In-the-Moment Feedback and Coaching: Improving R2C2 for a New Context

There is growing evidence that effective coaching improves hands-on skills in the operating room, nontechnical skills such as patient communication, leadership and teamwork, and physician well-being in residency.^1–6  Since residents may reject feedback that conflicts with their self-assessments due to a perceived lack of credibility or lack clarity on how to use it, coaching techniques can enable supervisors (ie, clinical supervisors or preceptors) and residents to collaboratively reflect on performance, focus on growth and development, and embrace performance gaps as catalysts for learning.^7–10 

The R2C2 resident formal model^3,4,11,12  was developed to facilitate formal feedback and coaching conversations, enable collaborative discussions between supervisors and residents, and establish a safe environment through a series of open-ended questions that emphasize reflection and continual improvement. When the model is used effectively, the resulting feedback conversation can be a dynamic approach to developing competence where attention is paid to building relationships (R), exploring reactions (R), exploring the content (C) of the feedback, and coaching for change (C), thereby enhancing feedback acceptance and use. This approach meets the contemporary conceptualization of feedback as “a dynamic and co-constructive interaction in the context of a safe and mutually respectful relationship for the purpose of challenging a learner's (and educator's) ways of thinking, acting, or being to support growth.”^13(p653) Faculty development in R2C2 has occurred through local, national, and international workshops, guided by trifold brochures along with demonstration videos and the development of a website.¹⁴  The brochures list each of the 4 R2C2 phases with sample phrases aimed at specific target populations and contexts.

The R2C2 model has been applied in a variety of health care settings with practicing physicians, nurse practitioners, and residents.^3,4,11,12,15  Residents and supervisors indicated that the open-ended questions facilitated self-reflection, and action plans enabled the dyad to co-create a concrete action plan that emphasized development and progress. Collectively, the research to date suggested that the R2C2 resident formal model can be used effectively for feedback and coaching with residents who are either excelling or struggling. However, the focus of research thus far has been on summative settings, where a variety of performance data are available for review from multiple points in time (eg, daily encounter forms, objective structured clinical examinations, multisource feedback), and participating supervisors questioned whether the model can also be adapted for in-the-moment feedback.^3,4,12 

The need for high-quality in-the-moment feedback is evident given research reports of performance improvement following immediate feedback, issues of recall post-feedback, and the association between depression and lack of timely and appropriate feedback.^16–19  It is of particular interest to those working in competency-based medical education (CBME) settings where learners are expected to meet milestones or complete entrustable professional activities (EPAs). CBME necessitates frequent feedback following direct observation of workplace-based performance with informal in-the-moment conversations at the end of a shift, clinic, or after a single patient encounter to focus on a specific skill or competence.²⁰  The R2C2 model highlights the importance of a credible relationship to optimize acceptance of feedback as well as leveraging that relationship to act as a coach for improvement, and it could be applicable to in-the-moment settings.

The purpose of this study was to explore how supervisors adapted the 4 phases of the R2C2 resident formal feedback and coaching model during informal in-the-moment conversations in the context of a brief clinical experience (eg, end of a clinic, shift, or operating room day; following observation of a learner). Specifically, we were interested in where supervisors used the model, how they adapted it for in-the-moment conversations, and specific phrases they used in each phase that could be incorporated into a new R2C2 in-the-moment trifold guide.

All interviewees were physicians who supervised residents in clinical settings. Using purposeful sampling, we recruited and invited clinical supervisors/preceptors who were currently using R2C2 and had adapted it for in-the-moment feedback conversations with residents. We believed these physicians could provide the best perspectives on the components of the model, including approaches to provision of more immediate feedback. The supervisors consented and were provided with copies of the R2C2 resident formal trifold.¹⁴ 

The 11 interviewees included 2 pediatricians, 2 medical oncologists, and 1 anesthesiologist, geriatrician, gastroenterologist, general internal medicine specialist, physiatrist, family physician, and a critical care and palliative medicine physician. Two interviewees were from the United States, and the rest were Canadian; 4 were members of the research team. Interviewees described varied contexts in which they used R2C2 in-the-moment model, including while working longitudinally with residents over a week or a month, or for a single shift or a few shifts or in conjunction with a single activity (eg, discussing bad news, performing a technical procedure). Others used it for discussions when concerns arose such as complaints from patients or nurses. It was also used when documentation needed to be completed for daily encounter sheets, field notes, or EPA forms. Settings included outpatient clinics, operating rooms, and hospital inpatient units.

One member of the research team (J.L.) conducted individual interviews with each participant using a semistructured interview protocol (provided as online supplemental material) based on the R2C2 resident formal trifold in November and December 2018. Interview questions addressed experiences and context using R2C2, participant suggestions for rewording of questions and prompts for each R2C2 phase, examples of how R2C2 was used, and other suggestions. Interviews were audio recorded and transcribed.

