Learner handoffs: Attitudes and practices of pediatric hospitalist educators

Introduction: Communication failures during frequent transitions in attending supervision can lead to delayed entrustment, uncoordinated feedback, and arrested professional development for trainees. With structured patient handoffs showing promise in patient outcomes, attention has been paid to learner handoffs (LHs), though little is known about their utilization and effectiveness. The aim of this study was to identify attitudes and practices of pediatric hospitalist educators towards LHs. Methods: An anonymous electronic survey was distributed to members of the American Academy of Pediatrics (AAP) Pediatric Hospital Medicine Educator Listserv. Data was analyzed using mixed-methods. Results: Of 584 eligible educators, 129 (22%) responded. Most respondents (78%) were familiar with LHs, though only 30% reported that they were an expectation of their pediatric hospitalist groups. Only 15% reported using a structured format, and only 10% of educators received any formal training. A handoff of each additional learner was reported to add less than 5 minutes to the overall handoff. The vast majority reported LHs improved their ability to give accurate feedback (93%) and to tailor teaching (89%), while almost half (46%) reported LHs improved patient care. Major qualitative themes for advantages of LHs included better understanding learner capabilities, learner goals/attending expectations, and optimizing feedback. Barriers to implementation included time constraints and potential for bias. Discussion: Despite overwhelmingly positive views regarding teaching and feedback, LHs often remain unstructured, unexpected, and delivered by untrained supervisors. The development of a structured LH tool, in conjunction with faculty buy-in and training, may accelerate trainee professional development, while mitigating any concerns for added bias.


Introduction
The additional work or re-work that is needed to counteract ineffective communication has been defined in the social sciences as "costs of coordination" (Becker and Murphy, 1992), with "costs" rising as processes become increasingly partitioned or segregated. As the number of clinical patient handoffs has increased, the risk of communication failures has become more pronounced (Joint Commission, 2020), leading to the application of this framework to reducing errors during clinical handoffs (Arora et al., 2008). The development and implementation of a structured handoff bundle has been shown to counteract communication failures and significantly reduce preventable adverse events (Starmer et al., 2012;Konkin and Suddards, 2017).
With asynchronous schedules (Seltz et al., 2014;Trowbridge et al., 2019;Goitein et al., 2008) and duty hour restrictions (Elnicki and Cooper, 2011;Sutkin et al., 2008) causing frequent transitions in attending supervision, coordination costs within medical education are also on the rise, leading to uncoordinated feedback, loss of entrustment, erosion of the student-teacher relationship, and, ultimately, delayed professional development (Holmboe, Ginsburg and Bernabeo, 2011;Schumaker et al., 2014). Learner development is longitudinal, and with the shift towards Entrustable Professional Activities (EPAs) and competency-based training models (Aschenbrener, Ast and Kirch, 2015;Englander and Carraccio, 2014;Ten Cate et al., 2016;Cleary, 2008), isolated assessments of learners are of limited utility (Sterkenburg et al., 2010). Entrustment, or the ability of the learner to complete specific "units of professional practice" without supervision (Englander and Carraccio, 2014;Cleary, 2008), is developed incrementally over time through deliberate faculty observations (Sterkenburg et al., 2010). However, with reports that inpatient attendings supervise for 5-7 days on average, and that it takes 5 or more days to determine initial entrustment (Fuchs et al., 2020), trainees may get stuck in a cycle of entrustment and non-entrustment if there is no communication between supervisor transitions.
One tool that has received increasing attention from medical educators to target these transitions is the "learner handoff" (LH) (Babbott, 2010), with a recent expert consensus recommending that LHs be an essential component of an inpatient teaching service (Gross et al., 2020). Similar to structured patient handoffs and defined as the sharing of information about learners between faculty supervisors (Humphrey-Murto et al., 2020), LHs are uniquely positioned to mitigate this 'discontinuity' through improved information transfer regarding learner performance (Warm et al., 2017), while also promoting entrustment and autonomy.
Given the frequency of inpatient faculty supervisor transitions, pediatric hospitalists frequently must decide whether and how to handover learners to the oncoming supervisor. However, little is known about the utilization and variability of LHs (Fuchs et al., 2020;Babbott, 2010;Warm et al., 2017;Gumuchian et al., 2020). With pediatric hospitalists' frontline role in pediatric education (Freed, Dunham and Lamarand, 2009;Fromme et al., 2010;Ottolini, 2014), our primary goal was to determine their experiences with and attitudes toward current LH practices during supervisory transitions.

