The debate continues over resident moonlighting, defined as "a second and independent job separate from a resident's primary position without regard for institutional control over outside activities" (1). In 1988, the Association of American Medical Colleges' Executive Council endorsed the position that "[a]ccrediting institutions, medical schools, teaching hospitals, residency program directors, and faculty should work actively to halt moonlighting" (2). This position is in contrast to the 1974 American Medical Association's (AMA's) House of Delegates resolution (which ceased to be policy in 1989 when the House of Delegates allowed the resolution to lapse) that stated that housestaff contracts should guarantee that "a member of the housestaff is free to use his off-duty hours as he/she sees fit, including engaging in outside employment" (3). In 1991, Keill (4) and Factor (5) elaborated the discussion for psychiatry residents in a debate published in Hospital and Community Psychiatry, in which Keill argued for monitoring of resident moonlighting for resident and patient safety, whereas Factor said educators should educate about, not monitor, moonlighting.
In reviewing the literature, more than 30 articles were located regarding house-officer moonlighting. Most were editorials and letters to the editor (4—21), with 6 conveying negative comments about resident moonlighting (6—11) and 12 conveying favorable comments (4,5,12—21). One-third of the articles reported survey data (22—33), and three described specific programs (1,34,35). The percentage of residents who hold a second job ranges from nearly one-third, according to the "AMA Survey of Resident Physicians" (22), to 40% of 181 internal medicine residents (23), to 92% of 73 family practice residents (24). In 1983, 44% of 1,013 psychiatric residents responded that they had moonlighted or were moonlighting (25). Little data on the moonlighting of psychiatric residents have been collected since the 1983 survey. With rising medical school educational debt, more residents may feel the need to moonlight (26—28). Strong concerns about the negative impacts of increasing housestaff working hours, especially in the aftermath of the Libby Zion case in New York, still exist (36). We felt it was time to gather additional data to assess the pros and cons of psychiatric resident moonlighting.
Plans for a questionnaire developed out of discussions of the "Subcommittee on Moonlighting of Residents," a section of the Association for Academic Psychiatry. The authors constituted the working group and constructed a 41-question instrument. Some questions obtained biographical data about the respondents and demographic information about the residency programs. Other questions sought details about moonlighting activities and affective responses by using a four-point scale to gauge the effects of moonlighting on the residency program and its activities. Some questions required assignment of percentages as a response. Sufficient information to identify specific programs was not asked, because we believed that an anonymous questionnaire would produce more accurate answers and a higher return rate.
The authors felt that the chief resident at a program would be more likely to know this information, since many chief residents served as coordinators of moonlighting activities or, at least, carried the oral tradition of conveying information about available moonlighting activities to junior residents. The authors also believed that a higher percentage of chief residents would respond to the survey than individual residents or program directors. Thus, a decision was made to mail the questionnaire to the chief resident at each Accreditation Council for Graduate Medical Education—approved psychiatry residency program.
In August 1991, the questionnaire was sent to the "chief resident," in care of the residency education director, of each program. By mailing the survey at this particular time, we hoped to capitalize on the enthusiasm of the recently installed chief residents.
We initially mailed 203 questionnaires. Over the next 3 months, 99 questionnaires were returned. A second mailing produced an additional 38 returned questionnaires, for a total response of 137 programs (67.5%). One program had ceased residency training, and the form was returned uncompleted. In five programs, the chief resident had another resident complete the form. Two programs indicated that they did not have a chief resident position, but the forms were completed, in one case by a senior resident and in the other by the residency education director. Statistical analysis was conducted on the 136 completed questionnaires from the 202 programs then conducting psychiatry residency training. Analysis of the responses used the "Test for Significance of Difference Between Two Proportions" (37).
Ninety-five percent of the survey respondents were chief residents (average age: 32.9 years). Men comprised 71% of the respondents and women 29%. Seventy-seven percent of the chief residents were currently involved in moonlighting activities.
The chief residents' descriptions of their programs were medical school-affiliated (70%), military (7%), public sector (11%), freestanding private (10%), and other (2%). One-quarter of the programs were in cities with populations under 250,000, one-quarter were in cities with populations between 250,000 and 750,000, and one-half of the programs were in cities with populations above 750,000. Since the questionnaire was anonymous, it was not possible to compare the responding with the nonresponding programs.
The average number of residents in the responding programs was 27, almost equally spread across the 4 postgraduate years (PGY): PGY-1: 7, PGY-2: 7, PGY-3: 7, PGY-4: 6). Average on-call nights (nonmoonlighting) per month were PGY-1: 4.7, PGY-2: 3.4, PGY-3: 1.9, and PGY-4: 0.6.
