Although burnout has been studied extensively among students and residents, in few studies have researchers examined burnout among fellowship trainees. We measured burnout among fellows in our freestanding children’s hospital and evaluated fellows' perceptions of stigma around (and willingness to seek treatment for) psychological distress. The objectives are as follows: to (1) measure burnout among pediatric fellows, (2) assess fellows' perceptions of stigma around help seeking for mental illness, and (3) examine the relationship between burnout and willingness to seek behavioral health counseling.
We distributed a 48-item inventory to all 288 fellows in our pediatric center. Items included the Maslach Burnout Inventory and Likert-type matrices to assess attitudes toward behavioral health treatment and associated stigma. We used 2-sampled t-tests to associate burnout with willingness to seek mental health treatment.
A total of 152 fellows (52%) responded, of whom 53% met the threshold for burnout. Most reported believing that their program directors (78%), attending physicians (72%), and patients (82%) hold negative attitudes about mental illness and its treatment; 68% believed that employers would reject their application if they knew they sought counseling. Fellows with burnout were more likely to believe that others in the clinical learning environment hold negative views of help seeking for behavioral health (odds ratio 1.2–1.9).
Just over one-half of the pediatric fellows in our center meet the threshold for burnout. They also experience significant workplace-based stigma around help seeking for psychological distress. Fellows with burnout are more likely than their peers to perceive significant stigma around help seeking for their distress, making them a particularly at-risk learner population.
Burnout is well documented in both pediatric residents and post-training pediatric practitioners. Among medical students and nonpediatric trainees, burnout is known to increase perceptions of stigma around help seeking for mental illness and/or emotional distress.
There are few studies in which researchers measure burnout specifically among pediatric fellowship trainees. To date, there are no studies in which researchers explore the relationship between burnout and help seeking in this group.
American physicians continue to experience symptoms of burnout at alarming rates, a trend that has resulted in tangible negative consequences across all fields of medicine.1–4 Literature over the past decade has linked burnout to increased medical errors, decreased physician productivity, increased rates of substance use disorder and other mental health conditions among physicians, and higher rates of physician motor vehicle and on-the-job unintentional injuries.5–13 Recognizing this crisis, multiple national organizations (including the Accreditation Council on Graduate Medical Education) have called for an increased focus on physician wellness and on ameliorating the systemic factors that lead to doctors’ job-related psychological distress.14–16
Building on previous data that symptoms of burnout peak during a physician’s training years,13,17,18 Kemper et al19 conducted a longitudinal national survey of burnout among pediatric residents, demonstrating that, between 2016 and 2018, more than one-half of pediatric residents completing the Maslach Burnout Inventory (MBI) met the criteria for burnout. Despite this recent focus on burnout among pediatric residency trainees, however, there is little literature exploring the rates of burnout among pediatricians during their years of subspecialty fellowship training. To better understand the prevalence of burnout among pediatric fellowship trainees at our own freestanding children’s hospital, in 2018 we conducted a cross-sectional assessment of fellows’ perceived stress levels and measured their degree of burnout by using the MBI Human Services Survey for Medical Personnel (HHS-MP).20
With the understanding that depression and burnout are distinct (but overlapping) constructs,21–26 we also aimed to explore whether pediatric subspecialty fellows perceive workplace stigma around help seeking for emotional distress or mental illness and whether the presence of burnout affects the degree to which this stigma is perceived. In their work with medical students and residents, Dyrbye et al27–34 extensively studied the intersection between burnout, depression, and help seeking among undergraduate medical trainees and resident physicians. In keeping with the findings of previous studies,35–37 Dyrbye et al27–34 found that trainees perceived significant stigma around help seeking for mental illness and noted that those meeting the criteria for burnout on the MBI were more likely to anticipate negative personal consequences for admitting to or seeking help for mental illness.38 In the current study, we adapted Dyrbye et al’s28 stigma assessment inventory for the context of pediatric fellowship trainees and sought to learn whether fellowship trainees with burnout experience similar concerns about stigma and help seeking.
In the fall of 2018, we distributed an anonymous electronic survey to all 288 subspecialty fellowship trainees across the 6 departments of our freestanding, tertiary-care children’s hospital. We invited fellows to participate in the study via an e-mail explaining the study and sent reminder e-mails both to the trainees and their program directors over the 6 weeks that the survey remained open. Both solicitation e-mails and the survey inventory itself reassured participants that their answers could not be traced back to their identity and that they would not be reported to their program directors on an individual level. This study was considered exempt from review by the Children’s Hospital of Philadelphia’s Institutional Review Board.
