Confronting Racism in Health Care: Proclamations to New Practices

Leaders at Penn Medicine, meanwhile, are encouraging staff members who have witnessed or experienced racism in the workplace to recognize and report it. They are piloting a digital platform, Lift Every Voice, that will give staff at two emergency departments — including physicians, nurses, technicians, environmental service workers, and unit clerks — an easy way to make anonymous reports when, for example, managers treat Black employees unfairly or Black patients receive disparate treatment from staff. The goal is to create a mechanism for employees to have their concerns heard without filing a formal complaint with human resources.

“We know that experiences of racism happen all the time. For some people, it’s daily and for others it is not as frequently as that, yet leaders often have a blind spot with regards to what is happening on the ground in the workplaces they oversee,” says Eugenia South, M.D., M.S.H.P., assistant professor and vice chair for inclusion, diversity, and equity, emergency medicine at Penn Medicine and one of the developers of the platform. 

By encouraging anonymous reports and asking respondents not to identify the people involved in an incident, the team developing the prototype hopes to call attention to the behavior itself and identify patterns that can be discussed among employees during huddles and at monthly meetings of a department’s senior leadership team. “We want to see if sharing these stories publicly can help people identify things that they might not understand or realize are forms of racism and act upon them,” says Lauren Hahn, M.B.A., innovation manager at the Penn Medicine Center for Digital Health.

A two-week pilot of the tool involving 15 people in one emergency department generated nine reports, including examples of disparaging remarks made about a Black patient and an instance of a manager reprimanding a Black employee but not a white one who did the same thing. If larger pilots of the tool are successful, the plan is to use it in all clinical departments at Penn Medicine and eventually offer it to other health systems and employers. “We hope to build a tool that will be relevant and useful to any workplace,” South says.

Lift Every Voice: Prototype

Penn Medicine has partnered with local artists of color to illustrate some of the reported incidents of racism. The illustrations will be posted on an internal Instagram-like platform through which users can scroll and leave comments. The goal is to spark dialogue about how best to respond to incidents like these. The artists have taken some creative license with the illustrations to preserve anonymity of the people reporting the incidents and to capture several themes at once.

This illustration, by Philadelphia artist Jude Marks, conveys the judgment of staff who couldn’t reach the mother of an emergency department patient in the middle of the night. Rather than assume she was sleeping, a member of the care team suggested she was just trying to get rid of her son.

At UCLA, leaders created a real-time reporting system for students, staff, and clinicians to report incidents of racist and discriminatory behavior. As part of training on equity, diversity, and inclusion principles for all staff, faculty, and students, UCLA Health trainers are teaching the appropriate steps to take when such behaviors occur. The Ohio State University Wexner Medical Center is developing a similar approach, called “Stop the Line,” to encourage patients, community members, students, and clinicians to speak up when they witness or experience racism. Massachusetts General Hospital recently created a patient code of conduct, which clarifies there is zero tolerance for discriminatory behavior toward staff and allows for termination of patients who repeatedly break this code. In addition to helping people feel heard, leaders say these tools will enable them to track what’s being reported, identify patterns, and respond with broader interventions to change the culture of their workplaces.

Listening to and Learning from Patients

Patient-reported outcome measures (PROMs) are another tool for identifying variation in health outcomes by race and ethnicity. PROMs are standardized questionnaires that measure patients’ perception of their symptoms, functioning, and quality of life that have been validated in studies of patients with particular conditions. Using PROMs makes it easier for clinicians to recognize when treatment decisions reflect bias, says Rachel Sisodia, M.D., senior medical director for specialty care and patient-reported outcomes at Massachusetts General Brigham. “It’s been proven over and over that Black women who have chest pain radiating down their arm are less likely to get referred for a cardiac catheterization than white men. When you have a validated tool where all patients answer the same questions about their symptoms, you impede a clinician’s ability to inadvertently hurt someone with implicit bias.“

PROMs can also uncover problems that might otherwise go unnoticed. When Mass General Brigham researchers began looking at outcome measures collected at a subset of clinics, they discovered Black patients in waiting rooms were given a tablet for reporting PROMs only 58 percent of the time while white patients received one 74 percent of the time. “The reason I know that this was due to implicit bias and not patient refusal is that it took one week to fix the problem once the front-desk staff were made aware of the discrepancy,” Sisodia says. “That’s how implicit bias works; we can hurt people without being aware we’re doing it.”

PROMs can also correct erroneous assumptions about the root causes of health disparities. Sisodia and colleagues found Black women receiving care at Mass General Brigham were significantly less likely to get knee replacement surgery, even for significant knee arthritis, than white men. This finding held true irrespective of patients’ health insurance status, socioeconomic status, age, or comorbidity. In addition to raising concerns about disparate treatment in medicine (e.g., that Black women may be less likely to be referred to specialists), the PROMs revealed that Black women had much less social support. Nearly half of the Black women surveyed said they “never” or “rarely” had someone to help them if they were bedbound, compared to less than 25 percent of white women. “Asking patients these questions allows you to get under the hood and understand inequities in ways we couldn’t before,” Sisodia says.

But there are challenges to using PROMs more broadly to detect disparate treatment. Most of these measures were developed among English-speaking populations, and most focus on the outcomes of surgical procedures, which offer clear before-and-after comparisons, rather than chronic conditions. “Many patients who have endured structural racism are not going to get their knee replaced but they are going to have a chronic kidney disease or diabetes,” Sisodia says. “We need more PROMs for diabetes, rheumatoid arthritis, heart failure, asthma, and chronic kidney disease, among others.” Sisodia and her colleagues’ efforts to build PROMs for non-English-speaking patients have been thwarted by the unwillingness of electronic health record vendors to build portals in other languages, she says.

