Explicit bias among fourth-year medical students

Abstract

The aim of this study was to analyze themes related to explicit bias in patient-doctor relationships among fourth-year medical students. Class cohorts between 2013 and 2016 taking an online elective, “Self and Culture,” submitted reflections about explicit bias. Thematic analysis was conducted on 283 student submissions totaling 849 entries until saturation. Themes included explicit bias toward patients with obesity, those who smoked, those from low-socioeconomic conditions, and, to a lesser extent, race/ethnicity. Themes related to the patient-doctor relationship included a negative impact on the relationship itself, trust, treatment of the patient, and patient experience. Themes related to making a positive impact included seeking positive treatment of the patient, understanding patients’ circumstances rather than making assumptions, partnering with the patient, and education. Furthermore, researchers noted external versus internal attribution of the bias. Some students used neutral language to explain explicit biases, whereas fewer used internal attribution language. Results demonstrated that this type of reflection promoted personal insight, and faculty members should be trained to ensure successful crucial conversations about the impact of assumptions and biases on patient treatment, care plans, and health disparities. Finally, the curriculum should be intentional, providing experiences with marginalized populations to develop cultural humility and empathy.

Explicit bias is a conscious positive or negative feeling and/or thought about groups or identity characteristics.¹ Because these attitudes are explicit in nature, they are espoused openly, through overt and deliberate thoughts and actions. In a national study with medical students, explicit weight bias was more prevalent than other types of bias, including racial bias²; however, weight-related biases are not the only concerns among physicians. Studies have shown physicians have implicit racial bias toward patients that influence their interactions through either a bias against people of color or a bias that favors white people.^3–6 Further, a national study with medical students found that nearly half reported explicit bias toward gay and lesbian individuals.⁷ However, it is also worth noting that another study found no racial or gender bias in senior medical students’ recommendations for procedures.⁸

Methods

In 2011, a fourth-year elective course was created for medical students titled “Self and Culture,” which helped satisfy cultural competencies required by the Liaison Committee on Medical Education. The course was designed as an online elective providing flexibility meeting the needs of fourth-year medical students. Students had the option of choosing 1, 2, or 3 weeks of the online elective. For 2 years, an assignment in the final week asked students to create a module that they believed would be helpful for future classes. Several submissions directly addressed bias in medical students. Beginning in the 2013–2014 academic year, with permission from two students, we adapted their modules to create a new module addressing explicit bias. The learning objective was to promote self-awareness about attitudes, beliefs, biases, and behaviors that may influence clinical care. After a lesson about bias and its impact on health care, students identified three biases, the impact that they could have on the physician-patient relationship, and how identifying the biases could change the way they approached their patients. The aim of this study was to explore potential themes in identified biases as well as themes in students’ reflections on how biases might impact patient care.

This was a retrospective qualitative study, using deidentified data from the bias module submission from the “Self and Culture” course. Fourth-year medical students voluntarily took the elective and submitted modules via Blackboard. There was neither in-person interaction with the instructors nor interaction with other peers in the course. Module material was not available to anyone outside of the instructors and the course coordinator. Students submitted written paragraphs addressing the following: (1) list three biases you have about people or cultures you may encounter as a physician; (2) discuss the impact that these biases could have on the physician-patient relationship; and (3) address how identifying your bias changes the way you approach your patients. We included submissions from three academic years spanning 2013 to 2016, totaling 283 students with three entries per student (849 total entries). There were 418 fourth-year medical students in these three cohorts; therefore, the sample accounted for 67.7% of all fourth-year students during this time frame. We used thematic analysis to assess reoccurring themes within the submissions. Exclusion criteria included those students who had not yet graduated. We (L.H. and B.A.W.) independently reviewed the entries until each reached theoretical saturation; saturation was reached after each reviewing author failed to find any new insights or information within the data.⁹ Following independent thematic analysis, themes were compared to ensure credibility to achieve trustworthiness.¹⁰

Results

The first assignment asked students to self-identify their explicit biases observed in the patient population. The explicit biases identified most often were toward patients with obesity and those who smoked. The accompanying stereotypes for obesity included laziness, disregard for self, and noncompliance, and smoking had similar descriptors. Within these biases and stereotypes was the common sentiment that patients did not care about their health. Additional biases and stereotypes captured were based on socioeconomic conditions such as homelessness and fewer opportunities for access to resources. In these instances, themes included patients being incapable of self-care, living unhealthy lifestyles, and being noncompliant. Other biases surrounded patients’ race/ethnicity and included themes of less education, noncompliance, and being difficult patients; however, these were not as frequent as biases related to obesity, smoking, and socioeconomic conditions. Although the submissions did not categorize one group as more compliant than another, they did identify compliance related to specific biases. Some of the examples cited a specific gender, race, or education level as more likely to be compliant.

