Do implicit bias trainings on race improve health care? Not yet – but incorporating the latest science can help hospitals treat all patients equitably

Nao Hagiwara, Professor of Public Health Sciences, Director of the Program on Health Disparities and Community Engagement Research, University of Virginia

Tiffany Green, Associate Professor of Population Health Sciences and Obstetrics and Gynecology, University of Wisconsin-Madison

There is increasing evidence that implicit bias – non-conscious attitudes toward specific groups – is a source of racial inequities in certain aspects of health care, and lawmakers are taking note.

Since the tragic murder of George Floyd in May 2020, wherein a Black man was killed by police, several U.S. federal and state legislators have introduced proposals declaring racism as a public health crisis. In March 2024, four U.S. senators led a resolution calling out the “implicit racial and ethnic biases within the health care system, which have an explicit impact on the quality of care experienced by members of racial and ethnic minority groups.”

Following this reasoning, states like California and Massachusetts have enacted legislation mandating implicit bias training for health care providers. Health institutions have also focused on addressing implicit bias among the next generation of providers. For example, the American Medical Association’s guidelines to address systemic racism in medicine includes requiring training that covers various forms of racial bias.

But is implicit bias training improving care quality for Black patients? We are a social and health psychologist and a health economist who are investigating the role that provider implicit bias plays in racial health care disparities. Our ongoing review of the existing evidence suggests the answer is: not yet.

What is implicit bias?

The first thing to understand is that implicit bias isn’t just one thing. It involves multiple interconnected components that govern how someone interacts with specific groups or its members: affect, behavior and cognition.

Psychologists sometimes refer to those components as the ABCs.

The affective component of bias, also known as prejudice, is defined as having negative feelings towards a group or its members. The behavioral component of bias, or discrimination, involves negative or harmful actions towards a group or its members. Lastly, the cognitive component of bias, also known as stereotyping, refers to expectations and beliefs about a group.

One common misunderstanding is that implicit bias is inherently unconscious and people are unaware of their own negative feelings, beliefs and behaviors. In fact, research suggests that people are remarkably accurate in perceiving their own levels of implicit bias.

Each component of bias can operate at implicit and explicit levels. At the implicit level, the ABCs arise spontaneously and effortlessly, while ABCs operating at the explicit level are intentional and effortful. For example, the unease someone may feel when encountering a large Black man at night is an emotion triggered at the implicit level. Actively making an effort to replace those feelings of unease with neutral or positive feelings are emotions activated at the explicit level.

Why does implicit bias matter in health care?

Black and white people experience stark differences in treatment during medical interactions. A December 2023 survey from the Kaiser Family Foundation found that nearly 1 in 5 Black people reported experiencing unfair or disrespectful treatment from their health care providers in the past three years because of their race. Only 3% of white respondents reported similar treatment. Researchers have seen similar health inequities across race and ethnicity.

Extensive research over the past two decades indicates racial inequities in patient-provider communication stem largely from implicit prejudice among health care providers. This implicit prejudice manifests during medical interactions with Black patients through a wide range of communication behaviors. These include nonverbal behaviors, or how people move their bodies – such as eye contact and hand movements – and paraverbal behaviors, or how people deliver speech – such as their tone and volume. Both of these behaviors typically occur spontaneously.

For example, providers with higher levels of implicit prejudice tend to talk more and spend less time evaluating Black patients. They also display less positive and more negative affect and more frequently use anxiety-related words like “worry,” “afraid” and “nervous.”

Importantly, Black patients are adept at discerning these subtle negative communication behaviors. “It’s petty, little things,” a Black patient told the Kaiser Family Foundation. “When they call the nurse, they rush to come see the white people. They don’t rush to see the Black people. I think it’s racist.” Consequently, Black patients report lower levels of satisfaction after interacting with providers with higher levels of implicit prejudice.

A common misconception is that implicit prejudice is a key driver of racial disparities in medical treatment. However, current research does not support the idea that providers with higher levels of implicit prejudice treat Black patients worse than white patients. Additionally, more research is needed to determine whether implicit stereotyping from providers – such as automatically associating the idea of being “medically uncooperative” with Black people – would also lead to negative communication behaviors or sub-optimal treatment decisions for Black patients.

What’s wrong with implicit bias training?

Many researchers and clinicians see implicit bias training as an essential component of medical education. However, current programs have shortcomings that undermine their effectiveness.

To understand what typical implicit bias training is like, our ongoing systematic review looks at 77 studies on implicit bias training programs in U.S. health care institutions. Although the majority of the programs were designed to address implicit racial bias, a significant number also addressed other forms of bias including gender identity, sexual orientation and socioeconomic status. Most programs aim to educate health care workers and trainees on implicit bias and how it may affect their patient care, as well as increase awareness about their own biases. Most are single sessions that last about 5.5 hours on average.

However, the design of these training programs does not align with current scientific knowledge about implicit bias.

First, while awareness of one’s biases is a necessary first step to mitigating implicit bias, it alone is not sufficient. Providers must also be personally invested in and have the mental capacity to address their biases.

Second, mitigating implicit bias requires repeated and consistent practice. Implicit bias is like a habit: it is deeply ingrained and operates without intentional control, making it challenging to recognize and change.

Third, training effectiveness is more accurately assessed through patient outcomes, such as care satisfaction, rather than self-reflection or implicit bias scores. Because providers may be concerned about how program facilitators will judge them, they may not provide honest feedback. Furthermore, changes in implicit bias scores do not necessarily result in decreased discriminatory behaviors, making it unclear how these programs can change the quality of care that Black patients experience.

How can health care systems better address implicit bias?

Developing and implementing effective implicit bias training in health care is a scientific endeavor that requires a strong supporting structure.

For example, the clinical and translational science, or CTS, framework, originally designed to help translate discoveries in the lab into treatments in the clinic, could also be applied to implicit bias training. This framework guides scientific progress across incremental stages, starting from confirming the mechanism behind an illness to developing and testing a new treatment for use in the broader community.

This framework is particularly relevant to the development and implementation of evidence-based implicit bias training. Researchers first focus on confirming the mechanisms that underlie implicit bias. Then, after developing and testing implicit bias training programs, they examine its effectiveness across institutions and among diverse health care professionals.

Applying a rigorous scientific process to the development of implicit bias training requires an institution’s long-term commitment, robust support and substantial resources. We believe this investment is a small price to pay for the invaluable progress it promises in reshaping health care for the better for everyone.