Physician and Trainee Experiences With Patient Bias

Key Points

Question  How do physicians and trainees perceive, react, and respond to incidents of biased patient behavior?

Findings  In this qualitative study of convenience samples of 50 trainees and physicians, participants reported a wide range of experiences with biased patient behavior ranging from belittling comments to outright rejection of care. Participants described a large negative effect on their emotional well-being and the clinical care environment, and many described uncertainty regarding appropriate and effective ways to respond to these encounters.

Meaning  Physicians’ and trainees’ reports of uncertainty, confusion, and pain associated with biased patient behavior indicate a need for training and institutional policies to deal with biased patients.

Importance  As the clinical workforce becomes more diverse, physicians encounter patients who demean them based on social characteristics. Little is known about physicians’ perspectives on these encounters and their effects. This knowledge would help develop policies and best practices for institutions and training programs.

Objective  To describe the range and importance of encounters with biased patients and the barriers and facilitators to effective responses.

Design, Setting, and Participants  This qualitative study recruited convenience samples of hospitalist attending physicians, internal medicine residents, and medical students from 3 campuses affiliated with 1 academic medical center. Data were collected from 50 individuals within 13 focus groups from May 9 through October 15, 2018. Focus groups were conducted using open-ended probes, audiotaped, and transcribed. Participants used their own definition of biased patient behavior. Each transcript was independently coded by at least 2 investigators. Data were analyzed from May 2018 through February 2019.

Main Outcomes and Measures  Major themes associated with types of encounter, importance to the participant, and barriers and facilitators to effective responses were abstracted through the constant comparative approach.

Results  Overall, 50 individuals (11 hospitalists, 26 residents, and 13 students) participated; 24 (48%) were nonwhite. At total of 26 participants (52%) identified as women; 22 (44%), as men; and 2 (4%), as gender nonconforming. Reports of biased behavior ranged from patient refusal of care and explicit racist, sexist, or homophobic remarks to belittling compliments or jokes. Targeted physicians reported an emotional toll that included exhaustion, self-doubt, and cynicism. Nontargeted bystanders reported moral distress and uncertainty about how to respond. Participant responses ranged from withdrawal from clinical role to a heightened determination to provide standard of care. Barriers to effective responses included lack of skills, insufficient support from senior colleagues and the institution, and perception of lack of utility associated with responding. Participants expressed a need for training on dealing with biased patients and for clear institutional policies to guide responses.

Conclusions and Relevance  In this qualitative study of physicians and medical students, encounters with demeaning patients ranged from refusal of care to belittling jokes and were highly challenging and painful. Addressing biased patient behavior will require a concerted effort from medical schools and hospitals to create policies and trainings conducive to a clinical environment that respects the diversity of patients and physicians alike.

The US health care workforce has become increasingly diverse: 28% of practicing physicians are foreign-born,¹ 51% are nonwhite,² 10% are from minority groups underrepresented in medicine,² and 34% are women.² Consequently, patients increasingly encounter physicians whose identity may not conform to their notion of a trustworthy, competent physician. In response, some patients reject or demean physicians based on their social characteristics. Although systematically collected data on the prevalence of these interactions are lacking to date, reports suggest they are common.^3-11 In an online survey, 59% of physicians reported having been demeaned by a patient within the last 5 years.¹¹ Minority and Asian physicians are more likely to face these experiences.^3,11 Physicians’ anecdotal reports portray these incidents as painful and often powerfully emotional.^3,8,11-15 These encounters compound the difficulties of a career with a high rate of burnout and may, in certain locales, pose challenges for physician recruitment.^3,8,11-14

Beyond reports from individual physicians,^3,5,15-22 little is currently known about patients’ biased behavior, the importance of these clinical interactions to physicians and clinical care, and how physicians respond. Prejudiced remarks and behaviors are difficult to confront in any setting; for physicians, these difficulties are heightened by the need to provide clinical care and the ethical and legal dictates of the medical profession.²³ Medical education often fails to prepare trainees for biased patients, leading even experienced physicians to flounder when responding to these interactions.^11,15,19,24 Lack of data on the nature of these encounters makes it difficult to create policies or training that can mitigate the effects of these experiences.