One last theme was around a lack of trust for certain groups such as homeless patients, which was expressed in phrases such as “drug seekers” and “frequent flyers.” This statement reflects this bias: “Individuals of lower socioeconomic status are just narcotic seekers.” Another common thread related to perceived level of value or care for their own health and compliance. When coupled with being directed toward traditionally marginalized groups and/or those with limited resources such as homeless individuals or those coming from low-socioeconomic situations, these stereotypes were especially concerning, because they could further harm or disenfranchise already oppressed populations. This reflection relays this thought:

 Homeless people are only looking for a handout and are not interested in bettering themselves. The circumstances of why someone ends up homeless or uninsured have an endless range; oftentimes it involves quite a bad hand dealt and a tragic story. The easier thing to do, rather than wrestle with any sobering details with the patients, is to chalk up their predicament as “they brought it on themselves” and they “deserve to be homeless.” Working through my month in the ER [emergency room] there were certain patients that were considered “frequent flyers,” and that reputation certainly had a negative connotation. Physicians viewed these patients as a nuisance, who just wanted a roof over their head for the night, and worst of all, weren’t even willing to pay money. This flippant attitude towards the indigent, uninsured community may create an atmosphere where only a surface level of medical treatment is given to those who have deep and chronic conditions.

The second question provided students the opportunity to reflect on how their self-identified biases may impact their patient-physician relationship. Common overall themes included concern that explicit biases could negatively impact the relationship and hinder trust, could negatively impact the care plan for the patient, could result in the patient being treated negatively, and could leave the patients with negative experiences as a result of their treatment. One reflection highlights the impact that a bias toward obese patients could have on the relationship:

 I held the bias about obese patients. It was not until I did an indigent care elective in pediatrics that I realized how multifaceted the issue of obesity is and how wrong it was for me to have this bias. I realized that by telling my patients and their parents a tirade over the importance of diet and exercise, I was losing their trust and respect as a health care provider. I was ignoring the other aspects of obesity. It was a mistake on my part to view these patients as being unwilling to work hard.

This reflection further highlights the theme of impacting the care plan by ignoring potential contributing factors, thus negatively impacting the overall care of the patient. These two themes were seen across a variety of identified biases.

Finally, themes emerged regarding how the bias could change the way they approached patients. Student acknowledgment of the effect of their biases and preconceived stereotypes on the quality of patient care (questions 1 and 2) provided an opportunity for consideration of possible remedies. Within clinical training, there is rarely an opportunity to reflect upon the personal perspectives that each trainee brings to a patient encounter and the ramifications of these perspectives. The themes that emerged included seeking positive treatment of the patient, seeking understanding of the patient’s life rather than making quick judgments or assumptions, partnering with the patient, and educating the patient. These included approaches such as being more empathetic, practicing more patience, treating all patients with respect and dignity, not judging patients, making an effort to listen to the patient’s story, working with the patient as a team, giving the patient a voice that reflects partnership, or taking responsibility as the provider. The reflection of providing further education to the patient as a solution to helping to overcome the bias shifted the responsibility back to the patient for the care provider’s own stated bias. For the bias that homeless individuals are “completely neglectful of their health,” one student proposed to address the bias by treating “this patient the same as anyone else and . . . make sure they are fully educated in how to take care of themselves.” This thought process and solution completely miss the multifaceted factors that a homeless person faces in meeting basic needs for survival, which likely come before and are impacted by having the necessary means to care for one’s health.

