differential within the dyad, which can lead to trainees experiencing fear of repercussions in terms of negative evaluations by their supervisors. If trainees intentionally fail to disclose ongoing problems, there can be negative consequences on patients’ progress and outcomes (Ladany et al., 1996), as well as on the supervisors who have legal responsibilities for their trainees’ patients (Cartwright, 2019).

Clearly, there is compelling evidence that the supervisory alliance is an important aspect of supervision. However, the supervisory alliance is only one ingredient of quality supervision, and other elements of CBS may be at least as important for promoting trainee learning. The extensive range of CBS interventions are captured and defined in Supervision: Adherence and Guidance Evaluation (SAGE; Milne and Reiser, 2008), an observational evaluation measure that reflects Milne and Reiser’s CBS tandem model. The SAGE emphasizes experiential learning as the primary mechanism by which CBT trainees acquire the necessary skills to practice therapy (Milne and Reiser, 2017). As such, CBS includes items that capture the basic structure and tools of CBT (e.g. agenda-setting, collaboration, feedback, case formulation) as well as many experiential supervision interventions, including the use of role-play and modelling, behavioural rehearsal, attention to trainee emotional awareness and experiencing, and experimental practice.

As in CBT, interventions in CBS are grounded in empirical evidence. However, little is known about whether trainees recognize these interventions as they occur, or whether they identify them as valuable for their learning process. Some studies examining trainee perspectives suggest that trainees generally appreciate supervisors who are attentive to their learning needs, who use a variety of teaching methods, and who structure supervision such that it parallels a CBT session (Kelly and Hassett, 2021; Murr et al., 2020). CBT trainees also report appreciating help in setting clear, appropriate goals, clarifying problem formulations, and planning future sessions with patients during supervision (Murr et al., 2020; Törnquist et al., 2018). Tornquist and colleagues reported that bi-directional feedback, especially receiving positive feedback from the supervisor and being able to discuss problems that arise with their supervisor, was valued in supervision. Murr and colleagues (2020) found that trainees want supervision experiences that help them link theory to practice (using audio-recordings, discussing case formulations, and using role-play and modelling) and that mirror the structure of CBT in supervision. They also reported appreciating having a knowledgeable supervisor who can differentiate CBT techniques from interventions used in other clinical orientations in order to better assimilate CBT material.

Although studies suggest that trainees may require a strong supervisory alliance to facilitate disclosure and process the content of supervision (Falender, 2014; O’Donovan and Kavanagh, 2014; Wilson et al., 2016), little is known about what aspects of CBS trainees consider most and least helpful for their learning. Some studies have reported strategies that trainees deem useful (e.g. Kelly and Hassett, 2021; Murr et al., 2020), but research examining trainee perspectives remains inadequate, and no firm conclusions can be made about what trainees want from supervision (Callahan and Love, 2020; Kelly and Hassett, 2021). Therefore, this study aimed to further explore what trainees believe is most and least helpful following each supervision meeting, and to determine whether trainees’ perceptions of helpful supervision interventions align with what the expert consensus purports to be effective CBT supervision.

Method

Participants

Participants were eight trainees completing a CBT rotation in an out-patient hospital clinic over the span of one academic year. All trainees who were invited to include their data in the study agreed to participate and provided written and informed consent for their participation. The trainees included four psychology doctoral practicum students, three psychiatry residents (including two 2nd-year and one 5th-year psychiatry resident), and one occupational therapist. All trainees except the 2nd-year residents self-selected to pursue their training in the clinic. The CBT rotation was mandatory for 2nd-year residents. All trainees were required to take a 15-week didactic CBT course to complement their prior training and concurrent individual supervision meetings.

Supervision was provided by licensed psychiatrists (n=2) and psychologists (n=2) with a minimum of 9 years of CBT experience and 3–19 years of CBT supervision experience. All supervisors had demonstrated competence in CBT, as evidenced by achieving passing scores on the Cognitive Therapy Scale - Revised (CTS-R; Blackburn et al., 2001). At the time of the study, two of the four supervisors had already obtained certification of CBT competence by the Canadian Association of Cognitive and Behavioural Therapies (CACBT) and the most junior supervisor’s application for certification was under review. All supervisors had also completed postgraduate continuing education in clinical supervision, including a CBT supervision workshop at the Beck Institute for Cognitive Therapy completed by three of the supervisors.