Methods
An anonymous electronic, cross-sectional survey was distributed three times in February and March 2020 to individuals who were members of the American Academy of Pediatrics Section on Hospital Medicine (AAP SOHM) Educator Subcommittee electronic mailing list, a self-identified group of pediatric hospitalists with an interest in medical education. This Subcommittee was selected with the expectation that it would represent a group of pediatric hospitalists invested in the education of medical trainees and would express a diverse range of perspectives regarding LHs. As of February 2020, the online directory listed 584 unique email addresses. The survey was determined to be exempt by the Children's Hospital of Boston institutional review board.

Survey Development
The authors, three of whom have extensive educational and pediatric inpatient medicine experience, drafted a survey including multiple-choice, Likert-scale, and free-text questions focusing on experiences/practices and attitudes towards LHs, along with baseline demographic data (Supplementary File 1: Learner Handoff Survey). The survey instrument was developed based on literature review, including a recent survey on medical student educational handoffs (Fuchs et al., 2020). The items were then developed through an iterative process among the authors and piloted via cognitive interviews with four pediatric educators at two institutions prior to inclusion in the final survey.

Data Analysis
Statistical analyses were performed with the electronic software Qualtrics (Qualtrics, Provo, UT). Standard descriptive summaries were analyzed and comparisons of categorical variables were examined with χ 2 -tests. Authors MF and BF used an inductive approach to analyze narrative responses to the free-text items regarding benefits and downsides to LHs. Emerging themes were discussed and revised until a consensus was achieved.

Results/Analysis
Response Rate There were 129 responses out of 584 eligible providers (22% response rate). There were 65 unique hospital programs reported, with 54 unique medical school affiliations, representing 35% of 155 total Association of American Medical Colleges programs, as well as all regions of the Association of Pediatric Program Directors.
The majority of respondents (74%) reported a formal educational role (including UME and GME roles; see Supplementary File 1), as well as primarily practicing in a university setting (86%). There was a relatively wide distribution with regard to years in practice and weeks per year spent on service with learners. Demographics are shown in Table 1.  Table 2. The majority of respondents (78%) reported familiarity with the concept of LHs. Only 30% reported that it was an expectation of their group of inpatient attendings. Even fewer (15%) reported using a structured format at least some of the time. When used, the most common structure included their institutional evaluation format (7/15; 47%). Educators reported rarely receiving formal training regarding LHs (10%). Educators identified "learner's performance" (100/107; 93%) as most commonly included in their LHs, followed by "learner ability to receive/incorporate feedback" (89/107; 83%), "self-identified learner goals" (80/107; 75%), and ACGME milestones (8/107; 7%).
Almost half (47%) of the respondents reported giving LHs to the oncoming supervisor >75% of the time, while less than one-third (30%) reported receiving LHs from the outgoing supervisor >75% of the time. The vast majority (92%) reported that the handoff of each additional learner added less than 5 minutes to their overall handoff, with almost two-thirds of those respondents (62%) reporting that each additional learner adds less than 2 minutes.

Learner Handoff Comments
Over half of all participants provided responses to the free-text items regarding benefits and downsides to LHs (56% and 60%, respectively). Seven themes for benefits of LHs were identified in descending order of frequency as follows: understanding learner capabilities, identifying learner goals/supervisor expectations, optimizing feedback, augmenting preparation, improving continuity, accelerating entrustment, and creating learner-specific interventions. Two themes for challenges regarding LHs were identified in descending order of frequency as follows: bias and time constraints. Themes and representative quotes are presented in Table 3.