The paucity of research in this area, the complexity of the subject matter, and the need to understand the perspectives of physicians^25,26 led us to conduct a qualitative investigation of the experiences of attending physicians, internal medicine residents, and medical students with patients who had demeaned them based on their social characteristics. We sought to capture the perspectives and experiences of those directly subjected to prejudiced behavior and that of witnesses to such encounters. We also aimed to determine the emotional and clinical importance of these interactions, the ways in which physicians and trainees responded, and the barriers and facilitators to responding effectively.

We conducted a qualitative investigation based on facilitated focus groups with convenience samples of internal medicine hospitalist physicians, internal medicine residents, and medical students affiliated with the University of California, San Francisco, School of Medicine (UCSF). Focus groups were selected over in-depth interviews for their potential in elucidating a greater range of themes and perspectives.^27,28 Using principles of grounded qualitative research, we aimed to generate hypotheses from the data rather than testing preexisting hypotheses.^26,29 The institutional review board of UCSF approved the study procedures, and all participants provided written informed consent.

Separate focus groups were held for faculty, residents, and students. Hospitalist faculty at Zuckerberg San Francisco General Hospital and the San Francisco Veterans Administration Medical Center participated in focus groups held in lieu of a routine faculty meeting. Residents and students were invited by email to join focus groups held after required didactic sessions. Trainees at UCSF rotate through multiple sites, including a public county hospital (Zuckerberg San Francisco General Hospital), 2 nonprofit academic hospitals, and the San Francisco Veterans Administration Medical Center. No attempt was made to recruit participants based on social characteristics.

Participants were provided a brief written study overview highlighting voluntary participation. Focus groups were facilitated by study investigators, audio recorded, and professionally transcribed. Participants were assigned a number used to identify them while speaking, which linked to a form where they self-reported their gender, race/ethnicity, and, for residents and medical students, year of training.

Focus group facilitators used a semistructured interview guide with open-ended questions (eMethods in the Supplement). Because we were interested in participants’ implicit working definitions of demeaning experiences, we did not provide a definition. Participants were encouraged by probes to describe the incident, to detail how they and others on their team responded, and (for trainees) whether they involved any supervisors. Probes also elicited details on personal and professional importance of the encounter and barriers and facilitators to effective responses. Focus groups were a mean of 43 minutes (range, 40-90 minutes) in duration. We conducted focus groups until thematic saturation was achieved.

Using focus group field notes, 5 members of the study team (M.W., S.d.B., A.P., A.M., and A.F.) generated a preliminary codebook. To ensure intercoder reliability, all investigators first independently analyzed 1 transcript using the existing codebook. We used the constant comparative method to expand existing themes and identify novel ones.^26,29 Through this method, a final coding list was created, reconciled within the group, and applied to all transcripts. Two investigators (M.W. and S.d.B.) read and coded all transcripts; other study members independently read and coded 1 to 4 transcripts (A.P., A.M., and A.F.), ensuring that every transcript was read a minimum of 3 times. ATLAS.ti software, version 8.0 (Scientific Software Development) was used for data analysis and retrieval. Data were analyzed from May 2018 through February 2019.

Fifty participants, including 11 hospitalists (22%), 26 residents (52%), and 13 medical students (26%), were interviewed in 13 focus groups ranging in size from 1 to 11 participants. Participants were diverse in gender (22 identified as men [44%]; 26 as women [52%]; and 2 as gender nonconforming [4%]) and racial/ethnic identification (26 white [52%], 8 Latinx [16%], 7 Asian [14%], 3 South Asian [6%], 1 Middle Eastern [2%], and 5 black [10%]) (eTable in the Supplement).

Incidents of biased patient behavior described by participants ranged from outright rejection of care and racist, sexist, or homophobic epithets to inappropriate compliments, flirtatious remarks, and jokes reflecting ethnic stereotypes (Table 1). Although infrequent, refusal of care was experienced or directly witnessed by some participants: “[The doctor] was Sikh so he had a long beard…the patient demanded, ‘I don’t want to be seen by a Muslim doctor’ [fourth year-medical student, man, South Asian].” Explicit biased remarks expressed through homophobic, racist, sexist, or Islamophobic sentiments were also described.

