Skip to main content
eScholarship
Open Access Publications from the University of California

Volume 20, Issue 1, 2019

Systematic Review

A Review of Natural Language Processing in Medical Education

Natural language processing (NLP) aims to program machines to interpret human language as humans do. It could quantify aspects of medical education that were previously amenable only to qualitative methods. The application of NLP to medical education has been accelerating over the past several years. This article has three aims. First, we introduce the reader to NLP. Second, we discuss the potential of NLP to help integrate FOAM (Free Open Access Medical Education) resources with more traditional curricular elements. Finally, we present the results of a systematic review. We identified 30 articles indexed by PubMed as relating to medical education and NLP, 14 of which were of sufficient quality to include in this review. We close by discussing potential future work using NLP to advance the field of medical education in emergency medicine.

Editorial

Show Me the Money: Successfully Obtaining Grant Funding in Medical Education

Obtaining grant funding is a fundamental component to achieving a successful research career.A successful grant application needs to meet specific mechanistic expectations of reviewersand funders. This paper provides an overview of the importance of grant funding within medicaleducation, followed by a stepwise discussion of strategies for creating a successful grant applicationfor medical education-based proposals. The last section includes a list of available medicaleducation research grants.

  • 1 supplemental file

Brief Research Report

Evaluation of an Intervention to Improve Quality of Single-best Answer Multiple-choice Questions

Introduction: Despite the ubiquity of single-best answer multiple-choice questions (MCQ) in assessments throughout medical education, question writers often receive little to no formal training, potentially decreasing the validity of assessments. While lengthy training opportunities in item writing exist, the availability of brief interventions is limited.

Methods: We developed and performed an initial validation of an item-quality assessment tool and measured the impact of a brief educational intervention on the quality of single-best answer MCQs.

Results: The item-quality assessment tool demonstrated moderate internal structure evidence when applied to the 20 practice questions (κ=.671, p<.001) and excellent internal structure when applied to the true dataset (κ=0.904, p<.001). Quality scale scores for pre-intervention questions ranged from 2-6 with a mean ± standard deviation (SD) of 3.79 ± 1.23, while post-intervention scores ranged from 4-6 with a mean ± SD of 5.42 ± 0.69. The post-intervention scores were significantly higher than the pre-intervention scores, x2(1) =38, p <0.001.

Conclusion: Our study demonstrated short-term improvement in single-best answer MCQ writing quality after a brief, open-access lecture, as measured by a simple, novel, grading rubric with reasonable validity evidence.

  • 1 supplemental file

Standardized Video Interviews Do Not Correlate to United States Medical Licensing Examination Step 1 and Step 2 Scores

Introduction: In 2017, the Standardized Video Interview (SVI) was required for applicants to emergency medicine (EM). The SVI contains six questions highlighting professionalism and interpersonal communication skills. The responses were scored (6-30). As it is a new metric, no information is available on correlation between SVI scores and other application data. This study was to determine if a correlation exists between applicants’ United States Medical Licensing Examination (USMLE) and SVI scores. We hypothesized that numeric USMLE Step 1 and Step 2 Clinical Knowledge (CK) scores would not correlate with the SVI score, but that performance on the Step 2 Clinical Skills (CS) portion may correlate with the SVI since both test communication skills. 

Methods: Nine EM residency sites participated in the study with data exported from an Electronic Residency Application Service (ERAS®) report. All applicants with both SVI and USMLE scores were included. We studied the correlation between SVI scores and USMLE scores. Predetermined subgroup analysis was performed based on applicants’ USMLE Step 1 and Step 2 CK scores as follows: (≥ 200, 201-220, 221-240, 241-260, >260). We used linear regression, the Kruskal-Wallis test and Mann-Whitney U test for statistical analyses. 

Results: 1,325 applicants had both Step 1 and SVI scores available, with no correlation between the overall scores (p=0.58) and no correlation between the scores across all Step 1 score ranges, (p=0.29). Both Step 2 CK and SVI scores were available for 1,275 applicants, with no correlation between the overall scores (p=0.56) and no correlation across all ranges, (p=0.10). The USMLE Step 2 CS and SVI scores were available for 1,000 applicants. Four applicants failed the CS test without any correlation to the SVI score (p=0.08). 

Conclusion: We found no correlation between the scores on any portion of the USMLE and the SVI; therefore, the SVI provides new information to application screeners.

Educational Advances

Asynchronous Curriculum “Socially Synchronized”: Learning Via Competition

Introduction: Now widespread in emergency medicine (EM) residency programs, asynchronous curriculum (AC) moves education outside of classic classrooms. Our program’s prior AC had residents learning in isolation, achieving completion via quizzes before advancing without the benefit of deliberate knowledge reinforcement. We sought to increase engagement and spaced repetition by creating a social AC using gamification. 

Methods: We created a website featuring monthly options from textbooks and open-access medical education. Residents selected four hours of material, and then submitted learning points. Using these learning points, trivia competitions were created. Residents competed in teams as “houses” during didactic conference, allowing for spaced repetition. Residents who were late in completing AC assignments caused their “house” to lose points, thus encouraging timely completion.

Results: Completion rates prior to deadline are now >95% compared to ~30% before intervention. Surveys show increased AC enjoyment with residents deeming it more valuable clinically and for EM board preparation.

