Is Implicit Bias Contributing to Time Disparities in Goals-of-Care Conversations With Minority Patients?

A Conversation With Cardinale B. Smith, MD, PhD

Get Permission


Jamie H. Von Roenn, MD, FASCO

Jamie H. Von Roenn, MD, FASCO

Addressing the evolving needs of cancer survivors at various stages of their illness and care, Palliative Care in Oncology is guest edited by Jamie H. Von Roenn, MD, FASCO. Dr. Von Roenn is ASCO’s Vice President of Education, Science, and Professional Development.

It has been well documented that minority patients with cancer have worse outcomes than white patients, even when their cancer is in an early stage. According to the results from a recent study of patients with early-stage lung and breast cancers, African American patients were less likely to complete treatment than white patients (79.8% vs 87.3%, respectively), even after taking into account the patients’ age, comorbidities, health insurance, income, and marital status. Among the reasons cited for worse treatment completion rates among African American patients was poor clinical communication.1

Newer information from a study presented at the 2018 Palliative and Supportive Care in Oncology Symposium by Cardinale B. Smith, MD, PhD, is showing that not only do minority patients have inadequate discussions with their oncologists regarding treatment and prognosis, they also spend fewer minutes with their oncologists discussing the goals of care, all of which can lead to substandard treatment, worse quality of life, and poorer survival than their white counterparts.2

Cardinale B. Smith, MD, PhD

Cardinale B. Smith, MD, PhD

The randomized controlled trial by Dr. Smith and her colleagues, evaluated the amount of time an oncologist spent with minority patients with advanced cancer compared with nonminority patients during a goals-of-care discussion visit. The study recruited 22 oncologists from 4 sites that included community, academic, municipal, and rural hospitals nationwide and 265 of their newly diagnosed patients with advanced solid tumors, including lung, gastrointestinal, castration-resistant prostate, and triple-negative breast cancers. The patients had a survival prognosis of 2 years or less. The study cohort was diverse and included 38% non-Hispanic white, 32% non-Hispanic black, and 19% Hispanic patients.

During the trial, the oncologists were randomly assigned to either an intervention group, which included a 1 hour of grand rounds, a 2-hour roleplay workshop with a simulated patient, and four 40-minute sessions with a communication coach, or usual care, which included 1 hour of grand rounds. The intervention was designed to improve physician/patient communication during a goal-of-care discussion after an imaging visit. Of these visits, 142 were audiotaped, and the median encounter time was recorded. Patients were then asked to fill out a questionnaire assessing the quality of these visits with their oncologists.

The study found that the median face-to-face time oncologists spent with their minority patients during a goals-of-care visit was 12 minutes vs 17 minutes with their nonminority patients. The median encounter times were longer when discussions involved disease progression, 18 minutes vs 13 minutes if there were no disease progression, and among these patients the time spent with both minority and nonminority patients was nearly equivalent, 16 minutes vs 18 minutes, respectively. After controlling for clustering of the patients within the hospitals represented and progression of disease, the time spent with minority patients remained less than with nonminority patients, 15 minutes vs 18 minutes, respectively.

The ASCO Post talked with Dr. Smith, Associate Professor of Medicine and Director of Quality for Cancer Services at Mount Sinai Health System in New York City, about the results of her study and the role implicit bias plays in disparities in oncology care.

Honing Communications Skills

In the intervention group of your randomized clinical trial, oncologists received communication skills training. What did the training entail?

We had a communications facilitator conduct a 2-hour roleplaying session, during which a simulated patient with advanced pancreatic cancer had a discussion with his oncologist about the results from his first imaging scan after initiation of chemotherapy. In one scenario, there was disease progression; in another, there was no disease progression, so the conversations were very different. If the imaging scan showed disease progression, we provided oncologists with various value statements to prompt the conversation with their patient. For example, “Knowing there is disease progression, what are you hoping for as you look to your future; what is most important to you?”


