Survey of Moonlighting Practices and Work Requirements of Emergency Medicine Residents

JAMES LI, MD,* RICHARD TABOR, MD,I AND MAURICIO MARTINEZ, MD:I:

We conducted an anonymous moonlighting and academic practice survey of all emergency medicine residents enrolled in accredited programs during 1997. Expanding on previous work, this survey included specific details and practice trends of moonlighting emergency medicine residents and for comparison also included academic work require- ments. The typical emergency residency program requires residents to work 204 hours monthly. However, the range of required work-hours is strikingly large (120-300). Half of emergency medicine residents moon- light. The typical moonlighting resident works as a solo emergency department practitioner in multiple facilities outside of residency- affiliated institutions. Moonlighting salaries generally double a resident's annual income and are used to pay off student loans and other debt. Residents with higher student debt are more likely to moonlight. Despite the fact that most residency programs restrict moonlighting, a majority of moonlighting residents have violated an Accreditation Council for Gradu- ate Medical Education prohibition restricting work within one period of a regular residency-scheduled shift. Half of all residents surveyed, whether involved in moonlighting practice or not, would violate a ban on the practice. Residents universally felt that moonlighting enhanced resi- dency performance and was a positive educational experience. Use of these data may aid in the development of formal guidelines regarding emergency medicine moonlighting practice. (Am J Emerg Med 2000;18: 147-151. Copyright © 2000 by W.B. Saunders Company)

Moonlighting among emergency medicine residents is relatively controversial because of its purported negative impact on educational responsibilities and academic endeav- ors. 1,2 Nevertheless, a 1994 survey sponsored by the Society for Academic Emergency Medicine (SAEM) confirmed moonlighting is widespread with half of all responding residents engaged in moonlighting practices. 3

Although broad in scope, that survey did not examine details of moonlighting pertaining to practice types, salaries, regional variations, moonlighting restrictions and enforce- ability of such restrictions, hourly schedules, patient work- loads, and benefits to emergency residency education. Indeed, not much is known about the specifics of resident moonlighting.

Prevalence of resident moonlighting has significant impli- cations for emergency medicine education. To further under- stand the specifics of moonlighting practices, we conducted a second recent and more detailed survey. For workload comparisons, residents were also surveyed on their clinical

From the Division of Emergency Medicine, Harvard Medical School, Mount Auburn Hospital, Cambridge, MA; the Department of Emergency Medicine, Pocono Medical Center, East Stroudsburg, PA; and the Section of Emergency Medicine, LSU Medical Center at Charity Hospital, New Orleans, LA.

Manuscript received July 14, 1999, returned August 17, 1999, revision received August 25, 1999, accepted October 12, 1999.

Reprints are not available. Key Words: Moonlighting, emergency medicine, residency, salary. Copyright © 2000 by W.B. Saunders Company 0735-6757/00/1802-0007510.00/0

academic work requirements. Use of these data may aid in the development of formal residency program or national guidelines regarding emergency medicine moonlighting practice.

METHODS

We administered a 29-point anonymous survey regarding moonlighting practices and perceptions to all 3,029 residents enrolled in emergency medicine programs accredited by the U.S. Accreditation Council for Graduate Medical Education (ACGME) during the 1996-1997 academic year.

The survey combined multiple-choice with fill-in re- sponses and was developed by the authors, all senior residents during the study. Survey distribution was per- formed by mailing to chief residents and program directors, with instructions to disburse surveys to individual residents. Postage paid return envelopes were provided with each survey. Funding for the project was donated by the authors and by the Louisiana State University emergency medicine program at Charity Hospital.

Statistical analysis of the results was performed using standard methods for single and comparative proportions and was reviewed by a statistical consultant. Confidence intervals were calculated and reported. 4

This study was undertaken to measure prevalence, so a cross sectional survey was the method of choice. Because of this, a low response rate was anticipated. However, because of lack of funding no effort to resurvey nonresponders Was possible. For this reason, validity of data was confirmed by comparing a subset of key results shared by this and the previous SAEM-sponsored survey, whose response rate exceeded half of those surveyed. These results included percent of respondents who had moonlighting practices, percent whose programs imposed moonlighting restrictions, average hourly moonlighting wages, and percent who had been sued in moonlighting cases. Similarity in key results was considered a marker of validity.