We used a type of thematic analysis called framework analysis²¹  to identify approaches to adapting and implementing R2C2 for in-the-moment feedback and key phrases that supervisors used within each phase. This approach was selected, as it provides a structure into which we could systematically reduce the data to analyze it by case and by code undergirded by the R2C2 structure.²¹  We followed the sequential steps of framework analysis: familiarization, coding, framework development, indexing, charting, and interpretation. Transcribed interview data were first reviewed by the interviewer (J.L.) for accuracy. Two members of the research team (J.L. and J.S.) coded transcripts from the first 2 interviews and developed a thematic framework comprising 6 framework categories: (1) participant experiences and context for using R2C2; (2) language and strategies for potential inclusion into a new in-the-moment trifold; (3) suggestions for phrases to exclude from the original R2C2 resident formal trifold; (4) overall suggestions for better wording; (5) examples of how the model was applied; and (6) other suggestions. A research associate (R.L-K.) created a spreadsheet with the 6 framework categories to index data from the 11 interviews and interviewer memos; no additional framework categories were identified in the indexing process. This research associate then undertook charting, summarizing data within each framework category. Finally, summaries were confirmed by 2 team members (J.L. and J.S.).

Two 2-hour consensus meetings were held with research team members to interpret and discuss the charted data, with the goal being to agree on the strategies and phrases to create a new trifold for R2C2 in-the-moment feedback and coaching model in February 2019. Meetings were audio recorded. They were followed by several e-mail discussions and document edits to clarify and obtain agreement on the wording for the trifold as well as other ideas on translating the model into practice. Data from the first consensus meeting were transcribed and made available to the group prior to the second meeting, along with an initial draft of the trifold for the first 3 phases. At the second meeting, the fourth phase was discussed, refinements were made, and clinical applications of the proposed tool were discussed. All members of the research team participated in both meetings. A draft of the trifold was then reviewed by each member of the research team, and additional edits were made with general agreement of the revisions.

The research team consisted of PhDs and MDs with diverse clinical backgrounds and years of clinical work, representing 3 countries. Members included those who had developed the R2C2 resident formal model and those new to R2C2 research but experienced users of R2C2 in their clinical workplace. At each step of data collation and analysis, there was a review and discussion of the data.

Once the team felt that the R2C2 in-the-moment trifold was complete, it was sent to researchers involved in developing previous versions of R2C2 and all interviewees.^4,11  Minor suggestions were made, primarily confirming that the trifold was appropriate. These suggestions were reviewed by e-mail discussion; however, no additional changes were made.

The University of Calgary Conjoint Health Research Ethics Board approved the study.

Supervisors were comfortable with the 4-phase approach to feedback and coaching provided by the R2C2 in-the-moment model and noted examples of phrases they used for each phase. These discussions were generally quite short (5 to 8 minutes) compared to the 30 to 60 minutes suggested for formal progress meeting sessions. Supervisors provided rich commentary on each R2C2 phase, outlined below. The Table provides exemplary quotes for each phase from the interviews.

In this phase, supervisors recognized the importance of distinguishing among residents they had not previously worked with and those they had worked with as part of a longitudinal experience (eg, ≥ 1 week). For the former group, they were welcoming and queried past experiences, asked residents to identify goals, and made their approach to feedback clear. For those residents with whom they had a longitudinal relationship, supervisors asked about previous and current goals. They queried these aspects as a way of engaging residents in the upcoming clinical experience and the feedback that would occur. When residents could not identify goals, supervisors took a more proactive approach to suggest types of learning the setting could provide or particular patients that might help achieve a specific learning goal, address an EPA, or enable the completion of direct observation of a procedure form. The importance of relationships was recognized throughout the clinical experience, and specifically was addressed again at the end where supervisors would query learners' perspectives on the experience and coach them.

Supervisors noted that feedback and coaching discussions in the R2C2 resident formal model used collated data included in formal assessment reports; while, for in-the-moment discussions, the supervisors drew on their own observations of residents as data. Supervisors sought residents' reactions to these observations, their perspectives, and self-assessments of the experience, while also providing their own observations. Additionally, they ensured that residents had time to reflect on and react to the information provided. For some residents, more directed probing was required to ensure supervisors could help them see patient perspectives or recall a patient or family member's reaction during the interaction.

Supervisors described how the R2C2 in-the-moment discussions differed from their usual R2C2 resident formal discussions. R2C2 in-the-moment discussions were constrained by time and truncated, while still achieving shared understanding of the experience and priorities. Supervisors described probing residents' understanding of what happened to ensure shared understanding. They noted that discussions about procedures were amenable to this type of probing as one could talk through each step of the procedure. In some cases, drawing on residents' experiences outside medicine was helpful, particularly with communication or professionalism issues where the learner lacked insight. One interviewee described creating a video of the resident performing a procedure, and watching it with that resident to help visualize what was suboptimal so that they could determine a way forward.

