Schindler, Barbara A. MD; Novack, Dennis H. MD; Cohen, Diane G.; Yager, Joel MD; Wang, Dora MD; Shaheen, Nicholas J. MD, MPH; Guze, Phyllis MD; Wilkerson, LuAnn EdD; Drossman, Douglas A. MD
The term “crisis” has been used to describe the state of American medical care for a number of years. The growth of managed care, escalating malpractice premiums and decreasing physician autonomy and reimbursements have stressed medical practitioners as never before.1 In a Kaiser Foundation national study, a majority of practicing physicians reported that their morale had waned in recent years.2 For instance, 45% of 2,608 physicians surveyed would not recommend the practice of medicine to young persons today and 76% said that managed care had a negative impact on how they practiced medicine. In all, 88% said they had decreased time to spend with their patients and 73% believed that health care quality has been compromised. Academic medical centers, the training grounds for future generations of physicians, have been affected by these same social and economic trends. Ludmerer, in Time to Heal,3 argues that teaching in academic medical centers and medical schools is in serious jeopardy because of decreasing funds from governmental and private sources, increasing emphasis on research, and pressure to generate clinical revenues. Although these changes are likely affecting faculty, trainees, and patients in adverse ways,4 there is a paucity of studies that support this perception.
Financial uncertainty and colleague attrition, noted by three of us (BS, DN, DC), appeared to contribute to increased physical and emotional morbidity and seemingly excessive mortality among faculty. This observation led us to ask if financial and other stressors were having similar effects in other medical schools. Job loss and financial uncertainty have been implicated in adverse effects on health.5–7 There is a growing literature on “burnout” among physicians and residents and its effects on physicians’ health and patient care.8–11 Less attention has been paid to faculty in academic medical centers. Gabbe’s study of obstetric and gynecology chairs12 indicated high levels of burnout, especially in younger chairs, and emphasized the importance of marital support in preventing burnout. Demmy et al.13 in comparing academic medical faculty who stayed at their institutions with those who left, found administrative frustration high on the list of reasons given for job dissatisfaction and for leaving academic medicine.
To our knowledge, no studies have compared the responses of a large number of academic medical faculty to the rapidly changing academic health care environment across several institutions and across disciplines. The most comprehensive study to date by Linn et al.14 evaluated the health status, job satisfaction, life satisfaction, and job stress among 211 volunteer and full-time academic faculty in internal medicine at one western public medical school in 1984.
We hypothesized that changes in the academic health care environment have created sufficient stress to negatively affect the professional and personal lives, as well as the physical and mental health, of full-time academic faculty; that measures of institutional instability such as colleague attrition and negative perceptions of institutional financial health would be associated with increased negative reports of faculty members’ health and well-being; that older faculty who had experienced the most changes in their academic careers would be more negatively affected by the changing environment, resulting in poorer mental and physical health and decreased job and life satisfaction; and that basic science faculty and medical specialties would be differently affected by these changes.
By modifying the survey instrument used in 1984 by Linn et al.14 the authors developed a 136-item self-administered questionnaire that was distributed to four medical schools: one private, three public, located in the east, the southwest, and the western part of the United States with interested researchers, representing a diversity of geographic locations, practice patterns, and revenue streams. One school was the same institution studied in 1984 by Linn and colleagues. Human subject approval was granted at the four participating institutions. Two of the four institutional review boards stipulated that the name of their institution not be disclosed; accordingly, we have withheld all four names.
We distributed the questionnaire, which took approximately 20 minutes to complete, to all 3,519 clinical and basic science faculty members employed at least half-time by the four schools between August 2000 and November 2001 in up to three mailings. The questionnaire was to be answered anonymously but included questions about sociodemographic characteristics, namely sex, marital/partner status, number of children, religiosity, race, employment data, and discipline. We included three scales used in the 1984 study: a revised 16-item work satisfaction scale (the Physician Job Satisfaction scale [PJSS], a five-point scale where 1 = “very dissatisfied” and 5 = “extremely satisfied”),14 the 10-item Rand Anxiety scale (a six-point scale where 1 = “never” and 6 = “always”),15 and a nine-item Life Satisfaction Scale (LSAT, a seven-point scale where 1 = “terrible” and 7 = “delighted”).16 We added to the questionnaire measures of physical and mental health status, medications use, behavioral response to stressors, and perception of institutional health. We also included Revicki’s 18-item Work-Related Strain Inventory (WRSI, a four-point scale where 1 = “applies” and 4 = “does not apply”),17 and the 20-item Center for Epidemiological Studies Depression Scale (CES-D, a three-point scale where 0 = “rarely” or “none of the time” and 3 = “most” or “all of the time”).18 We asked participants to provide specific information on colleague attrition and its impact, and encouraged them to make additional open-ended comments on their current work or life situations. Data return, entry, and analysis were centralized at one institution. The questionnaire is available from the authors upon request.
