Dunn, Laura B. MD; Green Hammond, Katherine A. PhD; Roberts, Laura Weiss MD, MA
Roughly 100,000 physicians engage in postgraduate training across the United States each year. Since the 1984 death of Libby Zion, an 18-year-old college student under the care of an exhausted resident physician in New York, efforts to ensure safe clinical practices by residents—and concerns about residents' stress and well-being—have emerged as vital issues in academic medicine. This said, relatively little is known about the personal health care attitudes and practices of residents. This gap in our understanding is worrisome because these early-career physicians comprise a large segment of our medical workforce, shoulder extensive responsibilities for patient care, and serve as the primary clinical teachers for the approximately 69,000 medical students1 annually enrolled in medical schools. Finally, as physicians-in-training drawn from nearly all developed and developing countries, residents represent the future of health care throughout the world.
We are learning that the personal health care views and habits of physicians are important to understand for a number of reasons. Intuitively, healthy physicians will have longer and more productive careers in the service of patients.2–7 Empirical work has shown, moreover, that physicians who practice healthy lifestyle habits are more likely to counsel patients about issues such as smoking, nutrition, exercise, and weight loss.8,9 Systematic quantitative data and narratives from physicians also indicate that their personal experiences with illness and as patients can profoundly affect their behaviors and attitudes toward their patients, often increasing their sensitivity and empathy.5–8,10,11 It is known that practicing physicians frequently postpone their own care, however, and many have no regular source of care.9,12 The retrospective stories of impaired physicians suggest that the roots of illness and compromised ability to fulfill duties as a practicing physician often begin during training.13,14 Moreover, many studies suggest that stigma greatly influences the perspectives and actions of physicians at all stages of professional development regarding self-care.3,5,14,15 Despite these known associations, residents' attitudes and behaviors regarding personal health care have received minimal inquiry.10,16
The most informative study to date surveyed internal medicine residents at four programs about their personal care.10 Approximately half of the participants lacked a primary care provider, and nearly half had not visited a physician in the prior year. Scheduling difficulties, residents reported, made it challenging to obtain health care; they also believed that residents should receive more care. Responses to open-ended questions reflected residents' concerns about privacy and worries about lack of sympathy from peers if they were to become ill or needed to be absent to attend medical appointments. These investigators also reported on residents' common self-prescription of medications.17
Data regarding residents postponing or avoiding care remain scarce.18,19 We examined the attitudes and behaviors of residents in primary care and specialty programs at one institution pertaining to their own health care. In this report, we focused on items related to postponing or avoiding care. We predicted that, overall, residents would endorse having health concerns and having previously avoided or postponed necessary care. We further hypothesized that individuals with greater concerns about confidentiality, jeopardy to training status, and perception of criticism from peers would be more likely to have avoided or postponed obtaining care because of these concerns, and that frequency of postponing or avoiding care would be associated with poorer self-rated health. Finally, based on our previous findings that women were more sensitive to the possibility of academic vulnerability in connection with personal health issues,20 we also hypothesized that these issues would have greater salience for women residents in our study.
In 2000–2001, a detailed written survey examining issues related to residents' personal health care was devised and pilot tested at the University of New Mexico (UNM) School of Medicine. The survey was based on previous work by one of us (L.R.) and a collaborative research group on medical student health care.20,21 Items included 11 demographic questions, including ethnicity and gender, 241 questions with scaled or counted responses, and 2 open-ended questions. Domains were related to personal health care behaviors and preferences, self-perceived current personal physical and mental health, special issues related to being a trainee and seeking care (e.g., confidentiality, multiple roles), attitudes reflecting compassion and empathy toward ill patients, and six vignettes tapping residents' possible views of stigma and professional jeopardy resulting from personal illness or treatment as a resident. Here, we report findings from 73 items regarding residents' responses about the following:
* Personal concern with 28 specific health issues during the previous year,
* Feelings of jeopardy to their training status if their training director or supervisor were to learn of a current, specific health problem,
* Avoidance or postponement of obtaining necessary health care during the previous year,
* Frequency of hearing other residents or attending physicians discussing a resident's health problems or absence from work because of illness during the previous year, and
* Treatment by peers and faculty after illness-related absence during the previous year.