Conclusion: Socially synchronized AC offers a previously undescribed method of increasing resident engagement via gamification.

The Council of Emergency Medicine Residency Directors Academy for Scholarship Coaching Program: Addressing the Needs of Academic Emergency Medicine Educators

Introduction: Didactic lectures remain fundamental in academic medicine; however, many faculty physicians do not receive formal training in instructional delivery. In order to design a program to instill and enhance lecture skills in academic emergency medicine (EM) physicians we must first understand the gap between the current and ideal states.

Methods: In 2012 the Council of Emergency Medicine Residency Directors (CORD) Academy for Scholarship designed a novel coaching program to improve teaching skills and foster career development for medical educators based on literature review and known teaching observation programs. In order to inform the refinement of the program, we performed a needs assessment of participants. Participants’ needs and prior teaching experiences were gathered from self-reflection forms completed prior to engaging in the coaching program. Two independent reviewers qualitatively analyzed data using a thematic approach.

Results: We analyzed data from 12 self-reflection forms. Thematic saturation was reached after nine forms. Overall inter-rater agreement was 91.5%. We categorized emerging themes into three domains: participant strengths and weaknesses; prior feedback with attempts to improve; and areas of desired mentorship. Several overlapping themes and subthemes emerged including factors pertaining to the lecturer, the audience/learner, and the content/delivery.

Conclusion: This study identified several areas of need from EM educators regarding lecture skills. These results may inform faculty development efforts in this area. The authors employed a three-phase, novel, national coaching program to meet these needs.

  • 1 supplemental PDF
  • 1 supplemental file

Recommendations from the Council of Emergency Medicine Residency Directors: Osteopathic Applicants

The Council of Emergency Medicine Residency Directors (CORD) Advising Students Committee(ASC-EM) has previously published student advising recommendations for general emergencymedicine (EM) applicants in an effort to disseminate standardized information to students andpotential advisors. As the shift to a single graduate medical education system occurs by 2020,osteopathic students will continue to represent a larger portion of matched EM applicants, but datashows that their match rate lags that of their allopathic peers, with many citing a lack of access toknowledge EM advisors as a major barrier. Based on available data and experiential information, asub-group of ASC-EM committee sought to provide quality, evidence-based advising resources forstudents, their advisors, and medical leadership. The recommendations advise osteopathic studentsto seek early mentorship and get involved in EM-specific organizations. Students should take Step 1of the United States Medical Licensing Exam and complete two EM rotations at academic institutionsto secure two Standardized Letters of Evaluation and consider regional and program-specific data onpercentage of active osteopathic residents.

Yogaman: An Inexpensive, Anatomically-detailed Chest Tube Placement Trainer

Introduction: Opportunities for chest tube placement in emergency medicine training programs have decreased, making competence development and maintenance with live patients problematic. Available trainers are expensive and may require costly maintenance.

Methods: We constructed an anatomically-detailed model using a Halloween skeleton thorax, dress form torso, and yoga mat. Participants in a trial session completed a survey regarding either their comfort with chest tube placement before and after the session or the realism of Yogaman vs. cadaver lab, depending on whether they had placed <10 or 10 or more chest tubes in live patients.

Results: Inexperienced providers reported an improvement in comfort after working with Yogaman, (comfort before 47 millimeters [mm] [interquartile ratio {IQR}, 20-53 mm]; comfort after 75 mm [IQR, 39-80 mm], p=0.01). Experienced providers rated realism of Yogaman and cadaver lab similarly (Yogaman 79 mm [IQR, 74-83 mm]; cadaver lab 78 mm [IQR, 76-89 mm], p=0.67). All evaluators either agreed or strongly agreed that Yogaman was useful for teaching chest tube placement in a residency program.

Conclusion: Our chest tube trainer allowed for landmark identification, tissue dissection, pleura puncture, lung palpation, and tube securing. It improved comfort of inexperienced providers and was rated similarly to cadaver lab in realism by experienced providers. It is easily reusable and, at $198, costs a fraction of the price of available commercial trainers.

Brief Educational Advances

Intern Passport: Orienting New Travelers to the Emergency Department

The objective of the Intern Passport (IP) curriculum was to implement a structured orientation for incoming interns that effectively defined and distinguished various personnel and assets within the emergency department (ED). The method of training was an on-the-job orientation that required interns to obtain “stamps” (signatures) on their passports during visits to eight “countries” (specialists) within the ED. Topics covered during the visit included introductions, tasks and capabilities, expectations, and pearls and pitfalls. Interns obtained stamps after spending 30-minute orientation visits with each country during the first four-week rotation of internship. The ED countries visited were Adult Nursing, Pediatric Nursing, Orthopedics Technician, Respiratory Therapy, Pharmacy, Psychiatry, Observation, and Radiology. Effectiveness was assessed by participant completion of an optional anonymous retrospective survey. The IP was a beneficial addition to our intern orientation curriculum. It effectively defined and distinguished various personnel and assets within the ED.

 

Asteroids® and Electrocardiograms: Proof of Concept of a Simulation for Task-Switching Training

Introduction: Emergency physicians are interrupted during patient care with such tasks as reading electrocardiograms (ECGs). This phenomenon is known as task-switching which may be a teachable skill. Our objective was to evaluate the potential of a video game for simulating the cognitive demands required of task-switching.

