Introduction
One of the ubiquitous challenges of clinical training is to maximize the benefits of direct patient contact for learners while minimizing the drawbacks of learner involvement for the healthcare system. Learners can slow down the process of care provision with important consequences for health economics [
1]. In an era of strained healthcare resources, effective educational strategies to address this problem are urgently needed.
For the emergency department (ED) physician working in an academic healthcare centre, this challenge has become paramount. Over the past 20 years the constant tension between the needs of the learner and patient flow has been further compounded by ED overcrowding, which has focused a great deal of institutional and government attention on the issue of patient flow, arguably at the expense of their educational mandate. In a recent Canadian study, Webster et al. [
2]. found that government policies aimed at reducing ED wait times left medical students and junior doctors with the impression that efficiency is more important than education.
A complicated relationship exists between the needs of ED learners and patient flow. While studies report that trainees prolong the length of stay of ED patients [
3‐
5], there is little known about the reciprocal effect of overcrowded EDs on the educational environment. What is known suggests differing perspectives from teachers and learners. Anecdotally, many teachers believe that poor patient flow has been detrimental to learning in the ED. This is supported by a study in which teachers perceived ED overcrowding to limit bedside teaching opportunities [
6], although lack of research on ED teaching prior to overcrowding makes comparative claims impossible [
7]. Concern has been expressed that in crowded EDs, learners may lower their standards for clinical care, patient privacy and professionalism [
8]. However, residents perceived minimal impact on educational value during times of overcrowding, although they saw fewer patients and performed fewer procedures [
9]. Analyses of learner evaluations of ED attending physicians suggested that overcrowding did not change learner satisfaction with teaching [
10]. Additionally, there are suggestions that poor patient flow may present affordances as well as threats; for instance, Shayne et al. [
11] argued that crowded EDs may offer learners in-depth opportunities to focus on multitasking, communication skills, professionalism and system-based competencies.
Because of the complicated relationship between ED learning and patient flow, solutions directed at one of these issues in isolation can have unintended negative consequences on the other. For example, ED Fast Track areas for low acuity patients [
12,
13], which have been shown to positively influence patient flow, often restrict or entirely exclude junior learners, thereby limiting their exposure to the assessment and procedure opportunities associated with this subgroup of patients.
Given the importance of the ED as a learning environment, we contend that it is imperative to develop effective educational solutions that are mindful of the issue of patient flow. ED teaching shifts may represent such a solution. An ED teaching shift is a shift scheduled for the sole purpose of learning, where students and teacher have no patient care responsibilities. The literature describes great variety in the composition and delivery of these teaching shifts [
14‐
16]based upon the needs of a particular programme. However, we have found no study to date examining how such a strategy affects ED throughput or the educational experience.
The aim of this study is to examine how one educational intervention, ED teaching shifts, relates to length of stay, learner self-efficacy and knowledge application. Our study addressed the following three questions:
1.
How do ED teaching shifts relate to emergency department length of stay?
Emergency department length of stay is often used as a metric of efficient ED patient flow and is defined as the time in minutes from arrival at triage or registration until discharge from the ED. It is dependent on timeliness of multiple interrelated elements such as offloading of ambulance patients, triage, nurse and physician assessment, investigations and availability of in-patient hospital beds.
2.
How do ED teaching shifts relate to the self-efficacy of clinical clerks?
Self-efficacy is defined as a personal judgment about how well someone can organize and execute a course of action required to deal with a prospective situation [
17]. Important features of the definition are context specificity and prospective action, which distinguish self-efficacy from other concepts such as self-esteem, self-assessment, self-concept and self-confidence. Increasing levels of perceived self-efficacy give rise to progressively higher performance accomplishment [
18]. Self-efficacy is a meaningful measure for a study of ED student learning, as it can affect rate of skill acquisition and performance mastery, which in turn can boost self-efficacy in a mutually enhancing process [
17].
3.
How do ED teaching shifts relate to knowledge application by clinical clerks?
Not yet at the point of sophisticated clinical reasoning, clinical clerks are challenged to apply the knowledge gained in their limited clinical experience and preclinical years to a variety of undifferentiated emergency patient presentations. Using a well-defined rubric, assessment of a post encounter exercise has shown validity in assessing medical content as well as logic and thought processes [
19].
Discussion
The competing goals of educating medical students while providing timely, quality patient care are felt acutely in the academic ED. New teaching innovations must take into account not only their relation to student education but also to the patient flow processes of the department. A programme that can show success in both areas increases its chances of acceptance and sustainability.
In our study, the overall median length of stay was 5 minutes shorter when clerks were in their teaching sessions and not caring for patients in the ED. This resonates with recent research by Delaney et al. [
3], who found that after adjustments for covariates, length of stay was increased by 9.5 minutes when medical students were involved in patient care. Other research has also shown improvement in length of stay when learners were not present in the ED [
4,
5]. The negative effect of students and residents on throughput is not unique to emergency medicine. Residents in orthopaedics [
23], opthalmology [
24] and anesthesia [
25] have been shown to prolong operating room time and cost. Denton and colleagues [
26] found that one medical student lengthened a half-day general internal medicine clinic by 15 %.
