Original Contribution

A study of the workforce in emergency medicine: 2007^?

Francis L. Counselman MD ^a, Catherine A. Marco MD ^b,?,

Vicki C. Patrick MS, APRN, ACNP-BC, CEN ^c, David A. McKenzie CAE ^d, Luke Monck PhD ^e, Frederick C. Blum MD ^f, Keith Borg MD, PhD ^g,

Marco Coppola DO ^h, W. Anthony Gerard MD ⁱ, Claudia Jorgenson RN, MSN ^j,

JoAnn Lazarus RN, MSN, CEN ^k, John Moorhead MD, MS ^l, John Proctor MD, MBA ^m,

Gillian R. Schmitz MD ⁿ, Sandra M. Schneider MD ^o

^aDepartment Emergency Medicine, Eastern Virginia Medical School and Emergency Physicians of Tidewater, Norfolk,

VA 23507, USA

^bDepartment of Emergency Medicine, University of Toledo College of Medicine, Toledo, OH 43614, USA ^cGraduate Program, School of Nursing, The University of Texas at Arlington, Arlington, TX 76019, USA ^dWork Force Study Liaison, American College of Emergency Physicians, Irving, TX 75261, USA

^eBooz Allen Hamilton and The City University of New York Graduate Center and Baruch College, New York, NY 22102, USA ^fDepartment of Emergency Medicine, West Virginia School of Medicine, Morgantown, WV 26506 and American College of Emergency Physicians, Dallas, TX, USA

^gEmergency Medicine, Medical University of South Carolina, Charleston, SC 29425, USA

^hDepartment of Emergency Medicine, Plaza Medical Center of Fort Worth, Fort Worth, TX 76104, USA

ⁱLebanon Emergency Physicians, Good Samaritan Hospital, Lebanon, PA 17042, USA ^jDepartment of Professional Practice, emergency nurses Association, Des Plaines, IL 60016, USA ^kEmergency Services, Blue Jay Consulting, LLC, Orlando, FL 32839, USA

^lDepartment of Emergency Medicine, Oregon Health and Science University, Portland, OR 97239, USA

^mSouthern Hills Medical Center, Regional Medical Director, Team Health, The University of Tennessee Health Science Center,

Nashville, TN 38163, USA

ⁿDepartment of Emergency Medicine, Wilford Hall Medical Center, San Antonio, TX 78236, USA

^oDepartment of Emergency Medicine, University of Rochester School of Medicine and Dentistry, Rochester, NY 14642, USA

Received 24 April 2009; accepted 15 May 2009

Abstract

Introduction: This study was undertaken to describe the current status of the emergency medicine workforce in the United States.

Methods: Surveys were distributed in 2008 to 2619 emergency department (ED) medical directors and nurse managers in hospitals in the 2006 American Hospital Association database.

Results: Among ED medical directors, 713 responded, for a 27.2% response rate. Currently, 65% of practicing emergency physicians are board certified by the American Board of Emergency Medicine or

^? Presented in part at the 5th Annual AAMC Physician Workforce Research Conference, Washington, DC, on April 2009.

* Corresponding author.

E-mail address: [email protected] (C.A. Marco).

0735-6757/$ - see front matter (C) 2009 doi:10.1016/j.ajem.2009.05.014

the American Osteopathic Board of Emergency Medicine. Among those leaving the practice, the most common reasons cited for departure include geographic relocation (46%) and better pay (29%). Approximately 12% of the ED physician workforce is expected to retire in the next 5 years.

Among nurse managers, 548 responded, for a 21% response rate. Many nurses (46%) have an associate degree as their highest level of education, 28% have a BSN, and 3% have a graduate degree (MSN or higher). Geographic relocation (44%) is the leading reason for changing employment.

Emergency department annual volumes have increased by 49% since 1997, with a mean ED volume of 32 281 in 2007. The average reported ED length of stay is 158 minutes from registration to discharge and 208 minutes from registration to admission. Emergency department spent an average of 49 hours per month in Ambulance diversion in 2007. Boarding is common practice, with an average of 318 hours of Patient boarding per month.

Conclusions: In the past 10 years, the number of practicing emergency physicians has grown to more than 42 000. The number of board-certified emergency physicians has increased. The number of annual ED visits has risen significantly.

(C) 2009

Introduction

Emergency medicine (EM) was officially recognized by the American Board of medical specialties in 1979. The first residency training program in EM was established in 1970. Also in 1970, the Emergency Department Nurses Association (now Emergency Nurses Association [ENA]) was founded.

In 2006, the number of patients in the United States seeking care in the ED reached 119 million annual visits, whereas the number of emergency departments (EDs) decreased [1-6]. From 1995 to 2005, the number of annual ED visits increased by 20%, whereas the number of EDs decreased by 10%, resulting in a mean ED increase in volume from 23 000 to 30 000 [4,6]. The challenges confronting emergency physicians and nurses in the United States have recently been highlighted by the Institute of Medicine’s publication Hospital-Based Emergency Care: At the Breaking Point [7].