Patients were aged 18 and older and most met DSM-5 (American Psychiatric Association, 2013) criteria for a principal diagnosis of an anxiety or anxiety-related disorder, a depressive disorder, a personality disorder, a psychotic disorder, or insomnia. The current study received ethical approval from the McGill University Health Centre Research Ethics Board (authorization no. 2021–6889).

Procedure

At the first supervision meeting, trainees reviewed and signed a supervision contract with their supervisor, outlining expectations for the rotation and the role of the supervisor and trainee. ¹ Supervision was delivered on an individual basis and trainees received 1–2 hours of weekly supervision depending on their needs and caseloads. All supervision meetings followed a CBT session structure. Meeting agendas were collaboratively constructed, learning goals for the supervision meeting were identified and feedback about learning was elicited at the end of each meeting.

Following each supervision meeting, trainees completed the Rating of Experiential Learning and Components of Teaching & Supervision (REACTS; Milne et al., 2011) as part of usual practice at the training site. The REACTS is a trainee-rated questionnaire about their supervision satisfaction and learning. It contains 11 items rated on a 5-point scale and two open-ended questions at the end of the questionnaire, specifically, ‘Of the events which occurred in this supervision session, which one do you feel was the most helpful for you personally? It might be something you said or did, or something your supervisor said or did’ and ‘Other comments? (e.g. unhelpful events, unresolved problems)’. For the purposes of the present study, only the two open-ended responses at the end of the questionnaire were analysed. To encourage trainee disclosure, responses were submitted to the research coordinator via email (author J.G.) and supervisors did not have access to their responses until rotations were completed and trainee evaluations were submitted. All written responses were de-identified and transferred to an Excel document for coding. This step ensured confidentiality of the trainee and supervisor.

Coding

To assess which CBS supervision interventions were most and least helpful, trainees’ written responses were coded for the presence and desirability of evidence-based CBS interventions. A comprehensive list of supervision interventions and processes was derived from the SAGE (Milne and Reiser, 2008), an observational evaluation tool originally designed to assess supervisor competence in delivering CBS. The SAGE consists of 23 items grouped into the common factors, the supervision cycle, and the supervisee’s learning, thus reflecting the CBS tandem model (see Table 1). The definitions used for each item were taken from the coding manual (Milne and Reiser, 2008) and a more detailed description of the items found in Milne and Reiser (2016). Clarifications were made to the definitions, as needed, throughout the coding process. Although the SAGE is an observational measure that is not designed to be used by trainees, using it as the coding scheme allowed for the comparison of trainees’ perspectives on helpful supervision with current expert-derived theory about supervisor competencies in CBS, including aspects related to the supervisory alliance.

Table 1. Frequencies of supervision strategies reported as ‘most helpful’ or ‘absent but desired’

* Agenda-setting was also reported three times as present but unwanted.

Each item on the SAGE was rated independently, and therefore, single responses could have been coded as multiple supervision interventions. In other words, each trainee response could be coded with multiple strategies, and strategies could be coded more than once due to overlap between some SAGE item definitions. Each response was coded with the presence or absence of all SAGE items as well as for the desirability of that intervention. As such, possible codes assigned were (a) absent and not desired, (b) absent and desired, (c) present and desired, or (d) present and not desired.

Inter-rater reliability

To increase the validity of the results, about 33% of the data (n=41 responses) was coded by two independent coders (J.G. and J.R.) to establish inter-rater reliability, as suggested by O’Conner and Joffe (2020). To reduce chance agreement, kappa statistics were calculated (McHugh, 2012). Two co-authors on this paper (J.R. and G.M.) are also clinical supervisors working in the clinic. Author J.R. also coded a randomly selected portion of the anonymized responses for reliability purposes. There was a noteworthy experience differential between the two reliability coders, which initially lowered inter-rater reliability. However, coders engaged in discussions to resolve disagreements and to ensure follow-up reliability. Adequate reliability was attained for all SAGE items following two rounds of coding (kappa=.90, p<.001). No kappa values were calculated for items that were not endorsed by trainees.