Biased patient interactions characterized as less extreme ranged widely and were often described by participants as microaggressions. These included questioning the physician’s or trainee’s role (eg, by assuming that women were nurses or that minority physicians were support staff). One resident recounted, “…then [the patient] turns to my attending and says, ‘Can you please tell the interpreter that I don’t want to leave’…my attending jumped in and said ‘she’s actually your senior doctor’ [third-year resident, woman, Latina].” Although most reports focused on verbal interactions, some participants noted nonverbal instances such as caring for patients with swastikas or other tattoos with racist connotations. References and aggressive inquiries into participants’ ethnic background were also often experienced as biased: “‘Oh, where’s that name from?’ I’m always 2 to 3 questions away from, ‘You’re Arabic’ [second-year resident, man, Syrian].” Finally, gendered or flirtatious comments or compliments directed at participants were described as pervasive: “At the VA, someone called me ‘honey’ at least once a day [fourth-year medical student, woman, white].”

These encounters were associated with self-reported negative and positive effects on the professional development of students, residents, and attending physicians (Table 2). Many participants described emotions of anger, confusion, and fear that were distracting and painful. One physician recounted: “Being in those situations…it makes the rest of the day painful to get through…and if [it] is happening to you every day, it stacks up [third-year resident, woman, Asian].” Some described withdrawing from their clinical roles: “I found myself visiting [a patient who had propositioned me] less than I would any other patients…[fourth-year medical student, woman, South Asian].” Others reported avoiding clinical sites where those encounters were reputedly common: “There are a lot of things I love about the VA, but I couldn’t work there because of the sexual harassment by patients [second-year resident, woman, white].” Most notable were the negative effects on learning and clinical practice: “[Biased interactions] take away from your ability to focus on learning or training or developing into a better clinician [first-year resident, woman, white].” Attending physicians had similar beliefs: “When it [a sexist or racist comment] happens to me…it makes my mind feel crowded, emotional, scared…it makes it harder to practice medicine [hospitalist, woman, white].” Finally, some participants reported accepting biased behavior as a part of the job: “The way that I deal with [biased behavior] is by accepting that this is how people view me…I don’t correct patients and they continue to call me ‘honey’ or ‘sweetie’ [fourth-year medical student, woman, Asian].”

In contrast, other participants reported a growth in self-efficacy from learning to respond to these encounters: “I’m very sensitive to being called not a doctor…that’s the one recurring microaggression that I feel very confident correcting [because] it happens a lot [third-year resident, woman, Latina].” Support from clinical team members enhanced feelings of security and inclusion and mitigated the effects of biased encounters. One participant recalled, “I was so grateful for [the attending’s] response....He [said], ‘Wow. I imagine this could feel very threatening and scary to you....’ I really felt like that attending had my back [fourth-year medical student, woman, white].” Finally, these experiences with biased patients propelled others to assume leadership roles to support their peers and mold medical culture: “I feel really motivated to support my colleagues in dealing with these difficult situations and I feel lucky that I’ve had good modeling for that [fourth-year medical student, gender nonconforming, white].”

Responses to these encounters could occur at the individual, team, and institutional levels (Table 3). Setting limits was a frequent tactic. For example, many individuals chose to reassert their clinical role, such as the participant who told the patient to “focus on the fact that she’s a really good doctor, and not on her appearance… [third-year resident, woman, white].” Creating a team plan to protect targeted individuals or debrief the incident was described as particularly helpful because it provided emotional support. One team, after days of enduring racial epithets and sexist comments, decided: “Our…policy…is going to be that if he uses that language with anyone, we will say, ‘We are going to step out of the room and come back and see you again when you use language that’s appropriate’ [third-year resident, woman, Latina].” A few participants reported avoiding confrontation with the patient by excusing or ignoring the behavior or withdrawing from patient engagement. Another strategy involved explaining the detrimental clinical consequences of the behavior because it usually created a delay in care.

On the institutional level, reassigning the patient to another clinical team or switching physicians within teams was a common response. Behavioral teams were available in some hospitals to help fashion interventions such as behavioral contracts or to negotiate transfer to other teams. In addition, some participants reported seeking out training on dealing with biased patients, a strategy they found helpful: “It’s much easier to have a plan in advance…because when you’re in the room, a lot of times, you’re a little bit shell shocked [and having had training] you have a better plan of what to do in real time [third-year resident, woman, Latina].”