Workforce studies are essential to the development and growth of any profession. Several studies have identified a projected shortage of physicians to meet the workforce needs in the United States [8,9]. Workforce projections have important implications for training, as well as consideration of other approaches to workforce planning. For example, because of the projected physician shortage, the Association of American Medical Colleges has recommended increased enrollment in US medical schools to contribute to the physician workforce [10]. Specifically, in emergency care, an accurate understanding of the nationwide workforce, includ- ing numbers, qualifications, and plans of the country’s emergency physicians and registered nurses (RNs), is vital to the ideal planning of professional endeavors, including training programs, professional planning, board certification, continuing education, administrative agenda, organizational support, legislative advocacy, and others.

Two previous studies conducted by the American College of Emergency Physicians (ACEP) have investigated the status of the workforce in EM in 1997 and again in 1999 [11,12]. Similar national benchmarking studies were con-

ducted by ENA, most recently in 2005 [13]. The most recent workforce study of 1999 estimated that approximately 32 000 physicians practice EM full-time and that training and board certification among these physicians varied widely. Between 1997 and 1999, the total number of ED visits increased during a time when the number of hospital EDs decreased. The workforce of emergency physicians has been dominated by white men and, as such, is not representative of the US population. Both previous work- force studies demonstrated a significant nationwide shortage of board-certified emergency physicians, despite dramatic increases in the number of Accreditation Council for Graduate Medical Education (ACGME)-approved EM residency programs [14]. Recently, ACEP has reaffirmed its statement that “there is currently a significant shortage of physicians appropriately trained and certified in emergency medicine [15,16].” A similar statement addressing the nursing shortage in EDs and its impact on patient care was issued by ENA[17]. There is wide variation across the nation regarding the type and size of EDs, as well as staffing, credentials, and qualifications [18].

This study is undertaken to describe the current status of the EM and nursing workforces in the United States, including both physician and nursing data, demographics, training and credentials, and other characteristics of emergency physicians, as well as patient characteristics, ED functions, nursing issues, and other factors influencing the practice of EM. Comparison to earlier ACEP workforce studies from 1997 to 1999 and other literature will provide important information in trends in EM care over the past decade. The addition of nursing data enriches the depth of workforce descriptions and supplies comparative data for future studies.

Methods

Sampling

As in previous ACEP workforce studies, the study sample was derived from US EDs. To approximate this population of

interest, ACEP obtained the American Hospital Association (AHA) database that was compiled from the 2006 AHA Annual Survey of Hospitals. This database contains information for 5125 hospitals. The database contained data for both hospitals that responded to the 2006 AHA Annual Survey of Hospitals (respondents) and those that did not (nonrespondents). Data for nonrespondents represent (a) data furnished by the hospital the last time it completed the survey, (b) data obtained through AHA research of public information, and/or (c) a combination of both (a) and (b).

To obtain a sample of respondents from which valid inferences could be made to the population of all EDs within the United States, a stratified random sample was used [19,20]. The methodology for sampling in this study was identical to that of the 2 previous workforce studies.

In the present workforce study, the stratified sampling approach included 2 strata based on hospital characteristics: institution type and teaching status. The first stratum, institution type, denotes the ownership and/or management of the hospital. Using the AHA database, hospitals were placed into 1 of 4 categories: federal (n = 194), other government (n = 1136), private/for profit (n = 955), and private/not for profit (n = 2840). The second stratum, teaching status, identified whether the hospital was an academic institution. The 3 levels within this stratum were academic teaching (n = 112), nonteaching (n = 3790), and other academic (n = 1223). These 2 strata were applied to both hospitals that had responded to the 2006 AHA Annual Survey of Hospitals (n = 4039) and those that did not (n = 1086).

To determine the necessary sample size for the study, the 2 strata were cross-tabulated to create a 4 x 3 table. This table was then used to determine the number of hospitals falling into each of the 12 possible categories in the sampling tables (eg, federal, academic teaching; federal, nonteaching; etc). Next, the necessary sample to support inferences at the 95% Confidence level and the +-5% confidence interval were calculated for each cell in the sampling table. Based on these calculations, it was determined that the sample size necessary to draw inferences at the 95% +- 5% confidence level/interval to each of the 12 cells in the sampling table was 1559. Because it was expected that roughly 50% of the sample contacted may not provide valid responses to the survey [21], the sample size of hospitals contacted was increased to 2789. For 7 of the 12 cells in the combined sampling table, this resulted in a census of all of the hospitals within the AHA database falling within that category. For 4 cells, this resulted in a doubling in the number of hospitals sampled; 1 cell had no hospitals (federal government, academic medical center). It was determined that 170 hospitals in our original sample did not have an ED; the final sample size for our study was therefore 2619.

Survey instrument design, development, and pilot

The survey instrument was previously developed and deployed in 1997 and 1999 for the first and second

workforce studies, respectively. For the third workforce study, elements of the instrument were refined and additional items were added, as recommended by the ACEP Workforce Technical Advisory Group, composed of emergency physicians, nurses, and ACEP staff with expertise in EM workforce issues and research design. The survey instrument was developed by members of the Technical Advisory Group in collaboration with Booz Allen Hamilton, a nationally recognized management consultant firm. The final survey was composed of 75 items. The survey included questions regarding ED and workforce characteristics in 2007.