Methods: Emergency medicine residents took a pretest on ECG interpretation and then a posttest while attending to a video game, Asteroids®.

Results: The 35 residents (63%) who participated, scored worse on the ECG posttest then they did on the pretest (p<.001; effect size=1.14). There were no differences between genders or training level.

Conclusion: Interpreting ECGs while playing the Asteroids® game significantly lowered ECG interpretation scores. This shows the potential of this activity for training residents in task-switching ability. The next phase of research will test whether ECG reading performance while task-switching improves with practice.

Implementation of a Departmental Female Emergency Medicine Physician Group

Gender disparities exist in academic emergency medicine (EM). We developed and implemented a female EM physician group – Women in Academic Emergency Medicine (WAM) – to support female EM residents, fellows, and faculty. The goal of WAM is to provide a support system through mentorship, education, and outreach. A targeted needs assessment was completed to identify goals and objectives specific to our department. In the first full year of implementation, WAM hosted eight events, including three topical dinners and one formal panel. Of 42 female faculty and residents, 40 (95%) attended at least one WAM event, and all (20/20) of the female faculty strongly supported WAM. WAM advocated for increased female physician representation on the department’s Physician Executive Leadership Group and preservation of dedicated lactation space in the emergency department. Using a needs assessment, the process of developing WAM can be replicated in any department to create a female physician group.

  • 1 supplemental file

Original Research

Randomized Controlled Trial of Simulation vs. Standard Training for Teaching Medical Students High-quality Cardiopulmonary Resuscitation

Introduction: Most medical schools teach cardiopulmonary resuscitation (CPR) during the final year in course curriculum to prepare students to manage the first minutes of clinical emergencies. Little is known regarding the optimal method of instruction for this critical skill. Simulation has been shown in similar settings to enhance performance and knowledge. We evaluated the comparative effectiveness of high-fidelity simulation training vs. standard manikin training for teaching medical students the American Heart Association (AHA) guidelines for high-quality CPR. 

Methods: This was a prospective, randomized, parallel-arm study of 70 fourth-year medical students to either simulation (SIM) or standard training (STD) over an eight-month period. SIM group learned the AHA guidelines for high-quality CPR via an hour session that included a PowerPoint lecture with training on a high-fidelity simulator. STD group learned identical content using a low-fidelity Resusci Anne® CPR manikin. All students managed a simulated cardiac arrest scenario with primary outcome based on the AHA guidelines definition of high-quality CPR (specifies metrics for compression rate, depth, recoil, and compression fraction). Secondary outcome was time to emergency medical services (EMS) activation. We analyzed data via Kruskal-Wallis rank sum test. Outcomes were performed on a simulated cardiac arrest case adapted from the AHA Advanced Cardiac Life Support (ACLS) SimMan® Scenario manual.

Results: Students in the SIM group performed CPR that more closely adhered to the AHA guidelines of compression depth and compression fraction. Mean compression depth was 4.57 centimeters (cm) (95% confidence interval [CI] [4.30-4.82]) for SIM and 3.89 cm (95% CI [3.50-4.27]) for STD, p=0.02.  Mean compression fraction was 0.724 (95% CI [0.699-0.751]) for SIM group and 0.679 (95% CI [0.655-0.702]) for STD, p=0.01. There was no difference for compression rate or recoil between groups. Time to EMS activation was 24.7 seconds (s) (95% CI [15.7-40.8]) for SIM group and 79.5 s (95% CI [44.8-119.6]) for STD group, p=0.007. 

Conclusion: High-fidelity simulation training is superior to low-fidelity CPR manikin training for teaching fourth-year medical students implementation of high-quality CPR for chest compression depth and compression fraction.

Behind the Curtain: The Nurses’ Voice in Assessment of Residents in the Emergency Department

Introduction: Feedback provides valuable input for improving physician performance. Conventionally, feedback is obtained from attending physicians; however, residents work in close contact with other members of the care team, especially nurses. Nurses may have more opportunity to directly observe trainees. In addition, they may value different behaviors and provide unique feedback. The objective of this study was to examine the nurse’s perspective of resident performance in the emergency department.

Methods: This was a retrospective, mixed-methods study of nursing assessments of residents using a  five-point scale from 1 (unsatisfactory) to 5 (outstanding) and providing comments. Analysis included descriptive statistics of the quantitative assessments and content analysis of the nursing comments by a group of attendings, residents, and nurses.

Results: Nurses assessed residents as above expectation or outstanding, especially for the categories of “How would you rate this resident’s attitude?” (65%) and “Is this resident a team player?” (64%). Content analysis of the comments yielded nine themes including being kind, communication with nurses, being a team player, work ethic and efficiency, and respect for other team members. Of the comments made, 50% provided positive feedback, and the majority of comments (80%) were determined to be actionable.

Conclusion: Our data indicate that nurses provide feedback on residents’ kindness, efficiency and communication. These two aspects of interacting in the healthcare setting may not be highlighted in conventional, attending provider feedback, yet they are clearly noted by the nurse’s voice.

The Impact of a Standardized Checklist on Transition of Care During Emergency Department Resident Physician Change of Shift

Introduction: Transitions of patient care during physicians’ change of shift introduce the potential for critical information to be missed or distorted, resulting in possible morbidity. The Joint Commission, the Accreditation Council for Graduate Medical Education, and the Society of Hospital Medicine jointly encourage a structured format for patient care sign-out. This study’s objective was to examine the impact of a standardized checklist on the quality of emergency medicine (EM) resident physicians’ patient-care transition at shift change.

