Coaching has emerged as a powerful tool in medical education, supporting learners in developing self-awareness, setting goals, and navigating the complexities of their professional journeys. ^1-6

When delivered effectively, coaching can enhance professional identity formation, foster adaptive learning strategies, and promote a growth mindset.^1,3,7 Yet, despite these benefits, access to coaching remains limited across undergraduate medical education (UME) programs.⁸ Longitudinal coaching relationships, while ideal, require sustained institutional investment, trained personnel, and structural support, all of which may be unavailable in resource-constrained environments.^9-11

Self-Determination Theory (SDT), a framework for understanding human motivation, forms the theoretical basis for coaching practices aimed at supporting learners’ basic psychological needs of autonomy, competence, and relatedness.¹² These needs are essential for fostering intrinsic motivation, well-being, and professional growth. ^12-14

Coaching conversations, particularly those grounded in reflection and goal setting, can foster autonomy by helping learners articulate personal aspirations; competence by highlighting strengths and planning for skill development; and relatedness through supportive dialogue with a trusted educator. These needs are especially salient during the transition to residency (TTR), a period marked by heightened uncertainty and vulnerability, making it an ideal target for interventions that support motivation, identity, and self-efficacy.^6,15

Although several longitudinal coaching models exist, there is currently no consensus-based, SDT-informed single-session coaching framework designed specifically for the TTR. This represents a critical knowledge gap, as most institutions lack the resources to implement sustained coaching programs. Our aim was to develop a consensus-based, SDT-informed single-session coaching conversation framework for the TTR and to evaluate its acceptability and content validity. The intended users are faculty across specialties, including those without formal coaching training, who support learners at this transition. We intentionally focused on a single-session framework to maximize accessibility, scalability, and generalizability.

We utilized a modified Delphi technique to establish consensus on the structure and core questions of a coaching conversation for learners transitioning to residency.^16,17 The modified Delphi technique is an iterative approach that gathers and refines expert input over successive rounds to achieve consensus. A team member (MW) with extensive experience in the modified Delphi method guided the methodology and facilitated the process but did not participate in voting.

We convened a multi-institutional panel of 13 medical educators with expertise in medical education coaching. Eligibility criteria included prior experience coaching undergraduate medical education (UME) or graduate medical education (GME) learners or designing institutional coaching programs. Individuals without coaching experience were excluded. Panelists were recruited through national coaching networks to ensure diversity in institutional type, geography, and coaching infrastructure. Six represented schools with formal coaching program, while others came from institutions with more limited resources. The final sample size (n=13) aligns with recommended Delphi panels of 10-18 experts to ensure robust consensus.¹⁶ All panelists participated in every round (100% retention).

The Delphi panel identified a set of powerful coaching questions and essential skills to support structured coaching conversations (Figure 1). All 13 panelists participated in each survey and meeting, with 100% retention across all rounds.

The initial nominal group process generated 125 questions distributed across the four stages of the coaching framework: 13 questions in the “opening the session” phase, 47 in the “exploring the topic and eliciting reflection” phase, 49 in the “exploring options and developing an action plan” phase, and 16 in the “closing” phase. Panelists also proposed 12 essential coaching skills. After deduplication and merging of overlapping submissions, the Round 1 survey contained 125 unique questions and 12 skills.

After Round 1, 17 questions and 2 skills reached high agreement (median = 9, IQR 0-1; >80% rating 9-10). In Round 1, Kendall’s coefficient of concordance (W) was calculated to assess the degree of agreement among panelists’ 1–10 ratings. The value of W = 0.78 indicated strong concordance, reflecting high stability in expert ratings at this stage. No items were excluded at this stage, and 108 questions with 10 skills were advanced to the next round.

In Round 2, panelists used yes/no voting to evaluate the remaining items. Two additional questions and one skill achieved >80% inclusion, while 62 questions and 2 skills were excluded.  Forty-four questions and seven skills progressed to Round 3. In this final rating round, six more questions and one additional skill reached consensus. Forty questions and two skills were excluded. Because Rounds two and three used binary yes/no voting rather than ranked data, stability was assessed using the proportion of panelists voting for inclusion or exclusion, consistent with Delphi methodology.

Overall, the process narrowed the list from 125 questions and 12 skills to a final framework of 25 powerful question prompts and 4 essential coaching skills. Skills were retained based on frequency of endorsement and consensus discussions emphasizing practicality and alignment with core coaching principles, including active listening, growth mindset, preparation, and minimizing distractions.

Qualitative feedback highlighted the importance of clarity, SDT alignment, and feasibility for non-coach faculty. Revisions included simplifying wording, removing redundancy, and ensuring prompts encouraged autonomy, competence, and relatedness. Of the final 25 questions, 9 primarily supported autonomy, 8 competence, and 8 relatedness, directly mapping to SDT domains.

These elements formed the final coaching conversation framework (Table 1), representing a strong consensus on practical strategies that faculty can use to guide effective coaching conversations during the transition to residency.

