Subscribe to eTOC

Analysis of Medicare Data Reveals Racial Disparities in Stroke Outcomes for Black Patients

Article In Brief

Black patients have lower mortality rates after stroke but tend to experience greater disability. Health care researchers and neurologists discuss what can be done to address the disparities in outcomes.

Black stroke patients have significantly lower mortality than white stroke patients but a higher level of short-term disability, according to a new analysis of Medicare data. The finding suggests that black patients and their families may have different care preferences during hospitalization, opting for life-sustaining treatments such as feeding tubes and intubation, while white patients may be less likely to choose such procedures and more likely to opt for hospice care.

Prior research has shown disparities by race in both stroke incidence and outcomes. While racial differences in pre-stroke disability seem to be small or nonexistent, post-stroke racial differences in outcome are notable, the authors of the new study published in the September 25 online issue of Neurology wrote.

“The reasons for these differences and why they arise is unknown,” they said.

The study authors suggest, however, that the divergence in outcomes can be detected at the first rehabilitative assessment and that differences in care preferences emerge during hospitalization that, in turn, affect mortality and functional outcome for those patients who survive.

The study, which was based on administrative Medicare data, was unable to determine whether cultural, societal or religious differences drive such differences or whether doctors present treatment and prognostic information differently to black patients and families than white ones, thus influencing the care decisions they make.

James F. Burke, MD, the study's lead author, said there is unlikely to be “one simply theory that is going to explain the disparities,” but he said the fact that they exist “should give us pause.”

“The study is an interesting starting point on the way to ‘why’ [there is] this divergence between mortality and disability and what might be driving those things,” said Dr. Burke, associate professor of neurology at the University of Michigan.

An accompanying editorial by Roland Faigle, MD, PhD, and Lisa A. Cooper, MD, MPH, suggested some concrete strategies to address the disparities, including offering training to doctors and the broader stroke community that focuses on communication skills and implicit bias.

“Racial disparities in stroke care cannot be solved by individual providers in its entirety, since the problem at hand is interwoven with structural and system-level inequities,” Dr. Faigle, assistant professor in the cerebrovascular division at Johns Hopkins University, told Neurology Today. “However, individual providers can employ strategies to mitigate disparities by leaving assumptions at the door, employing structure to patient encounters and care discussions, and communicating effectively with patients and surrogate decision-makers in order to elicit patient preferences.”

Dr. Cooper, the James F. Fries professor of medicine and director of the Johns Hopkins Center for Health Equity, said in an email to Neurology Today, that the 2018 National Healthcare Quality and Disparities Report, which used data from a wide variety of national surveys and databases, academic institutions, and professional organizations, showed that although 70 percent of person-centered care measures have improved over time, disparities between whites and all ethnic minority groups (African-Americans, Asian Americans, American Indians/Alaska Natives) except Hispanics, either persist or have worsened for most person-centered care measures.

Dr. Cooper was awarded a MacArthur fellowship in 2007 for her research and work in addressing racial, cultural, and ethnic differences between physicians and their patients. She said person-centered care measures include physicians explaining things in a way that the patient understands, listening carefully to the patient, showing respect for what the patient has to say, and asking the patient to help decide when there was a choice between treatments.


“The study is an interesting starting point on the way to ‘why’ [there is] this divergence between mortality and disability and what might be driving those things.”—DR. JAMES F. BURKE

Dr. Cooper said that some patient-focused strategies that could help reduce health disparities include the use of health coaches, community health workers, or other patient advocates to allow patients to be more actively involved in their care. In addition, she said, more time could be allocated for visits of patients with greater social needs. Training for health care professionals on communication skills, cultural competency, and bias and stereotyping reduction would be beneficial, she said, as would greater inclusion and diversity among health care professionals.

“My colleagues and I at the Johns Hopkins Center for Health Equity, and several colleagues across the country, have tested these approaches in clinical trials and found many to be effective,” she said. “We are now working in partnership with patient advocacy groups, health care systems, technology companies, and community-based organizations, to better understand the resources and organizational changes that will make it possible to implement and spread these evidence-based approaches in real-world settings.”

Study Design

The new study, which was retrospective, included 390,251 Medicare fee-for-service beneficiaries hospitalized with primary ischemic stroke or intracerebral hemorrhage diagnosed from 2011 to 2014. Of the total cohort, 87 percent were white and 13 percent were black. Racial differences in post-stroke disability were measured in the initial post-acute care setting (inpatient rehabilitation facility, skilled nursing facility or home health care) using the Pseudo-Functional Independence Measure. The 18-item measure—each item is scored on a scale of 1 to 7—provides an assessment of functioning in activities of daily living and motor functions, with lower scores representing greater disability and increased dependence. The assessment focuses on self-care, mobility, bowel/bladder management, and cognition.

The study also factored in racial differences in the type of setting patients were sent after hospitalization, since those assignments may not have been random. Race was identified using Medicare data, which is drawn from Social Security data. Patients were excluded from the analysis if their reported race/ethnicity was other than white or black.

The study found overall that black patients were 30 percent less likely than white patients to die within 30 days of stroke—18.4 percent versus 12.6 percent—and a 3 percent point difference in mortality persisted at one year. But “this mortality benefit is offset by greater post-stroke disability” the study found.

“At the initial assessment, black patients with stroke had greater disability than white patients with stroke across all three post-acute care settings,” the researchers reported. The differences were most pronounced in skilled nursing facilities—where blacks patients scored 1.8 points worse than whites on the disability scale, an 11 percent difference.

The researchers said that there are likely many factors that contribute to racial differences in stroke mortality and disability outcomes. The Medicare administrative data they used for the analysis did not provide, for instance, a window into the patient's socioeconomic status, where they received care (rural, urban, stroke center), or family structure.