Methods: Investigators developed a standardized sign-out checklist for EM residents to complete prior to sign out. This checklist included topics of diagnoses, patient-care tasks to do, patient disposition, admission team, and patient code status. Two EM attending physicians, the incoming and departing, assessed the quality of transitions of care at this shift change using a standardized assessment form. This form also assessed overall quality of sign-out using a visual analog scale (VAS), based on a 10-centimeter scale. For two months, we collected initial, status quo data (pre-checklist [PCL] cohort) followed by two months of residents using the checklist (post-checklist [CL] cohort).

Results: We collected data for 77 days (July 1, 2015 – November 11, 2015), 38 days of status quo sign-out followed by 39 days of checklist utilization, comprised of 1,245 attending assessments.  Global assessment of sign-out for the CL was 8 compared to 7.5 for the PCL. Aspects of transition of care that implementation of the sign-out checklist impacted included the following (reported as a frequency): “To Do” (PCL 84.3%, CL 97.8%); “Disposition” (PCL 97.2%, CL 99.4%); “Admit Team” (67.1%, CL 76.2%); and “Attending Add” (PCL 23.4%, CL 11.3%).

Conclusion: Implementation of a sign-out checklist enhanced EM resident physician transition of care at shift end by increasing the frequency of discussion of critical tasks remaining for patient care, disposition status, and subjective assessment of quality of sign-out.

Integration of Entrustable Professional Activities with the Milestones for Emergency Medicine Residents

Introduction: Medical education is moving toward a competency-based framework with a focus on assessment using the Accreditation Council for Graduate Medical Education Milestones. Assessment of individual competencies through milestones can be challenging. While competencies describe characteristics of the person, the entrustable professional activities (EPAs) concept refers to work-related activities. EPAs would not replace the milestones but would be linked to them, integrating these frameworks. Many core specialties have already defined EPAs for resident trainees, but EPAs have not yet been created for emergency medicine (EM). This paper describes the development of milestone-linked EPAs for EM.

Methods: Ten EM educators from across North America formed a consensus working group to draft EM EPAs, using a modified Glaser state-of-the-art approach. A reactor panel with EPA experts from the United States, Canada and the Netherlands was created, and an iterative process with multiple revisions was performed based on reactor panel input. Following this, the EPAs were sent to the Council of Residency Directors for EM (CORD-EM) listserv for additional feedback.

Results: The product was 11 core EPAs that every trainee from every EM program should be able to perform independently by the time of graduation. Each EPA has associated knowledge, skills, attitudes and behaviors (KSAB), which are either milestones themselves or KSABs linked to individual milestones. We recognize that individual programs may have additional focus areas or work-based activities they want their trainees to achieve by graduation; therefore, programs are also encouraged to create additional program-specific EPAs.

Conclusion: This set of 11 core, EM-resident EPAs can be used as an assessment tool by EM residency programs, allowing supervising physicians to document the multiple entrustment decisions they are already making during clinical shifts with trainees. The KSAB list within each EPA could assist supervisors in giving specific, actionable feedback to trainees and allow trainees to use this list as an assessment-for-learning tool. Linking each KSAB to individual EM milestones allows EPAs to directly inform milestone assessment for clinical competency committees. These EPAs serve as another option for workplace-based assessment, and are linked to the milestones to create an integrated framework.

  • 2 supplemental files

Defining the “Problem Resident” and the Implications of the Unfixable Problem: The Rationale for a “Front-door” Solution

Introduction: Problem residents are common in graduate medical education, yet little is known about their characteristics, deficits, and the consequences for emergency medicine (EM) residencies. The American Board of Internal Medicine (ABIM) defines a problem resident as “a trainee who demonstrates a significant enough problem that requires intervention by someone of authority, usually the program director [PD] or chief resident.” Although this is a comprehensive definition, it lacks specificity. Our study seeks to add granularity and nuance to the definition of “problem resident,” which can be used to guide the recruitment, selection, and training of residents. 

Methods: We conducted semi-structured interviews with a convenience sample of EM PDs between 2011 and 2012. We performed qualitative analysis of the resulting transcripts with our thematic analysis based on the principles of grounded theory. We reached thematic sufficiency after 17 interviews. Interviews were coded as a team through consensus. 

Results: The analysis identified diversity in the type, severity, fixability, and attribution of problems among problem residents. PDs applied a variety of thresholds to define a problem resident with many directly rejecting the ABIM definition. There was consistency in defining academic problems and some medical problems as “fixable.” In contrast, personality problems were consistently defined as “non-fixable.” Despite the diversity of the definition, there was consensus that residents who caused “turbulence” were problem residents.

Conclusion: The ABIM definition of the problem resident captures trainees who many PDs do not consider problem residents. We propose that an alternative definition of the problem resident would be “a resident with a negative sphere of influence beyond their personal struggle.” This combination acknowledges the identified themes of turbulence and the diversity of threshold. Further, the combination of PDs’ unwillingness to terminate trainees and the presence of non-fixable problems implies the need for a “front-door” solution that emphasizes personality issues at the potential expense of academic potential. This “front-door” solution depends on the commitment of all stakeholders including medical schools, the Association of American Medical Colleges, and PDs.