In this study, we used a structured nominal group technique followed by a modified Delphi process to develop a structured coaching framework to support learners during the transition to residency (TTR). From an initial pool of 125 questions and 12 skills, the panel distilled a final framework of 25 questions and 4 essential coaching skills. These prompts are designed to guide reflection, self-assessment, and action planning across the natural arc of a coaching conversation.

Importantly, this study addresses a specific gap: although longitudinal coaching models are well described, no consensus-derived, SDT-informed single-session framework for the TTR has previously been developed. Our results demonstrate the feasibility of reaching strong expert consensus on such a framework, offering a pragmatic tool for use across diverse educational settings.

The final framework offers an accessible entry point for educators, even those without formal coaching training, to engage learners in meaningful, future-focused conversations. Its adaptability and low resource requirements make it especially relevant for institutions unable to support longitudinal coaching programs.

Self-Determination Theory (SDT), a well-established theory of human motivation, provides the theoretical underpinning for our study by explaining the value of coaching to support learners’ basic psychological needs for autonomy, competence, and relatedness—key drivers of intrinsic motivation, well-being, and professional development.¹²  Coaching interactions grounded in reflective dialogue and goal-setting naturally promote autonomy by centering the learner’s personal values and aspirations, competence by helping learners recognize their strengths and make actionable plans, and relatedness by fostering supportive, learner-centered relationships. These psychological needs become particularly salient during transitions such as the shift from medical school to residency, when learners often experience increased vulnerability, uncertainty, and self-doubt. By mapping prompts to the basic psychological needs described in SDT, the framework is theoretically grounded in motivation science and designed to support learner engagement and professional identity formation during a critical phase of training.

The TTR is widely recognized as a period of heightened stress and uncertainty, with national bodies such as the Undergraduate Medical Education - Graduate Medical Education Review Committee (UGRC) highlighting the importance of coaching during this transition.⁶ Our findings align with these recommendations by providing a structured, consensus-derived approach that institutions can adopt. Because transitions occur throughout medical training, the framework may also be useful beyond TTR. By emphasizing exploration of uncertainty, surfacing internal barriers to growth, and building actionable plans, the frameworksupports conversations that normalize struggle, counter imposter syndrome, and encourage reflection on strengths.

Such dialogue fosters a growth mindset and professional identity formation, both of which are essential during critical career transitions. The final months of medical school offer a rare window for intentional reflection, and structured coaching conversations at this stage may enhance readiness for residency while supporting learners’ long-term professional development.

This study has several limitations. First, the panel was small (n=13) and limited to expert educators, so learner perspectives were not included. Second, although we calculated Kendall’s coefficient of concordance (W) to demonstrate strong agreement in Round 1, subsequent rounds used binary voting, which did not permit calculation of stability statistics. Agreement in those rounds was therefore assessed by proportion of inclusion or exclusion, which may provide a less nuanced assessment of consensus. Third, consensus rules shifted between rounds, from a 1–10 scale to yes/no voting, to streamline the process. While prespecified, this change may have influenced outcomes. Fourth, although anonymity was maintained for survey ratings, live Zoom discussions may have introduced potential for dominance bias despite structured facilitation and safeguards. In addition, the reduction of coaching skills from 12 to 4 was based on panel discussion and endorsement frequency rather than formal psychometric analysis. Finally, the framework has not yet been tested in real-world coaching sessions, so its acceptability and effectiveness for learners remain unproven.

These limitations temper the claims that can be made. Rather than a definitive model, the framework should be viewed as a pragmatic starting point that requires local adaptation and brief faculty development.

Future research should focus on cognitive interviews with learners to assess clarity and acceptability of prompts, pilot feasibility and fidelity studies of faculty use, mapping of prompts to observed SDT outcomes such as self-efficacy and preparedness, and multi-site trials to evaluate generalizability. Additional work should also explore whether use of the framework supports long-term outcomes such as resilience, well-being, and professional identity formation.

Despite these limitations, this study contributes a practical and scalable framework for coaching at the transition to residency. By broadening access to high-quality, SDT-informed coaching conversations at a pivotal moment in training, this framework may help learners navigate uncertainty, strengthen self-awareness, and prepare more confidently for the demands of early residency.

We developed a practical, consensus-based coaching conversation framework to support learners during the transition from medical school to residency. Designed for ease of implementation, this framework offers institutions a scalable, low-resource approach to facilitating reflective conversations that foster self-awareness, goal-setting, and professional growth. Grounded in SDT, the framework supports learners’ psychological needs for autonomy, competence, and relatedness. While promising, the framework should be viewed as a pragmatic starting point that requires local adaptation and further study. Future work is needed to examine its feasibility, fidelity, and impact on learner outcomes. By broadening access to coaching at a pivotal moment in training, this framework may help enhance learner readiness and well-being during one of the most critical transitions in medical education.