Comparing Performance Outcomes of Emergency Medicine-Trained vs. Non-Emergency Medicine-Trained Physicians in Emergency Departments

Abstract

Background

Non-emergency medicine (EM)-trained physicians comprise a notable proportion of the emergency medicine workforce in Taiwan and many other countries. The possible performance differences at the emergency department (ED) between EM-trained and non-EM-trained physicians have not been evaluated before.

Methods

This retrospective observational study was conducted between August 2018 and July 2020 at a regional hospital in Taiwan. We compared the two physician groups for quality-of-care outcomes, including waiting time, rate of failing to visit patients within the required time, length of ED stay, admission rate, intensive care unit admission rate, unscheduled return visit, return of spontaneous circulation rate in out-of-hospital cardiac arrest patients, in-hospital cardiac arrest incidence, referral rate, and computed tomography (CT) scan utilization.

Results

A total of 37,013 ED visits were included. When compared to the non-EM-trained physicians, patients managed by the seven EM-trained physicians had shorter waiting time (6.1 min vs. 9.2 min, p < 0.001), shorter ED stay (146.5 min vs. 176.1 min, p < 0.001), lower rate of failing to visit patients within the required time (0.8% vs. 1.1%, p = 0.010), lower unscheduled return visit rate (4.9% vs. 5.4%, p = 0.043), and lower CT scan utilization (0.16 [times/patient/visit] vs. 0.18 [times/patient/visit], p < 0.001).

Conclusion

The EM-trained and non-EM-trained physicians’ performance at a regional hospital ED differed. Our findings could be used as a reference for healthcare policy-makers and hospital management.

Introduction

With the increase in emergency department (ED) visits in Taiwan, from four million in 1998 to 12 million in 2018 based on Ministry of Health and Welfare-National Health Insurance-Statistics & Publications, there is a growing demand for emergency medicine (EM)-trained physicians. However, there is a significant shortage of properly trained and certified EM physicians. EM has become an independent specialty in Taiwan since 1998.

A “Model of the Clinical Practice of Emergency Medicine” was introduced to the EM-training program and has become a recognized feature among other specialties. 1 , 2 The completion of an EM residency program was required for eligibility for EM board certification by the Ministry of Health and Welfare. Before 1998, medical doctors trained in non-EM specialties could become certified EM physicians after passing the board examination.

Currently, many non-EM-trained physicians practice in district hospital EDs, comprising a notable proportion of the EM workforce in Taiwan and many other countries, including Japan, India, Thailand, and Hong Kong. 3 , 4 Nearly 40% of the EM clinicians in the United States are not EM-trained physicians (14.3% are non-EM-trained physicians, and 24.5% are advanced practice providers). 5 Many European countries have adopted the Franco-German system, sending doctors in ambulances to the field, with only a basic organization of hospital-based EM. 6 The primary reason for non-EM-trained physicians to practice at the ED is the increase in patient visit volume over the last four decades that has created workforce shortages despite the increase in the number and size of EM residency training programs. 7 Consequently, ED staffing gaps are often filled by non-EM-trained physicians.

Although several studies have discussed the emergency physician’s workforce, previous papers mainly focus on forecasting future workforce supply and demand. 8 Few have investigated the performance of non-EM-trained physicians at the ED. Furthermore, only a limited number of studies have evaluated the performance of physicians after EM specialty training. We aimed to evaluate the potential differences in performance between EM-trained and non-EM-trained physicians in an ED in Taiwan.

Methods

Study Design and Setting

We conducted a retrospective observational study involving 37,013 ED visits at a regional hospital in rural Taiwan between August 2018 and July 2020. The study ED only have several consultants on duty include: surgeon, internal medicine and orthopedics. The ED also had trauma team, which could provide care in major trauma patients.