RESULTS

Respondents, Extent, and Characteristics of Moonlighting

Of 3,029 potential respondents 887 nonblank responses were returned for a response rate of 29%. Approximately a quarter of the total respondents were distributed to each of four U.S. regions: Northeast, Midwest, South, and West.

Half (53%, 95% confidence interval [CI] 47% to 56%) of all emergency medicine residents responding had moonlight- ing practices. This proportion increased to over three- quarters in senior residents (R3 and above). Moonlighting

147

148 AMERICAN JOURNAL OF EMERGENCY MEDICINE • Volume 18, Number 2 • March 2000

TABLE 1. Respondents, Extent, and Characteristics of Moonlighting

Percent Year Percent

By region Northeast 28 By year 1 20 Midwest 28 2 31 South 21 3 37 West 23 > = 4 12 Total 1 O0 Total 100

Moonlighters Percent 95% CI

Unadjusted 53 47-56 By postgrad year 1 12 7-17

2 42 36-48 3 77 72-82 4 80 72-88 >=2 64 60-68 > = 3 78 74-82

Moonlighting Practices Percent 95% CI

ED setting 74 70-78 Multiple facilities 81 78-85 Outside residency facility 67 63-71 Within residency facility 28 24-32 Clinic setting 41 37-45 Other (H&Ps) 11 8-14 Contract through small group 27 23-31

through large contract company 24 20-28 through individual hospital 28 24-32 through residency-affiliation 17 14-20

Solo coverage only 58 53-63

NOTE. Noninteger'values have been rounded to the nearest integer.

residents worked primarily in emergency departments (74%, CI 70% to 78%) and in multiple facilities (81%, CI 78% to 85%). Most but not all worked outside of residency- affiliated institutions (67%, CI 63% to 71%). Non-ED moonlighting practices included clinic work (41%, CI 37% to 45%) and physical examination services (11%, CI 8% to 14%). Approximately one-quarter of moonlighters each contracted through small groups (27%, CI 23% to 31%), large contract management groups (24%, CI 20% to 28%), and individual hospitals (28%, CI 24% to 32%), respec- tively.

Over half of moonlighting residents provided solo cover- age while working (58%, CI 53% to 63%). This proportion of residents providing solo coverage did not vary by training level, Residents estimated treating an average of 2.8 patients per hour while moonlighting (CI 2.7 to 2.9).

Moonlighting Restrictions and Enforceability of Such Restrictions

Most residents reported that their emergency residency programs allowed moonlighting in some form (91%, CI 89% to 93%). However, the majority of programs had preestablished moonlighting restrictions (73%, CI 70% to 76%). The three most common of these included restriction of moonlighting privileges to upper level residents, restric- tion to residents who had achieved passing scores on in-service examinations, and restriction to residents who had obtained written approval.

Many respondents commented on additional moonlight- ing restrictions. The most common example restricted

residents to a limited number of moonlighting hours per month. These monthly limits included 24, 36, 40, 48, 60, 72, and 96 hours. Several programs reemphasized the require- ment by the ACGME that no moonlighting occur within one work period before or after a regularly scheduled residency shift. Others restricted moonlighting to residents who had completed their research requirements. All military pro- grams expressly prohibited moonlighting. One program restricted moonlighting to residents who had completed trauma and orthopedic rotations.

Moonlighting was important enough to residents that roughly half, including those who did not currently moon- light, would violate a program-directed ban prohibiting the practice. Most moonlighting residents had violated the requirement by the ACGME prohibiting work within one period of a regular residency-scheduled shift (69%, CI 65% to 73%). 5

Motivations and Attitudes Toward Moonlighting

The most frequently cited motivation for moonlighting among moonlighting residents was the educational experi- ence it afforded (88%, CI 85% to 91%). Second was the supplement it provided to residency income (66%, CI 63% to 69%). In addition, most moonlighting residents simply enjoyed their practices (72%, C168% to 76%).