Results

Approximately 92% (n=127) of the 138 REACTS completed by trainees contained written responses for at least one of the two open-ended questions. Completed responses varied in length, ranging from two to 239 words (M=37.26, SD=35.98). Overall, the qualitative analyses showed that trainees found multiple aspects of their supervision meetings helpful and they rarely identified supervision interventions or events that were unhelpful or that could be improved (see Table 1). Additionally, there were only three instances of trainees reporting SAGE items as present, but unwanted, and these reports were all related to a preference for less structured supervision meetings. Every SAGE item was reported as most helpful at least three times, suggesting diversity in trainees’ reports across supervision meetings and/or across trainees.

Five aspects of supervision were endorsed as most helpful (i.e. coded as present and desired) particularly often: didactic teaching, planning actions following supervision, clarifying case formulations, using experiential activities (training/experiencing), and being evaluated on their work. The features of supervision that were coded least often (i.e. less than five times overall) as helpful and desired by trainees were supervisor’s use of prompting, listening, questioning, and supervisor receiving feedback. Experimenting, in which the supervisee engages in actions to seek greater understanding, problem-solve, or rehearse outside of supervision, was also infrequently coded.

In terms of negative supervision events or aspects of supervision that could have been improved, only three aspects stood out across trainees’ responses (i.e. coded as present but not desired, or absent but desired): collaboration, managing, and agenda-setting. In most cases, when trainees expressed that one of these items was unhelpful, trainees also described issues with one of the other aspects of CBS.

Most helpful supervision interventions

Teaching

The most helpful supervision intervention, noted by all trainees, was didactic teaching. Specifically, trainees identified specific examples of receiving concrete advice or guidance (e.g. how to conduct certain interventions), suggested readings, explanations of concepts, theory and techniques, or suggestions to use handouts and worksheets with patients. For example, one trainee wrote, ‘I really appreciated the suggested readings and the more theoretical part about the different diagnoses and how they may present in therapy’. Another trainee expressed appreciation for how theoretical principles can be applied to a specific patient, ‘[I] appreciated that my supervisor took the time to discuss how core beliefs develop from early experiences and how identifying them could help unlock stuck points in therapy. The discussion was like having a condensed lecture on core beliefs!’.

Planning

The second most frequently endorsed supervision intervention was planning. Planning was coded when it was clear that trainees attempted to solve problems and make decisions about their future actions, either independently or cooperatively with the supervisor. Planning was often identified as helpful in the context of other CBS strategies, such as teaching, formulating, and using experiential techniques. For example, one trainee noted that ‘The role play we did was helpful. It forced me to articulate a plan and to practise shaping interventions when it gets tricky’. Another response highlighted the value of discussing a specific patient’s case formulation to inform their plan for their work with that patient, ‘My supervisor collaboratively guided me to review and complete [the patient’s] case formulation with the patient’s new info. Based on that, we decided on a plan for next session and how to set-up [a] behavioral experiment to target the new assumptions’).

Formulating

We coded formulating when supervisors actively helped trainees to develop an individualized case conceptualization. Formulating was also often coded alongside other CBS strategies, such as discussing and teaching. Trainees often identified formulating as helpful when they were experiencing obstacles or difficulties in their work with their patients. It was also often described when trainees learned through supervision that their own case formulations were lacking or under-developed. For example, one trainee wrote, ‘Having my supervisor’s input was very helpful to realize I was drifting from the patient’s initial goal. I now see how he avoids talking about his anxiety and catastrophizing, and instead externalizes on others [and] intellectualizes his problems’. Another trainee explained, ‘Getting direction from [my supervisor] about how to work on PTSD symptoms that don’t include intrusions (i.e. re-experiencing, flashbacks, etc.) was helpful for me in my treatment planning and case conceptualization for one of my patients’.

Training/experimenting

Training and experimenting were coded when trainees described their supervisors’ use of experiential methods of teaching. Several experiential interventions were described as helpful. Supervisor demonstrations of specific CBT skills (e.g. Socratic questioning, setting an agenda) were especially appreciated. The use of role play was also often reported as helpful in that it appeared to enable trainees to feel more confident in how to manage difficult situations with patients in future sessions. For example, one trainee wrote, ‘We role played how to get the client focused without being dismissive/invalidating when they go off on a tangent. Seeing the types of things my supervisor said to bring me back in the role play was very helpful for me’. Another common experiential supervision intervention that trainees viewed as helpful was reviewing video clips of session recordings in supervision and getting immediate feedback. Trainees noted that reviewing videos helped them identify specific skills that they could improve or how to better manage moments in sessions where they felt ‘stuck’ with a patient. They also felt that watching videos provided the supervisor with more context in understanding the patient presentation and issues related to the therapeutic processes, which ultimately resulted in supervisors providing more contextualized feedback to the trainee.