To test the psychometric properties and solicit qualitative feedback from a sample of potential respondents, the revised survey instrument was pilot tested. The pilot survey was deployed as an online survey that was accessed via a URL and was open from April 21, 2008, through May 16, 2008. Although the low response rate (5%) precluded a quantitative assessment of the psychometric qualities of the survey tool, pilot participants provided qualitative feedback, which was used to revise the survey. The 19 EDs that participated in the pilot were subsequently removed from the sampling frame and were thus ineligible for potential participation in the full study.

Survey deployment

Because research suggests that multimode survey admin- istration may yield higher aggregate response rates because of survey respondents being partial to one mode of response vs another (ie, paper vs online) [22,23], the third workforce study was conducted using both an online and a paper version of the survey. The survey period ran from June 23 through September 30, 2008.

At the outset of the study in June 2008, letters announcing the survey were sent to all ED medical directors and nurse managers in the sample as well as the hospital administrators. Included in the letters were the following: (1) a letter explaining the survey, its purpose, and approximate length of time it would take to complete it; (2) the URL for the online survey; (3) a unique participant identification for each respondent, which would allow access to the URL; (4) a paper-based version of the survey; and (5) a preaddressed, postage-paid envelope to return the paper survey.

Communication and incentive plan

To increase response rate, ACEP and ENA developed a robust communication and incentive plan to support the study. Several modes of communication were used to maximize the number of contacts potential respondents received about the survey. Various conduits of available communication were used by ACEP and ENA to publicize the survey and disseminate the message of its importance. The study was publicized via several organizational Web

Federal, nonteaching	58	1.17	6	0.86	8	1.17	-2	-0.31	1.354
Federal, other teaching	124	2.5	29	4.18	17	2.5	12	1.68	0.599
Other government,	44	0.89	24	3.46	6	0.89	18	2.57	0.257

Other government, 120	2.42	35	5.04	17	2.42	18	2.62	0.480
other teaching
Private for-profit, 6	0.12	2	0.29	1	0.12	1	0.17	0.420
academic teaching Private for-profit, 728	14.69	82	11.82	102	14.69	-20	-2.88	1.243
nonteaching
Private for-profit, 98 other teaching	1.98	27	3.89	14	1.98	13	1.91	0.508
Private not-for-profit, 62	1.25	43	6.2	9	1.25	34	4.94	0.202
academic teaching Private not-for-profit, 1902	38.39	154	22.19	266	38.39	-112	-16.2	1.730
nonteaching
Private not-for-profit, 851	17.17	210	30.26	119	17.17	91	13.08	0.568
other teaching Total 4955	100	694 a	100	694	100			1

sites, newsletters, and e-mails. Letters were sent to hospital administrators, medical directors, and nurse managers at each institution. Several EM contract management groups were contacted. State chapters of ACEP participated with personal contacts within the state. A telemarketing firm was employed to make calls to nonresponding hospital EDs. Several small incentives were also offered to those who responded to the survey. Similar actions were initiated by ENA to improve participation.

Table 2 Demographics of emergency physician respondents ^a

Physician 1997 1999

information (%) (%)

%

2007

?-1997 (%)

?-1999 (%)

Nursing information (%)

Table 1 Weighting based on 12 sampling strata

Type of institution Population Observed

Expected

Residual (observed- expected)

df

?2

(? = .05)

P

Weight

No. %

No. %

No.

0

%

No. %

Federal, academic teaching

0

0

0

0

0

0

0

10 365.2

.000 NA

academic teaching Other government,

nonteaching

962 19.41

82 11.82 135 19.41 -53 -7.6

1.643

^a Excludes 19 surveys that did not contain any identification information.

Sample bias and weighting

Data were weighted by 12 sampling strata of practice settings. Bias was detected using a ?2 test to determine whether the sample obtained was representative of the population under study vis-a-vis the sampling strata. A significant ?2 value (?2 = 365.2, P = .000) indicated that the sample was disproportionate and that sample bias did exist (ie, that some groups were overrepresented, whereas others

Male sex	83	83	73	-10	-10	18
Female sex	17	17	21	4	4	82
White	81	82	80	-1	-2	80
African American	6	5	5	-1	0	4
Hispanic	2	3	2	0	-1	3
Native American	0	0	0	0	0	0
Asian/Pacific	5	6	5	0	-1	5
Islander Other	5	4	3	-2	-1	1
?-1997, difference from 1997 to 2007; ?-1999, difference from 1999 to 2007. a Totals may not equal 100 because not all respondents answered each question.

were underrepresented). Accordingly, the survey sample was weighted to boost the weight of underrepresented groups (ie, private not-for profit, nonteaching hospitals) and reduce the weight of overrepresented groups (ie, private not-for profit, other teaching hospitals) [24]. The weighting procedure used in this study adjusted the influence of each survey response by its expected, proportionate influence. In practice, the sample weight for each response was the expected propor- tionate sample size for that response based on its group memberships divided by actual sample size achieved for those group(s).

This study was reviewed and determined to be exempt by the institutional review board of the University of Toledo College of Medicine.