Personal reflections about how the students felt were not as ubiquitous as those that reflected in a more neutral tone about the bias. Many used words such as “we” to describe what physicians in general could do rather than taking ownership themselves. This powerful notion gives agency to all physicians, rather than being accountable for one’s own thoughts and ideas. One entry relayed this realization:

 Since many practitioners don’t feel comfortable addressing the holistic needs of their patient, . . . I think it can be a coping mechanism to shift complete responsibility back to the patient; our inadequacy as healers fuels the stigma to some degree, in my mind. Until this tendency is addressed, I fear individual physicians may struggle to cultivate a therapeutic relationship and fully address the needs of their patients.

Another example of what seemed like cognitive dissonance, distancing, or justification for an identified bias was related through this reflection, “This is a bias that is not unique to myself, but I would venture to say permeates the medical profession.” When these themes became apparent, there were essentially two categories of responses, which were internal or external attribution. Those who looked internally found ways to be proactive and self-reflected on how to truly address any concerns, whereas those who used justification language projected the issue back onto the culture of the profession or the patient.

Discussion

In the seminal monograph by the Institute of Medicine, Unequal Treatment,¹¹ the authors concluded:

 Bias, stereotyping, prejudice, and clinical uncertainty on the part of health care providers may contribute to the racial and ethnic disparities in healthcare. While indirect evidence from several lines of research supports this statement, a greater understanding of the prevalence and influence of these processes is needed and should be sought through research.

To respond to this pressing need and extend understanding of bias as it relates to training physicians, we sought insight regarding explicit bias and opportunities to reframe understanding to enhance patient care. As medical educators, it is important to assist students in considering all identity characteristics that patients bring into the clinical encounter, including race, gender, sexual orientation, age, and social factors, including socioeconomic conditions and obesity. However, these factors do not stand alone and must be considered along with patient history and presenting ailments. Although holistic medicine is a philosophy that a majority of physicians aspire to, the educational experiences often fall short. Student learning and pedagogy are closely linked to educational attainment. Students are taught key identifiers when looking at health disparities instead of broadly considering the cultural factors that lead to disparity. This reflects an endemic problem with cultural competence, in that within many clinical encounters, cultural competence is linked to specific illness. The idea of race-based medicine has been consistently challenged, because it can lead to misdiagnosis and further erode the trust of patient populations that are already skeptical of the medical establishment. Instead, tools such as Kleinman’s questions¹² can help recenter the patient in the medical history and provide more accurate and germane information about the patient’s experiences.

A primary limiting factor in this study was the inability of the researchers to interact with the subjects to ascertain meaning, context, and depth relative to the students’ responses. As such, their concrete written reflections, at times, leave questions unanswered. Another potential limiting factor may be found in the subjects that students were willing to engage. Within the context of our society, race, gender, and sexual orientation continue to be deeply divisive topics, ones that students seem reluctant to address and reflect upon. By allowing space for students to be vulnerable to discuss their own explicit biases, there is an acknowledgment that we will be limited by what students feel comfortable to explore, particularly with those who are assessed and graded. Finally, students were required to identify three explicit biases for this assignment and may have felt forced to identify something that was not true for themselves.

In conclusion, students who utilized ownership language were much more reflective and proactive in their submissions; thus, it is possible that the vulnerability needed for this type of assignment was uncomfortable for some. A recent study by Wellbery et al, who taught a two-semester populations health course to first-year medical students and had them write reflective writing assignments about caring for vulnerable populations, found that this brief time only had a limited impact on social empathy.¹³ Though personal empathy is a critical characteristic in medicine, there could also be structural or environmental barriers for cultivating social empathy. For example, first-year medical students identified aspects such as observing callous behavior within the institution that demonstrated a focus on self by the providers and indifference to others, setting the tone for what new medical students should model.¹³ Like Wellbery et al, we would recommend that more needs to be done to demonstrate to medical students that explicit biases could result in lack of empathy and mistreatment of patients. Medical educators cannot assume that students understand the critical nature of this topic, and though reflection is important, applied experiences early in their medical education may also be valuable. These could include exposure to disadvantaged, underserved, and marginalized populations through clinical encounters, volunteering, and coursework. Additionally, in all components of education the institution needs to be mindful of messaging about negative treatment of patients in its climate and modeling. The responsibility should be shared among educators, mentors, and clinicians to help students dispel erroneous assumptions through demonstrating empathy and reflecting on how biases can negatively impact care.