The electronic inventory included a single item measuring self-assessment of work-related stress on a 10-point scale, an assessment of burnout using the MBI, and Likert-type agreement matrices, assessing fellows’ perceptions of stigma around mental illness, their concerns about workplace disclosure of help seeking for mental illness, and their likelihood of seeking help for psychological distress. Demographic items on the survey included self-identified gender, age, postgraduate training year (PGY) level, training program (including department and division), and time taken between residency and fellowship training.
Assessment of Burnout
We used the MBI HSS-MP inventory to measure burnout among fellows.20 The inventory includes 3 subscales, the psychometric properties of which have been well studied in the medical community: emotional exhaustion (EE), depersonalization, and low personal accomplishment (PA).20 For this study, we used Maslach and Leiter’s20 standardized (z) values for medical personnel to calculate whether an individual respondent met the threshold for burnout. We used the following critical boundaries as recommended by the fourth edition of the MBI User’s Manual, using means and standard deviations from the Maslach and Leiter20 sample: high exhaustion (EE) at z = mean + (SD * 0.5); high cynicism (depersonalization) at z = mean + (SD * 1.25); high professional efficacy (PA) at z = mean + (SD * 0.10). Respondents were considered burnt out if any of their scores were above the threshold.
Perceptions of Stigma and Fears of Disclosure
To measure fellows’ perceptions of stigma related to help seeking for mental health, we adapted Dyrbye et al’s28 inventory with permission to make the items relevant to the fellowship training context. Survey items on stigma included measurement of fellows’ internal attitudes about mental illness, as well as their perceptions of the attitudes of others, including their peers, supervisors, and patients.
Descriptive summary statistics, by using percentages and frequencies, were run for the variables of department, division, age, residency-to-fellowship interval, training level, and gender. Means and standard deviations were calculated for individual MBI domains and fellows’ self-reported stress level. Independent sample t-tests were used to test for significant differences in perceptions of stigma relative to respondents’ burnout. We ran Spearman’s correlations to test for relationships between demographics, stress, help-seeking behaviors, and the 3 MBI domains (EE, depersonalization, and PA). We used independent sample t-tests to look for differences in perceptions of stigma relative to respondents’ burnout, and we calculated odds ratios to analyze the significance of those perception differences.
A total of 147 fellows (51%) completed the inventory in its entirety; 152 fellows (52%) answered all questions except the Likert-type agreement matrix. Table 1 shows respondents’ demographic characteristics, which were reflective of the overall study population. The majority of respondents (62%) identified as female, and fellows were evenly distributed between the first 3 years of fellowship training (PGY 4–6), with a smaller number (16%) in their seventh year of training or beyond. The majority (74%) went straight from their residency training into their fellowship without time off in between. The fellows’ mean reported stress level was 6.8 out of 10, and the prevalence of burnout in the study group was 53%. Table 2 details fellows’ self-reported levels of stress and their aggregate scores on each of the 3 MBI subscales.
Spearman Correlation Between Stress, Demographics, and MBI Domains
We found moderate positive relationships between EE and depersonalization (0.61) and between self-reported stress and EE (0.53). We found weak positive relationships between depersonalization and self-reported stress (0.17) and between PGY level and PA (0.19). Finally, we found weak inverse relationships between PA and both EE (–0.22) and depersonalization (–0.29). There were no significant relationships relative to trainee age.
Perceived Stigma, Fear of Discrimination, and Disclosure of Mental Illness
Table 4 details respondents’ perceptions of stigma surrounding mental health issues and help-seeking behaviors for emotional distress. Most fellows reported believing that their program directors or chairs (78%), attending physicians (72%), and patients (82%) hold negative attitudes about mental illness and its treatment. A total of 68% of fellows agreed or strongly agreed that potential employers would pass over their application if they were aware that they had sought help for a mental health problem during their training. When asked whether respondents would hide the fact of any mental health treatment they had received, 75% of fellows said they would. Just over one-half (or 56%) agreed or strongly agreed that patients would not want them as their doctor if they were aware that they had been treated for a mental health problem.
Effect of Burnout on Perceptions of Stigma, Fears of Discrimination, and Concerns About Disclosure
We also analyzed fellows’ perceptions of stigma were by whether they met the criteria for burnout on the MBI. Respondents with burnout were significantly more likely to believe that their supervisors, patients, potential employers, and peers held negative attitudes about mental illness and its treatment (OR 1.2-1.9; all P < .05; see Fig 1). A total of 79% (62 of 79) of fellows with burnout agreed or strongly agreed that future employers would pass over their application if they knew that they had sought help for a mental health problem, and 59% of fellows with burnout (47 of 79) agreed that “most people think less of a person who has sought mental health care.”