Whether they had moonlighting practices or not, residents universally believed that moonlighting provided a positive educational experience (98%, CI 97% to 99%). Typical comments received indicated that moonlighting provided opportunities "to help decide a future type of practice," "to ride on helicopter runs," and to see a "fun and different atmosphere apart from residency." In addition, moonlight- ing and nonmoonlighting residents universally believed that moonlighting enhanced residency academic performance (94%, CI 92% to 96%). In addition to survey choices, residents also commented on beneficial experiences not obtained from residency in areas of orthopedics, obstetrics and gynecology, otolaryngology, cardiology, and ultrasonog- raphy. Many commented additionally that moonlighting experiences "enhanced strengths" obtained in residency.

Most residents reported a primary motivation for moon- lighting was to pay off student loans (61%, CI 56% to 66%) or other debt (66%, CI 63% to 69%). Approximately half of

TABLE 2. Moonlighting Restrictions and Enforceability of Such Restrictions

Percent % CI

Specific restrictions Only upper level residents 75 71-79 Must pass in-service exams 52 48-56 Only by written approval 51 47-55 In-house only moonlighting 8 6-10

Respondents who would defy a moon- lighting ban

Moonlighters 55 50-60 Nonmoonlighters 40 32-48

Respondents who have worked within 12 hours of an ED shift 69 65-73

LI, TABOR, AND MARTINEZ • MOONLIGHTING SURVEY 149

TABLE 3. Motivations and Attitudes Toward Moonlighting

Percent 95% Cl

Respondents who enjoy moonlighting 72 68-76 Respondents who tolerate moonlighting 28 24-32 Respondents who hate moonlighting 1 0-2

Moonlighters Nonmoonlighters

Percent 95% CI Percent 95% CI

"Definitely" qualified to moonlight 78 74-82 40 34-46

"Sometimes" quali- fied to moonlight 21 17-25 46 60-52

"Not" qualified to moonlight 1 0-2 14 10-18

Moonlighting educa- tionally positive 99 98-10O 97 95-99

Enhances residency performance 97 95-99 89 85-93

Respondents who believe moonlighting fills residency voids in the following areas

Percent 95% CI

Independent decision making 73 69-77 Private attending relations 39 35-43 Private patient relations 39 35-43 Rural hospital medicine 30 26-34 Pediatrics 26 22-30 Resuscitation skills 9 6-12 Suturing skills 8 6-10 Airway procedures 4 2-6 Trauma 4 2-6

residents who moonlighted required the additional salary it provided to meet "basic expenses" (41%, CI 37% to 45%). Child and housing expenses were the two most common write-in responses.

Self-selection was a factor for moonlighters. The majority of moonlighting emergency medicine residents felt strongly qualified to do so. Of nonmoonlighting residents, most felt poorly or partly qualified to moonlight.

Student Debt, Moonlighting Salaries, and Malpractice Issues

Residents with higher student debt were more likely to moonlight. Excluding interns, who generally did not moon- light, 64% of moonlighters versus 48% of nonmoonlighters owed $60,000 or more in student debt (absolute difference 16%, P < .0001). (When interns were included in the analysis, the absolute difference, 8%, was still significant at P = .03.)

The average annual moonlighting salary was in the same range as residents' training salaries. A quarter of moonlight- ing respondents earned more in these practices than they did as residents (28%, CI 24% to 32%). Ten respondents earned over $75,000 annually from their moonlighting practices (one R2, six R3, and three R4 residents). Three southern residents, all in the R3 year, earned $100,000 in their most recent year of moonlighting practice.

Malpractice coverage was widely perceived as adequate.