Results

Among 2619 ED medical directors, 713 participated (27.2% response rate). The response rate for nurse managers was 21% (548/2619). Seventy-one percent of ED medical directors completed the survey online, whereas 29% used paper. Seventy-eight percent of nurse managers completed the online survey, whereas 22% completed the paper version. Not all respondents answered every question. The total number responding to each question is reported throughout the results. Table 1 shows, by strata, the number of hospitals in the population, the sample, the response rate, and the weighting.

Physician demographics

The mean age of emergency physicians is 43.7 years. There has been little change in the ethnicity of emergency physicians since 1997, with whites representing the largest

Table 3 Board certification and training of US emergency physicians

Training and certification 1997 1999

information (%) (%)

2007

%

?-1997 (%)

?-1999 (%)

group (80%). Most (87%) hold an MD degree, and 13% hold a DO degree (no significant change from 1997; 88% and 12%, respectively). Most (87%) graduated from a US medical school.

Although still a male-dominated field, more women practiced EM in 2007 (21%) compared with 1997 (17%) (Table 2). Similar to the 1999 study, the highest percentage (27%) of female EM physicians practice in academic medical centers.

Training and board certification

Training in EM residencies has increased significantly since 1997 and 1999. The number of physicians certified by the American Board of Emergency Medicine (ABEM) and the American Osteopathic Board of Emergency Medicine (AOBEM) has increased. Certification by ABEM has increased significantly since 1997 and 1999 (Table 3). Currently, 58% of practicing emergency physicians are board certified by ABEM; 7%, by AOBEM; and 4%, by the Board of Certification in Emergency Medicine. Seventy-two percent of EM physicians practicing in an urban environment are board certified in EM, compared with only 13.5% in rural areas. Some practicing emergency physicians are residency trained in a second specialty, most commonly internal medicine (6%, down from 14% in 1997), Family Medicine

(5%, down from 9% in 1997), pediatrics (2%, no change

since 1997), and other (1%, down from 2% in 1997). There has been a significant decrease in practicing emergency physicians who are solely board certified in non-EM specialties, including family medicine (13%, down from 32% in 1997), internal medicine (9%, down from 28% in

1997), and pediatrics (3%, down from 4% in 1997). Physicians who are not board certified in any specialty were not separately identified.

EM residency trained	37%	42	60	23	18
Board certified	54	58	68	14	10
ABEM certification	48	50	58	10	8
AOBEM certification	4	3	7	3	4
BCEM certification	3	2	4	1	2
Family medicine certification	32	37	13	-19	-24
Internal medicine certification	28	30	9	-19	-21
Pediatrics certification	4	8	3	-1	-5
Surgery certification	7	5	1	-6	-4
Other certification	7	4	1	-6	-3
No report of board certification a	26	16	5	-21	-11
BCEM indicates Board of Certification in Emergency Medicine. a One interpretation of this number is that up to 5% of practicing emergency physicians are not board certified in any specialty. Another interpretation of this number is that up to 5% of practicing emergency physicians indicated “none” of the above categories, for any reason. Physicians who are not board certified in any specialty were not separately identified.

Fig. 1 Types of institutions. (A) Urban compared with suburban and rural institutions. (B) Teaching institutions compared with nonteaching institutions.

Emergency department nurse demographics and staffing characteristics

Most ED nurses are women (82%) and white (80%), followed by African American (4%), Asian-Pacific (4%), and Hispanic (3%). Many (46%) nurses have an associate degree as their highest level of education, 28% have a BSN, and 3% have a graduate degree (MSN or higher). The mean age is 39.7 years. The largest proportion (11%) of board- certified RNs hold the certified emergency nurse (CEN) credential. The CEN process is a validation of an emergency nurse’s expertise and demonstrates a commitment to emergency nursing. Unlike the certification process for physicians, it is an added qualification that all emergency nurses may pursue. Most CENs are employed in academic teaching hospitals (12.5%).

Registered nurses dedicated to the ED are the predomi- nant staff workforce. Relatively small numbers per ED were reported in the licensed vocational nurse/licensed practical nurse (LVN/LPN) (1.21), float pool (1.12), and agency/ traveler (0.91) categories. New-graduate RN positions average 1.93 per ED. An average of 4.46 RN staff nurses left their EDs in 2007, and an average of 3 full-time staff RN positions per ED are currently available. Respondents who

had RN staff nurses who left reported that it took an average of 57 days to fill these vacancies.

Nurse managers anticipate that 6% of staff RNs will retire in the next 5 years.

Geographic relocation (44%), better pay (31%), and less stressful environment (30%) were the most common reasons cited for staff nurse resignations. Ongoing education is an important characteristic of ED nurses. The most common verifications include advanced cardiac life support (95%), pediatric advanced life support (74.8%), and trauma nursing core course (50.4%).

Physician and nursing collaboration

Most participants reported positive physician-nurse inter- actions. The majority (88% of physicians, 86% of nurses) strongly agreed or agreed that there is effective communica- tion between physicians and staff nurses in their ED. The majority (87% of physicians and 86% of nurses) agreed (“strongly agree” or “agree”) that there is effective collabora- tion regarding patient care between physicians and staff nurses. The majority (92% of physicians and 85% of nurses) reported mutual respect (“strongly agree” or “agree”).