Evaluating

Trainees also reported on the helpfulness of receiving summative and formative evaluations of their work in supervision meetings. One frequently reported evaluation intervention that trainees rated as most beneficial for their learning was having their supervisor watch a video clip of specific moments in a session and receiving corrective feedback on CBT specific skills. In addition, all trainees reported that formal evaluation of their CBT competence using the CTS-R (Blackburn et al., 2001) was helpful. Trainees felt that the CTS-R review facilitated in-depth discussions with their supervisors about their strengths and weaknesses. When strengths were addressed, trainees reported feeling ‘reinforce[d]’ regarding what they should keep doing (i.e. solidified existing competences). Discussion of weaknesses helped trainees identify and learn how to address areas for further development.

According to one trainee:

‘Doing the CTS-R and having [my supervisor] review a full encounter has hands down been the most helpful learning experience to date in terms of supervision. I came out of the session with better understanding of areas to improve [and received] tangible feedback. Feedback was sobering in positive way but highlights the importance of regular review of full encounters – I feel like this is really the only way to concretely build competency but I’m not sure how realistic this would be. As newcomers to CBT, we have been launched into it with little to no experience – today’s CTS-R feedback highlights that depth of supervision supersedes quantity of supervision. I would much rather have a full session reviewed [every] 2 weeks than weekly supervision sessions.’

Trainee suggestions for improvement

The following section describes the most commonly, yet still infrequently, reported supervision interventions that trainees viewed as absent, but desired or that occurred, but were not desired.

Managing

When trainees expressed that an aspect of supervision was not helpful, they most frequently reported a desire for longer or more frequent supervision meetings. Some trainees did not feel they had enough time to discuss all their patients in every supervision meeting. Interestingly, although some of the same trainees noted that their supervision time had been increased, they still felt that they could benefit from additional supervision time after the adjustment. Another management issue that arose was that some trainees wished that their supervisor did more to prepare before supervision meetings. For example, one trainee wrote that they ‘would appreciate if supervisor reviewed notes before supervision as a significant amount of most sessions is spent bringing supervisor up to speed with the context of the encounter’.

Collaborating

Some trainees indicated that their supervisor could have been more collaborative. An absence of sufficient collaboration was coded when the supervisor’s expectations of what the trainee should do with the patient seemed unclear or unrealistic, or when trainees seemed to report a general lack of productive teamwork. For example, one trainee reported:

‘[My supervisor was] jumping in with suggestions/feedback before asking the approach I took or how I tried handling the situation in the session. A lot of the time in supervision was spent on suggestions/pointers about interventions that I had actually tried in the session, so the time wasn’t used as helpfully as it could have been.’

Agenda-setting

Agenda-setting was the only CBS item that was reported as both ‘absent but desired’ after some supervision sessions and ‘present and unwanted’ after others. In addition, when trainees reported a desire for more agenda-setting, it often overlapped with a lack of management and/or collaboration. For example, while one trainee wanted more structure in supervision, ‘Added structure by supervisor (e.g. setting agenda at start of encounter) would also be beneficial’, another trainee thought supervision meetings were too structured:

‘I also wish there could be less focus on “what’s your supervision question” and more time to freely discuss cases. I understand that questions help keep things focused, but I think it would also be beneficial to have some openness to less structured ways of discussing patients and sessions. Otherwise, I feel that it leaves little room for “I was happy with this intervention” – not sure why there would be a question attached to this type of discussion, for example.’

In terms of trainees desiring less structure, one trainee wrote that they ‘would appreciate more formal teaching integrated, [and] for supervision to feel less like a CBT session (i.e. setting agenda, reviewing HW/supervision question)’.

Discussion

The aim of this study was to identify the CBS interventions that trainees view as most and least helpful for their learning. To answer this question, we examined trainees’ written responses to two open-ended questions regarding their supervision satisfaction (REACTS; Milne et al., 2011). We coded trainees’ responses for the presence and desirability of CBS features using criteria derived from a measure of CBS competence (SAGE; Milne and Reiser, 2016).