Selection of Participants

This study was approved by the Institutional Review Board, which waived the need to obtain individual consent as all identifying data were encrypted. We excluded all ED visits due to obstetric problems and those of patients aged < 12 years, which were managed by on-duty obstetricians or pediatricians, respectively. We also excluded ED visits managed by physicians with over three years of experience at the ED after residency training. EM-trained physicians were physicians who successfully completed EM training and passed the examination administered by the Taiwan Society of Emergency Medicine. Physicians with multiple specialties were included in this group if one of their specialties was in EM. Non-EM-trained physicians were physicians with any specialty other than EM that have practiced in the ED for less than three years. Such physicians did not work full-time in the ED. The ED shifts were of 12 hours. The day shift was from 08:00 to 20:00, and the night shift was from 20:00 to 08:00. One physician and three nurses were on duty during every shift.

ED patients were assessed by a triage nurse to identify their acuteness level based on the five levels of the Taiwan Triage and Acuity Scale 9 : Level 1, the patient requires immediate management; Level 2, the patient should be seen within ten minutes; Level 3, the patient should be assessed within 30 minutes; Level 4, the patient should be assessed within 60 minutes; Level 5, the patient should be assessed within 120 minutes.

Outcomes

The physicians’ ED performance measurement was assessed based on the following parameters: waiting time, the time from arrival for triage to the initial physician assessment; length of ED stay, the time from arrival for triage to departure from the ED; unscheduled return visit rate, a patient presenting for the same chief complaint within 72 hours of the last discharge from the ED 10 ; return of spontaneous circulation achievement rate in out-of-hospital cardiac arrest patients, excluding patients who had signed a do-not-resuscitate order; referral of patients to any other health care unit, regardless of the reasons, including advanced center referral or any health care unit of the patients’ preference; computed tomography (CT) scan utilization, defined as the number of CT scan orders per patient per visit.

Analysis

The study data were summarized using descriptive statistics. Chi-squared tests and t-tests compared the two physician groups for multiple ED quality indices, including waiting time, rate of failing to visit patients within the required time, length of ED stay, admission rate, intensive care unit admission rate, unscheduled return visit rate, admission rate after unscheduled return visit, discharge against medical advice (DAMA) rate, return of spontaneous circulation rate in out-of-hospital cardiac arrest patients, in-hospital cardiac arrest rate, referral rate, and CT scan utilization. All data were analyzed using IBM SPSS Statistics for Windows, Version 22.0 (IBM Corp., Armonk, NY, USA). Differences were considered statistically significant at p < 0.05.

Results

Characteristics of Study Subjects

As shown in Fig. 1 , there were 37,013 ED visits from August 2018 to July 2020. We excluded 7,072 visits due to obstetric problems or those of patients aged under 12 years. Thirty-one visits were excluded due to incomplete data. We excluded 2,267 visits managed by physicians with over three years of experience in the ED after residency training. Of the 27,643 included visits, 11,097 were managed by EM-trained physicians (n = 7) and 16,546 by non-EM-trained physicians (n = 26). Table 1 presents the general characteristics of the two physician groups. Most physicians were males in their thirties. Physicians specializing in surgery or internal medicine constituted nearly 80% of the non-EM-trained physician group. Table 2 shows the patient acuteness level distribution managed by each of the physician groups. Those assigned triage Level 3 (urgent) constituted over 50% of the patients, while triage Levels 1 (resuscitation) and 2 (emergent) combined were less than 10%.

Fig. 1 . Study flow diagram.

ED: emergency department; OBS: obstetric; y/o: years old; EM: emergency medicine.

Table 1 . Characteristics of emergency medicine-trained and non-emergency medicine-trained physicians .

EM-trained: emergency medicine-trained physician; ER: emergency room; NA: not available; Non-EM-trained: non-emergency medicine-trained physician.