“One possible unifying theory that may explain the diverge between mortality and disability are patient preferences for life-sustaining treatment,” they said. “Black patients generally have more aggressive preferences for care in the face of survival with severe disability, and receive more life-sustaining treatment after stroke, a finding replicated in our study.”

For instance, 5.9 percent of white stroke patients had a gastronomy tube insertion compared to 9.7 percent of blacks. Intubation occurred in 2.2 percent of whites comparted to 3.2 percent of blacks. At the same time, more whites than blacks got tissue plasminogen activator: 6.1 percent versus 4.8 percent, which could have increased the chances of lower disability in survival.

The researchers noted that their findings that black stroke patients had longer hospitalization and were less likely to be discharged to hospice compared with white patients support the hypothesis that care preferences are the driving factor. But it could also be that doctors communicate differently with patients of different races or make assumptions about what patients and their families want or want to hear.


“Racial disparities in stroke care cannot be solved by individual providers in its entirety, since the problem at hand is interwoven with structural and system-level inequities.”—DR. ROLAND FAIGLE


“We are now working in partnership with patient advocacy groups, health care systems, technology companies, and community-based organizations, to better understand the resources and organizational changes that will make it possible to implement and spread these evidence-based approaches in real-world settings.”—DR. LISA A. COOPER


“In medicine and especially in end-of-life situations, often it isnt obvious what the ‘right’ answer is.”—DR. DAWN O. KLEINDORFER

Dr. Burke noted that doctor-patient discussions around stroke prognosis and treatment decisions—whether to escalate care or to hold back—are complex.

Most patients do not have advanced directives or other clearly stated preferences pertaining to goals of care after stroke, Dr. Faigle wrote in the accompanying editorial, so doctors often rely on families to communicate what their loved ones would want.

Dr. Burke said it is for a doctor to conclusively answer what every patient would like to know: “What will my life be like?” At the same time, “To stop doing something really requires the ultimate trust. You have to believe the doctor who is saying ‘the situation here is really dire.’”

Implicit bias and stereotyping can creep into patient-doctor discussions, Drs. Faigle and Cohen wrote in their editorial, perhaps leading to assumptions about whether an aggressive treatment should or should not be undertaken.

“Despite efforts to increase diversity among health care professionals, most black patients or surrogate decision-makers will find themselves in a race-discordant encounter with their providers,” the editorial said.

In addition to individual doctor training on communication strategies and implicit bias, a system-wide use of more structured approaches to discussing care options could help reduce racial disparities, the editorial suggested.

Expert Commentary

Dawn O. Kleindorfer, MD, FAAN, professor of neurology and rehabilitation medicine at the University of Cincinnati College of Medicine, said the new study was useful in calling greater attention to the issues of racial disparities in health care, although the data could not explain why that is the case.

She said her own experience in stroke care, which mirrors the literature, is that black patients and families withdraw care less often than whites, but she said it's likely oversimplifying to say that race onto itself is the driving factor. She said the differences could be due to “less trust of the system, cultural beliefs, and religious beliefs.” Family support can also be an issue, she said.

“In medicine and especially in end-of-life situations, often it isn't obvious what the ‘right’ answer is,” Dr. Kleindorfer said, but she said doctors may unknowingly present the same scenario to patients from different backgrounds differently based on their own implicit bias.

Michael Dobbs, MD, professor and chair of neurology at the University of Texas Health Rio Grande Valley, said he participated in an implicit bias course at the University of Kentcuky and now makes a conscious effort to make sure “I am not projecting my wishes and biases on families and influencing them in ways that I may not have realized before.” Dr. Dobbs said, for instance, that when he takes a call on the stroke service, he doesn't want to know the race or ethnicity of the patient because that could introduce bias.

“Our biases can be deep-seeded and subtle and insidious,” he said.

Dr. Dobbs said the culture of an institution—whether it's an aggressive stroke service that does everything possible—can also introduce bias into decision-making and ultimately influence outcomes.

Dr. Dobbs said it takes time and focus to develop a family's trust. “Sometimes in stroke a patient's brain is so damaged they are unlikely to be able to have a conversation, and they must rely on family or friends to communicate their preferences,” he said. In those cases, to avoid bias, Dr. Dobbs said he strives to learn as much as he can about the patient and family and their values and to present information in a specific and neutral way.

He said that having a provider, or at least a translator, who can communicate in the family's spoken language is also important.

Virginia Howard, PhD, professor of epidemiology at the School of Public Health at the University of Alabama at Birmingham, said the new study was well done and raised important questions that need to be further explored. She said it is critical to look beyond the usual categories—white, black, man, woman—when trying to explain disparities in mortality and outcomes in stroke. A better understanding of the factors that precede stroke, for example, such as a person's social networks and family support, could be revealing, she pointed out, though except for marital status that information is not found in most databases.

“Having a large family or church group to lean on could make patients or their family or friends say “Yes, we should do this. We are going to help you get through this,” said Dr. Howard, a researcher on stroke who is one of the lead epidemiologists for the REGARDS study, which examines reasons for geographic and racial differences in stroke.


Drs. Burke, Faigle, and Howard had no disclosures. Dr. Dobbs has received royalties from Elsevier for a neurotoxicology textbook.

Link Up for More Information

• Burke JF, Feng C, Skolarus LE. Divergent poststroke outcomes for black patients. Lower mortality, but greater disability Neurology 2019; Epub 2019 Sep 25.
    • Faigle R, Cooper LA. Editorial-Explaining and addressing racial disparities in stroke care and outcomes: A puzzle to solve now Neurology 2019; Epub 2019 Sep 25.