Andrew J. Halvorsen

Background Internal medicine (IM) resident physicians spend a considerable amount of time managing their inbox as part of their longitudinal continuity clinic experience. There are no standardized guidelines for how programs should train, monitor, or supervise residents in this type of patient care. Objective To understand how IM residency programs educate, monitor, and supervise resident electronic health record (EHR) inbox management as part of their longitudinal continuity clinic and determine whether patient safety events have occurred due to EHR inbox-related patient care decisions made by unsupervised resident physicians. Methods In August 2021, 439 program directors at accredited US IM residency programs who were members of the Association of Program Directors in Internal Medicine (APDIM) were asked 12 questions developed by the study authors and APDIM survey committee members regarding resident EHR inbox management as part of the annual APDIM survey. Results Two hundred and sixty-seven (61%) PDs responded. The majority (224 of 267, 84%) of programs provided guidelines for expected message response times; less than half (115, 43%) monitored timeliness metrics. Only half (135; 51%) of programs required faculty supervision of inbox messages for all residents; 28% (76) did not require supervision for any residents. Twenty-one percent of PDs (56) reported awareness of a patient safety event occurring due to an unsupervised resident inbox-related patient care decision. Conclusions Substantial variability exists in how IM residency programs train, monitor, supervise, and provide coverage for resident inbox work. Program directors are aware of patient safety events resulting from unsupervised resident inbox management.

Background: The unique traits of residents who matriculate into subspecialty fellowships are poorly understood. We sought to identify characteristics of internal medicine (IM) residents who match into cardiovascular (CV) fellowships. Methods: We conducted a retrospective cohort study of 8 classes of IM residents who matriculated into residency from 2007 to 2014. The primary outcome was successful match to a CV fellowship within 1 year of completing IM residency. Independent variables included residents’ licensing exam scores, research publications, medical school reputation, Alpha Omega Alpha (AOA) membership, declaration of intent to pursue CV in the residency application personal statement, clinical evaluation scores, mini-clinical evaluation exercise scores, in-training examination (ITE) performance, and exposure to CV during residency. Results: Of the 339 included residents (59% male; mean age 27) from 120 medical schools, 73 (22%) matched to CV fellowship. At the time of residency application, 104 (31%) had ≥1 publication, 38 (11%) declared intention to pursue CV in their residency application personal statement, and 104 (31%) were members of AOA. Prior to fellowship application, 111 (33%) completed a CV elective rotation. At the completion of residency training, 108 (32%) had ≥3 publications. In an adjusted logistic regression analysis, declaration of intention to pursue CV (OR 6.4, 99% CI 1.7–23.4; p < 0.001), completion of a CV elective (OR 7.3, 99% CI 2.8–19.0; p < 0.001), score on the CV portion of the PGY-2 ITE (OR 1.05, 99% CI 1.02–1.08; p < 0.001), and publication of ≥3 manuscripts (OR 4.7, 99% CI 1.1–20.5; p = 0.007) were positively associated with matching to a CV fellowship. Overall PGY-2 ITE score was negatively associated (OR 0.93, 99% CI 0.90–0.97; p < 0.001) with matching to a CV fellowship. Conclusions: Residents’ matriculation into CV fellowships was associated with declaration of CV career intent, completion of a CV elective rotation, CV medical knowledge, and research publications during residency. These findings may be useful when advising residents about pursuing careers in CV. They may also help residents understand factors associated with a successful match to a CV fellowship. The negative association between matching into CV fellowship and overall ITE score may indicate excessive subspecialty focus during IM residency.

Introduction: Learners may subconsciously change their behavior once they know they are being observed, and this Hawthorne effect should be considered when designing assessments of learner behavior. While there is a growing body of literature to suggest direct observation is the ideal standard for formative assessment, the best method to directly observe learners is unknown. We explored scheduled and unscheduled methods of direct observation among internal medicine residents in the outpatient continuity clinic to advance the understanding of both observation methods. Methods: We conducted a thematic analysis of faculty and internal medicine residents in an outpatient clinic setting. A semi-structured interview guide for focus group sessions was created. Focus groups were used to explore the internal medicine resident and core teaching faculty perceptions of the scheduled and unscheduled direct observation methods in the outpatient clinc. An experienced qualitative research interviewer external to the internal medicine residency was moderating the sessions. Eight peer focus groups were held. Abstraction of themes from focus group transcripts identified resident and faculty perceptions of the different observation methods. Results: Focus groups had 14 resident participants and 14 faculty participants. Unscheduled observations were felt to be more authentic than scheduled observations since residents perceived their behavior to be unmodified. Unscheduled observations allowed for increased numbers of observations per resident, which permitted more frequent formative assessments. Residents and faculty preferred remote video observation compared to in-room observation. Participants found direct observation a useful learning tool for high-yield, specific feedback. Conclusions: Unscheduled remote direct observation captures authentic clinical encounters while minimizing learner behavior modification. An unscheduled observation approach results in more frequent formative assessment and therefore in more instances of valuable feedback compared to scheduled observations. These findings can help guide the best practice approaches to direct clinical observation in order to enhance residents learning and experience.