Responses regarding the site of seeking care (outside versus within the training institution) and the stigma of the patient role for residents have been reported elsewhere.16,22
The voluntary and confidential self-report survey, which was reviewed and approved by the UNM Health Sciences Center institutional review board, was distributed in 2000–2001 to all 217 residents who were in their second or third postgraduate year (PGY) at UNM School of Medicine through their mailboxes. Two follow-up mailings were sent to nonresponders. For their time and effort, respondents received $50.
For conceptually related sets of items (using nine-point scales) regarding experiences and attitudes, we conducted repeated-measures Item (within-subjects repeated measures) × Gender (between subjects) × Residency Area (primary care [internal medicine, family practice, pediatrics] versus specialty [all other areas], between subjects) × Training Level (PGY 2 versus 3 or 4, between subjects) MANOVAs. In addition, several item sets with additional within-subjects structure were subjected to Category (within subjects) × Item × Gender × Residency Area × Training Level MANOVAs. Single items were analyzed using Gender × Residency Area × Training Level ANOVAs. We found no consistent effects of Residency Area or Training Level. To test our hypotheses, we also examined correlations among personal health ratings, attitudes, and experiences using SPSS 14.0.2.
The response rate was 71% (155 surveys returned). Of the surveys returned, 14 were excluded for missing responses to all items in at least one set reported here. Ethnicity differed by gender (P < .06); 24% of women versus 9% of men were Hispanic, whereas 62% of women versus 74% of men were white. Men reported better mental health than did women (respective means = 7.5 versus 7.1, P < .01; scaled from 1 = “poor” to 9 = “excellent”) (see Table 1).
Residents' personal health concerns
Residents rated their personal concern with 28 specific health issues during the previous 12 months (1 = “no concern” to 5 = “some concern” to 9 = “great concern”). Residents overall reported some concern with health maintenance and prevention (mean = 4.84) and lower concern with fatigue, depression, minor infections, marital or relationship problems, pain, anxiety, gynecological or urological problems, and sleep-related problems (means = 3.02–3.71). Residents overall reported very low concern with dermatological problems, ophthalmic problems, personal counseling issues, physical injury, gastrointestinal problems, allergy, eating-related problems, HIV, other STDs, other serious infections, pregnancy-related issues, elective surgery, arthritis, cancer, chest pain, and diabetes (means = 1.46–2.75) and almost no problems with alcohol, prescription drugs, or other drugs (means = 1.13–1.29; Health Issue Item main effect F(27,107) = 19.69, P < .0001, maximum Cohen d = 1.83). Women reported greater concerns overall than did men (means = 2.61 versus 2.15, Gender main effect F(1,133) = 19.24, P < .01, d = 0.51).
There was substantial variation in individuals' recent health concerns, with responses across the scale for most issues considered and large SDs (2.13–2.80) for half of the specific issues. More than one fourth of respondents (26%–43%, n = 41–64) indicated some to great personal concern (responses 5, 6, 7, 8, or 9) during the past year with depression, anxiety, marital or relationship problems (32%, n = 48) or sleep-related problems, fatigue, gynecological or urological problems, pain, or minor infections. Additionally, 11% to 24% (n = 16–37) indicated some to great personal concern with HIV, other serious infections, personal counseling, eating- or pregnancy-related issues, injury, ophthalmic, gastrointestinal, or dermatological problems, elective surgery, or allergies.
Residents' concerns about health issues and jeopardy to training status
In rating concern (1 = “no concern” to 5 = “some concern” to 9 = “great concern”) about potential jeopardy to their training status if their training director or supervisor were to learn of a current specific health problem, high levels of concern were expressed about problems with alcohol, prescription, or other drugs (means = 6.45–6.69). Moderately high concern was shown for HIV (mean = 5.56); moderate concern for depression, anxiety, personal counseling issues, eating-related problems, cancer, and STDs other than HIV (means 3.60 to 4.72); low concern for serious infections, marital or relationships problems, sleep-related problems, fatigue, and pregnancy-related issues (means = 3.18–3.39); and very low concern for the remaining health issues (means = 1.38–2.53; Item main effect F(27,107) = 20.84, P < .0001, maximum d = 2.42). Women expressed more concern than did men for potential jeopardy to training status from pregnancy-related issues (respective means = 4.35 versus 2.01, d = 1.01), cancer (means = 4.27 versus 3.15, d = 0.48), and alcohol problems (means = 6.97 versus 5.92, d = 0.45) but similar levels of concern for other health issues (Item × Gender interaction, F(27,107) = 2.15, P < .01) (see Figure 1).