We made comparisons with continuous variables using the t-test. We compared data from the four schools using analyses of variance (ANOVAs). Multivariate analyses were used to test covariance of age and sex; we considered a p value less than .05 significant. All analyses were performed using SPSS version 11.5 (Chicago, IL: SPSS Inc., 2002).
A total of 1,951 of 3,519 basic science faculty and academic physicians returned the questionnaire, for an overall 54.3% response rate, with response rates varying between medical schools (school 1 = 42.1%, school 2 = 53.8%, school 3 = 54.4%, school 4 = 66.8%). We present responses to the major components of the survey, including respondents’ assessment of their physical and mental health, life satisfaction, and job satisfaction, and report differences between basic scientists and academic physicians and across specialties, as well as significant sex and age/academic rank differences, marital status differences, and differences between the four institutions. Tables 1–3 illustrate differences in respondents’ work characteristics, faculty responses to current stressors, and differences across disciplines and specialties.
In all, 66% (1,284) of respondents were male and 34% (660) female. Respondents had a mean age of 47.4 years; the mean age for men was 49.0 and for women was 44.3 (p < .001). The sample was similar in male/female composition to the population of faculty members in academic medicine reported by the Association of American Medical Colleges (AAMC) in 2001.19 Ninety percent (1,695) of respondents reported a stable marital or committed relationship, with significantly more men (94.8% [1,210]) than women (83.1% ) married (p < .001). Seventy-seven percent (1,516) of respondents had children, with a mean of 2.25 children. Most respondents (80% [1,550]) were Caucasian, 49 (2.5%) were African American, 62 (3.2%) were Hispanic, and 203 (10%) were Asian American. Five hundred forty nine respondents (28.1%) were Protestant, 372 (19.5%) Catholic, 368 (19.3%) Jewish, 30 (1.6%) Hindu, 25 (1.3%) Muslim, and 150 (7.9%) other, whereas 414 (21.7%) did not identify a religious preference.
Most respondents (1,457 [75%]) were academic physicians spread across academic ranks: 621 (32.2%) professors, 540 (28%) associate professors, 701 (36.4%) assistant professors, and 58 (3%) instructors. The mean number of years in academic medicine was 17.0, with a mean of 12.8 years at the current institution. Respondents’ clinical disciplines represented were 484 (27.7%) in internal medicine, 232 (13.3%) in pediatrics, 131 (7.5%) in surgery, 71 (4.1%) in family practice, 98 (5.6%) in psychiatry, 69 (4.0%) in anesthesiology, 67 (3.8%) in radiology, and 53 (3.0%) in obstetrics and gynecology. A number of respondents did not identify their disciplines or academic ranks.
Work characteristics of academic physicians were compared across schools, with the 1984 cohort,14 and by sex, with a mean of 60.8 hours worked per week. Male academic physicians reported working a mean of 61.1 hours per week, whereas female academic physicians worked a mean of 55.1 hours (p < .001).
Academic physicians reported working more hours per week than their basic science colleagues (60.8 hours compared to 54.7 hours) but took more vacation days in the year preceding this study. Basic scientists reported significantly more time devoted to research (46.1% compared to 14.7%) and teaching (20.5% compared to 11.4%) than their clinician colleagues (see Table 1). Table 1 also compares work characteristics for academic physicians between the 1984 cohort and the present sample. Both cohorts worked similar hours, but the present sample has considerably more clinical responsibilities and less time for teaching and research.
