You are here
Racial discrimination was a key feature at a 2-day summit on pain management and the opioid crisis, hosted by the National Institutes of Health on Thursday and Friday.
Asheley Cockrell Skinner, PhD, of the Duke Clinical Research Institute in Durham, North Carolina, provided a broad overview of racial bias in opioid prescribing.
"Pain is not treated in minorities, particularly African Americans ... to the same extent as it is for white patients ... Our implicit biases affect our behaviors," she said of prescribers.
In emergency rooms, white patients were almost twice as likely to receive opioids for migraine as black patients were, she said. White patients were also significantly more likely to receive opioids for back pain than black patients were, Cockrell Skinner added, citing a study by Joshua Tamayo-Sarver, MD, PhD, and colleagues in the American Journal of Public Health
However, for long bone fractures, in this study race was not a significant factor, and differences, across all three categories of pain, were "less and nonsignificant" when comparing whites and Latinos, Tamayo-Sarver et al noted.
What researchers learned is that when pain is "poorly defined," black patients are less likely to receive opioids than white patients, Cockrell Skinner added.
She noted that these same biases also extend to younger patients. In a study published in JAMA Pediatrics in 2015, white children were more likely to receive opioid analgesics than black children for treatment of an appendicitis in the emergency room.
In addition, among patients in the Department of Veterans Affairs, white patients under age 65 with moderate to high levels of chronic noncancer pain were more likely than black patients to receive opioids, she said. That study was published in The Journal of Pain in April 2014.
"All of our attempts to improve safe prescribing practices run the risk of increasing a disparity that is already very large," she concluded.
Adam Hirsh, PhD, of the School of Science at Indiana University-Purdue University Indianapolis, shared the preliminary results of his research on bias in prescribing habits.
Using "interactive virtual patients" and physician interviews, Hirsh and his colleagues examined how race influences prescribing patterns for opioids and what can be done about it.
Of the 502 physicians in the study, approximately 126 showed racial and socioeconomic status (SES) bias, 45 showed race-only bias, and 52 showed SES-only bias, Hirsh noted.
In an online environment, providers were introduced to different virtual patients whose socioeconomic background and race were made clear, visually and through a biographical sketch. Providers were then asked how they would care for each patient.
Specifically, providers were asked how much pain they believe patients are in, how likely they are to prescribe opioid therapy, and how likely they are to suggest patients take time off of work.
A scoring system captured providers' responses in real time and created individual "disparity scores." The non-biased providers are asked to leave, and biased providers -- those who displayed systematic differences in prescribing -- continue on. Biased individuals were randomized into a control group and an intervention.
In the intervention group, providers who showed bias against black patients participated in a tailored interaction with a black virtual patient, and those who showed bias against poor patients engaged in a similar activity with a poor virtual patient.
In each virtual encounter, providers were prompted to ask certain questions of their virtual patients. Patients responded by sharing personal stories that explain how their pain affects their life.
In addition, at "key moments" virtual videos began playing -- showing, for example, a patient stuck on the sidelines as his child plays a sport he can't participate in because of pain, Hirsh said.
Providers were given individualized feedback at the end of the intervention. For instance, those who gave worse care to black patients than whites, would be told so.
Finally, all providers were asked to return a week later to re-do the activity.
After the second encounter, providers in the intervention group showed 73% lower odds of being biased against low-SES patients and 85% lower odds of being biased against black patients, Hirsh said.
Providers in the intervention group were also more likely to recommend work changes, to use non-pharmacological therapies, and to show more compassion towards their patients.
Asked about any differences in bias related to a provider's race, ethnic group, gender, or specialty, Hirsh said, "We didn't find much there" -- acknowledging, however, that he and his colleagues had completed only a "cursory" analysis to date.