Participants described many barriers to confronting biased patient behavior (Table 4). Clinical priorities often dictated a participant’s decision not to respond. Derogatory remarks from intoxicated or mentally ill patients, for example, were considered part of the illness and much less likely to be addressed: “I don’t try to address [biased behavior] if I think the patient…is completely psychotic…but it’s hard to know what to do…because those words still hurt [second-year resident, man, white].” In addition, concerns about preserving the therapeutic alliance prevented some participants from responding to biased comments.

Lack of skills was frequently cited as a reason for not addressing offensive comments. Uncertainty about who on the team should respond was also widespread. One resident asked rhetorically, “[I]s it the most senior person who should be saying something? Is it the person to whom the comment is being directed who should say something? Is it the person who is not of the group being targeted who should say something [second-year resident, woman, white]?” Uncertainty over appropriate responses to incidents perceived as less extreme and unintended to harm also impeded addressing biased behavior. One resident explained: “There’s a contrast between the stuff that you can clearly state ‘That’s racist’ and ‘That’s racist but he didn’t mean ill by it.’ That’s something [the latter] I almost struggle more with [third-year resident, woman, Latina].”

Residents and students perceiving a lack of support from their attendings or the institution were reluctant to address biased behavior with the patient. One participant contrasted the response of 2 attendings: “The patient ask[ed] me to join him in his bed…I felt disappointed with the way the attending handled it because he laughed and said ‘That’s not her job description’…the message was that I needed to get over it to take care of the patient…But…with another attending this patient made a similar comment…and this attending said, ‘We are not going to continue taking care of you if you can’t respect our trainees’ and I thought, ‘Oh, there is a different way to deal with this type of treatment from a patient’ [fourth-year medical student, woman, South Asian].”

Absent or unclear guidance resulting from lack of knowledge of institutional policies caused difficulties, particularly when confronting ethically complex situations such as weighing the duty to provide patient care against the desire to protect themselves or colleagues. In such cases, the sense of professional duty often dominated. One participant reflected: “I didn’t know if I could excuse myself from taking further care of that patient, or if by doing so I was passing off that dangerous situation to my co-intern [second-year resident, woman, Latina].”

An assumption by some participants that professional duty includes “rising above” these interactions served as a considerable barrier. Others worried that responding would reflect poorly on their professional competence: “There is this fear on the part of women or on the part of the people from different ethnic groups, different sexual orientations, [that in responding they will] be perceived as hysterical [hospitalist, woman, Latina].” Trainees in particular were concerned that their responses would be perceived as unprofessional and affect their evaluations.

A sense of futility led others not to respond. One resident reflected, “What’s the goal here? Are we really trying to correct these veterans’ racist beliefs [third year resident, man, white]?” Strong emotions beyond futility also drove nonresponsiveness: “If I processed everything…I’d go crazy…So I block it out and keep working [third-year resident, woman, black].”

Participants’ decisions on how or whether to respond were determined by personal values and perceptions of support (Table 5). A sense of professional responsibility for others, often driven by a position of relative power in the medical hierarchy, and personal investment in values of equity spurred some to respond: “The majority of the time, I don’t say anything. I think the reason I did [say something in this case], was mostly motivated by being really angry…I saw the injustice of him judging his nurse, so I felt empowered to say something [third-year resident, woman, Latina].” Participants who had seen others model appropriate responses were also more likely to choose to respond. Finally, explicitly egregious behavior for some participants shifted the ethical balance from risking the therapeutic alliance to the need to protect themselves and others: “When something is really blatant, it’s really easy to respond…so if someone makes an inappropriate comment…about something overtly sexual it’s easy to say, ‘That’s inappropriate. And this is my professional role, and we’ll have a respectful relationship.’ [first-year resident, woman, white].”

To better understand the experiences of an increasingly diverse physician workforce, we undertook a qualitative study into patient encounters that demean physicians based on social characteristics such as race, ethnicity, sex, gender conformity, sexual orientation, and religion. We found that the nature of these experiences ranged broadly, spanning outright refusal of care to demeaning quips and belittling language. We also found that they are painful, often confusing, and influence the professional development of students and residents. Finally, we found that there are multiple barriers and facilitators to effective professional responses that can serve as a foundation for training and policies to enhance physician welfare while balancing the rights of patients.