Physical and mental health status
Using the Center for Epidemiology Study Depression Scale (CES-D), 368 (20.5%) faculty reported symptoms consistent with clinical depression (a score of 16 or above).20 According to this measure, there were only small sex differences in CES-D reported depressive symptoms, with 231 (19.8%) men and 137 (22.3%) women scoring 16 or above (p = .05). In contrast, when respondents used the self-reported symptom inventory, 132 (21%) women reported experiencing “depression” in the past 5 years as compared to 136 (11.0%) men. Younger faculty members (27–35 years) reported higher total CES-D scores (mean score of 11.23) than did older faculty (56–65 years), who had a mean score of 8.89 (p = .03).
On the Rand Anxiety Scale, with possible scores ranging from 0–100, 72.5% (1,369) of physicians had scores in the normal range (score, <39), 292 (15.5%) had scores indicating mild anxiety (score, 40–50), 194 (10.3%) indicated moderate anxiety (score, 51–72), and 34 indicated (1.7%) severe anxiety (score, ≥73). The mean score for all schools was 25.42 (standard deviation [SD] = 8.36). Age differences were again noted, with younger faculty having a mean Rand Anxiety score of 36.01, whereas older faculty had a mean score of 26.80 (p < .001).
Reported alcohol consumption increased by age, with 24% (210/878) of younger faculty (age 27–45) drinking daily to several times per week, compared to 45.3% (159/351) of older faculty (age 56+). Thirty-six percent (462/1,263) of men drank daily to several times per week, compared to 27.2% (178/655) of women.
Health-related problems, both new onset (within 18 months) and past (within five years), showed no significant differences between basic scientists and academic physicians. In the past five years, approximately 72% (1,400) of respondents experienced one or more health problems, the most common being back pain (29.7% ), weight gain or loss (24.9% ), headache (22.6% ), gastritis (17.6% ), and hypertension (10.7% ). New-onset problems included weight gain or loss (8% ), back pain (6.3% ), gastritis (4.2% ), depression (3.1% ), and headache (2.8% ). In general, respondents reported minimal use of medications, including antihypertensives (4.6% ), antidepressants (4.0% ), gastrointestinal medications (4.3% ), pain medications (4.3% ), and anxiolytics (0.4% ). Women reported higher rates than did men of anxiety/panic (p = .02), constipation, depression, headache, hypertension, and weight changes (all p < .001), and diabetes (p =. 001) over the past five years.
An average of 1,260 academic physicians completed each item on the PJSS, a five-point job satisfaction scale. They rated themselves as moderately satisfied (2.93–3.17) on the following scale measures: “role in making organizational/administrative decisions,” “opportunities for promotion and/or increasing success,” “ability to meet the needs and demands of your work,” and “salary/income.” These respondents rated other aspects of their work somewhat higher: “degree of status and prestige associated with your work” (3.44), “ability to remain knowledgeable and current” (3.53), and “ability to derive personal gratification from your work” (3.70). Physicians were most satisfied with “the degree to which your work is educationally stimulating” (3.96). Physicians were less satisfied with “the manpower resources available to you” (2.52) and on “being mentored in a way that furthers your career” (2.64). An average of 390 basic science faculty reported small but significantly higher levels of job satisfaction than academic physicians on six of the scale measures (p < .005), including “being mentored,” “ability to derive personal gratification from your work,” “the manpower resources available to you,” “time allocated to teaching,” “work volume,” and “ability to derive personal gratification from your work.”
Faculty responses to the Work Related Strain Inventory (WRSI) revealed some significant sources of work strain, including interference with family life (35% ) and being “more edgy than in the past” (30% ). At the same time, 720 (37%) reported that their job expectations were being realized, 866 (44%) that they were contributing to the same degree that they had previously, 644 (33%) that their productivity has increased, and 788 (40%) reported that their professional growth and skills continued to improve. There were no significant differences in the mean total scores for the WRSI between academic physicians and basic science faculty. However, on eight of 18 questions, academic physicians indicated small but significantly higher levels of work strain than did basic science faculty. There were no significant correlations between work strain and frequency of physical symptoms.
