How to Start Redressing Racial Bias and Reducing Health Care Inequities

In recent weeks there have been many cries for the health care system to finally address racial inequities. Now is the moment to harness that energy toward a process of substantive change.

Value-Based Health Care is not achievable without addressing racial inequities that drive costs and poor outcomes: patient disengagement, higher risk factors, greater admissions, and emergency room usage. Fully acknowledging the issues is the first step. Creating better methods to evaluate how race affects decisions and to innovate change is the next.

Racial inequities in health care in the U.S. are well documented, including:

• Higher maternal mortality, with Black women three to four times more likely to die in childbirth or from its complications than their white counterparts;
• Evidence of dismissal of pain symptoms for Black women (this bias runs across all races and ethnicities) and men; likely misdiagnoses of women with cardiovascular disease;
• Higher incidence and more deaths from cancers of almost all types;
• Across many conditions, diseases caught later, with greater progression and associated risks, and higher mortality;
• Much shorter life expectancy in the Black population;
• Disproportionate number of deaths from COVID-19 .

Racial Inequities Result from Racial Bias

Health outcomes and mortality across all medical conditions are poorer for people of color. Socio-economic disparities—lack of access to quality health care, lack of adequate health insurance, lack of resources to pay for health care—account for some of these disparities. A disproportionate percentage of high risk factors such as obesity and hypertension within this population also leads to poorer outcomes.

But addressing financial coverage and access to care won’t correct disparities. These inequities are driven by racial (and gender) bias in the provision of care itself. That’s the hard part for providers—and health system administrators—to address. Physicians go into medicine to help people and earnestly believe they are living up to the Hippocratic oath. But racism is invisible and misunderstood. Unconscious and unintended racial attitudes bias risk assessments, diagnoses, treatment algorithms, and continuing care for patients of color.

Racial bias in health care has also created a false narrative about race and health care, especially to the detriment of Black Americans. That narrative includes using race as a proxy for genetics and biological differences that in reality do not exist. For example:

• Allegations that hypertension in Black Americans is a completely different disease, referencing more severe clinical status and a widely-touted (but debunked) “genetic salt-sensitivity,” even involving ocean crossings into slavery, among other contributing theories. The fact is that many humans, regardless of race, have a higher sensitivity to salt. If Black Americans present later with disease due to other barriers to care, the health ramifications will be greater.
• Racial bias in diagnoses of mental illness. Black Americans are disproportionately diagnosed with schizophrenia and are overrepresented in state psychiatric hospitals. Yet such diagnoses represents interpretations of symptoms and not, necessarily, real schizophrenia. A number of studies indicate significant racial bias in diagnosis, with Black people three to four times more likely to be diagnosed with psychotic disorders than whites. According to one recent Rutgers study, clinicians performing assessment of symptoms for severe depression were more likely to diagnose Black patients with schizophrenia while attributing white patients’ symptoms to severe depression, by overlooking mood symptoms.
• Characterizations that Black patients are non-compliant, unreliable, late, neglectful of their health, and fail to control their own risk factors, by physician practices, health administrators, and even politicians arguing for higher consumer share of health care costs. For example, in a study of provider decisions to recommend total knee replacement (TKR) as a cost-effective treatment of severe osteoarthritis, providers were more likely to assume that white patients would be more medically cooperative than African American patients. Although those implicit biases were not predictive of treatment recommendations in the study, the researchers concluded that those attitudes “may have influenced treatment decisions.”
• Inclusion of race in specialty treatment risk assessments and algorithms that problematically use race as a proxy factor for genetic risk, a practice called out by a recent New England Journal of Medicine article. Despite a wealth of data proving that race does not confer distinct genetics, these methods result in restricting Black patients’ eligibility for certain courses of treatment, and raise risk assessments.

How Do We “Fix” Racial Bias?

Because we perceive our feelings as reality, informed by experiences and the influence of others, unconscious bias is almost inevitable. Whether that bias is racial, cultural, or gender-based, providers are not immune. Biases enter into the exchange with the patient in ways that determine belief about the patient’s reliability, symptoms, and behaviors, and influence diagnosis, treatment options, and ongoing care.

But a critical, punitive approach to addressing racial attitudes will only put clinicians on the defensive. Such attitudes are often unconscious and unintended, and are not exclusive to clinicians. Solutions must be directed throughout the health care organization to help providers and administrators alike. Every industry should be doing work to promote self- and structured education, conversations, and activities to bring racism to an end.

Health care, however, cannot reach accountable care goals without measuring its performance. Quality and cost measures warrant expansion to incorporate equity of care criteria. Since both race and gender have substantial evidence of disparities, key decision points should be illuminated through data that focuses on medical decision processes to reveal those disparities, such as:

• Constellation of symptoms and the resulting diagnosis for patients by gender and race;
• Diagnosis and prescribed/fulfilled treatments by gender and race, including pharmaceutical, surgical, and other types of interventions;
• For high risk individuals, volume, cost, and variety of care inputs by gender and race;
• For high risk individuals, intermediate and long term outcomes in relation to care provided.

Episodes of care may be a helpful model that can help to provide a method to fairly assess these issues. In a future article, we’ll examine how we might use episodes to create the building blocks to evaluate care to patients in populations, not just as individuals.

Founded in 2002, Roji Health Intelligence guides health care systems, providers and patients on the path to better health through Solutions that help providers improve their value and succeed in Risk.