Overall, we found that trainees described many more aspects of CBS that were helpful and desired compared with supervision experiences that were unhelpful or lacking. Additionally, all 23 items of CBS outlined in the SAGE were rated as helpful and desired by at least one trainee in the study, suggesting that trainees appreciated a variety of CBS interventions. This finding is consistent with theory and research supporting the use of multiple teaching methods in promoting trainee satisfaction and better learning outcomes (e.g. Kelly and Hassett, 2021; Milne et al., 2003; Murr et al., 2020; Newman, 2013; Prasko et al., 2012).

With respect to the aspects of supervision that were most often reported as helpful by trainees, the CBS interventions that supported skill development, namely teaching, planning, formulating, training/experimenting and evaluating, were most frequently identified. This finding may reflect trainees’ need for supervision activities that favour didactic learning experiences (i.e. teaching) and discussions about case formulation to fill in knowledge gaps and provide opportunities to practise or observe supervisor CBT skill demonstration (i.e. training/experimenting). The presence of these more educational CBS interventions may have contributed to trainees’ capacity to develop a plan of action (i.e. planning) that imbued them with confidence for their next session with a patient. Indeed, trainees’ descriptions of the most helpful aspects of supervision often included multiple CBS strategies that parallel the supervision cycle described by Milne and Reiser’s tandem model (2017).

One of the most frequently reported aspects of supervision deemed helpful for learning was the use of evaluation. The perceived value of evaluation amongst trainees was initially surprising to us given that trainees typically experience anxiety about the evaluative nature of supervision (Bernard and Goodyear, 2014; Ellis, 2010; Inman et al., 2014) and research has shown that most trainees conceal important information related to their work from supervisors (Cook et al., 2020; Mehr et al., 2010; Murr et al., 2020). However, our findings are consistent with other studies showing that trainees are generally more satisfied with supervision meetings when their work is evaluated, especially when therapy sessions are reviewed and rated for CBT competence (Rakovshik and McManus, 2013; Törnquist et al., 2018). Although time-consuming, the use of evaluation, especially of entire sessions, may help supervisors to more accurately assess a supervisee’s needs and tailor supervision interventions to provide more effective learning opportunities and enhance the collaborative nature of the supervision. Indeed, every single trainee expressed appreciation for CTS-R feedback, which suggests that trainees’ motivation for development may outweigh any performance anxiety they may experience during the process. Of note, trainees in this study were presented with supervision contracts describing the nature of the supervision process and outlining the types and frequency of evaluations. In this regard, the use of a contract may have attenuated some of the anxiety that supervisees usually experience by clarifying expectations, which has been shown to reduce anxiety in junior trainees (Ellis et al., 2015).

Although all strategies were reported as helpful, the CBS strategies that were least often reported by trainees as most helpful include: prompting, listening, questioning, providing supervisor feedback, and experimenting. There are several possible reasons why trainees did not report these strategies as most helpful. For example, it may be trainees were unintentionally primed to write about specific themes when recalling the perceived helpfulness of supervision interventions because they provided their open-ended responses after answering other questions about supervision satisfaction on the REACTS. Another possibility is that some CBS interventions may only be noticed if there is an evident lack of them, especially if the problem persists across multiple supervision sessions. For example, most supervision meetings are likely to include at least some questioning, listening and agenda-setting, which trainees may take for granted. In addition, trainees may not notice when supervisors model more subtle behaviours (Falender, 2014) and may therefore tend to overlook these supervision interventions as especially helpful. Given that the major benefit of CBS is to model the CBT approach to enhance skill acquisition (Milne and Reiser, 2017), supervisors should consider more explicit signposting when using more subtle supervision interventions.