TABLE 4. Student Debt, Moonlighting Salaries, and Malpractice Issues

Moonlighters Nonmoonlighters

Student Debt Percent 95% CI Percent 95% CI

$ 0 18 15-21 29 23-35 $ 30,000 17 14-20 22 17-27 $ 60,000 24 20-28 17 12-22 $ 90,000 21 17-25 14 10-18 $120,000 12 9-15 12 8-16 $150,000 8 6-10 6 3-9

Excluding interns 64 60-68 48 42-55 >=$ 60,000

Moonlighting Salaries $ (mean) SD Range

Average wage/hr 65 16 20-125 Highest wage/hr 82 31 20-250 Annual income (K) 25 3 0-100

Malpractice Percent 95% CI

Extra coverage provided 95 93-97 Coverage perceived adequate 95 93-97 Respondents sued moon- lighting 2 1-3

However, one in fifty moonlighting residents had been sued in cases involving their practices (2%, CI 1% to 3%).

Moonlighting Hours and Residency Work Requirements

Emergency medicine residency programs imposed a wide variation in hourly resident work requirements. The average residency workload when performing emergency depart- ment rotations was 19 11-honr shifts (rounded figures).

This point bears emphasis. Although residents enrolled in typical programs worked an average of 204 hours per month, ED shift work in all programs ranged from 120 to 300 hours. Thus, residents enrolled in programs with the lowest hourly clinical requirements worked 84 hours less, and residents in the highest requirement programs 96 hours more, than those enrolled in programs with average requirements. Strikingly, those in the lowest hourly requirement programs worked 180 hours less per month than those enrolled in the highest requirement programs.

TABLE 5. Moonlighting Hours and Residency Work Requirements

Mean SD Range

Moonlighting Hours/month 28 18 0-125 Longest continuous

stretch 19 16 6-150 Moonlighters who work

more than 12 hours in a row 28% (95% CI, 24-32)

Residency Mean Hours/month 204 28 120-300 Shifts/month 19 3 10-27 Hours/shift 11 1 8-16

150 AMERICAN JOURNAL OF EMERGENCY MEDICINE • Volume 18, Number 2 • March 2000

In comparison, emergency medicine residents moon- lighted an average of 27 hours per month, with a range of between 0 and 125 hours. A quarter of moonlighting residents routinely worked over 12 hours in a row (28%, CI 24% to 32%) and an additional quarter of moonlighters had worked 24 or more hours in a row (24%, CI 20% to 28%). Four respondents from southern programs reported moon- lighting over 100 hours in a row.

Validity of Data

To test the present study for validity, results for four key questions from the previous SAEM-sponsored survey were compared with results from the current survey. The percent of moonlighting residents reported by the previous survey was 49%, compared with 53% (CI 47% to 56%) in this survey. (No confidence intervals were reported in the previous survey.) Formal residency moonlighting policies were previously reported by 75% of respondents, compared with 73% (CI 70% to 76%) in this survey. Average hourly moonlighting wage for residents in their R3 and R4 years was reported previously as $64 and $67, respectively. In comparison, residents in their R3 and R4 years reported earning an average of $67 (CI 65 to 69) and $68 (CI 66 to 71) per hour in this survey. Finally, 2% of residents previously reported being sued in moonlighting cases, compared with 2% (CI 1% to 3%) in the current survey.

DISCUSSION

Many previous studies indicate that emergency depart- ment moonlighting is common practice for residents from many specialty training programs. However, emergency medicine residents occupy a unique role in this practice, because moonlighting is perceived to expand their educa- tional experience, introduce them to future practices, and influence professional career decisions. ~4 Despite this role, moonlighting has never been formally recognized as an integral part of emergency medical training, and national guidelines are absent.

This survey was limited by a low response rate which may have led to selection bias. Despite this limitation, the survey was not based on a statistical sampling model, but was sent to the entire population of emergency medicine residents with the belief that selection bias would be preferable to sampling bias. In addition, validity was confirmed in all key comparisons to the previous SAEM survey, which had a response rate of 76%.

The present survey portrays both the pervasiveness and perceived importance of moonlighting practice among emer- gency medicine residents. The perception among residents that moonlighting provides both additional income vitally needed to overcome debt and practical experience essential to complete residency education is strongly held, and may act as a relative barrier to formal policies which would curtail or eliminate this practice. This is further evidenced by the number of residents who reportedly would defy a ban on moonlighting.

