It is a call day, and you are the medical student/resident/attending working up a patient being admitted to the general medicine service. You are reviewing emergency department (ED) records and previous notes in the electronic medical record (EMR). The patient is described as “difficult” and “uncooperative.” Additionally, the term “substance abuser” is seen throughout the chart. How will you approach this patient?
The pesky pager goes off again. Another ED admission. You open the chart. The HPI starts with patient reporting headaches starting 2 days ago. The HPI starts with, “27-year-old Black female sickle cell patient with repeated hospital admissions for ‘headaches.’ Patient states headache is 10/10 and is requesting dilaudid.” Do you already have positive or negative thoughts about this patient?
It is switch day. You started looking at the handoff for your first patient. “81-year-old elderly obese female being admitted for abdominal pain.” Notes say she is hard of hearing and a “poor historian.” What are your expectations when you meet this patient?
These scenarios are not uncommon on a busy inpatient service. The days are busy, what with balancing patient care, admissions and discharges, and many interruptions and handoffs. Therefore, we rely heavily on our EMRs to prepare background information to assist with an efficient encounter. How many of us have done the eyeroll while reading a patient’s chart or handoff? The deep breath and squaring of shoulders before proceeding to a patient room based on what we have read? My hand is raised. This unconscious and implicit bias is very common; however, it does our patients and us a disservice. Although there can be appropriately raised red flags and safety concerns in some cases, we find ourselves often perpetuating preconceived thoughts based on what has been written in the chart.
Why is it important for us to recognize that words can be stigmatizing and perpetuate bias? Because it interferes with our ability to listen to patients openly and actively. In the hospital, this is essential where trust and rapport need to be developed rapidly—these patients do not know us and we do not know them. What we read can influence our perceptions positively or negatively, leading to diagnostic error, inappropriate use of resources, propagation of healthcare disparities, and poor patient experience.
Biased Language and Testimonial Injustice
While there are few studies that evaluate the effect of how language propagates clinician bias through the chart, Goddu et al1 assessed whether stigmatizing language in the EMR is associated with trainees’ attitudes towards the patient and clinical decision making. A randomized vignette study of two charts employing stigmatizing versus neutral language to describe a hypothetical 28-year-old male with sickle cell disease demonstrated that stigmatizing language used in the medical record influenced trainees’ attitudes towards the patient and affected their medication prescribing behavior. The chart utilizing stereotyping and negative language cast doubt on the patient’s pain and influenced the patient’s pain regimen. Increased negative attitudes towards patients were demonstrated by residents versus medical students, attributed to experiential bias and ethical erosion over time. This study highlights the “powerful role of language in influencing clinician attitudes and behaviors.”1
We also need awareness of how language can be responsible for perpetuating bias based on social disparities. Testimonial injustice is defined as “that which occurs when a speaker receives an unfair deficit of credibility due to prejudice on the part of the hearer.”2 Beach et al3 examined medical records for testimonial injustice, identifying certain linguistic features that can be markers of disbelief in medical records of black compared to white patients. This suggested the presence of clinician bias around credibility. For example, using factual sentence structure versus evidential statements, such as “patient has a pain score of 5/10” versus “patient claims/complains pain is a 7/10,” or using quotes or judgment words seemed to be highly correlated with physician disbelief. Himmelstein et al demonstrated in a cross- sectional study, analyzing admission notes, stigmatizing language varied by medical condition and more often used to describe non-Hispanic Black patients.5
Beach et al3 also highlighted two possible reasons for casting doubt on a patient’s credibility. The first surrounds concerns about competency. Can the patient interpret the situation correctly and convey it with accuracy? The second reason is sincerity. Do we as clinicians believe we are being deliberately deceived? These questions are often biased by a patient’s background and what has been recorded in the chart. These doubts can be readily carried forward in the chart without confirming the accuracy or dismantling the bias later in the documentation.
Regarding quotations, the few studies done state that quotations from patients are not inherently negative, but thought must be given to the context and possible interpretation.