Examining the Use of Race in Clinical Diagnoses

Many health system leaders are taking a critical look at how patients’ race or ethnicity are used in clinical algorithms — often prompted by medical students and residents who are surprised when they hear about these tools. Clyde Yancy, M.D., chief of cardiology and vice dean for diversity and inclusion at Northwestern University’s Feinberg School of Medicine, notes a measure of pulmonary function still in use today dates to the 19th century, when it was used by eugenicists and some physicians to argue that enslaved Africans “were biologically adapted for work and biologically flawed for cognitive pursuits.” The measure — which adjusts results of spirometer tests based on the assumption that Black people have naturally lower lung function than whites — could cause serious problems to be missed.

Not only do these race-based formulas propagate the false view of race as a biological, rather than a social construct, they can perpetuate racial disparities in access to care. For example, two formulas used to estimate how well someone’s kidneys are functioning report higher values (i.e., better functioning) for Black than white patients who have the same creatinine measurements. This means that Black patients whose kidneys are deteriorating may get referred later than white patients for specialist care or transplants, potentially affecting their quality of life and health outcomes. One study of nearly 10,000 Black patients found removing the race “correction” from the estimated glomerular filtration rate substantially changed diagnoses: the prevalence of chronic kidney disease (CKD) increased from 14.9 percent to 18.4 percent, and 29.1 percent of those with existing CKD were reclassified to more severe stages of the disease.

Several health systems have recently stopped using the race “correction” in kidney assessments; one is Massachusetts General Brigham, where leaders are also reviewing other instances of “racialized medicine.”

In 2020, the House Ways and Means Committee put out a request for information about the use of race in clinical decision-making. The 19 responses they received from medical specialty societies and individuals illustrate broad consensus that race is often misused in diagnoses and risk assessments. But they also reveal some disagreement about how to do better, including concerns about the potential unintended consequences of too precipitously removing race “correctors.”

“When you asked health equity experts about this issue, universally they said we need to stop race correction,” says Michelle Morse, M.D., then a Robert Wood Johnson Health Policy Fellow working with Committee Chairman Richard Neal and now chief medical officer and deputy commissioner at the Center for Health Equity and Community Wellness at the New York City Department of Health. “And then when you ask more mainstream professional societies, they tended to think, well, you know, removing these race corrections might do harm and more research is needed. But my feeling is that we know race is a social and political, not biological, category. It’s time to move on. We can create better tools that don’t normalize racial inequities.”

The Association of Black Cardiologists called for smarter approaches, such as broadening the pool of patients included in clinical algorithms and doing more to take into account ways that patients’ social and clinical circumstances affect their health.

As more clinical decision-making leverages artificial intelligence (AI), critics worry that racist assumptions are getting baked into AI tools, too. A 2019 study found that one algorithm used by health plans and systems to comb through population health data to identify patients who may need extra supports, like nurse care managers or health coaches, relies on medical costs as a proxy for health needs. The problem: on average, Black and other patients of color with chronic conditions use fewer medical services than white patients with similar conditions — not because they aren’t as sick but because they lack health insurance, don’t trust the system, or face other barriers. Use of the algorithm could therefore worsen health disparities between Black and white patients. In a commentary, one of the author’s studies suggests health systems should designate staff members to be stewards of how clinical algorithms and AI tools are used, regularly auditing their impacts to ensure accuracy and fairness.

Building Racially Diverse Leadership and Boards

In recent decades, many health systems hired chief diversity officers to help recruit and promote more clinicians and leaders of color. But there’s clearly work to be done. In 2019, when Black and other racial minority groups made up about 39 percent of U.S. residents, they comprised only 19 percent of hospital boards and 16 percent of executive leadership roles.

Last year, leaders at University of Chicago Medicine changed their leadership recruitment approach: instead of just asking recruitment firms to identify minority candidates, they put out a request for information for search firms that had succeeded in placing diverse candidates. They eventually hired two new firms and, in fiscal year 2020–21, increased the percentage of senior leaders of color from 22 percent to 23 percent. By fiscal year 2025, the health system expects that 35 percent of senior leaders will be Black, Indigenous, or other people of color.

After recruiting a diverse pool of applicants, the health system took steps to reduce unconscious bias. Hiring decisions are now made using a blind process in which applicants’ names and other details not related to their experiences and qualifications are removed. University of Chicago Medicine also introduced unconscious bias training for managers who conduct job interviews. “One of the key problems in the process of promoting or hiring people of color is bias in the interview, from inconsistency in questions and using bias to make decisions to judging people based upon who’s more like me,” says Brenda Battle, M.B.A., R.N., senior vice president of community health transformation and chief diversity, inclusion, and equity officer at University of Chicago Medicine. To encourage promotion of internal candidates, the health system tied financial incentives for vice presidents to their success in achieving diversity recruitment, development, and promotion metrics.

Some health systems have created systemwide equity councils or workgroups in an effort to make antiracism efforts part of their clinical, educational, and research enterprises. Last year, Penn Medicine increased the number of vice chairs for inclusion, diversity, and equity from four to 18, representing all clinical departments. “The goal of deputizing these individuals is to help ensure that antiracism becomes a part of culture and that there’s a point person responsible for tracking metrics and being engaged,” says Raina Merchant, M.D., M.S.H.P., professor of emergency medicine at Penn Medicine’s Perelman School of Medicine. Each vice chair is determining the steps their departments will take to achieve the health system and medical school’s joint Action for Cultural Transformation. For example, the emergency department will be tracking results of a new program in which residents and fellows from racial or ethnic groups that are underrepresented in medicine receive mentoring from faculty peers.