The effects of these encounters were meaningful. Physicians and medical students described an emotional toll that undermined well-being, learning, and occasionally patient care. Emotions linked to physician and trainee burnout, such as emotional exhaustion, fear, self-doubt, and increased cynicism,^30-32 were common responses, and the emotions lingered long after the event. Because some students described avoiding rotations and clinical sites to protect themselves, medical schools should develop policies to guide clerkship directors in formulating strategies for dealing with biased patients. Institutional policies that address egregious patient behavior, such as refusal of care, are also necessary to protect residents and attending physicians. Nontargeted bystanders reported moral distress from not knowing how to protect their colleagues. As such, institutional guidance for these encounters will improve the working environment for all.

Reports on barriers and facilitators to effectively navigating these encounters underscore the need to incorporate training on dealing with biased patients into medical curricula. We found that team support was crucial in transforming painful experiences into learning moments that enhanced inclusiveness and self-efficacy. Consequently, training should focus on both individual and team responses and include setting limits, appropriate deflection, team plans, debriefings, and support for the offended physician or trainee. For academic settings, faculty development initiatives will be necessary. Faculty members with little personal experience in these matters may need to discuss how these incidents differ from other difficult encounters with patients, such as the patient who, faced with a newly diagnosed cancer, lashes out at the treating physician. We believe this distinction is key. Professionalism requires physicians to accept a broad spectrum of human behavior in response to illness: physicians accept the need to deal with difficult patient situations. However, professionalism does not require physicians to readily accept attacks on fundamental aspects of their identity and self-worth.

Our findings should be interpreted in light of several limitations. As with any qualitative study with a convenience sample, we do not know the views of those who did not participate in interviews. In this study, participants were affiliated with a single medical school in the San Francisco Bay Area with a diverse population and a culture of supporting diversity. Our findings may not reflect the experiences of physicians in other geographic regions. We also did not quantify how often these experiences occurred. Qualitative and quantitative studies of physicians and trainees from different institutions and regions are needed to fully understand the clinical and workforce effects of encounters with demeaning patients.

Despite these limitations, this study revealed that physician, resident, and medical student experiences with biased patient behavior range broadly and have a powerful effect on witnesses to the behavior and those targeted. Diversifying the physician workforce remains a national priority.³³ Despite slow progress in increasing the numbers of African American and Latinx physicians, the clinical workforce has become increasingly diverse, and demeaning patient interactions are likely more common.³⁴ Addressing demeaning behavior from patients will require a concerted effort from medical schools and hospital leadership to create an environment that respects the diversity of patients and physicians alike.

Accepted for Publication: July 28, 2019.

Corresponding Author: Alicia Fernandez, MD, Division of General Internal Medicine, Zuckerberg San Francisco General Hospital, UCSF, PO Box 1364, San Francisco, CA 94143 (alicia.fernandez@ucsf.edu).

Published Online: October 28, 2019. doi:10.1001/jamainternmed.2019.4122

Author Contributions: Drs Wheeler and Fernandez had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Wheeler, de Bourmont, Paul-Emile, Fernandez.

Acquisition, analysis, or interpretation of data: Wheeler, de Bourmont, Pfeffinger, McMullen, Critchfield, Fernandez.

Drafting of the manuscript: Wheeler, de Bourmont, Fernandez.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Wheeler.

Obtained funding: Wheeler, Paul-Emile, Fernandez.

Administrative, technical, or material support: Wheeler, de Bourmont, Pfeffinger, McMullen.

Supervision: Wheeler, Pfeffinger, Fernandez.

Conflict of Interest Disclosures: None reported.

Funding/Support: This work was supported by the Greenwall Foundation.

Role of the Funder/Sponsor: The sponsor had no role in the study design; collection, management, analysis, and interpretation of the data; preparation, review or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: Gurpreet Dhaliwal, MD (UCSF), and Lawrence Haber, MD (UCSF), assisted in organizing hospitalist focus groups, for which they were not compensated. Eric Steinbrook, BA (University of Michigan School of Medicine), assisted in acquiring grant funding and providing logistical support for focus groups. He was compensated as a staff member. We thank the students, residents, and faculty who shared their experiences with us.