Responses were received from programs located in all seven of the American Psychiatric Association's (APA's) geographic regions. Ninety percent of the programs permitted moonlighting. Only 46% of the responding programs had a written moonlighting policy. Each region contained programs that indicated that moonlighting was allowed. Fourteen programs (10%) indicated that moonlighting was not allowed. Nine of these programs denied that moonlighting occurred. Of these nine, two programs were in the New York/Puerto Rico region, two in the Mid-Atlantic region, one in the Midwest region, and four in the Southeast region. Five of these programs identified themselves as military, three as medical school—affiliated, and one as a state hospital. Five chief residents reported that moonlighting occurred among residents in opposition to their program's policy prohibiting moonlighting.
There was a steady increase through the residency years in the percentage of residents who were moonlighting. Less than 2% of the interns were involved in moonlighting. Roughly one-quarter of the resident pool began to moonlight during each of the succeeding years, with 26% of PGY-2, 54% of PGY-3, and 69% of PGY-4 residents moonlighting (F1).
When viewed by each APA region, the percentage of moonlighting residents by residency year varied considerably. Most moonlighting PGY-1 residents came from the Southeast region. Considerably more Southeast-region and California-region residents were involved in moonlighting in the PGY-2 year than those from other regions. The percentage of PGY-3 moonlighting residents varied, from a low of 33.9% in the Mid-Atlantic region to a high of 72.6% in the California region. By the PGY-4, the number of residents involved in moonlighting was fairly consistent across all regions (F2).
Monthly moonlighting income rose steadily through the residency years. Average reported income per month for residents by year was as follows: PGY-1: $517, PGY-2: $813, PGY-3: $1,198, and PGY-4: $1,402.
Overall, each moonlighting resident spent 31.4 hours per month, or about one 8-hour shift per week, moonlighting. The Northeast region reported the highest (36.6 hours/month), and the California region reported the lowest (19.1 hours/month) (F3). The exact period of moonlighting hours across the week or month was not obtained, but several penciled-in comments suggest mostly evening or overnight ("graveyard") moonlighting shifts.
Weekday daytime moonlighting shifts were reported for 10% of the residents. One-third reported working weekday night shifts. One-half of the residents worked on weekday evening shifts, and a similar amount on weekend day/night shifts. On average, the residents traveled 17 miles (one way) to moonlight.
The major reason for resident moonlighting, by rank-order responses from a list of options, including "other, please specify," was the payment of living expenses (58%), followed by payment of student educational loans (24%). Together, these two reasons accounted for five-sixths of resident moonlighting. The traditional reasons for moonlighting— paying for psychotherapy (6%), educational benefits (5%), and advancing career goals (2%)— were rarely given (F4).
Because most residents were moonlighting in the specialty area in which they were also receiving education/training, the provision of supervision for these activities was queried. On-site supervision was provided to only 10% of the moonlighting residents, while no supervision was provided to 22% of the moonlighting residents. Required telephone supervision occurred for 14% of the moonlighting residents. The most common system of supervision (51%) was the availability of telephone supervision, if desired by the resident (F5).
The following definitions were used to determine each program's approach to moonlighting: 1) "APPROVED" moonlighting: moonlighting in which the residency program officially permits moonlighting by residents; 2) "MONITORED" moonlighting: moonlighting in which the residency program permits moonlighting and monitors the location, hours, and general scope of resident moonlighting; and 3) "SPONSORED" moonlighting: moonlighting in which the residency program permits and monitors moonlighting, as well as sponsors moonlighting assignments by arranging and recruiting residents who want to moonlight.
One hundred three (76%) of the responding programs reported that moonlighting activities were approved. Fourteen programs (10%) did not allow moonlighting, and 19 programs (14%) were aware that their residents were moonlighting, but these programs had no established policy or oversight procedures.
In addition to the approval process, 58 (43%) of the responding programs monitored the resident's moonlighting activities. Similarly 31 (23%) of the responding programs were involved in actually sponsoring moonlighting activities, in addition to the approval and monitoring process. Sixty-nine programs (51%) identified a specific person who coordinated the moonlighting activities, while 67 programs (49%) had no formal coordinator. Of the programs with a coordinator, the chief resident had this role in 27 programs (39.1%), a resident who was not chief in 16 programs (23.2%), the residency training director in 15 programs (21.7%), and another unspecified individual in the remaining programs.
In a series of questions about preferred location sites for moonlighting, the chief residents preferred moonlighting in a setting outside the training institution, followed by the preference to moonlight within their own training program. The residents' rank-ordered preferences for moonlighting activities were covering private inpatient wards, working at community mental health centers, performing private histories and physicals at private hospitals, and working in psychiatric emergency rooms. However, the most common moonlighting sites, in order of occurrence, were psychiatric emergency rooms, private hospitals, public hospital wards, and community mental health centers.