Substantial variability was noted in ratings of potential jeopardy to training status for the 14 health issues of greatest concern (SDs = 2.59–2.91), with responses ranging across the scale for almost all health issues considered. Greater ratings for potential jeopardy to training status were associated with greater recent personal concern with depression (r = 0.31, P < .01), anxiety (r = 0.28, P < .01), personal counseling (r = 0.30, P < .01), eating-related problems (r = 0.20, P < .02), sleep-related problems (r = 0.19, P < .03), diabetes (r = 0.37, P < .01), HIV (r = 0.22, P < .01), other serious infections (r = 0.19, P < .03), ophthalmic problems, dermatological problems, elective surgery (all r = 0.20, P < .02), and cancer and gynecological or urological problems (both r = 0.17, P < .05).
Avoidance or postponement of seeking health care
Residents reported how often they had had physical or mental health concerns, but had not sought professional attention for them, during the past year (1 = “never” to 5 = “sometimes” to 9 = “many times”). Overall, residents had sometimes avoided seeking health care (mean = 4.13); 18% (n = 26) had frequently avoided seeking care (responses of 7, 8, or 9). Women had avoided seeking health care more often than had men (respective means = 4.78 versus 3.48; Gender main effect F(1,140) = 11.04, P < .01, d = 0.56). More frequent avoidance was associated with poorer self-rated mental health (r = −0.45, P < .01) and physical health (r = −0.23, P < .01) and greater recent personal concern with health maintenance/prevention (r = 0.35, P < .01). More frequent avoidance was also associated with greater personal concern with alcohol problems, depression, anxiety, fatigue, eating- and sleep-related problems, marital or relationship problems and other personal counseling issues, STDs, serious infections, gynecological, urological, or gastrointestinal problems, pain, injury, ophthalmic problems, and dermatological problems (r = 0.17–0.45, mean r = 0.29, all P < .04) (see Table 2).
Respondents reported how frequently during the past year (1 = “never” to 5 = “sometimes” to 9 = “many times”) they had avoided or postponed getting health care, and how frequently they had observed other residents avoiding or postponing care, because of concerns that their privacy might be violated, that a supervisor might find out, or for any other reason. Residents overall (combining responses for themselves and observations of other residents, which did not differ significantly) infrequently avoided or postponed getting care because of concern that a supervisor might find out (mean = 2.35), concern about violation of privacy (mean = 3.19), and sometimes for other reasons (mean = 3.66; Item main effect F(2,132) = 24.86, P < .0001, maximum d = 0.57). Women avoided or postponed care for the listed reasons more often than did men (means = 3.74 versus 2.39, Gender main effect F(1,133) = 16.00, d = 0.68). Residents did not report different rates of avoiding or postponing health care for other residents than for themselves (P < .87). Greater frequency of avoiding or postponing health care for all the reasons considered was associated with poorer mental health (r = −0.30 to −0.36, mean −0.32, all P < .001), and greater frequency of avoiding or postponing health care for “any other reason” was also associated with poorer physical health (r = −0.27, P < .01). Greater frequency of avoiding or postponing health care because a supervisor might find out was associated with greater concern about potential jeopardy to training status for all 28 health issues considered (r = 0.15–0.43, mean r = 0.26, all P < .09) and because of concerns about privacy violations for 24 of the 28 health issues considered (r = 0.15–0.37, mean r = 0.25, all P < .09).
Discussion of residents' health issues by other residents and attending physicians
Residents indicated how often (1 = “never” to 5 = “sometimes” to 9 = “many times”) they had heard other residents or attending physicians discussing residents' health problems or absences from work due to illness. Residents heard discussions of others' health problems or absences from work by other residents sometimes (mean = 4.35) and by attendings infrequently (mean = 3.03; Resident versus Attending Category main effect F(1,133) = 105.05, P < .0001, d = 0.64). Residents heard residents' absences from work discussed more often than residents' health problems (means = 3.95 versus 3.42, for discussion by residents and by attendings combined, Health Problem versus Work Absence Item main effect F(1,133) = 13.63, P < .001, d = 0.26). Compared with men, women reported more frequently hearing residents discuss peers' health issues or absences (means = 4.81 versus 3.88, d = 0.46), but men and women reported similar frequencies of hearing attendings discuss residents' health or absences (Resident versus Attending Category × Gender interaction, F(1,133) = 9.15, P < .01). Greater frequency of hearing residents discuss peers' illnesses or absences was associated with greater frequency of avoiding or postponing care because of concerns about privacy, a supervisor finding out, or other reasons (r = 0.22–0.34, mean r = 0.29, all P < .01) (see Table 3).