ACGME policy pertaining to moonlighting in emergency medicine programs is arguably vague. Many programs justify moonlighting restrictions based on an ACGME directive mandating "at least an equivalent period of

continuous time off between scheduled work per iods . . , and no more than 60 scheduled hours per week seeing patients in the emergency department . . . whether spent within or outside the educational program. ''5 However, such language is subject to interpretation and the definition of "time off" has never been formally challenged. In addition, our data show that a third of residents moonlight in residency affiliations, suggesting that distinctions are made between "scheduled hours" and extracurricular hours counted and paid separately by programs with manpower shortages.

As a final note on the issue of residents' work hours, the finding of widespread variation in educational duty hours is startling. Residents in programs failing into the extremes of range appear to be both under and overworked, and both extremes have detrimental implications for emergency medi- cal education.

Moonlighting emergency medicine residents are exposed to significant liability, evidenced by the large proportion of residents who work in solo practice, and the significant number of residents who have been involved in malpractice cases associated with moonlighting. Such considerations also raise patient care questions, particularly when neophyte moonlighting residents are practicing without sufficient backup resources. The majority of residents who moonlight, however, reported feeling qualified to do so, and often commented that being specifically trained in emergency medicine, they provided superior patient care than could be afforded by moonlighting residents from other specialties .9

In light of the widespread prevalence of emergency medicine moonlighting, wide variations in practice, issues of liability, and lack of specific assistance to residents involved in moonlighting, we believe that this topic is important to the future of emergency medicine training. As such, we urge individual program directors, national resi- dents' associations, and national emergency medicine groups to develop guidelines which address the role of emergency medicine residents in the larger workforce of emergency medical care. Such guidelines could directly address issues of minimum competency required for moonlighting prac- tice, provide education in contractual negotiations, recom- mend minimum needs for liability and malpractice cover- age, and put into place resources for residents who require assistance while engaged in moonlighting practice.

We believe that guidelines such as these would enhance patient care by reducing the number of moonlighting residents poorly qualified to do so, by addressing backup resources for moonlighting use, and, through resident educa- tion, by diminishing activities of contractual agencies who provide questionable patient care and liability coverage. Despite the fact that residents are still in training, patient care is most likely enhanced when qualified emergency residents replace moonlighters from specialty programs outside of emergency medicine.9.1°,J 1

The balance between patient care, resident education, and resident debt is a delicate one. Such a balance has thus far been dictated by chance and a free market. We believe that formal recognition of emergency medicine resident moon- lighting practice would provide stability to this balance, and be of benefit to residency programs, to residents, and ultimately to the patients for whom they care.

LI, TABOR, AND MARTINEZ • MOONLIGHTING SURVEY 151

The authors would like to offer grateful appreciation to Linda Hill and Evelyn LeBlanc for their invaluable assistance in performing this survey, as well as Jon Dwyer, for his assistance with statistical analysis.

REFERENCES

1. Majikian AM, Brinker MR, Rice JC, et al: Moonlighting in a university surgical training program. South Med J 1993;86:441-446

2. Krome RL: Moonlighting. Acad News Views 1992;8:3 3. Langdorf MI, Ritter MS, Bearie B, et al: National survey of

emergency medicine moonlighting. Acad Emerg Med 1995;2:308- 314

4. Kirkwood BR: Essentials of Medical Statistics. Oxford, Black- well Scientific Publications, 1988, pp 1-234

5. ACGME: Program requirements for residency education in emergency medicine. 1995;IV:7

6. Banahan BF Jr, Anderson RL, Banahan BF III, et al: Physicians' evaluation of their moonlighting during residency training. J Med Educ 1987;62:351-353

7. Cohen SN, Leeds MP: The moonlighting dilemma: Balancing education, service, and quality care while limiting risk exposure. JAMA 1989;262:529-531

8. Chisholm C: Moonlighting: An appropriate EM resident en- deavor. Acad New Views 1992;8:2

9. Kellerman AL: Moonlighting. Ann Emerg Med 1995;26:83-84 10. Kasman DL: When a heart stops. Ann Intern Med 1994;120:

432-433 11. Wilson LA, Cummins RO: Moonlighting revisited. Ann Emerg

Med 1996;26:102