Twenty-eight percent of medical directors and 35% of nurse managers reported that their model of patient care involves the same providers working consistently with the same nurses. Among those, 73% are satisfied with the patient care model.

Hospital and ED characteristics

Annual ED volumes have increased by 49% since 1997. The mean annual ED volume in 1997 was 21 667; for 1999,

the number was 23 912; for 2006, it was 29 800; and for

2007, it was 32 281. Since 1997, there has been an increase in urban institutions, compared with suburban and rural, and the percentage of teaching institutions has not changed (Fig. 1). Many EDs are divided into service areas, with the largest increase in urgent-care/fast-track areas (Table 4). Staffing of separate service areas varied by area.

Medical Screening examinations are performed most commonly by physicians (51%), followed by RNs (30%) and midlevel providers (20%). Referral of care from ED triage to treatment settings outside the ED is relatively

Table 4 Common separate service areas in EDs

ED separate service areas

1997 1999 2007

(%) (%)

% ?-1997 (%) ?-1999 (%)

Pediatric	19	29	10 -9	-19
Psychiatric	17	-	10 -7	-
Clinical decision unit/	18	23	9 -9	-14
observation Urgent care/fast track 37		86	47 10	-39
Other 4		-	6 2	-

staffing models”>common (16.6% of EDs). Of those who did not refer patients, 15% plan to implement a referral of care program. The average reported length of stay is 158 minutes from registration to discharge and 208 minutes from registration to admission. Nurse managers reported an average of 86 patients per month with an unexpected return to the ED within 72 hours. In 2007, an average of 78 patients per month left without treatment, and 29 patients per month left against medical advice. Emergency departments spent an average of

49 hours per month in ambulance diversion in 2007.

Boarding in the ED of inpatients awaiting an available bed is common practice; nurse managers reported an average of 318 hours of patient boarding per month. The majority (95%) of boarded patients are cared for by ED nurses. Most nurse managers (93%) indicated that boarding or crowding inter- fered with the provision of Quality patient care (25%, to a great extent; 42%, to some extent; 26%, little extent; 7%, no extent).

Physician ED staffing models

The mean number of attending physicians scheduled per month is 11.22. Assuming a 40-hour work week (the number used in previous studies), we found a physician-FTE ratio of 1.24, compared with 1.47 in the 1999 study. The average number of patients seen per hour during March 2007 was 2 patients per hour per attending physician. Most participants expected to expand the physician staff (73%, increase from 40% in 1997) and midlevel provider staff (65%, increase

from 33% in 1997) within the next 5 years.

physician staffing estimates”>Physician staffing estimates

In the previous studies, estimates were made regarding the number of EM physicians. The results of this analysis can be

Table 5 Key physician staffing measures

seen in Table 5. According to the AHA database, in the United States, there are 4587 hospitals with an ED [6]. We found that the average number of FTEs per institution was

9.03. The average number of physicians working in each ED was 11.22. The total number of FTEs in the population is 41 421. If one assumes that physicians work in only 1 ED, the number of physicians practicing clinical EM is 51 362. However, we found that 22% of EM physicians work in more than 1 ED (vs 25% in 1999). Therefore, the number of unique or unduplicated physicians working in EDs is reduced to 42 100. The estimated shortfall of board- certified/board-prepared EM physicians needed in the United States to fill all FTEs is 14 735 (35% of 42 100).

Approximately 1.3 physicians per ED leave an EM practice annually. The most common reasons cited for departure include geographic relocation (46%), better pay (29%), less stressful environment (23%), electronic medical record (19%), retirement (11%), expanded responsibility

(11%), changed specialties (10%), nonvoluntary separation (9%), academic positions (7%), and concern about mal- practice lawsuits (3%). Respondents were asked to indicate all reasons contributing to the decision to leave.

Approximately 12% of the ED physician workforce is expected to retire in the next 5 years, as compared with physician assistants (3%) and nurse practitioners (2%).

Discussion

Demographics of emergency physicians

There has been little change in age, ethnicity, and sex in the past 10 years. The typical emergency physician is a 43.7- year-old white man. In the 1999 and 1997 studies, the typical

Item	Calculation	Result a
A. Total no. of institutions responding to survey with physicians schedules		692
B. Average no. of physicians scheduled per institution		11.22 (0.32)
C. Total no. of physicians scheduled at responding institutions	692 (A) x 11.22 (B)	7764
D. Total no. of clinical hours scheduled during March 2007		1 027 247
among survey respondents
E. Average clinical hours per institution in March 2007	1 027 247 (D)/657 (no. of institutions	1563 (162)
	responding with this information)
F. FTE clinical hours per 1-month period	1 FTE = 40 h/wk x 4.33 wk/mo	173.20
G. Total no. of FTEs	1 027 247 (D)/173.20 (F)	5931
H. Average no. of FTEs per institution	1563 (E)/173.20 (F)	9.03 (0.94)
I. Physician-FTE ratio	11.22 (B)/9.03 (H)	1.24
J. Total no. of hospitals with an ED [25]		4587
K. Total no. of FTEs in US EDs	9.03 (H) x 4587 (J)	41 421
L. Total no. of EM physicians needed in the United States to fill all FTEs	41 421 (K) x 1.24 (I)	51 362
M. Unduplicated EM physicians (accounting for 22% of EM physicians	51 361 (L)/1.22	42 100
working in more than 1 ED)
a Numbers in parentheses indicate the standard error.

emergency physician was a 42.6-year-old and 42-year-old white man, respectively. Although more women practice EM currently (21%) compared with 1997 (17%), EM still demonstrates a disproportionate representation of men, compared with medical students, of which approximately 50% are women [25].