Development of a Clinical Teaching Evaluation and Feedback Tool for Faculty

Introduction: Formative evaluations of clinical teaching for emergency medicine (EM) faculty are limited. The goal of this study was to develop a behaviorally-based tool for evaluating and providing feedback to EM faculty based on their clinical teaching skills during a shift. 

Methods: We used a three-phase structured development process. Phase 1 used the nominal group technique with a group of faculty first and then with residents to generate potential evaluation items. Phase 2 included separate focus groups and used a modified Delphi technique with faculty and residents, as well as a group of experts to evaluate the items generated in Phase 1. Following this, residents classified the items into novice, intermediate, and advanced educator skills. Once items were determined for inclusion and subsequently ranked they were built into the tool by the investigators (Phase 3). 

Results: The final instrument, the “Faculty Shift Card,” is a behaviorally-anchored evaluation and feedback tool used to facilitate feedback to EM faculty about their teaching skills during a shift. The tool has four domains: teaching clinical decision-making; teaching interpersonal skills; teaching procedural skills; and general teaching strategies. Each domain contains novice, intermediate, and advanced sections with 2-5 concrete examples for each level of performance. 

Conclusion: This structured process resulted in a well-grounded and systematically developed evaluation tool for EM faculty that can provide real-time actionable feedback to faculty and support improved clinical teaching.

  • 1 supplemental PDF

Factors Affecting Entrustment and Autonomy in Emergency Medicine: How much rope do I give them?

Introduction: During residency, the faculty’s role is to provide supervision while granting the trainee autonomy. This concept is termed entrustment. The goal is appropriate progression from supervision to autonomy while decreasing oversight as residents train. The objective of this study was to better understand the factors affecting the degree of autonomy or supervision faculty choose to provide residents.

Methods: This was a qualitative study of resident and faculty perceptions. We conducted two faculty and two resident focus groups. We then transcribed the transcripts of the audiotaped discussions and coded them using grounded theory.

Results: Analysis of the transcripts yielded four major factors affecting entrustment of residents. Patient Factors included the acuity of the patient, sociomedical issues of patient/family, and complexity of risk with patient or procedure. For example, “sometimes there are families and patients who are exceedingly difficult that immediately sort of force me [to allow less autonomy].” Environmental Factors included patient volume and systems protocols (i.e., trauma). “If you’re very busy and you have a resident that you already trust, you will give them more rope because you’re trying to juggle more balls.”Resident Factors included the year of training, resident performance, clinical direct observation, and patient presentations. “But if you have a resident that you do not trust […] I tell them you’re going to do this, this, this, this, this.”Faculty Factors included confidence in his/her own practice, risk-averse attitude, degree of ownership of the patient, commitment to education, and personality (e.g., micro-manager). Significant variability in entrustment by faculty existed, from being “micromanagers” to not seeing the patients. One resident noted: “There are some attendings, no matter how much they like you and how much you’ve worked with them, they’re always going to be in your face in the trauma bay. And there’s some attendings that are going to be ghosts.”

Conclusion: Multiple factors affect the amount of autonomy and entrustment given to residents and their level of supervision by faculty, leading to wide variability in entrustment. In the end, regardless of resident, patient, or environment, some faculty are more likely to entrust than others.

Assessment of Emergency Medicine Resident Performance in an Adult Simulation Using a Multisource Feedback Approach

Introduction: The Accreditation Council for Graduate Medical Education (ACGME) specifically notes multisource feedback (MSF) as a recommended means of resident assessment in the emergency medicine (EM) Milestones. High-fidelity simulation is an environment wherein residents can receive MSF from various types of healthcare professionals. Previously, the Queen’s Simulation Assessment Tool (QSAT) has been validated for faculty to assess residents in five categories: assessment; diagnostic actions; therapeutic actions; interpersonal communication, and overall assessment. We sought to determine whether the QSAT could be used to provide MSF using a standardized simulation case.

Methods: Prospectively after institutional review board approval, residents from a dual ACGME/osteopathic-approved postgraduate years (PGY) 1-4 EM residency were consented for participation. We developed a standardized resuscitation after overdose case with specific 1-5 Likert anchors used by the QSAT. A PGY 2-4 resident participated in the role of team leader, who completed a QSAT as self-assessment. The team consisted of a PGY-1 peer, an emergency medical services (EMS) provider, and a nurse. Two core faculty were present to administer the simulation case and assess. Demographics were gathered from all participants completing QSATs. We analyzed QSATs by each category and on cumulative score. Hypothesis testing was performed using intraclass correlation coefficients (ICC), with 95% confidence intervals. Interpretation of ICC results was based on previously published definitions.

Results: We enrolled 34 team leader residents along with 34 nurses. A single PGY-1, a single EMS provider and two faculty were also enrolled. Faculty provided higher cumulative QSAT scores than the other sources of MSF. QSAT scores did not increase with team leader PGY level. ICC for inter-rater reliability for all sources of MSF was 0.754 (0.572-0.867). Removing the self-evaluation scores increased inter-rater reliability to 0.838 (0.733-0.910). There was lesser agreement between faculty and nurse evaluations than from the EMS or peer evaluation.

Conclusion: In this single-site cohort using an internally developed simulation case, the QSAT provided MSF with excellent reliability. Self-assessment decreases the reliability of the MSF, and our data suggest self-assessment should not be a component of MSF. Use of the QSAT for MSF may be considered as a source of data for clinical competency committees.