Response to stressors
Table 2 displays respondents’ answers to a series of questions about their behavioral response to stressors. The majority reported engaging in a variety of health-promoting behaviors, such as eating regularly, getting adequate sleep, exercising, engaging in pleasurable social and sexual activities, and reading nonwork-related materials. However, 619 (32.2%) respondents never exercise or exercise up to several times a month. Only 446 (23.2%) reported getting adequate amounts of sleep daily. Twenty-nine percent (578) withdraw emotionally from family and friends up to several times a month. The majority of respondents did not routinely discuss the stressful nature of work with friends or family. We noted some significant differences between basic science faculty and academic physicians. Basic science faculty were more consistent in eating three meals a day, in getting adequate sleep, and in drinking less alcohol. Academic physicians discussed the stressful nature of their work more often with family or friends than basic science faculty.
On the Life Satisfaction Scale (LSAT), respondents were asked to rate various aspects of their lives, such as work, family life, finances, social relations, and physical health. For the most part, mean scores ranged between five (mostly satisfied) and six (pleased).
There were several questions relating to perceptions of job stability. When asked “to what extent does the stability of your salary, now and in the future, closely depend on your clinical and research productivity,” 467 (24%) responded that their salaries were “not very” or “not at all” dependent on productivity, whereas 493 (25%) responded that their salaries were “completely dependent.” To the question, “how likely is it that you will be forced to leave your institution in the next two to three years because of institutional changes,” 245 (12.6%) respondents said that it was “more likely than not” or “very likely,” whereas 638 (32.7%) responded “not at all likely.” When asked “how often in the recent past have you considered leaving academia,” 530 (27.7%) reported thinking about it “more often than not” or “often,” whereas 740 (38%) “rarely” or “never” thought about it. To the question, “how often in the recent past have you considered taking early retirement,” 381 (20%) reported thinking about it “often” or “more often than not,” whereas 1,109 (58%) “never” or “rarely” thought about it.
Institutional financial health and faculty well-being
Respondents were asked to describe the financial health of their institutions on a five-point scale, where 1 = “we’re in trouble” and 5 = “the best it’s ever been.” Two hundred and seventy one (14%) responded “we’re in trouble,” 385 (20%) stated “we could be in trouble if things continue this way,” and 902 (47.2%) chose “there may be difficult times ahead.” Only 352 (18.5%) chose either “we’re in pretty good shape” or “the best it’s ever been.” To investigate whether concerns about institutional financial health were related to measures of faculty well-being, we correlated responses to the above question about financial health to the five main scale measures (life satisfaction, job satisfaction, work strain, depression, and anxiety). At all four schools, there were small but significant negative correlations between respondents’ concerns about institutional financial health and life satisfaction, job satisfaction, and work strain; and positive correlations with depression and anxiety. Concerns about institutional financial health correlated most with job satisfaction across schools (r = –.29 to –.41, p < .001). There were no correlations between concerns about institutional financial health and frequency of physical symptoms.
Colleague attrition and faculty well-being
Respondents documented the number of colleagues who quit, were terminated, retired, left because of illness, or died in the 18 months prior to the survey. Mean numbers were low (<1.0) and comparable across schools, with the exception of school 1, where respondents reported six times the mean of colleagues who quit and about four times the mean of those who had been terminated. To correlate the perceived effect on respondents of these losses with responses on the various other scales, we asked respondents to rank the severity of these colleague losses on them, where 1 = “no effect,” 2 = “moderate effect,” and 3 = “very negative effect.” To evaluate the potential emotional impact of all of these losses, we created a “total effect score,” the sum of all the responses to the effect questions in each of the categories of faculty attrition. For the entire sample, the mean total effect score showed modest but significant correlations with the five main scales (p < .001, r = .200 to .301), the most significant with physicians’ job satisfaction (r = –.300) and work-related strain (r = –.301).
Effects of sex, age, marital status, and discipline on scale measures
We determined some significant differences in responses based on respondents’ sex, age, marital status, and discipline. On the Rand Anxiety Scale, women reported slightly higher levels of anxiety than did men (mean score of 32.68 compared to 30.06, p < .001), and in the CES-D, women reported slightly higher rates of depression than did men (mean score of 10.46 compared to 9.67, p = .053). There were small but significant correlations between younger age and the five main scales: CES-D, –.127; Rand, –.229; PJSS, .105; WRSI, .132; LSAT, .162; all p < .001. With regard to early retirement, 111 (15.5%) respondents aged 36–45 years and 153 (25%) of those aged 46–55 years reported thinking about early retirement “often” or “more often than not.” Ninety percent (1,759) of respondents who answered the question about their marital status were married. Of the 1,739 who described the quality of their marital relationship, 90.1% (1,569) reported being in a “fairly” or “very” supportive relationship. In looking at the five main measures, those who were not married had significantly more depressive symptoms than married respondents (CES-D mean score of 11.6 compared to 9.8, p = .007), lower job satisfaction (PJSS mean score of 49.6 compared to 51.8, p = .01), and life satisfaction (LSAT mean score of 44.8 compared to 48.8, p < .001).