Theory: Networking has positive effects on career development; however, personal characteristics of group members such as gender or diversity may foster or hinder member connectedness. Social network analysis explores interrelationships between people in groups by measuring the strength of connection between all possible pairs in a given network. Social network analysis has rarely been used to examine network connections among members in an academic medical society. This study seeks to ascertain the strength of connection between program directors in the Association of Program Directors in Internal Medicine (APDIM) and its Education Innovations Project subgroup and to examine possible associations between connectedness and characteristics of program directors and programs. Hypotheses: We hypothesize that connectedness will be measurable within a large academic medical society and will vary significantly for program directors with certain measurable characteristics (e.g., age, gender, rank, location, burnout levels, desire to resign). Method: APDIM program directors described levels of connectedness to one another on the 2012 APDIM survey. Using social network analysis, we ascertained program director connectedness by measuring out-degree centrality, in-degree centrality, and eigenvector centrality, all common measures of connectedness. Results: Higher centrality was associated with completion of the APDIM survey, being in a university-based program, Educational Innovations Project participation, and higher academic rank. Centrality did not vary by gender; international medical graduate status; previous chief resident status; program region; or levels of reported program director burnout, callousness, or desire to resign. Conclusions: In this social network analysis of program directors within a large academic medical society, we found that connectedness was related to higher academic rank and certain program characteristics but not to other program director characteristics like gender or international medical graduate status. Further research is needed to optimize our understanding of connection in organizations such as these and to determine which strategies promote valuable connections.

Purpose: To begin to quantify and understand the use of the flipped classroom (FC)—a progressive, effective, curricular model—in internal medicine (IM) education in relation to residency program and program director (PD) characteristics. Method: The authors conducted a survey that included the Flipped Classroom Perception Instrument (FCPI) in 2015 regarding programs’ use and PDs’ perceptions of the FC model. Results: Among the 368 IM residency programs, PDs at 227 (61.7%) responded to the survey and 206 (56.0%) completed the FCPI. Regarding how often programs used the FC model, 34 of the 206 PDs (16.5%) reported “never”; 44 (21.4%) reported “very rarely”; another 44 (21.4%) reported “somewhat rarely”; 59 (28.6%) reported “sometimes”; 16 (7.8%) reported “somewhat often”; and 9 (4.4%) reported “very often.” The mean FCPI score (standard deviation [SD]) for the in-class application factor (4.11 [0.68]) was higher (i.e., more favorable) than for the preclass activity factor (3.94 [0.65]) ( P < .001). FC perceptions (mean [SD]) were higher among younger PDs (≤ 50 years, 4.12 [0.62]; > 50 years, 3.94 [0.61]; P = .04) and women compared with men (4.28 [0.56] vs. 3.91 [0.62]; P < .001). PDs with better perceptions of FCs had higher odds of using FCs (odds ratio, 4.768; P < .001). Conclusions: Most IM programs use the FC model at least to some extent, and PDs prefer the interactive in-class components over the independent preclass activities. PDs who are women and younger perceived the model more favorably.

Alignment between institutions and graduate medical education (GME) regarding quality and safety initiatives (QI) has not been measured. The objective was to determine US internal medicine residency program directors’ (IM PDs) perceived resourcing for QI and alignment between GME and their institutions. A national survey of IM PDs was conducted in the Fall of 2013. Multivariable linear regression was used to test association between a novel Integration Score (IS) measuring alignment between GME and the institution via PD perceptions. The response rate was 72.6% (265/365). According to PDs, residents were highly engaged in QI (82%), but adequate funding (14%) and support personnel (37% to 61%) were lower. Higher IS correlated to reports of funding for QI (76.3% vs 54.5%, P = .012), QI personnel (67.3% vs 41.1%, P < .001), research experts (70.5% vs 50.0%, P < .001), and computer experts (69.0% vs 45.8%, P < .001) for QI assistance. Apparent mismatch between GME and institutional resources exists, and the IS may be useful in measuring GME–institutional leadership alignment in QI.

E-learning—the use of Internet technologies to enhance knowledge and performance—has become a widely accepted instructional approach. Little is known about the current use of e-learning in postgraduate medical education. To determine utilization of e-learning by United States internal medicine residency programs, program director (PD) perceptions of e-learning, and associations between e-learning use and residency program characteristics. We conducted a national survey in collaboration with the Association of Program Directors in Internal Medicine of all United States internal medicine residency programs. Of the 368 PDs, 214 (58.2%) completed the e-learning survey. Use of synchronous e-learning at least sometimes, somewhat often, or very often was reported by 85 (39.7%); 153 programs (71.5%) use asynchronous e-learning at least sometimes, somewhat often, or very often. Most programs (168; 79%) do not have a budget to integrate e-learning. Mean (SD) scores for the PD perceptions of e-learning ranged from 3.01 (0.94) to 3.86 (0.72) on a 5-point scale. The odds of synchronous e-learning use were higher in programs with a budget for its implementation (odds ratio, 3.0 [95% CI, 1.04–8.7]; P = .04). Residency programs could be better resourced to integrate e-learning technologies. Asynchronous e-learning was used more than synchronous, which may be to accommodate busy resident schedules and duty-hour restrictions. PD perceptions of e-learning are relatively moderate and future research should determine whether PD reluctance to adopt e-learning is based on unawareness of the evidence, perceptions that e-learning is expensive, or judgments about value versus effectiveness.