Training and credentials of emergency physicians

The results of this study indicate a steady increase in the number of emergency physicians board certified/board prepared by ABEM. Since 1980, more than 25 000 physicians have completed the ABEM board-certification process [26]. Some debate exists regarding the appropriate training and qualifications of physicians practicing EM [27]. In addition, the distribution of qualified emergency physicians nationwide may not be optimal, especially in lower volume EDs [28]. This is supported by our finding that only 13.5% of physicians are board certified in EM in Rural hospital EDs compared with 72% of those working in urban EDs.

These results are comparable with those reported by Daniel Stern and Associates [29], who reported that 89% of participating emergency physicians are board certified in EM; 10%, in family medicine; 6%, in internal medicine; and fewer than 6%, other specialties.

Our data support the need for more emergency physicians to meet the workforce needs in the United States. The Institute of Medicine report and other authors suggest that in addition to an increase in residency training programs, additional approaches to staffing EDs and training EM physicians may be needed because it will likely be many years before the entire workforce is EM residency trained. Innovative approaches to improve longevity and unique solutions to staffing rural EDs may be necessary to fill current workforce needs [30-33].

Others have looked at EM workforce needs. Camargo et al [34] attempted to estimate emergency physician workforce needs using 2005 data. The authors determined, based on a series of staffing assumptions, that 40 030 emergency physicians would be needed to staff all EDs. Despite the fact that the authors approached the problem of workforce needs from a completely different perspective than ours, their estimate of 40 030 emergency physicians in 2005 is very similar to our estimate of 42 100 in 2007. When one takes into account that approximately 1 350 EM residency-trained, board-certified emergency physicians join the workforce each year (one of the assumptions of Camargo et al [34]), our estimates are nearly identical.

Nursing characteristics

A recent report described the current shortfall of RNs in the nursing workforce nationally despite some significant improvements [35]. Other studies have described the impact

of the nursing shortage on patient care and vulnerability of the ED to shortage, citing intensity of care, overcrowding, and controversies over staffing levels as significant factors [7,18]. These factors are validated in the 2008 workforce study. This study also showed growing stability of the ED nursing workforce, with low usage of float pool, agency, and travel nurses. Position vacancies have dropped; however, recruiting for an experienced nurse continues to be a major issue, with almost 2 months required to fill positions. With new graduates filling many positions, inexperience is a significant challenge in emergency nursing. The nursing workforce is aging, which may contribute to reasons for voluntary separation such as changing specialties, better scheduling, and less stressful environment.

Hospital and ED characteristics

Emergency department volumes have continued to increase significantly since 1997. Workforce projections must take into account increasing ED volumes. The common use of separate ED service areas is an important descriptor unique to the third workforce study. An understanding of these areas and typical staffing patterns is important to workforce projections.

Referral of care to treatment settings outside the ED is a relatively common practice and growing in scope. The efficient and ethical practice of referral of care is essential to patient safety and patient satisfaction. The policy of ACEP, “Medical Screening of Emergency Department Patients,” recently approved in 2006 and 2007, states that “ACEP strongly opposed deferral of care for patients presenting to the ED.” The policy also states that “in situations in which it is required that patients be deferred, very specific and concrete standards must be adopted…to ensure patient access…and timely, appropriate treatment [36].”

Emergency department crowding, largely a result of increasing patient visits in an environment of fewer numbers of EDs, has a dramatic influence on the practice of EM [37]. Boarding has been cited as another primary factor contribut- ing to crowding [38]. Boarding is common practice; nurse managers reported an average of 318 hours of patient boarding per month, with most (95%) of boarded patients cared for by ED nurses. The perception of interference of provision of quality patient care by most nurse managers (93%) indicates the significant impact of this practice on Patient care and safety.

Staffing estimations

Numerous factors influence trends in the EM workforce, including the number of training programs, retirement, change of specialty or type of practice, full- or part-time status, stress, patient volumes, Job satisfaction, practice environment, and numerous others [39-41]. The longevity of emergency physicians is an important area of concern in our specialty. According to Daniel Stern and Associates,

55% of emergency physicians have considered leaving the specialty. Previous authors have estimated an annual attrition rate of 2.5% to 12% among emergency physicians [42,43]. Among the most common reasons cited in this study for physician departure from practice are geographic relocation, better pay, less stressful environment, electronic medical record, retirement, expanded responsibility, chan- ged specialties, nonvoluntary separation, academic posi- tions, and concern about malpractice lawsuits. These reasons represent important areas to address to increase physician longevity and job satisfaction.

Given that there are 4587 hospitals with EDs, and given that the data indicate that there are 9.03 FTEs per ED, the total number of FTEs is projected to be 41 421. Because the data indicate a physician-FTE ratio of 1.24:1, we conclude that there are 51 362 physicians needed to staff those FTEs. When adjusted for physicians working at more than 1 ED, that number is reduced to 42 100.

