[reposted: Vidya Viswanathan, Matthew Seigerman, Edward Manning, and Jaya Aysola. Examining Provider Bias In Health Care Through Implicit Bias Rounds, Health Affairs Blog, July 17, 2017. http://healthaffairs.org/blog/2017/07/17/examining-provider-bias-in-health-care-through-implicit-bias-rounds/: Copyright ©2017 Health Affairs by Project HOPE – The People-to-People Health Foundation, Inc.]

In 2015, a 27-year-old patient presented to our primary care resident practice in intractable pain, having been recently discharged from the hospital following surgery for a complex shoulder fracture. The orthopedic surgeons evaluated him the day before and scheduled a second surgery but did not adequately treat his pain. The inpatient nurse had told him he would be discharged with the oral pain regimen he had been taking for the past day or so within the hospital. But upon discharge, he found himself without those prescriptions and came to our primary care practice in severe pain. When we reviewed his inpatient record to determine the reason for this discrepancy, the attending physician discovered the phrase “drug-seeking” in the record. The rationale for this statement was not provided, nor the context. When questioned by his new primary care provider about this, the patient was shocked. He tried to recollect what he may have said to result in that assumption. He had no prior history of documented substance or prescription drug abuse.

The patient in question was a young black male and the victim of a drive-by shooting by a stranger. He had been sitting in the passenger seat of a stationary car when it happened. Standard practice in this type of case involves long-acting oral opioid medication, with gradual adjustments of a medication regimen tailored to meet the needs of the patient. But the patient didn’t receive the standard of care, and we naturally wondered why. The answer may be implicit bias.

The literature suggests that he would be more likely to be perceived as drug-seeking when requesting pain relief, compared to his white counterpart. Bias is particularly well-documented in pain management, with black children and adults receiving less adequate pain treatment than their white counterparts in the emergency department for the same presenting condition, even when accounting for insurance status and severity of pain. Longitudinal, national data on 156,729 pain-related emergency department visits found that even among those presenting with the same condition, non-Hispanic white patients were significantly more likely to receive an opioid than all other ethnic minorities examined. Researchers using an instrument to assess implicit bias in more than 2,500 physicians found a significant implicit preference for white Americans relative to black Americans among physicians of all racial/ethnic groups except for black physicians. Another study found that physicians were twice as likely to underestimate pain in black patients compared with all other ethnicities combined and also more likely to overestimate pain in nonblack patients than in black patients.

To address the case of our patient who was inadequately treated for pain based on apparently false assumptions—and other patients who have experienced a different standard of care due to implicit bias—we believe there needs to be formal discussion of this source of clinical errors at institutions. We propose the initiation of a new kind of case conference—“Implicit Bias Rounds”—to specifically identify and discuss these cases.

How Does Bias Occur?

We conceptualized Implicit Bias Rounds based on theories on why disparities in care occur despite well-intentioned providers and despite the recognition of the importance of cognitive error as a source of diagnostic error. Providers, when faced with the need to make complicated judgments quickly and with insufficient and imperfect information, may rely on assumptions associated with a patient’s social categories to fill in the gaps with information that may be relevant to diagnosis and treatment. Physicians are at risk for relying on stereotypes or assumptions for efficient decision making, even when attempting to be objective. In addition to the assumptions providers may make about patients that are dissimilar to them, they may also unconsciously favor patients whose identity they relate to. Such affinity bias may cause a provider not to consider the possibility of a drug problem in an adolescent that appears similar to him, despite a positive urine screen for marijuana. Current efforts in medicine to combat bias may also serve to perpetuate them: Physician-anthropologist Arthur Kleinman states that one problem with traditional cultural competency training is that it may erroneously characterize culture as static and cultural understanding as a technical skill.

It is not enough to merely consider potential sources of provider bias without considering proposed strategies to mitigate that bias. Evidence tells us that simply adjusting the explicit medical curricula is not enough to change implicit bias; increasing positive role modeling for medical trainees is more effective. Strategies proposed to combat implicit bias include consciously thinking of the patient’s perspective and approaching each provider-patient interaction as a shared negotiation between worldviews. Focusing on specific and unique details about an individual, instead of his or her social category, serves to combat biases by diminishing stereotyping and promoting empathy building. Clinicians who are trained to consider the unique perspectives and experiences of their patients are more likely to show empathy toward them, the study suggests. Priming physicians with information about the relevance or irrelevance of sociocultural factors in medical care can combat cognitive errors that stem from stereotyping. A regular intervention such as Implicit Bias Rounds would serve to implement these strategies on a consistent basis.

Evaluating Bias Within Health Care Institutions

In medicine, when patients are given inadequate or erroneous treatment, hospitals often hold morbidity and mortality (M&M) conferences after the fact, in which physicians identify their mistakes and take colleagues through the causes of the medical error. The classic format involves identification of a medical error to generate discussion on lessons learned, elucidate systemic errors and areas for quality improvement, and reinforce accountability in care. Such conferences typically focus on quality and safety, with no specific lens on provider bias as a potential cause of error.

At present, medical training environments lack a formal educational forum for providers to examine their role in the complex issues of unequal care and unconscious bias. Implicit Bias Rounds would fill this gap, focusing on discussing cases in which bias may have altered the care of a patient. Such conferences could address unequal treatment documented on the basis of any patient characteristic, such as race/ethnicity, gender identity, sexual orientation, disability status, or mental health status. Depending on institutional resources and time constraints, Implicit Bias Rounds could either comprise a subset of existing M&M conferences and ethics rounds, or exist as a stand-alone educational initiative. Just as with discussions of medical error and breaches of patient safety, a conversation about implicit bias might be uncomfortable for the providers involved, and the causes of error may sometimes be ambiguous. Implicit Bias Rounds could offer a safe forum for dialogue on a sensitive subject and provide an important teaching moment when unequal treatment is perpetuated under our care.

In our patient’s case, we envision that Implicit Bias Rounds for his case could be conducted in the format of an M&M conference, following steps such as these:

  1. Walk through the case to determine what factors led to inadequate pain treatment, including the factors that potentially led to the erroneous documentation of the patient as “drug-seeking.” Discuss how this label was carried forward in the chart and how it influenced treatment decisions.
  2. Discuss the events that led staff to correct this error and what could be done in the future to correct the error sooner or prevent it from occurring.
  3. Perform a thought exercise, varying the reason for the injury and personal characteristics of patient and provider, and discuss how stereotypes can influence our assumptions and actions as providers.
  4. Discuss the literature on racial/ethnic disparities, focusing on post-surgical pain management of orthopedic injuries. Review the standard-of-care pain management.
  5. Determine how this error should be discussed with the patient.
  6. Discuss follow-up information: How is the patient doing now, since the error was corrected?

In the case of our patient, the error was noticed by the primary care resident who took on his case and was brought to the attention of an attending physician. The patient was then put on an effective pain management regimen consisting of opiates for his peri-operative pain. As of last follow-up with his primary care physician, he has been successfully tapering his opiate dose and will soon be off this class of pain medication.

Given the national discussion of systemic racial discrimination, our call for implementing strategies that attempt to mitigate bias in medicine is timely. When we read about racial inequality in pain management, it may be easy to assume that it only happens at other institutions or with other providers. But as physicians dedicated to delivering equitable care, we have to recognize the possibility that we may act with bias unconsciously and in contrast to our personal values. Recognition of our own implicit biases can start with reflecting upon and discussing a particular case at our own institution.