Discussion
Previous studies have shown increasing interest in LHs as a means to improve learner education (Fuchs et al., 2020;Babbott, 2010). However, there is still little known about their utilization and perceived effectiveness in practice. Our study, the first of front-line pediatric inpatient educators, found that the vast majority of respondents felt that LHs improved the delivery of accurate feedback and the ability to tailor teaching to the learner. Additionally, almost half of our respondents reported an improvement in delivering patient care. To the best of our knowledge, this study is the first to report specific practical benefits of LHs within pediatric residency training. And while further studies are needed to expand on this, the potential that LHs may have an impact on clinical patient care aligns with the notion that optimal medical education can and should "improve learner outcomes in order to improve patient outcomes" (Babbott, 2010). These findings are in line with a recent single institution multi-specialty study in Canada of clinical supervisors that reported potential benefits of LHs in tailored teaching, improved assessments, and enhanced patient safety (Gumuchian et al., 2020). Our data adds further validity and generalizability to these findings, while also quantifying specific practices of educators with regard to delivery, receipt, and institutional practices towards LHs.
Despite these overwhelmingly positive perceptions, LHs remain infrequently utilized, with less than one-third of respondents reporting regularly receiving LHs from the oncoming attending during clinical handoffs. Additionally, less than one-third of our respondents reported that LHs were a formal expectation of their group of faculty. Despite sampling pediatric educators, who have likely received more faculty development than their non-educator colleagues, a minority reported using a structured format for their LHs, with even fewer having received any form of formal training. While our study shows that LHs are being used more frequently when compared to a 2016 survey of pediatric undergraduate medical student educators (Fuchs et al., 2020), our results further suggest that few individuals or faculty groups currently employ a structured or coordinated method for learner handovers.
In a recent set of semi-structured faculty interviews, Humphrey-Murto and colleagues (2020) found that LHs occur both formally and informally, serving multiple purposes for both the supervisor and the learner. Despite reporting being primarily motivated by learner benefit and patient safety, faculty also reported that LHs informally served a therapeutic purpose, allowing a place to gossip or vent about learners and their performance. Additionally, a previous review regarding forward feeding of information found an assimilation effect with prior performance information influencing ratings of current performance (Humphrey-Murto et al., 2019). These findings contribute to the previous debate in the literature regarding the risks of sharing information, specifically with regards to the struggling learner (Pangaro, 2008). Namely, a concern for forward feeding bias, or that sharing information about trainees in the clinical setting may positively or negatively influence the oncoming supervisor, is a large barrier to LH utilization (Gumuchian et al., 2020).
Respondents in our study echoed similar concerns, tempering the demonstrated positive perceptions with concerns for added bias, with one educator stating that LHs "may bias our assessment of student depending on colleague's impressions, which may not be structured or based on milestones." However, the potential benefits to learners, patients, and even faculty (Humphrey-Murto et al., 2020) from LHs might serve to counterbalance these fears (Fuchs et al., 2020;Pangaro, 2008). In fact, a recent randomized controlled study of clinical supervisors demonstrated that the sharing of information between supervisors improved targeted feedback without influencing scores (Dory et al., 2020). Additionally, differences in the vocabulary and expectations of supervisors may be a large contributor to these concerns (Pangaro, 2008). Most importantly, the variability in intention and purpose of LHs, both from the delivery and receiving perspective, allows for significant subjectivity and potential bias.
The significant lack of expectations, structure, and formal training surrounding LHs may negatively contribute to these fears. The development of a structured LH tool, in conjunction with faculty development, could serve to mitigate or avoid forward feeding bias, while retaining and enhancing the demonstrated benefits by creating a shared mental model. With regard to the challenge of time constraints, we found that each additional learner contributes minimal added time to the total clinical handoff, with the majority reporting each learner taking less than 2 minutes to handoff.