This single-center study reveals that approximately one-half of pediatric subspecialty fellowship trainees (like their more junior counterparts in residency training)19 meet the criteria for burnout on the MBI HSS-MP. This finding is in keeping with literature across the continuum of medical education and suggests that training progression toward one’s specialty of choice is not itself sufficient to ward off the EE and depersonalization experienced by even more advanced trainees in medicine.
Data from this study also demonstrate that pediatric fellowship trainees perceive both internal and workplace stigma around issues of mental illness and help seeking for psychological distress. Over one-half of the participants in this study reported believing that those around them (from their supervisors to their peers and even their patients) would think less of them if they were to seek professional behavioral health counseling, and more than two-thirds of respondents reported believing that a known history of help seeking for mental illness would harm their chances at future employment. Most noted that they would hide the fact that they had sought professional help for mental illness, suggesting that there may be cultural factors both locally in our institution and more broadly across the field of medicine that have potentially influenced fellows’ beliefs on this subject.
Perhaps more concerning, however, is our finding that the presence of burnout among fellowship trainees is associated with an increase in fellows’ perception of workplace help-seeking stigma and is correlated with the belief that patients and employers would not want to receive care from (or hire) a physician who had sought care for mental illness. It is well documented in the mental health literature that stigma, generated both internally and from others, is a multifaceted phenomenon that discourages those with emotional distress from seeking professional help for their symptoms.35,36,39–42 Our finding that trainees with burnout experience amplified perceptions of workplace stigma holds implications for program directors seeking to bolster and sustain the wellness of their fellows; it is possible that many of the cultural workplace phenomena that lead to burnout also underscore trainees’ fears about psychological safety in the learning environment. It is similarly likely that fellows with burnout are correct that workplace stigma around mental health and its treatment is highly prevalent. With our findings in mind, program directors should be cognizant of the fact that those trainees with the greatest need for psychological support may be the least inclined to seek it out for themselves because of concerns about stigma and disclosure.
Our study has several limitations. First, our response rate (51%) captures mostly those trainees in pediatrics and anesthesiology and critical care medicine. It is likely, therefore, that our results are influenced to a degree by nonresponse bias. Second, fellows were not always willing to identify the division in which they were training, and sample sizes in each division category were small, limiting our ability to evaluate the relationship between training division, burnout, and help seeking. We had only 1 respondent each from the departments of pathology and radiology, which represents a loss of potentially meaningful data because trainees in fields with less patient-facing time may face different stressors than those with more front-line clinical exposure. In addition, sample sizes in the demographic categories of gender and PGY were too small to allow for meaningful inferential analysis, and our survey did not ask respondents to identify their race or ethnicity, limiting the degree to which we can draw conclusions about the role of trainee identity in burnout and perceptions of help-seeking stigma. Nevertheless, our results are in keeping with those of other studies examining rates of burnout across specialties in the United States.1,2,6,9,18 Finally, whereas our results mirror those in other studies of burnout around the country, our single-center methodology has the potential to limit the generalizability of our findings because there are undoubtedly local cultural factors that influenced our results. Further study in this field will be needed to more fully elucidate the drivers of burnout among subspecialty fellowship trainees across specialties and geographic regions.
Moving forward (as physicians and as medical educators) it is important for us to normalize and to model help seeking and self-care for trainees and peers. For example, many residency programs, including ours, have embedded psychologists or social workers for their trainees to access, many with opt-out models that seek to reach even those trainees who might not otherwise be inclined to seek help on their own. Our residency program and others have incorporated scheduled half-days for making medical and therapy appointments and other self-care needs. Our findings suggest that similar services and scheduling for personal health may be necessary at the fellowship level as well.
Furthermore, more explicit psychoeducational mentorship may be beneficial for senior trainees as they navigate through advanced training at a time when many are also balancing various life-transitions (ie, marriage, divorce, parenting, and/or caring for aging parents). Knowing that a respected mentor has also struggled with burnout or with mental health issues may help form a supportive alliance that supports trainees in seeking the help they need to be most successful.
We thank Drs Stephen Ludwig and Joseph St Geme III for their support of this work and the fellowship trainees of the Children’s Hospital of Philadelphia for their participation.
Dr Weiss conceptualized and designed the study, designed the data collection instrument, coordinated and supervised data collection, drafted the initial manuscript, and reviewed and revised the manuscript; Drs Mehta, Quinn, and Wiens conceptualized and designed the study, designed the data collection instrument, and reviewed and revised the manuscript; Dr Danley performed statistical analysis, drafted critical portions of the methods and results, and critically reviewed the final manuscript for important intellectual content; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
FUNDING: No external funding.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they no potential conflicts of interest relevant to this article to disclose.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.