The clinical implications of these improvements in length of stay are difficult to gauge given the approximately 4‑hour median length of stay. It is unlikely that the 5‑minute average reduction would be seen as a meaningful difference to individual patients. However, when this 5‑minute difference is applied to all patients seen during the Monday, Tuesday and Wednesday time periods when clerks were out of the ED, this translates into an annual savings of over 140 hours per month for our two hospital study sites, whose pre-intervention wait times were in the bottom 5 % of the 74 provincial EDs measured (Dreyer J. South West Local Health Integration Network ED Lead, Personal e‑mail communication with L. Shepherd, 30 March 2014.) The 20-minute reduction in block 1 translates into length of stay being shortened by approximately 615 hours per month over this three-month period.
The financial implications of these length of stay reductions are difficult to determine within the Canadian healthcare system. We were unable to acquire reliable cost estimates at the local or provincial level to calculate costs associated with changes in length of stay. However, the cost of running the programme is easily calculated. Four hundred and eighty hours of faculty instruction made up most of the cost of this programme, with programme development being primarily uncompensated. Therefore, it is impossible to argue the case for teaching shifts based solely on a cost/benefit analysis comparing faculty costs with length of stay savings.
The educational implications of the change in ED length of stay are interesting. The block 1 results may reflect the degree of clerk inexperience early in the clerkship year, when naïve clerks require substantial time and effort to integrate into the efficient provision of patient care in any clinical rotation. This problem may be exacerbated in the ED due to patient flow pressures and the lack of a teaching team structure to help absorb the impact of naïve clinical clerks on patient care. Given these factors, some might argue that our results suggest that very junior learners are an unacceptable drag to the system and should be removed from the ED altogether. We contend that this is a slippery slope in clinical training, because a temporal effect of clerks’ impact on patient care has been reported in other settings. For instance, in surgical resident training, Hosler et al. [
24].found that the operative time and cost increased with trainee participation during the first half of the academic year; these increases disappeared in the second half of the year and were attributable to the trainees’ learning curve. Our results show an improvement in length of stay in blocks 2 (non-significant) and 3 (significant) with clinical clerks working in the department, which is consistent with Hosler’s results later in the academic year. Interestingly, this progress did not continue into block 4. One possible explanation would be that clinical clerks in this block, who self-selected this late in the year scheduling, had less interest and inclination towards emergency medicine and therefore worked less efficiently. However, further study is required to understand these changes.
Furthermore, the ED offers a unique learning opportunity that medical students urgently need: an opportunity to assess and manage the range of undifferentiated patient presentations that is unequalled in most other services. With this undergraduate education perspective, we would argue that this unique learning opportunity outweighs the extra time required by clinical clerks in the ED. Rather than removing them to reduce length of stay, we would argue that dedicated teaching shifts offer a viable strategy, particularly early in the clerkship year when clerks are most naïve.
Demonstrating improved ED length of stay will please ED and hospital administrators, but ED educators will also want to understand the implications of teaching shifts in their institutions. Following the teaching shifts, the self-efficacy of clinical clerks significantly increased across each of the four factors identified, with team efficacy demonstrating the most significant improvement. We expect that this was influenced by the simulation exercises in the teaching shifts which served to supplement the students’ prior minimal exposure to active team participation. The importance of demonstrating improved self-efficacy relates to the positive effect on future performance rather than the accuracy of the prediction [
27]. This is supported in the surgical literature with self-efficacy relating to laparoscopic performance in obstetrics and gynaecology residents [
28] and simulator performance among surgical residents [
29]. Second year medical students with higher self-efficacy were more likely to perform better on an observed structured clinical exam (OSCE) than less efficacious students [
30].
Recognizing the undifferentiated nature of ED patient presentations, we were especially interested in examining practical, broad-based student performance and influences.
The assessment of knowledge application had practical origins and applications. Students and faculty both found the paper-based cases a useful exercise in providing a realistic introduction to approaching the non-resuscitative ED patient. This was reflected in the significant improvement of the scores post teaching shifts. Although Boulet’s scoring rubric has been well validated in the international medical graduate cohort applying to United States graduate training programmes [
20], we acknowledge that applying this method of assessment to knowledge application is novel and requires further validation. However, we believe that this application holds promise as evidenced by the qualitative review of the exercise which illustrates clear improvement in these students’ ability to generate a differential diagnosis. There was substantive improvement in both capture of critical diagnosis and increased system inclusion in the generation of differential diagnoses. Proposed investigations and treatments also showed qualitative improvement in the paper cases.
There are several potential limitations of this investigation. It was a single-centre study where medical students spend two weeks in their emergency medicine rotation and length of stay at baseline is high. Future research is required to understand the impact of dedicated teaching shifts in a centre where rotation length is four weeks (common throughout North America) and ED length of stay is lower. As in all studies exploring ED throughput times, it is impossible to control for all of the variables that effect length of stay. By examining data from Mondays, Tuesdays and Wednesdays with the same medical students over two different weeks, we were able to control many variables. However, it is possible that other variables, such as staffing patterns and availability of ancillary and consulting services, would have been different during the weekend and other days of the week, potentially altering the results. Test/retest effect of the self-efficacy questionnaire may have falsely inflated the results. It would have been ideal to conduct a multi-level analysis of these self-efficacy results to examine teacher effect but we were limited by our small group size. Finally, our exploration of knowledge application was a proof-of-concept assessment and requires validation for future use. It was not designed to objectively assess the extent or quality of student learning during the teaching shifts, and therefore this study cannot speak to those important issues. With departmental buy-in secured by this first study’s results, our future research will examine student learning in the context of our new ED clerkship format.