Open Notes: Patients Reading Clinician Bias
Consider the impact of biased language in charts that patients can now read due to open notes. Even if not overtly biased, some outdated terminology can be considered offensive or judgmental (e.g., “obesity” or “in distress”) even if they had been previously acceptable for inclusion in documentation. Reading such language can lead to anger, distrust of the clinician and the medical system, or even negative implications for the patient’s own outlook of their health. Patients may also perceive error, labeling, or evidence of respect based on a study done by Fernandez et al.4 Disclaimers at the bottom of chart notes stating that these are meant to be communication between clinicians are insufficient and cannot negate their impact on patient perceptions.
I also experienced this sense of mistrust when reading my father’s hospital records (SR). My father had a fear of surgery and chose to postpone recommendations for surgery. As an engineer, he weighed the pros and cons of the procedure and felt he still needed additional information. Although I was frustrated by this delay and his need to thoroughly analyze each part of the decision, I also had to let him process his fears of the medical system and his own mortality. When I read notes from his second hospitalization, the notes stated that my dad was declining treatment. I felt that this reflected his treating physicians’ dismissal of my father as not “compliant” despite all his adherence to the recommended treatment plan otherwise. I was frustrated by the lack of acknowledgement in the clinical notes of my father’s fear of surgery and his need to understand. Reading the cryptic notes, I recognized the biased attitudes of his physicians. Yet, this experience also made me question how often I am unconsciously doing this and how this influences other clinicians. Or, how often is reading others’ biased language influencing me?
Striving for Better
I hope I am improving at recognizing and reflecting on biased language for myself and my learners. Sometimes, I pause a trainee’s presentation to ask what they felt when they read the chart notes before they met the patient. I ask them to reflect on whether their impressions changed after meeting the patient, or I have another member of the team reflect on what they heard in the presentation. If we can take a moment to recognize when we are negatively influenced before meeting a patient, we can debrief about our own biases and strategize ways to be more conscious of it in the future. We can also take the time to observe if our medical decision making was impacted by our initial beliefs. This does take insight and time and I am sure that I am missing it more frequently than I am aware.
More studies and strategies are needed to determine how to document to communicate accurately, effectively, and without bias. Hopefully, we can continue to bring awareness to what and how we write to break down barriers to patients’ trust and improve the quality of care delivered.
The next time the pager beeps for a new admission, it is time to review or write a new chart note, or it is time for handoff, take a moment to recognize and mitigate potential biases in the language written.
- Goddu AP, O’Conor KJ, Lanzkron S, et al. Do words matter? Stigmatizing language and the transmission of bias in the medical record. J Gen Intern Med. 2018 May;33(5):685-691. doi:10.1007/s11606-017-4289-2. Epub 2018 Jan 26.
- Fricker M. Epistemic Injustice: Power and the Ethics of Knowing. Oxford, UK: Oxford University Press; 2009.
- Beach MC, Saha S, Park J, et al. Testimonial injustice: Linguistic bias in the medical records of black patients and women. J Gen Intern Med. 2021 Jun;36(6):1708-1714. doi:10.1007/s11606-021-06682-z. Epub 2021 Mar 22.
- Fernandez L, Fossa A, Dong Z, et al. Words matter: What do patients find judgmental or offensive in outpatient notes? J Gen Internal Med. 2021 Sep;36(9):2571-2578. doi:10.1007/s11606-020-06432-7. Epub 2021 Feb 2.
- Himmelstein G, Bates D, Zhou Li. Examination of stigmatizing language in the electronic medical record. JAMA Netw Open. 2022 Jan ;5(1):e2144967. doi:10.1001/jamanetworkopen.2021.44967.
Clinical Informatics & Health IT, Clinical Practice, Health Equity, Medical Education, SGIM
UNINTENDED BIAS IN STUDENT EVALUATIONS: A KEY TOPIC FOR INCLUSION IN INTERNAL MEDICINE RESIDENCY CURRICULUM
A female medical student reviews narrative feedback from the senior resident she…
UNFORGETTABLE: PATIENTS’ PERSPECTIVES ON THE UNINTENDED (AND THE UNEXPECTED) IMPACTS OF PHYSICIANS’ SPEECH
Shakespeare famously sought immortality through his words (“When in eternal lines to…