The possibility that trainees may not attend to aspects of supervision that are less explicit may also explain why the items from the SAGE that pertain most closely to the supervisory alliance (i.e. relating, managing, collaboration) were not among the most frequently described helpful aspects of supervision. While previous studies have shown that trainees’ rate the supervisory alliance as critical for their experience, our results show that supervision interventions that favour trainee skill development were more frequently endorsed as beneficial. As Milne and Reiser’s ‘supervision cycle’ would predict, the ‘common factors’ may set the stage for trainees’ ability to profit from supervision interventions (Milne and Reiser, 2017). When the supervisory relationship is positive, trainees may be better able to take advantage of learning opportunities. In contrast, when the relationship is strained, trainees may feel overly anxious and concerned with evaluation which negatively impacts their ability to disclose and benefit from the supervision interventions that promote their skill development. It may also be that when supervisors are effective in teaching and modelling CBT skills, trainees view the supervisory relationship more positively due their increased confidence in the supervisor and their own ability to be effective with their patients.

In support of the importance of the supervisory alliance in promoting satisfaction with supervision meetings, we found that when trainees identified a supervision event or intervention as unhelpful, they were most likely to note a desire for better management and greater collaboration, and a preference for either more or less structure when setting the agenda. This finding is consistent with the possibility that trainees may take certain aspects of their supervision experiences for granted when they are present, but their absence impacts the supervision experience. In terms of management, an interesting finding was that trainees often reported wanting more supervision time. Although not directly a lack of the management per se, the frequency of these reports could imply a lack of effective use of supervision time. Alternatively, as Sudak and Reiser (2021) have suggested, it may be due to other factors such as high caseloads or other situations requiring more intensive supervision. However, trainees who received longer supervision meetings still reported desiring more supervision time. Therefore, there may be other issues underlying these reports, such as a lack of trainee confidence or an appreciation of their interactions with their supervisor. Furthermore, because agenda-setting was often reported as absent but wanted in conjunction with lack of management, added collaborative structuring to meetings may improve time management and trainee satisfaction.

Strengths and limitations

A strength of this study was its intensive and naturalistic approach which reflected trainees’ perspectives on a weekly basis in an ecologically valid way. Importantly, the methodology allowed for the exploration of many CBS strategies beyond the supervisory alliance, which has been disproportionally emphasized in previous studies (Watkins, 2014). In addition, although supervision research is typically a time-consuming endeavour, administering the REACTS every week was an efficient way to collect reflective data on trainees’ perspectives about supervision over time. The use of brief trainee measures following supervision meetings may be useful in future studies on supervision to help detect patterns of supervision satisfaction over time and help identify the links between specific supervision interventions and trainee skill acquisition and patient improvement.

Future studies could also benefit from the use of larger sample sizes, which would enable researchers to examine whether there are any individual characteristics or trainee–supervisor dyad differences that influence supervision satisfaction or CBS intervention preferences. For example, it may be that more junior trainees prefer didactic teaching prior to engaging in experiential activities (e.g. role-play). In addition, our limited sample size also did not allow us to identify whether there the results were affected by trainee or supervisor competence, trainee or supervisor discipline or level of experience, a better supervisor–trainee match, or chance. Examining these questions might contextualize some of the findings on the importance of the supervisory alliance and trainee satisfaction.

Finally, we examined the question of what trainees appreciate in CBS. Although we specifically measured the concept of ‘helpfulness’ in our study, we believe that satisfaction with supervision is in large part determined by trainees’ perceptions of the extent to which they perceive a supervision meeting as helpful. Furthermore, understanding what trainees view as beneficial for their learning may help supervisors tailor supervision to maximize engagement and learning which would hopefully predict better patient outcomes. However, we did not examine the associations between the aspects of supervision that trainees found helpful and trainee competence development or patient outcomes. Given their inexperience, trainees may not be able to adequately judge what they need to learn or how to effectively develop new CBT skills. Therefore, what trainees believe is most helpful may not be the best indicator of successful supervision. However, it may also be that trainees recognize when skilled supervisors can help them identify their own learning needs. To answer these questions, future studies ought to identify whether the CBS interventions that are most valued by trainees actually promote objective outcomes of competence.

Conclusion

In conclusion, our study shows that trainees evaluated a variety of CBS interventions as helpful, especially interventions that address knowledge gaps and provide a model for how to implement CBT interventions with patients. Although the putative link between trainees’ perspectives of helpful supervision practices and trainee learning has not yet been established, the focus on what trainees want from supervision may help guide supervisors’ interventions to increase satisfaction, which may lower burn-out and distress in trainees (Livni et al., 2012; Milne, 2020; Prasko et al., 2012).