Generally, coaching was done at the end of a session and involved collaboratively identifying one specific change that residents could address. Supervisors described how this was similar to the R2C2 resident formal model, in that residents were encouraged to co-develop the plan by identifying how the goal would be achieved, resources needed, implementation plan, timeline, and how results would be evaluated. Ensuring a follow-up plan was seen as important, whether the supervisor was able to work with the resident subsequently or the resident needed to seek the help of a different supervisor. Documenting the plan was seen as challenging in some instances, due to time constraints or when supervisors did not want to formally record negative or developmental information.

Others used the coaching to complete required program-specific assessment forms. If residents had difficulty coming up with a goal, supervisors noted they sometimes identified an area and then sought agreement on the goal. Rather than vague goals such as “read more” that were rarely helpful, participants described how they worked with residents to develop a plan for a specific goal so residents could apply what they learned to future clinical encounters.

Participants' suggestions and sample phrases were used to develop the final R2C2 in-the-moment trifold, included on the R2C2 website.¹⁴  The Box provides an abbreviated version of the trifold.

Our interviews with supervisors contributed to our knowledge about ways to provide feedback and coaching in the moment, a need identified by earlier research and CBME.^16–20  They identified approaches and specific phrases to adapt the R2C2 model for in-the-moment use to allow shorter and more frequent feedback conversations. In doing so, they maintained the core elements of the R2C2 resident formal model and found that the structured approach to building a relationship, exploring reactions, determining content, and coaching for change was a practical way to have a discussion even when time was limited or the focus was on a specific task.

As expected, there were differences and similarities in use of the R2C2 in-the-moment model compared to its use in formal meetings with physicians and residents in discussions that drew on collated data.^3,4,11,12  Of necessity, in-the-moment discussions were shorter, yet participants indicated that all 4 phases of the model remained important. They pointed out the need to use communication approaches and phrases within each phase that would engage the learner and yield an impactful conversation in the brief period of time available in busy clinical settings. This process required engaging residents in self-reflection and facilitating insight into their performance and encouraging them to commit to specific changes. Coaching in the moment requires specific communication and facilitative strategies to engage the learner and ensure commitment, similar to more longitudinal coaching.^12,22,23 

Specific differences arose in wording for phase 1, where relationship building differed between residents who worked with the supervisor for a brief time and those with a longitudinal relationship. As reported in earlier research related to feedback,^4,12,24,25  participants confirmed that relationship building was critical to success of the learning conversation. Building the relationship occurred during the various phases depending on whether or not the supervisor had a prior relationship with the learner and the nature of that relationship.

The iterative interplay between phases 2 and 3 (reaction and content phases of the model, which focused reactions to the activity, reflections on the activity, and self-assessment leading to discussions of specific content) were more pronounced in the moment. Supervisors described exploring both residents' views and their own, actively reflecting on and responding to residents' responses, and moving toward a shared understanding of the event and identification of a goal moving forward.

Coaching for improvement was a critical element of the feedback discussion, yet action planning needed to be specific and realistic, so residents were clear on what they should do in a short period of time about a specific skill or activity. While coaching and planning for improvement enabled facilitating by specific goals,^1,22  the time pressures for the coaching conversation and the opportunities for residents to enact the change and receive follow-up feedback highlight the importance of this step for in-the-moment clinical coaching.

There are limitations to the study, as a small number of interviews with supervisors were conducted. We did not obtain residents' perspectives. Four of the interviewees were members of the research team who would have been able to recognize their own data in the spreadsheet and in the summary data provided. We recognize this adds a level of bias, although in the consensus discussions, it didn't appear that anyone's particular interview or data became a focus for the final content of the trifold.

This research will continue with the recruitment of dyads of supervisors and residents/clinical clerks who will be audiotaped during a feedback session with each member of the dyad participating in interviews soon after the feedback session. This will help assess viability and utility of this version of R2C2 with a larger and more diverse population in various clinical settings.

Perspectives of clinician educators have informed development of an R2C2 model for in-the-moment feedback and coaching. The model emphasizes key principles of earlier R2C2 models and of feedback and coaching in general (ie, importance of relationships, engaging learners, fostering reflection, seeking their views, identifying goals, and collaboratively developing an action plan). It appears this model's 4 phases help structure in-the-moment feedback in varied teaching contexts. Specific phrases and adaptations within each phase were identified to enable effective coaching interactions in time-constrained in-the-moment encounters.