Variables	EM-trained (n = 7)	Non-EM-trained (n = 26)
Age, mean ± SD (years)	33.4 ± 2.0	34.8 ± 2.6
Male (%)	100	92.4
Graduated from residency, mean ± SD (years)	1.9 ± 0.9	1.8 ± 0.7
Total shifts, n	1,238	2,506
Night shift ratio (%)	52.2	36.2
ER visits per shift, mean ± SD	21.0 ± 5.2	20.5 ± 5.4
Residency training, n (%)
Sole Emergency Medicine	6 (85.7)	NA
Internal Medicine	0 (0)	9 (34.6)
Surgery	1 (14.3)	11 (42.3)
Family Medicine	0 (0)	4 (15.4)
Orthopedics	0 (0)	2 (7.7)

Table 2 . Patient flow in physicians with different specialties .

EM: emergency medicine.

	Managed by EM-trained physician (n = 11,097)	Managed by non-EM-trained physician (n = 16,546)
Acuity scale, n (%)
Level 1 (Resuscitation)	89 (0.8)	127 (0.8)
Level 2 (Emergent)	547 (4.9)	450 (2.8)
Level 3 (Urgent)	7,243 (65.3)	9,176 (55.5)
Level 4 (Less urgent)	3,014 (27.2)	6,317 (38.3)
Level 5 (Not urgent)	204 (1.8)	416 (2.6)

Main Results

The performance characteristics of emergency physicians with different specialties are presented in Table 3 . When compared to those managed by non-EM-trained physicians, patients treated by EM-trained physicians had significantly shorter waiting time (6.1 min vs. 9.2 min, p < 0.001), shorter ED stay (146.5 min vs. 176.1 min, p < 0.001), with a lower rate of failing to visit patients within the required time based on the Taiwan Triage and Acuity Scale (0.8% vs. 1.1%, p = 0.010), a lower rate of unscheduled return visit to the ED (4.9% vs. 5.4%, p = 0.043), lower CT utilization (0.16 [times/patient/visit] vs. 0.18 [times/patient/visit], p < 0.001) and higher DAMA rate (2.8% vs. 1.5%, p < 0.001). However, no differences were seen in the other assessed characteristics, including admission rate, intensive care unit admission rate, admission rate after unscheduled return visit, return of spontaneous circulation rate in out-of-hospital cardiac arrest patients, in-hospital cardiac arrest rate, and referral rate.

Table 3 . Performance characteristics of physicians with and without emergency medicine specialty .

^* Asterisk indicates statistically significant difference at p < 0.05.

Variables	EM-trained	Non-EM-trained	p -value
Waiting time (min)	6.1 ± 6.8	9.2 ± 9.0	< 0.001 *
Rate of failing to visit patients within the required time (%)	0.8	1.1	0.010 *
Length of ER stay, mean ± SD, (min)	146.5 ± 327.6	176.1 ± 375.7	< 0.001 *
Day shift	155.2 ± 331.3	174.3 ± 380.8	< 0.001 *
Night shift	142.1 ± 313.4	186.9 ± 363.0	< 0.001 *
Admission rate (%)	11.0	11.3	0.465
ICU admission rate (%)	12.5	12.2	0.694
Unscheduled return visit rate (%)	4.9	5.4	0.043 *
Admission rate after unscheduled return visit (%)	15.8	12.1	0.055
Discharge against medical advice rate (%)	2.8	1.5	< 0.001 *
ROSC in OHCA rate (%)	20.8	21.6	0.965
IHCA rate (%)	0.06	0.08	0.678
Referral rate (%)	0.8	0.8	0.958
CT scan utilization, mean ± SD, (times/patient/visit)	0.16 ± 0.54	0.18 ± 0.49	< 0.001 *

Table 4 . Length of stay with different specialties in sub-analysis .

^* Asterisk indicates statistically significant difference at p < 0.05.

Acuity scale	Managed by EM-trained physician (minutes)	Managed by non-EM-trained physician (minutes)	p -value
Level 1–2	283.75	448.31	< 0.001 *
Level 3–5	146.11	172.84	< 0.001 *

Discussion

Through this retrospective observational study in a single regional hospital in Taiwan, we have demonstrated the difference in ED performance between EM-trained and non-EM-trained physicians. Patients treated by EM-trained physicians had significantly shorter waiting and ED stay time, and their rate of failing to visit patients within the required time based on the triage scale was lower.