There were significant differences on scale measures across some disciplines (see Table 3). Among the 15 specialties responding to the survey, those who identified themselves as surgeons reported the most depressive symptoms, the highest level of anxiety, the least job satisfaction, and the most work strain. Those who identified themselves as orthopedic surgeons reported the least depressive symptoms, the lowest anxiety, and the least work strain. Ophthalmologists reported the highest job satisfaction, whereas psychiatrists reported the highest life satisfaction.
We noted a number of differences on some measures across all four schools. Respondents’ health-related problems were similar across schools, and the CES-D scores did not vary between schools. Yet, the Rand Anxiety scores at school 4 were higher than at other schools (p = .001). School 3 had the youngest faculty, whereas school 4 had the oldest faculty, who also had also been employed at the same institution the longest. Work characteristics varied across schools as well: respondents at school 1 worked longer hours than at school 2 (61.14 hours per week compared to 57.32 hours, p < .001) and at school 3 (61.14 hours compared to 56.96, p = .001). Respondents at school 2 devoted the most time to teaching, and those at school 3 the least. Faculty at school 2 reported being considerably less concerned about their school’s financial health than respondents at schools 1 or 4 (p = .001) and school 3 (p = .002).
The results of our survey of academic medical faculty confirmed a number of the concerns we raised in our hypotheses, while refuting others. Significant findings include higher than expected rates of depression and anxiety, especially in younger faculty and in surgeons, with minimal differences according to sex or school. In addition, the results demonstrate higher levels of work strain, depression, and anxiety associated with faculty perceptions of their institution’s financial instability. We note some differences between basic science and academic physician faculty and across specialties in mood, job satisfaction, and life satisfaction. Somewhat surprising was that we found no significant relationships between respondents’ perceptions of financial instability, work strain, and physical symptoms. In general, faculty reported high levels of life satisfaction; and being married mitigated some of the negative effects of the academic environment. Older faculty, who had experienced the most changes in academic medicine over the course of their careers, were the least affected by various stressors.
Our overall response rate of 54.3% is consistent with the average response rate in other physician surveys.21 Indeed, the response rate is higher than we anticipated, given the length of the questionnaire and the sensitive nature of many of the questions. These responses from almost 2,000 medical school faculty raise questions that need to be further studied and addressed. Although the data in our sample are limited to responses from four medical schools in the United States, the overall sample size of 1,951 faculty, the varied backgrounds of each of the participating institutions, and the lack of interschool differences on many measures leads us to believe that our sample is representative of the larger cohort of faculty at U.S. medical schools.
The results of our study add to the growing evidence that American medicine is in trouble. A recent report from the Institute of Medicine delivered a broad indictment of our health care system,22 and physician discontent with the practice of medicine is rising.1 The future of medicine resides in the medical students of today and tomorrow, who in turn depend upon the quality of faculty and teaching in our nation’s medical schools. This study raises the concern that current medical students are being taught by faculty who are increasingly stressed and dispirited. One in five faculty members had significant depressive symptoms at the time of completing the questionnaire, yielding a point prevalence of approximately 20%. Our study indicates that men and women faculty members had comparable rates of depression, whereas more generally women are reported to have up to twice the rate of depressive disorders as men.23 The higher rates of depression and anxiety reported by younger faculty are equally disconcerting. Although we are not aware of more recent studies using the CES-D in community samples, older studies using the CES-D and the Zung Depression Scale,24 which we did not use but that measures similar self-reported symptoms, suggest that the prevalence of clinically significant depressive symptoms in a comparable nonpatient population should be about 9%.25,26 The National Comorbidity Survey of a large national sample aged 15–54 years found a 12-month prevalence of major depression and dysthymia combined to be 12.8%.27 Yet according to our study, nearly 21% of respondents indicated symptoms consistent with a depressive disorder.