Continuity, a centerpiece for clinical education reform (Hirsh et al., 2007), has become challenging, with educators reporting concerns with uncoordinated feedback (Trowbridge et al., 2019;Kilminster and Jolly, 2000), delayed skill acquisition (Babbott, 2010), delayed entrustment (Cleary, 2008;Fuchs et al., 2020), lack of targeted teaching, and arrested professional development (Schumaker et al., 2014;Warm et al., 2017) due to frequent transitions. With hospitalists often responsible for general inpatient services at a majority of training programs (Freed, Dunham and Lamarand, 2009), our narrative data suggests that LHs may be uniquely suited to ameliorate these concerns, with specific benefits in understanding learner capabilities, identifying learner goals, improving feedback, accelerating entrustment, and increasing continuity.
Just as a structured, shared mental model for patient handoffs has been shown to improve patient care by minimizing communication failures and decreasing coordination costs (Becker and Murphy, 1992;Arora et al., 2008;Konkin and Suddards, 2017), a structured, shared mental model for learner handoffs has the potential to do the same, both for learner outcomes, and, ultimately, patient outcomes (Warm et al., 2017). As such, guided by previous literature and models (Starmer et al., 2012;Fuchs et al., 2020;Babbott ,2010;Warm et al., 2017 should incorporate: (1) identified capabilities and competencies of the learner to facilitate entrustment and aid in the preparation of the oncoming supervisor, and (2) self-identified learner goals in order to establish connections and shared goals to foster coaching, help tailor teaching, and give accurate feedback. By creating a consistent and more objective tool, we also hope to focus the polymotivation for engaging in LHs (Humphrey-Murto et al., 2020) as well as to limit barriers to honest feedback (Pangaro, 2008) and thus minimize any potential negative consequences of sharing information. The role of the leaner in the handoff process, as well as who should have access to the LH information, must also be determined, echoing comments found in the recent literature (Gumuchian et al., 2020).
Our study was limited by response rate, which could allow for participation and recall bias. However, our responding population represents a diverse group based on geography, educational role, and program/hospital affiliations. These biases could be amplified by the decision to survey a population of pediatric hospital medicine educators, potentially limiting generalizability to other hospitalists. Pediatricians who primarily practiced in the outpatient setting were not included and could potentially have unique perspectives regarding trainee transitions in the ambulatory context. However, the sample population was chosen specifically to elucidate attitudes and practices of those at the frontline of inpatient medical education. Finally, our survey focused on "learners on the medical team", and thus is not able to specifically comment on any potential differences between medical students or residents. In line with this, and similar to several recent studies (Fuchs et al., 2020;Humphrey-Murto et al., 2020), our findings did not include learner input, a crucial perspective on the role and impact of learner handoffs on their learning and professional development.

Conclusion
Our study builds on previous work and adds further data regarding the attitudes and practices of inpatient pediatric educators regarding LHs. Despite our findings that LHs often remain unexpected, unstructured, and provided by untrained supervisors, respondents reported overwhelmingly positive benefits of LHs with regards to teaching and feedback, with potential positive impact on patient care. This strongly suggests that a thoughtfully crafted LH tool, in conjunction with faculty buy-in and appropriate training on mitigating bias, may serve to promote professional development by improving continuity in supervision and thus mitigating uncoordinated feedback, delayed entrustment, and forward feeding bias, as well as strengthening faculty-trainee relationships and potentially enhancing clinical performance. Further studies are needed to delineate the essential components of a structured LH tool, describe implementation strategies, direct faculty development, and understand the impact on learners and patients.

Take Home Messages
Learner handoffs can improve directed feedback and targeted teaching towards learners Learner handoffs add minimal time to a clinical handoff There is concern about possible 'forward feeding' bias with handoffs Learner handoffs currently are unexpected, unstructured, and provided by untrained supervisors A structured learner handoff has the potential to improve trainee professional development while mitigating concerns for bias