Views of residents' missing work due to illness
Respondents rated residents' treatment by peers and faculty after illness-related absences (1 = “much more negatively” to 5 = “no differently than before their absence” to 9 = “much more positively”). Respondents believed that residents were treated only slightly more negatively (mean = 4.40, different from 5 = “no differently” at P < .01 by Fisher least significant difference), with no difference in perceived treatment by residents versus faculty (Item main effect P < .31). Compared with men, women perceived that residents were treated more negatively after illness-related absences (means = 4.21 versus 4.59, Gender main effect F(1,133) = 3.71, P < .06, d = −0.33). More negative perceptions of residents' postillness treatment were associated with greater frequency of avoiding or postponing care because of concerns about privacy, a supervisor finding out, or other reasons (r = −0.23 to −0.31, mean r −0.25, all P < .01) (see Table 4).
In response to items about degree of empathy for residents missing work because of illness (1 = “not at all” to 5 = “some” to 9 = “very much”), residents indicated that other residents show some empathy (mean = 4.97) and that attendings show less, but still some, empathy (mean = 4.65; Item main effect F(1,133) = 7.01, P < .01, d = 0.15). Respondents rated how much residents and attendings ostracize residents who miss work because of illness (1 = “not at all” to 5 = “some” to 9 = “very much”). Residents indicated a moderately low degree of ostracism by other residents (mean = 4.06) and a low degree by attendings (mean = 3.46; Item main effect F(1,133) = 2.15, P < .0001, d = 0.30). Greater perceptions of ostracism by other residents were associated with greater frequency of avoiding or postponing getting care because of concerns about privacy or a supervisor finding out (both r = 0.21, P < .02).
The perspectives of residents regarding their personal health have received little attention in the past. In this small but novel hypothesis-driven study, we found that residents endorsed at least some need for care for a variety of health issues and that they had delayed or avoided necessary personal health care in the prior year. Negative perceptions related to highly sensitive health concerns such as worry over academic jeopardy, being the subject of discussion by colleagues, and special issues related to substance abuse, mental health, sexual health, and/or relationship problems seem to shape care-seeking behaviors, as reported by residents. These patterns overall were accentuated for women and, not unexpectedly, for individuals with poorer health and greater health needs. Taken together, these data suggest the need for better understanding of the health-related issues of residents and, importantly, of the adverse impact of stigma on their views and self-care practices. The implications of this early study for residency training and for future inquiry are discussed here.
Residents' perspectives on personal health needs and care-seeking
Given that the respondents were young adults who had undertaken what is known to be a physically and mentally taxing career path, we did not expect widespread concern about health issues among the residents in this study. Nevertheless, we found that a substantial minority (i.e., 11%–43%) reported significant personal concern in the prior year related to diverse and sensitive health issues, and that these concerns were greater among women trainees. Our data corroborate those of prior studies indicating the salient role of depression, anxiety, fatigue, and sleep issues for residents. 8–10,12,16–25 Our findings also align with published data from one residency assistance program designed to provide accessible, confidential care for residents at a university-based institution.26 In that program's first eight years, of 193 cases seen, 102 (52.8%) of presenting problems were related to depression, anxiety, adjustment issues, and stress. Similarly, in our study, nearly one third of respondents endorsed marital and relationship difficulties as being of recent personal concern, and the residency assistance program reported that marital difficulty was the presenting concern in 15% of cases.27
Our data on the health needs and care-seeking practices reported by residents suggest that both the nature and personal salience of a health issue, together or separately, may be especially important. For instance, as a group, residents viewed many different health issues as potentially jeopardizing a resident's academic training status. Our respondents expressed most concern about highly sensitive health issues, such as mental health and substance use disorders, rather than the most physically disabling conditions. Individual residents also expressed greater concern about academic jeopardy when the specific health issue had more direct relevance to their lives. This was true for the physical health items (e.g., eye and skin problems) as well as mental health (e.g., depression) or stress-related issues (e.g., fatigue). Similarly, overall, residents reported having sometimes delayed or avoided care for both physical and mental health concerns. A minority (18%) reported having done so frequently. Beyond delayed care, poorer self-rated mental health and greater personal concern with mental and physical health issues were associated with avoidance of care. Delaying or avoiding care may lead to several deleterious outcomes, from lack of detection and care to spread of infection to clinical deterioration. At the extreme, avoidance of care for specific psychiatric conditions, such as depression, can lead to suicidal thoughts and behavior.24