  • 1 supplemental file

Qualitative Analysis of Well-being Preparedness at an Emergency Medicine Residency Program

Introduction: There is significant variability in the preparedness of incoming interns at the start of residency training with regard to medical knowledge, procedural skills, and attitudes. Specialty-specific preparatory courses aimed at improving clinical skills exist; however, no preparatory courses targeting wellness promotion or burnout prevention have previously been described. Resident well-being has gained increasing attention from the Accreditation Council for Graduate Medical Education, and numerous studies have demonstrated high levels of burnout among resident physicians. The American Medical Association (AMA) divides resident well-being into the following six categories: nutrition, fitness, emotional health, financial health, preventative care, and mindset and behavioral adaptability. Using the AMA’s conceptual framework for well-being in residency, we performed a targeted needs assessment to support the development of a “pre-residency” well-being curriculum. Our aim was to discover what current residents and faculty felt were the perceived areas of under-preparedness, in relation to resident well-being, for incoming interns at the start of their residency training.

Methods: Using a grounded theory approach, we conducted a series of semi-structured, focus group interviews. Focus groups consisted of junior residents (postgraduate years [PGY] 1-3), senior residents (PGY-4), and current faculty members. A standardized interview guide was used to prompt discussion and themes were identified from audio recording. We modified theories based on latent and manifest content analysis, and we performed member checking and an external audit to improve validity.

Results: Participants noted variable exposure to both formal and informal well-being training prior to residency. Regardless, participants uniformly agreed that their past experiences did not adequately prepare them for the challenges, specific to burnout prevention, faced during residency training. Of the six domains of resident well-being described by the AMA, emotional health, mindset and behavioral adaptability, and financial health were the domains most cited for interns to be underprepared for at the start of residency training.

Conclusion: Despite variability in prior medical school and life experiences, incoming interns were underprepared in several domains of well-being, including emotional health, mindset and behavioral adaptability, and financial health. Targeted interventions toward these areas of well-being should be piloted and studied further for their potential to mitigate effects of burnout among resident physicians.

  • 2 supplemental files

Predictors of an Initial Position in Emergency Medicine

Introduction: Each year, emergency medicine (EM) residency graduates enter a variety of community and academic positions. For some training programs, the potential for an academic career is a consideration during the interview process; however, no studies have looked at factors that might predict an academic career. Our goal was to identify variables present during the EM application cycle that predict an initial academic position.

Methods: We retrospectively reviewed application materials from 211 EM graduates at Emory University from 2003-2013. We analyzed biographical variables, board scores, personal statements, and both undergraduate and medical school research experience and publications. An academic position was defined as working at a site with residents rotating in the emergency department, full or part-time appointment at a medical school, or a position with research required for promotion. We used a logistic regression model to determine the impact of these predictors on obtaining an initial academic position.

Results: A total of 79 (37%) graduates initially chose an academic job, and 132 (63%) took a community position. We identified the following statistically significant variables: younger age (odds ratio [OR] [0.79], 95% confidence interval [CI] [0.67-0.93], p=0.01); undergraduate publications (OR [1.41], 95% CI [1.08-1.83], p=0.01); and medical school publications (OR [3.39], 95% CI [1.66-6.94], p<0.001). Of note, mention of an academic career in the personal statement showed no statistical correlation (p = 0.41).

Conclusion: Younger age, and undergraduate and medical school publications were the variables most associated with an initial academic position. As this is a single-institution study, more studies are needed to validate these findings.

 

Assessing Residency Applicants’ Communication and Professionalism: Standardized Video Interview Scores Compared to Faculty Gestalt

Introduction: The Association of American Medical Colleges has introduced the Standardized Video Interview (SVI) to assess the communication and professionalism skills of residency applicants to allow a more holistic view of applicants beyond academic performance. Initial data suggests scores are not correlated with academic performance and provide a new measure of applicant attributes. It is not currently known how the SVI compares to existing metrics for assessing communication and professionalism during the interview process.

Methods: Applicants to the University of Wisconsin Emergency Medicine Residency program were invited and interviewed without use of the SVI scores or videos. All faculty interviewers were blinded to applicants’ SVI information and asked to rate each applicant on their communication and professionalism on a scale from 1-25 (faculty gestalt score), analogous to the 6-30 scoring used by the SVI. We transformed SVI scores to our 1-25 system (transformed SVI score) for ease of comparison and compared them to faculty gestalt scores as well as applicants’ overall score for all components of their interview day (interview score).

Results: We collected data for 125 residency candidates. Each applicant received a faculty gestalt score from up to four faculty interviewers. There was no significant correlation of SVI scores with faculty gestalt scores (Spearman’s rank correlation coefficient [rs] (123)=0.09, p=0.30) and no correlation with the overall interview score (rs(123)=0.01, p=0.93). Faculty gestalt scores were correlated positively with interview scores (rs(123)=0.65, p<0.01).

Conclusion: SVI scores show no significant correlation with faculty gestalt scores of communication and professionalism. This could relate to bias introduced by knowledge of an applicant’s academic performance, different types of questions being asked by faculty interviewers, or lack of uniform criteria by which faculty assess these competencies. Further research is needed to determine whether SVI scores or faculty gestalt correlate with performance during residency.