Approximately 12% of the ED physician workforce is expected to retire in the next 5 years. Currently, there are 147 ACGME-approved EM residencies [44] and 37 osteopathic EM residencies nationwide [45]. It may be surmised that over time, with increasing numbers of residencies and increasing numbers of graduates entering the workforce, the gap in board-certified emergency physicians will be expected to decrease. However, this will depend on numerous factors, including demographic changes of the US population, emergency physician retirement rates, and number of residency programs and graduates. Camargo et al [34], using a best-, intermediate-, and worst-case scenario, estimated the time it would take to staff all EDs with board-certified emergency physicians. They that found this would not occur until 2019, 2038, and never, respectively, based on the 3 scenarios [34].

Reasons for ED nurses leaving practice are similar to those for physicians and require new models to increase longevity. In addition, an area of controversy that continues to need evidence-based research is redesigning nursing staffing patterns to meet job satisfaction, patient needs, and efficiency requirements [7,18]. Also requiring additional study and further recommendations are standardized meth- ods for staffing boarding areas, clinical decision units, and fast-track/urgent-care areas. An additional area of study, related to staffing, is needed to determine the impact and value of the CEN credential as related to recruitment, retention, and quality patient care.

Limitations

The response rate in this study was lower than that in previous workforce studies. The reasons for the decline in response rate are not clear but may be related to errors in the AHA database, busy schedules, a more complicated and time-consuming survey instrument, and/or declining interest in survey participation or in the topic of workforce characteristics. Because of the extensive methods used to

boost response rate, there may be significant sampling bias. Examples of potential bias include personal contacts, variation in state ACEP chapter involvement, and/or interest in incentives offered among potential participants.

Given our low response rate, the characteristics of nonresponders are important. We do not know if they were similar to or differ systematically from respondents. We did not attempt to contact nonresponders. Given our relatively low response rate, one must be careful not to come to conclusions based solely on our data. It is interesting, however, that despite our low response rate and completely different methodology, we came to essentially the same conclusions regarding estimated number of EM physicians needed as Camargo et al did. The significant advantage of our study over the Camargo et al study is that we collected data from working emergency physicians, whereas their study was based solely on a series of staffing assumptions.

Conclusions

In the past 10 years, the number of practicing emergency physicians has grown significantly to more than 42 000. The demographics of practicing emergency physicians have remained stable. The number of physi- cians certified by ABEM and AOBEM has increased. It will likely be many years before the entire workforce is fully EM residency trained. Increases in residency training and new approaches to staffing EDs and training EM physicians may be needed. The number of emergency nurses continues to grow. The number of annual ED visits has risen significantly. As ED volumes have increased significantly since 1999, the number of physicians needed to staff EDs has also increased.

Acknowledgments

The authors gratefully acknowledge the financial support of The American College of Emergency Physicians, Council of Emergency Medicine Residency Directors, Emergency Medicine Foundation, Emergency Medicine Residents’ Association, ENA, GE Healthcare, and The Society for Academic Emergency Medicine.

References

1. Sullivan AF, Richman IB, Ahn CJ, et al. A profile of US emergency departments in 2001. Ann Emerg Med 2006;48:694-701.
2. Nawar EW, Niska RW, Xu J. National hospital ambulatory medical care survey: 2005 emergency department summary. Adv Data 2007; 386:1-32.
3. http://www.cdc.gov/nchs/about/major/ahcd/ercharts.htm.
4. Nawar EW, Niska RW, Xu J. National Hospital Ambulatory Medical Care Survey: 2005 emergency department summary. Advance data

from vital and health statistics; no. 386. Hyattsville, MD: National Center for Health Statistics; 2007.