A previous study revealed that compared to emergency physicians with under 10 years of work experience, senior emergency physicians took longer to order prescription and patient disposition, used fewer diagnostic investigations, particularly in non-urgent patients, and were associated with a lower ED mortality rate. 11 In this study, we excluded physicians with over three years of work experience at the ED to avoid bias. Studies have revealed that patient satisfaction was lower in cases of longer waiting time and lengthy ED stay. 12 There is also a close correlation between unscheduled returns to the ED and dissatisfaction of the returning patients. 13 With ED being the main gateway to access care in the hospital, the medical care received at the ED could contribute considerably to the perception of the hospital. Furthermore, longer waiting and ED stay time might worsen ED crowding, 14 a universal concern in recent decades that was associated with adverse patient outcomes while in the ED. Therefore, we believe that the ED performance differences between the two physician groups have significant implications for improving the quality control of ED care.

The outbreak of COVID-19 has reinforced the critical role of the ED as a frontline of public health. The ED staff, particularly the emergency physicians, serve an essential function in identifying undifferentiated patients while providing timely medical care. Emergency physicians with thorough EM training and preparedness, personal protective equipment, and following established triage processes would be more competent in this global pandemic.

The differences in ED performance might be attributed to the distinct core value differences among specialties. Disposition, the ultimate endpoint for all ED visits, is also one of the most valued goals of EM training. EM training programs help trainees cogitate the patients’ disposition from the moment they enter the ED. Additionally, EM training programs in Taiwan are based on the “Model of the Clinical Practice of Emergency Medicine,” which is composed of (1) an assessment of the patient acuity; (2) a description of the tasks that must be performed to provide appropriate emergency medical care; (3) a list of medical knowledge, patient care, and procedural skills. 11 The EM-trained physicians’ initial approach is determined by the acuteness of the patient’s presentation, be it critical, emergent, or low acuity. 15 Unlike EM, other specialties might promote other primary goals, such as making a differential diagnosis. While assessing a patient, physicians complete a series of tasks, collect information, and select the most likely cause of the patients’ problem from the list of conditions and components. The physician arrives at the most probable diagnosis and implements a treatment plan for the patient by simultaneously applying all three components of the model, which are interrelated and applied concurrently in EM practice. 16

Furthermore, EM-trained physicians might have more of a “home advantage.” This could explain the differences in unscheduled return visit rate and CT scan utilization between the two physicians’ groups. EM-trained physicians had spent years of residency training at the ED, making them more familiar with the unexpected or urgent scenarios encountered there. Another concern is the so-called “man with a hammer” syndrome, derived from the saying that “to the man with a hammer, every problem looks like a nail.” 17 EM-trained physicians are adept at routine diagnosis and high-speed emergency care. Skills and extensive knowledge of all branches of medicine are required to recognize the most serious cases while stabilizing the patients’ condition. Non-EM-trained physicians underwent residency training in one specific specialty, so they might encounter some scenes with which they are unfamiliar. The solution is to educate the man with the hammer that many other tools are also available. 18 Likewise, to deal with a myriad of scenarios at the ED, ED physicians ought to have broad knowledge and experience in all professions. Besides, EM-trained physicians are usually working under pressure to deal with diverse patients, it’s important to be able to multitask and prioritize. Therefore, it is often necessary to apply a dual process model of reasoning and medical decision making. 19 This may explain why EM-trained physicians could determine patient disposition in a shorter length of ED stay than non-EM-trained physician.

Patient treated by emergency medicine-trained physician had a higher DAMA rate. Previous study focus on the DAMA patient revealed that main reason for DAMA included refusal of operation, long ED waiting time and subjective improvement. 20 In the emergency medicine-training course, a complete observation period (i.e., 4–6 hours) is usually emphasized for ensure no complication developed in the emergency stay period. In contrast to non-EM-trained physician, physician trained by emergency medicine may foresee complication and unscheduled return visit based on residency training course. This is evident by there is a lower unscheduled return visit rate in the emergency medicine-trained group.