Heightened concerns about privacy were also associated with negative self-care practices for the majority of the health issues. Residents with poorer self-rated health (mental or physical) and those with greater recent personal concern with health maintenance or prevention were more likely to have delayed care because of concerns about privacy or a supervisor finding out. Residents who endorsed greater frequency of avoiding or delaying care because of worries about academic jeopardy, for instance, also reported greater sensitivity to negative social messages about being or becoming ill—that is, more frequently hearing colleagues discuss residents' illnesses or absences, having more negative perceptions about how residents are treated after absences, and perceiving more ostracism by other residents after illness-related absences. These findings are consistent with a British study that reported that most junior doctors had not taken any leave for the majority of illness episodes during the prior 6 months and that a majority of them had expressed concerns about colleagues having to cover.19
Residents were asked whether colleagues who must miss work because of health issues were treated with empathy or ostracism. They endorsed both. This mixed response speaks to the lack of coherence across the “lessons” to be found in the “hidden curriculum”—the culture of medical training, with its distinct narratives and role modeling.28 The hidden curriculum is believed to be critical and formative in trainees' experience, affecting not only attitudes and behaviors toward patients and colleagues but also the physician's self-orientation. Basic work attitudes, including whether to take leave when ill, whether to seek care if it means missing work, and whether to admit to having a personal health issue, are affected by a nuanced set of cultural “rules,” both spoken and unspoken, which vary among specialties and institutions. Recognition of unintended consequences of the cultural milieu of residency training is an initial step toward implementing changes that will positively influence physicians-in-training. The endorsement of both empathy and ostracism also may reflect the emergence of a more progressive stance in academic medicine, or it may be a marker of the more enduring cultural complexity that residents experience. This topic warrants further study. The positive implications of new knowledge in this area may be of great value to those who despair of the cynicism-engendering nature of medical training,29,30 and the experience of being ill and of being cared for in a health care system may impart deep and indelible lessons of compassion and professionalism.31–33
Our overall set of findings suggests several plausible interpretations. It is possible that delaying or avoiding care may be a marker for some other factor, such as general distress or feeling overwhelmed. It is also possible that those residents who endorsed delaying care were, in some way, more honest, forthcoming, or conscientious about their health concerns, because they acknowledged the need to seek care. Stated more negatively, these residents may tend to fret more, both about their own health but also about repercussions of people finding out about their health concerns; such worries could overlap with endorsement of poorer mental and physical health and more health concerns of various types. Without further information about the residents' personal characteristics, such relationships remain speculative. It is also possible that some residents had themselves either witnessed or experienced adverse effects of having health problems, seeking care, or having their privacy violated and, therefore, assigned greater importance to these experiences or “risks.” Finally, it is possible that the measured variables could have tapped overlapping constructs, in that awareness of not receiving proper health maintenance or care for chronic conditions may have influenced self-rated health.
Gender differences in residents' attitudes toward care-seeking
Among our findings, a number of gender differences emerged. Compared with men, women reported worse mental health and greater health concerns overall, expressed greater concerns about potential jeopardy to training status from pregnancy-related issues, cancer, and alcohol-related issues, and had avoided seeking health care more often. Concerns about privacy violations, a supervisor finding out, or any other reason for avoiding care were greater for women than men as related to postponing care. Also, women noted more frequently having heard residents discussing other residents' health problems or absences from work and perceived postabsence treatment of residents more negatively than did men.