The difficult question is whether current stressors in the health care environment have led to increased depressive symptoms or whether the depressive symptoms have led to faculty’s increasingly negative perception of their work environment. The majority of respondents indicated that their initial job expectations were not being realized, that they were not the contributors they used to be, and that their productivity was decreasing. Significant numbers of faculty felt unsupported and “edgy.” More than one in four faculty members frequently thought about leaving academia, and one in five “often thinks” about early retirement.
It is not surprising that academic faculty are discontent. Given today’s economic realities, faculty face the dual pressures of decreased support for teaching and increased demands for clinical and research productivity. In our cohort, concerns about the financial health of their institutions correlated with respondents’ increased work strain and decreased job and life satisfaction, as well as increased depression and anxiety. For many faculty, the lure of academia seems to have been the relatively stable mix of patient care, research, and teaching. Yet our study supports Ludmerer’s assertions3 that in recent years patient care responsibilities have burgeoned, crowding out time for teaching and research. In comparison with Linn et al.’s 1984 study of internists,14 our current cohort has less time to supervise students and residents, and their research time has been cut in half.
Though we did not have access to the original data set in the 1984 study by Linn and colleagues,14 and therefore could not make statistical comparisons between it and our current sample, it appears there are important differences. Using the Zung Depression Scale comparable to the CES-D,24,28 the 1984 study revealed 14% of faculty with clinically significant depressive symptoms, in contrast with 20% in the 2001 sample. In general, academic physicians in our cohort appeared to be less satisfied according to almost every measure of job satisfaction. On the LSAT, the 1984 academic physicians were somewhat more satisfied with work and their physical and mental health than were the 2001 sample. Although the work hours of both groups were comparable, current faculty are spending less time teaching and doing research and more time doing clinical work than the 1984 cohort.
There are limitations to our study. We do not have information on nonresponders. Although we have data from four geographically diverse regions, we cannot assert that our sample is widely representative of medical school faculty throughout the nation. Clearly, more studies are needed. If, however, our data prove generalizable, today’s medical students are interacting with faculty who have less time to teach them, who are increasingly discontent with their jobs, and who, especially among younger faculty, are increasingly depressed. Can such faculty be the role models they want to be, provide a quality educational experience, and be productive, let alone convey the excitement and privilege of caring for patients? And if faculty have less time for research, how will they continue to address the pressing needs for new clinical knowledge, achieve promotions, and to enhance their institutions with their experience and wisdom?
Although our study reveals some disconcerting facts about academic medicine, there is also some good news as well. On average, faculty respondents were “mostly satisfied” with their work and life. Indeed, there were no differences in respondents’ assessments of “life in general” between the 1984 and 2001 cohorts, the decreases in satisfaction with “work” perhaps being offset by the higher scores in such categories as “family life,” “leisure activities,” and “social situation.” Medical school faculty still work hard, but many may be paying more attention to balancing their work and home lives. Ninety percent were in a stable marital or committed relationship, and most seem to attend to their health and engage in pleasurable social activities, and rarely argue with or withdraw emotionally from family members.
Though there is a reservoir of satisfaction and well-being among medical school faculty, the increasing stresses and discontent challenge us, especially in mentoring younger faculty. We must find ways to support and encourage faculty in their personal and professional development through mentoring and faculty development programs. We must identify and offer assistance and support to our colleagues who are dispirited or may be clinically depressed, a real challenge as physicians generally are poor at detecting and adequately treating depression in their patients.29 Both department chairs and colleagues need to intervene and provide appropriate referrals for treatment. We need additional research and more continuing medical education programs on the pathogenesis of mood and anxiety disorders in health professionals from clinical, biological, environmental, and systems perspectives. New funding approaches need to be developed to ensure that medical school faculty can carry out essential clinical research as well as medical student and resident education. We must find ways of ensuring that our students are exposed to high-quality, enthusiastic educators who have enough time to teach and enough contact to serve as role models to medical students. The attractiveness of our profession to young people and the quality of medical care now and in the future will depend on our resourcefulness and our abilities to meet these challenges.
The authors wish to thank Edward Gracely, PhD, Drexel University College of Medicine, for his help with the statistical analysis of the data. There was no external funding source for this research. Individual schools defrayed the cost of mailing the surveys.