During the emergency medicine residency training period, complication and comorbidity conference, unscheduled ED return visit and mortality conference for EM residents only, which usually demonstrate higher risks and undesirable results of some cases, hence have those physicians highly aware of the unpredictable risks subconsciously. These will all cause a tendency that once the patient does not meet the physician’s expectation anywhere, they will be discharged against medical advice. This “once bitten, twice shy” effect may able to explain why patient treated by EM-trained physician had a high DAMA rate. 21

Despite the evident differences in certain performance characteristics, the return of spontaneous circulation rate in out-of-hospital cardiac arrest patients and the incidence of in-hospital cardiac arrest remained similar. Possible explanations for this similarity include (1) multidisciplinary teamwork. Although emergency physicians are a key component in the overall ED workforce, successful resuscitation could only be achieved by close collaboration between paramedics, nurses, nurse practitioners, and physicians; (2) the requirement for an Advanced Cardiac Life Support certification every two years from every healthcare staff in Taiwan. EM-trained physicians are no longer the only ones familiar with advanced cardiac life support. Timely and accurate resuscitation of cardiac arrest patients could be performed by physicians of all specialties.

Non-EM-trained physicians are still an important part of the ED workforce in many countries. Based on our study, their performance differs from that of EM-trained physicians. Although EM-trained physicians achieved shorter waiting and ED stay time and lower rate of failing to visit patients within the required time and unscheduled return visits rate, the return of spontaneous circulation rate in out-of-hospital cardiac arrest patients and in-hospital cardiac arrest rate were similar in the two groups. Continuous and frequent educational training of advanced cardiac life support could explain why all physicians provided similar resuscitation quality during cardiac arrest. However, training sessions based on the EM model for non-EM-trained physicians are irregular. Further educational initiatives for non-EM-trained physicians working in the ED would be welcome.

Optimizing physician scheduling systems can improve ED efficiency. By assigning higher acuity patients (Level 1–2) to EM specialists, who are more proficient in handling critical cases, patient waiting time can be reduced. A balanced shift system should also be implemented, ensuring these specialists work day shifts, especially during off-peak or high-demand hours. For lower-acuity patients (Level 3–5), a fast-track system led by ED specialists can enhance overall throughput. A multilayered triage system further supports quick patient identification and categorization. Collaboration across specialties, multidisciplinary rounds, and focused training for non-ED physicians are essential for complex cases. Data-driven decision support systems, including predictive analytics and AI-driven patient tracking, can streamline workflow and optimize resource allocation in real time.

This study has some limitations. First, this study was conducted in a single regional hospital with relatively small daily ED volume (approximately 50 patients/day). Second, the study did not further analyze the prognosis of patients after admission or those with time-sensitive diseases such as ST-segment elevation myocardial infarction, or acute ischemic stroke. Third, the study was conducted in a rural area, and it is known that disparities exist between areas and countries in the emergency physician workforce. Last, work in the ED is dynamic; therefore, more than one physician would sometimes participate in diagnosing and treating the same patient. This could bias the physicians’ ED performance evaluation. The primary reason for conducting this study on physicians with under three years of practice at the ED at the regional hospital level rather than at a medical center level was to try to reduce the chance of such a bias due to: (1) the presence of other EM-trained physicians and rotating residents or intern doctors at the same shift in the ED of medical centers; (2) the availability of a variety of other specialty physicians to consult at medical centers; (3) a longer ED working experience could compromise the possible.

In conclusion, we found certain differences in performance between EM-trained and non-EM-trained physicians in the ED at a regional hospital in Taiwan. Our findings could be used as a reference for healthcare policy-makers and hospital management.

Acknowledgments

This study was supported by grants from the Tri-Service General Hospital, National Defense Medical Center in Taipei, Taiwan (TSGH-E-109219, MAB-109-035, TSGH-E-110225, TSGH-PH-D-110008). We are grateful for data collection from Mr. Chia-Hau Li at the PengHu branch, Tri-Service General Hospital.