The chief residents reported that 62% of their moonlighting time was in clinical activities, with the only exception being the residents from the Northeast region, who reported only 43% of their moonlighting time devoted to clinical activities. The nature of the nonclinical activities was not assessed. The predominant specialty area in which the moonlighting residents worked was psychiatry, with all other areas totaling only 13%.
Moonlighting activities, as all clinical activities in medicine, require a state medical license. Twenty-three (17%) programs have residents who use the institutional training permit license for moonlighting within the institution. All respondents report that an unrestricted state license may be used for moonlighting in their venue, either inside or outside the institution. The respondents were not asked to compare moonlighting activities for residents with and without unrestricted state licenses.
Thirty-one percent of the residents who moonlight had arranged for their own personal professional liability insurance, whether or not the residency program's insurance covered their moonlighting activities. The average cost was $1,484 per year, with a high of $2,088 in the APA's Midwest region and a low of $700 in the APA's Northeast region. Fifty-eight percent of the residents believed they were covered by some form of institutional liability insurance. Of particular interest and concern were the 2.9% of residents who were moonlighting with no professional liability insurance coverage, either institutional or personal. Unfortunately, the anonymous nature of the survey did not allow this issue to be pursued and addressed. Type of coverage, amount of coverage, and full-time/part-time coverage of professional liability insurance, including presence or absence of "tail" coverage, were not assessed.
Chief Residents' Attitudes About Moonlighting
Six questions were asked that used a strongly agree/agree/disagree/strongly disagree (a four-point modified Likert scale) format to assess the attitudes of the chief residents about moonlighting's impact on residents and the residency program. With the limited number of respondents, and to make the data more understandable, the results were collapsed into an agree/disagree dyad for statistical and reporting purposes.
The study sample was divided into moonlighting chief residents vs. nonmoonlighting chief residents, men vs. women, and age (those older vs. those younger than the median age of 31). The male-vs.-female and age comparisons produced no significant differences for the questions that assessed the chief residents' opinions. When the respondents who moonlight were compared with those who do not, the test for significance of difference between two proportions, however, produced significant differences at the 0.05 confidence level for four questions (T1).
Most chief residents (62%) indicated that they would not join a residency program in which moonlighting was prohibited. They further reported that about 47% of medical student applicants to psychiatry residency programs inquired about the availability of moonlighting experiences at their program.
There are obvious limitations to using survey data sent solely to chief residents. However, the authors felt that any drawbacks would be offset by a potentially higher response rate than if residents had been surveyed directly. Our decision was on target and is reinforced when we compare our 68% response rate with the 27% response rate for a similar 1983 study in which residents were polled directly (25).
Seventy-five percent of the chief residents surveyed were currently moonlighting and, hence, are representative of the PGY-4 residents who moonlight. Our finding that 52.3% of PGY-2 through PGY-4 residents moonlight was comparable to 1976 and 1983 data, further suggesting that our findings are representative (29,25).
Seventy-six percent of the psychiatric residency programs who responded to our survey permitted moonlighting, while an additional 14% did not officially approve but did not intervene to stop residents from moonlighting. This latter figure compares with approval rates of 70% in pediatric programs and 97% in nonmilitary family practice programs found in other studies (30,31).
The AMA's Accreditation Council for Graduate Medical Education General Requirements for Residencies specifies that "each resident be offered for acceptance a written agreement encompassing the following…practice privileges and other activities outside the educational program" (38). The Special Requirements in Psychiatry further states, "the [residency] program should carefully monitor any activity outside the residency that interferes with education, performance, or clinical responsibility. The program should carefully monitor all on-call schedules and hours within and outside of the residency to prevent interference with education, performance, or clinical responsibility."
Yet a 1989 American Association of Directors of Psychiatric Residency Training survey revealed that only 44% of programs monitored their residents' moonlighting activities (32). This percentage is almost identical to our survey result of 43% of responding programs who reported monitoring their residents' outside activities. Compliance with the Residency Review Committee's requirement does not appear to be a high priority among program directors.
Roughly one-quarter of programs actually sponsor moonlighting activities. It would be interesting to document what these programs actually do. Cohen and Leeds have reported on the University of Massachusetts Medical Center's "extended employment" experience, in which moonlighting was brought within the control of the institution for better supervisory control and for the increased ease of coordination for the residents, as well as lower professional liability-coverage rates through the institution (1). Cohen has described the University of Pittsburgh's role as a clearinghouse and service-contract provider (34). Yingling et al. (35) describe a similar system at the University of Cincinnati, where residents conduct peer-review sessions for each other's activities. Programs such as these probably reflect the fact that hospital affiliates still increasingly require on-site medical personnel on evenings and weekends. Housestaff provide this service at a less expensive rate than fully trained physicians.