1. Pitts SR, Niska RW, Xu J, Burt CW. National Hospital Ambulatory Medical Care Survey; 2006 emergency department survey. National health statistics reports; no.7. Hyattsville, MD: National Center for Health Statistics; 2008. http://www.cdc.gov/ nchs/data/nhsr/nhs007.pdf.
2. American Hospital Association. Trendwatch chartbook 2008, Spring 2008, Appendix 3, Table 3.3 Emergency department visits, emergency department visits per 1,000, and number of emergency departments, 1991-2006. www.aha.org/aha/trendwatch/chartbook/2008/08appen- dix3.pdf.
3. Committee on the Future of Emergency Care in the United States Health System. Hospital-based emergency care: at the breaking point. Washington DC: National Academies Press; 2007.
4. Kirch DG, Vernon DJ. Confronting the complexity of the physician workforce equation. JAMA 2008;299:2680-2.
5. Cooper RA, Getzen TE, MdKee HF, et al. Economic and demographic trends signal an impending physician shortage. Health Aff (Millwood) 2002;21:140-54.
6. Association of American Medical Colleges. AAMC statement on the physician workforce. Washington, DC: AAMC; 2006.
7. Moorhead JC, Gallery ME, Mannle T, et al. A study of the workforce in emergency medicine. Ann Emerg Med 1998;31(5):595-607.
8. Moorhead JC, Gallery ME, Hirshkorn C, et al. A study of the workforce in emergency medicine: 1999. Ann Emerg Med 2002;40(1):3-15.
9. Emergency Nurses Association. ENA national benchmark guide: emergency departments. Des Plaines: Emergency Nurses Association; 2005. p. 2005.
10. American Board of Emergency Medicine Task Force on Residency Training Information. Report of the Task Force on Residency Training Information, American Board of Emergency Medicine. Ann Emerg Med 1998;31:608-25.
11. Emergency Medicine Workforce. Ann Emerg Med 2001;38:485-6.
12. Emergency Medicine Workforce. Ann Emerg Med 2006;48:510.
13. ENAwritten testimony. Emergency Nurses Association Written Testimony IOM capstone workshop; 2006 [Washington, DC].
14. Robinson K, Jagim M, Ray C. Nursing workforce and trends affecting emergency departments. Top Emerg Med 2004;26:276-86.
15. Kish L. Survey sampling. New York, NY: John Wiley & Sons, Inc.; 1995.
16. Rea LM, Parker RA. Designing and conducting Survey research: a comprehensive guide. San Francisco, CA: Jossey-Bass; 2005.
17. Babbie ER. Survey research methods. Belmont, CA: Wadsworth; 1990.
18. Groves RM, Fowler FJ, Couper MP, Lepkowski JM, Singer E, Tourangeau R. survey methodology. New York, NY: John Wiley & Sons, Inc.; 2004.
19. Couper MP, Blair J, Tripplett T. A comparison of mail and e-mail for a survey of employees in federal statistical agencies. J Off Stat 1999;15: 39-56.
20. Pfeffermann D. The use of sampling weights for survey data analysis. Stat Methods Med Res 1996;5:239-61.
21. American Association of Medical Colleges. FACTS-applicants, matriculants and graduates. http://www.aamc.org/data/facts/ 2003/2003summary.htm accessed 12/8/08.
22. American Board of Emergency Medicine Exam Statistics. http://abem. org/public/portal/alias Rainbow/lang en-US/tabID 3373/Desk- topDefault.aspx, accessed 3/17/08.
23. McKenna M. IOM report ignites new debate on who should practice emergency medicine. Ann Emerg Med 2007;49:614-7.
24. Moorhead JC. Distribution of emergency medicine residency gradu- ates. Ann Amerg Med 1998;32:509-10.
25. Daniel Stern and Associates. Emergency physician salary and benefit survey. Daniel Stern and Associates; 2008.
26. Williams JM, Ehrlich PF, Prescott JE. Emergency medical care in rural America. Ann Emerg Med 2001;38:323-7.
27. family physicians help meet the emergency care needs of rural America. Am Fam Physician 2006;73:1163.
28. Critical challenges: family physicians in emergency medicine 2006, AAFP position paper. http://www.aafp.org/online/en/home/policy/ policies/e/emposition.html accessed 12/31/08.
29. Rapp MT, Podgorny G. Reflections on becoming a specialty and its impact on global emergency medical care: our challenge for the future. Emeg Med Clin North Am 2005;23:259-69.
30. Camargo C, Ginde AA, Singer AH, et al. Assessment of emergency physician workforce needs in the United States: 2005. Acad Emerg Med 2008;15:1-4.
31. Buerhaus PI. Current and future state of the US nursing workforce. JAMA 2008;300:2422-4.
32. American College of Emergency Physicians. Medical screening of emergency department patients. http://acep.org/practres.aspx? id=29572 Accesses 12/8/08.
33. Institute of Medicine. The future of emergency care in the United States health system. Ann Emerg Med 2007;48:115-20.
34. Falvo T, Grove L, Stachura R, et al. The opportunity loss of boarding admitted patients in the emergency department. Acad Emerg Med 2007;14(4):332-7.
35. Holliman CJ, Wuerz RC, Hirshberg AJ. Analysis of factors affecting US emergency physician workforce projections. SAEM Workforce Task Force. Acad Emerg Med 1997;4:648-50.
36. Moorhead JC, Schafermeyer RW, Rorrie CC. Key issues in emergency medicine workforce planning. Ann Emerg Med 1997;30: 765-72.
37. American Board of Emergency Medicine Physician satisfaction Study. http://abem.org/public/_Rainbow/Documents/Physician%20Satisfac- tion%20Study.pdf, accessed 3/17/08.
38. Gallery ME, Whitley TW, Kolnis LK, et al. A study of occupational stress and depression among emergency physicians. Ann Emerg Med 1992;21:70-6.
39. Goldberg R, Boss RW, Chan L, et al. Burnout and its correlates in emergency physicians: four years’ experience with a wellness booth. Acad Emerg Med 1996;3:1156-64.
40. Accreditation Council for Graduate Medical Education: List of ACGME Accredited Programs and Sponsoring Institutions. http:// www.acgme.org/adspublic/, accessed 3/17/08.
41. Society for Academic Emergency Medicine Residency Catalog: Osteopathic Residency Programs. http://www.saem.org/saemdnn/ Home/Communities/MedicalStudents/ResidencyCatalog/Osteopathi- cEMResidencyPrograms/tabid/279/Default.aspx, accessed 3/3/09.