For more on cultural factors that contribute to disparities in cancer care, see an interview with Karen M. Winkfield, MD, PhD, on The ASCO Post Newsreels at

To improve effective and empathetic communication when delivering bad news, we used patient-centered communication techniques, including the SPIKES (Setting up, Perception, Invitation, Knowledge, Emotions With Empathy, and Strategy or Summary) and NURSE (Naming the Problem or Emotion, Understanding Patients’ Feelings, Respect, Support, and Explore) protocols (See the sidebar on “Becoming a More Effective Communicator” on page 113). After the roleplaying session, the oncologists received four 40-minute visits in their practice with a communications coach before meeting with a patient. During that time, the coach and the oncologist would discuss the results from the patient’s scan; if there was disease progression, they would brainstorm how best to apply the skills the oncologist learned during the roleplay session with the patient.

Did the patient’s gender influence how the oncologist communicated with the patient?

Not that we can tell from our initial analysis, but in terms of the length of time oncologists spent discussing goals of care, there was no difference between male and female patients. In this study, we audio recorded about 140 of those visits, and we are now doing a qualitative analysis of the content and plan to publish additional information on the study soon.

Determining the Causes of Time Disparities

In total, oncologists spent 3 minutes longer with nonminority patients than with minority patients. What do you think accounts for the time difference?

My initial hypothesis is implicit bias. Research suggests that implicit bias in physicians is a factor in health-care disparities, especially in primary care. However, I think implicit bias exists in oncology care, too, and that it played a role in the amount of time oncologists spent discussing goals of care with their patients.

When there is racial or ethnic concordance between patients and physicians, patients report greater trust and satisfaction in their oncologist and tend to be more adherent with their treatment regimen.
— Cardinale B. Smith, MD, PhD

Tweet this quote

However, there are other factors to consider. Almost 30% of our patient population were on Medicaid, and I wonder what role health literacy and social support may have played in our results. For example, if there were several people in the room asking questions when these discussions were taking place, did that account for the longer time spent with nonminority patients? We need further analysis of the audio recordings to answer the question.

Achieving Equity in Health Care

Studies show that lack of exposure to an ethnically diverse oncology workforce is a barrier to good care. Would having a more diverse oncology workforce achieve greater health equity and potentially improve outcomes in minority patients with cancer?

Yes, I definitely think it would. When there is racial or ethnic concordance between patients and physicians, patients report greater trust and satisfaction in their oncologist and tend to be more adherent with their treatment regimen. Although blacks and African Americans make up 13% of the U.S. population and Hispanic or Latino Americans make up 18% of the population, just 2% of the oncology workforce self-identifies as black/African American and 3%, as Hispanic/Latino.3,4 Thus, having a more diverse oncology workforce would make a huge difference in the care of minority patients.

Maintaining Proficiency in Patient-Centered Communication

It is incumbent upon oncologists to communicate more effectively with their patients. How can they improve their communication skills?

I hear all the time from oncologists that they are already good at patient communication and don’t need more coaching. And my answer always is, “You are already good at prescribing current chemotherapy regimens, but you have to keep improving that skill to become proficient in prescribing new therapies. It’s the same with patient communication.”

Good patient-centered communication is a skill that has to be constantly maintained and improved upon. Unfortunately, there is no board certification requirement for proficiency in communication, so there is no incentive for oncologists to make it a priority. However, patient-centered communication and shared decision-making are critically important components in the practice of oncology and to the delivery of the best-possible patient care. 

DISCLOSURE: Dr. Smith has received honoraria from Teva, has served as a consultant or advisor for Teva, and has served on a speakers bureau for Teva.


1. Cykert S, Eng E, Manning MA, et al: A multi-faceted intervention aimed at black-white disparities in the treatment of early stage cancer: The ACCURE pragmatic quality improvement trial. J Natl Med Assoc. March 27, 2019 (early release online).

2. Smith CB, Pintova S, Adelson KB, et al: Disparities in length of goals of care conversations between oncologists and patients with advanced cancer. 2018 Palliative and Supportive Care in Oncology Symposium. Abstract 19. Presented November 17, 2018.

3. U.S. Census Bureau: Quick facts. Available at Accessed September 19, 2019.

4. Association of American Medical Colleges: Diversity in the physician workforce: Facts and figures 2014. Available at Accessed September 19, 2019.