Does Implementation of a Corporate Wellness Initiative Improve Burnout?

Introduction: Burnout affects over 50% of all physicians. Nearly 70% of emergency physicians are affected, and it has been found to be as high as 76% in resident physicians overall. Previous wellness initiatives have yielded variable results; therefore, we looked for interventions that could potentially be effective at reversing this trend. We explored effective wellness programs originating from other industries. Our objective was to implement a corporate wellness program with previous evidence of success in other healthcare provider populations. We aimed to investigate whether this program would be effective in decreasing burnout in emergency medicine (EM) residents.

Methods: This program was conducted during required EM resident conference hours from 2016-2017. The Maslach Burnout Inventory was completed before and after the series of sessions, and we collected reactions-level data following completion of the six sessions.

Results: Post-intervention scores revealed a small trend toward increased emotional exhaustion and depersonalization scores, and with increased personal accomplishment scores. The overall satisfaction rating for this program was low, at 1.5 on a 5-point scale. Forty-three percent of residents stated that this intervention subjectively worsened their overall burnout, with another 39% stating it did not improve their burnout at all. A similar trend was seen for effects on wellness.

Conclusion: We found that a corporate wellness intervention that had previously been shown to be successful with other types of healthcare providers did not objectively improve burnout and was subjectively perceived as paradoxically worsening burnout for many residents. This result may be related to the type of intervention chosen (individual vs. systems-focused), the design of the intervention itself, or the unique stressors faced by the resident population.

Simulation-Based Remediation in Emergency Medicine Residency Training- A Consensus Study.

Introduction: Resident remediation is a pressing topic in emergency medicine (EM) training programs. Simulation has become a prominent educational tool in EM training and been recommended for identification of learning gaps and resident remediation. Despite the ubiquitous need for formalized remediation, there is a dearth of literature regarding best practices for simulation-based remediation (SBR).

Methods: We conducted a literature search on SBR practices using the terms “simulation,” “remediation,” and “simulation based remediation.” We identified relevant themes and used them to develop an open-ended questionnaire that was distributed to EM programs with experience in SBR. Thematic analysis was performed on all subsequent responses and used to develop survey instruments, which were then used in a modified two-round Delphi panel to derive a set of consensus statements on the use of SBR from an aggregate of 41 experts in simulation and remediation in EM.

Results: Faculty representing 30 programs across North America composed the consensus group with 66% of participants identifying themselves as simulation faculty, 32% as program directors, and 2% as core faculty. The results from our study highlight a strong agreement across many areas of SBR in EM training. SBR is appropriate for a range of deficits, including procedural, medical knowledge application, clinical reasoning/decision-making, communication, teamwork, and crisis resource management. Simulation can be used both diagnostically and therapeutically in remediation, although SBR should be part of a larger remediation plan constructed by the residency leadership team or a faculty expert in remediation, and not the only component. Although summative assessment can have a role in SBR, it needs to be very clearly delineated and transparent to everyone involved.

Conclusion: Simulation may be used for remediation purposes for certain specific kinds of competencies as long as it is carried out in a transparent manner to all those involved.

July Phenomenon Impacts Efficiency of Emergency Care

Introduction: The “July effect” describes the period in which new interns begin learning patient care while senior residents take on additional responsibility in an academic hospital setting. The annual change in staffing creates inefficiencies in patient care, which may negatively impact quality of care. Our objective was to evaluate the impact of the annual resident turnover on emergency department (ED) efficiency in a teaching hospital.

Methods: This was an institutional review board-approved retrospective chart review spanning two academic years analyzing 79,921 records. We grouped July and August into the period of least experience (PLE) and May and June into the period of most experience (PME). Outcomes included faculty and resident productivity, ED door-to-doctor time, and time to disposition.

Results: Patients were evaluated by 117 emergency residents and 73 emergency faculty. We excluded patient records for 35 off-service residents. Residents saw 15.8% more patients in the PME compared to the PLE (p<0.0001). The residents’ average door-to-doctor time during the PLE was 45.63 minutes (standard deviation [SD] 33.01, median 36) compared to 34.69 minutes (SD 25.22, median 28) during the PME, with a decrease in time by 21.3% (p=0.0203). The residents’ average time to disposition during the PLE was 304.6 minutes (SD 308, median 217) compared to 269.0 minutes (SD 282, median 194) during the PME, decreasing by 12.4% (p=0.0001). Residents had an average ED length of stay for discharged patients of 358.5 minutes (SD 374.6, median 238) during the PLE compared to 309.9 minutes (SD 346.4, median 209) during the PME, decreasing 13.7% for discharged patients (p=0.0017).

Conclusion: Annual turnover of resident staffing has a significant impact on common ED efficiency metrics. EDs should consider interventions to mitigate the impact of these expected inefficiencies.

Randomized Evaluation of Videoconference Meetings for Medical Students’ Mid-clerkship Feedback Sessions

Introduction: Videoconferencing has been employed in numerous medical education settings ranging from remote supervision of medical trainees to conducting residency interviews. However, no studies have yet documented the utility of and student response to videoconference meetings for mid-clerkship feedback (MCF) sessions required by the Liaison Committee on Medical Education (LCME).