These findings suggest that at least some women may be more attuned to their residency environment/culture as it relates to personal health care and/or that women tend to respond differently than men when surveyed about these issues. When women residents delay or avoid seeking care, for instance, such delays may likely be influenced by greater sensitivity to perceptions of the training culture (e.g., implicitly discouraging care-seeking), by residents' own mental and physical health and personal traits (e.g., the desire to appear in control, high levels of conscientiousness),34 and by negative messages associated with some of the most common concerns (e.g., depression and anxiety), which can impair even the most competent residents.35 Indeed, similar to the general public,36 residents in our study clearly have concerns about the privacy of their personal medical information. Women residents may be especially aware of these issues, because they simultaneously fulfill multiple roles in their health care settings—as trainees, employees, physicians, and sometimes as patients. Our findings thus illustrate that residents are sensitive to possibly deleterious effects of their health concerns becoming known to those in positions of authority—training directors or clinical supervisors. In one more example of the importance of gender-related perspectives in residency, it has been documented that women surgery residents hold divergent views from men on numerous items about duty hours restrictions37—for instance, half of male residents but only one quarter of female residents agreed that duty hours restrictions had decreased faculty expectations of surgical residents. The authors speculated that women's differing views may reflect the ways in which women are “more nearly outsiders than insiders” in the surgical environment. Because women currently comprise approximately 44% of the total number of residents in graduate medical education programs—including more than half of those in psychiatry and family medicine and nearly three fourths in pediatrics38—our findings, in the context of data on women physicians' health concerns,39,40 suggest that further exploration of gender issues in resident well-being is warranted.
Some strengths of this hypothesis-driven study are its use of an instrument evaluated in prior empirical projects, its inclusion of residents from different specialty programs, and a high rate of response for a written survey study on a potentially sensitive topic. Nevertheless, in interpreting our data, it is important to understand the limitations of this work. First, this project took place before revisions in residency requirements, including duty hours restrictions and the implementation of core competencies. Residents may now have more time to obtain care. Still, graduate medical training has not been radically changed regarding the key issues of this study, because residents continue to report high levels of stress, heavy workloads, and little personal time.26 No systematic efforts to deconstruct the influence of stigma in medical training have been undertaken. Because nearly 50% of residents are now women, the specific issues we examined as they relate to residents' personal care and overall well-being may be more critical than in the past.
Additionally, the study included only one training institution and, to allow for greater uniformity across programs, we confined our survey to PGY 2 and PGY 3 participants. Consequently, the views of this sample of residents may not be generalizable to residents elsewhere. The 71% response rate, however, and the participation of primary care as well as specialty residents, help mitigate the likelihood of substantial bias. We also did not inquire about a full range of reasons that might have contributed to delaying care, because this set of questions intentionally focused on privacy and jeopardy concerns. Responses falling into the “other” category indicate that there were likely numerous reasons or barriers that we did not specifically assess (e.g., logistical or scheduling constraints). Thus, as with any survey, our interpretations must be tempered by an awareness that the selection of content areas for study as well as phrasing of questions likely affected responses. The lack of a control group of similarly aged people from the general population or in another professional training setting (e.g., law students) also makes us unable to draw any inferences about whether the issues described here are specific to medical students or may extend to young adults or people in professional training in general. Additional limitations were the self-report nature of responses and the potential for recall bias.
Greater attention to residents' health issues and care-seeking will benefit our early-career colleagues and represent a valuable effort on behalf of the patients to whom their lives will be entrusted. Steps that programs and institutions can take to improve the likelihood that residents will seek appropriate care include recognizing and affirming the essential health care needs of residents; promoting work environments that explicitly identify the importance of both routine and urgent health care, as well as the overall importance of resident well-being41; building accessible, efficient avenues for care (e.g., with quick turnaround, walk-in services, extended hours); actively discouraging students and residents from working through illness; encouraging and rewarding role modeling of appropriate self-care by attending clinicians and other teachers of residents; and making residents aware of existing policies and safeguards related to self-care and confidentiality.
The authors gratefully acknowledge support from the Arnold P. Gold Foundation for this project (Dr. Roberts, principal investigator). Dr. Roberts is funded through the Research for a Healthier Tomorrow-Program Development Fund, a component of the Advancing a Healthier Wisconsin endowment at the Medical College of Wisconsin, and she also expresses her appreciation to the NIMH for support through her Career Development Award (1K02MH0298). Dr. Dunn gratefully acknowledges support from the NIMH for her Mentored Career Development Award (K23MH66062).