Four studies have cited loan indebtedness as a major factor prompting residents to moonlight (22,23,27,28). With 81% of 1989 graduating medical students incurring average loan debts totalling $50,000 (39), it is logical to surmise that repaying medical school debt is a driving force spurring residents to seek additional income (27). Yet only 24% of our respondents indicated that payment of student educational loans was the major reason for moonlighting. The majority of residents indicated that paying living expenses was the primary reason for moonlighting, which is also consistent with Buch and Swanson's results (25). Compared with their nonmoonlighting counterparts, moonlighting residents tend to have larger families, nonworking spouses, higher noneducational debts, and home mortgages. Researchers speculate that maintaining a higher standard of living is the principal motivation for moonlighting (27,28). Since psychiatric residents have educational debt levels similar to nonpsychiatric residents (27,39), one can only speculate as to why debt does not appear to be a dominating factor. As medical education costs continue to rise, this factor may yet become an important consideration.
The observation that few first-year residents moonlight is probably due to the fact that, in most states, they do not have unrestricted licenses to moonlight outside the site of their institutional permit or temporary training license. Hence, any moonlighting by PGY-1s most likely occurs in program-sponsored activities. The time-intense experience and demands of the first-year internship may also be factors. Descriptions of training programs that have sponsored moonlighting activities for their residents would be valuable, including the mechanics of the activity, resident responses, and effect on recruiting.
The increase in income earned by ascending residency year is probably caused by residents working more moonlighting hours in each successive residency year. This gradual income increase also reflects the movement of residents into higher-paying activities, as they find these options available when senior residents have completed the program and move into higher-paying jobs.
As a group, chief residents who moonlight view the experience in a more positive light than those chiefs who do not moonlight. The only equivocal response involved the question on whether moonlighting affected the energy available for residency duties. Only one-third of moonlighting chief residents admitted that their energy level was affected, whereas two-thirds of the nonmoonlighting chief residents group felt this to be true. The nonmoonlighting chief residents were also four times more likely to believe that moonlighting seriously interfered with family life and outside social activities.
While the residents did not appear to moonlight excessively (averaging one 8-hour shift per week, although it is not clear if this time is in a single block or in 1- or 2-hour increments), 10% reportedly worked on weekdays. Interference and conflicts with the residency program are of concern. The study finding that sleep deprivation enhances performance among surgery residents notwithstanding (40), most psychiatrists will attest that it is difficult to conduct therapy after a sleepless night.
The findings on supervision of moonlighting residents are a source of some concern. Almost one-quarter of the moonlighting residents were not supervised. While in years past it may have been considered part of training for doctors to make independent clinical decisions, today's ethos reinforces the resident's role as trainee, requiring supervision or consultation to ensure that patients receive appropriate care. Three percent of the residents who moonlighted were doing so without malpractice insurance and were jeopardizing their own financial futures. If the number of family practice residents who moonlight is any indication, many residents are not even knowledgeable about the distinction between occurrence and claims-made malpractice insurance and the importance (and expense) of tail coverage (33).
Perhaps the time has arrived for training programs to incorporate remunerated service positions into their residencies. This change might be the only way to ensure supervision and monitorship for the allowed maximum hours worked each week (as required in New York State). However, increased involvement by residency training directors in managing and supervising moonlighting activities may produce greater liability exposure for the residency education director regarding resident practice in areas in which the director may have little administrative control. Yet fulfilling affiliate hospitals' medical personnel needs has been successfully done in the programs described in the literature (1,34,35). In doing so, the term "moonlighting" would be technically eliminated also.
Our findings have updated and confirmed prior surveys' findings. Moonlighting occurs among a substantial number of residents, and it even occurs when the activity is specifically proscribed. These data can serve as the starting point for a more informed and balanced discussion about the educational issues and mechanics involved with moonlighting, rather than as a basis for further editorials that condemn the evils of it. The lessons of prohibition should not be lost. A better approach may be to truly monitor and guide the moonlighting experience rather than simply ban it, ignore it, or wish it away.
Data from this manuscript were presented at 1) the Annual Meeting of the American Psychiatric Association, San Francisco, CA, May 1993; 2) the Mid-Winter Meeting of the American Association of Directors of Psychiatric Residency Training, San Diego, CA, January 1993; and 3) the Annual Meeting of the Association for Academic Psychiatry, Charleston, SC, March 1993.