Methods: From March 2017 to June 2018, third-year medical students rotating through the mandatory, four-week emergency medicine (EM) clerkship at a single medical school were randomly assigned either to a web-based videoconference meeting via Google Hangouts, or to a traditional in-person meeting for their MCF session. To compare students’ MCF experiences we sent out an electronic survey afterward to assess the following using a 0-100 sliding scale: overall satisfaction with the meeting; the effectiveness of communication; the helpfulness of the meeting; their stress levels, and the convenience of their meeting location. The survey also collected data on these demographic variables: the name of the faculty member with whom the student met; student gender, age, and interest in EM; location prior to meeting; meeting-method preference; and number of EM shifts completed.

Results: During the study period, 133 third-year medical students responded to the survey. When comparing survey responses between individuals who met online and in person, we did not detect a difference in demographics with the exception of preferred meeting method (p=0.0225). We found no significant differences in the overall experience, helpfulness of the meeting, or stress levels of the meeting between those who met via videoconference vs. in-person (p=0.9909; p=0.8420; p=0.2352, respectively). However, individuals who met in-person with a faculty member rated effectiveness of communication higher than those who met via videoconference (p=0.0002), while those who met online rated convenience higher than those who met in-person (p<0.0001). Both effects remained significant after controlling for preferred meeting method (p<0.0001 and p=0.0003, respectively) and among EM-bound students (p=.0423 and p<0.0110, respectively).

Conclusion: Our results suggest that LCME-required MCF sessions can be successfully conducted via web-based programs such as Google Hangouts without jeopardizing overall meeting experience. While the convenience of the meetings was improved, it is also important for clerkship directors to note the perceived deficit in the effectiveness of communication with videoconferencing.

 

  • 4 supplemental files

Emergency Physicians’ Familiarity with the Safe Handling of Firearms

Introduction: Emergency physicians (EP) experience high rates of workplace violence, the risks of which increase with the presence of weapons. Up to 25% of trauma patients brought to the emergency department (ED) have been found to carry weapons. Given these risks, we conducted an educational needs assessment to characterize EPs’ knowledge of firearms, frequency of encountering firearms in the ED, and level of confidence with safely removing firearms from patient care settings.

Methods: This was a survey study of attending and resident EPs at two academic and four community hospitals in the Midwest and Northeast. A 26-item questionnaire was emailed to all EPs at the six institutions. Questions pertained to EPs’ knowledge of firearms, experience with handling firearms, and exposure to firearms while at work. We calculated response proportions and p-values.

Results: Of 243 recipients who received the survey, 149 (61.3%) completed it. Thirty-three respondents (22.0%) reported encountering firearms in the workplace, 91 (60.7%) reported never handling firearms, and 25 (16.7%) reported handling firearms at least once per year. Thirty-six respondents (24.0%) reported formal firearms training, and 63 (42.3%) reported no firearms training. There were no significant regional differences regarding firearms training or exposure. Residents from the Northeast were more likely to be moderately confident that they could safely handle a firearm prior to law enforcement involvement (p=0.043), while residents from the Midwest were more likely to be not at all confident (p=0.018).

Conclusion: The majority of surveyed attending and resident EPs reported little experience with handling firearms. Among resident EPs, there was a regional difference in confidence in handling firearms prior to law enforcement involvement. Given the realities of workplace violence and the frequency with which firearms are encountered in the ED, further investigation is needed to evaluate provider competence in safely handling them. EPs may benefit from training on this topic.

  • 1 supplemental PDF

Accuracy Screening for ST Elevation Myocardial Infarction in a Task-switching Simulation

Introduction: Interruptions in the emergency department (ED) are associated with clinical errors, yet are important when providing care to multiple patients. Screening triage electrocardiograms (ECG) for ST-segment elevation myocardial infarction (STEMI) represent a critical interrupting task that emergency physicians (EP) frequently encounter. To address interruptions such as ECG interpretation, many EPs engage in task switching, pausing their primary task to address an interrupting task. The impact of task switching on clinical errors in interpreting screening ECGs for STEMI remains unknown.

Methods: Resident and attending EPs were invited to participate in a crossover simulation trial. Physicians first completed a task-switching simulation in which they viewed patient presentations interrupted by clinical tasks, including screening ECGs requiring immediate interpretation before resuming the patient presentation. Participants then completed an uninterrupted simulation in which patient presentations and clinical tasks were completed sequentially without interruption. The primary outcome was accuracy of ECG interpretation for STEMI during task switching and uninterrupted simulations.

Results: Thirty-five participants completed the study. We found no significant difference in accuracy of ECG interpretation for STEMI (task switching 0.89, uninterrupted 0.91, paired t-test p=0.21). Attending physician status (odds ratio [OR] [2.56], confidence interval [CI] [1.66-3.94], p<0.01) and inferior STEMI (OR [0.08], CI [0.04-0.14], p<0.01) were associated with increased and decreased odds of correct interpretation, respectively. Low self-reported confidence in interpretation was associated with decreased odds of correct interpretation in the task-switching simulation, but not in the uninterrupted simulation (interaction p=0.02).

Conclusion: In our simulation, task switching was not associated with overall accuracy of ECG interpretation for STEMI. However, odds of correct interpretation decreased with inferior STEMI ECGs and when participants self-reported low confidence when interrupted. Our study highlights opportunities to improve through focused ECG training, as well as self-identification of “high-risk” screening ECGs prone to error during interrupted clinical workflow.

  • 1 supplemental file