Pitt researchers find racial disparities in pediatric medical imaging

Dr. Jennifer Marin and a team of 13 other researchers analyzed 13 million hospital visits from 44 hospitals across the country. Their results — published in late January — found that non-white children are less likely to get diagnostic imaging compared to white children.

Specifically, Black children were 18% less likely to get medical imaging — such as CT scans, X-rays, MRIs and ultrasounds — than white children, and Hispanic children were 13% less likely.

“Truth be told, when we started, we went into [the study] with a pretty good idea of the conclusion,” Marin said. “This was not novel, that we knew, to some extent, that for certain conditions in certain settings, white children were getting imaged more than non-white children.”

Marin, an associate professor of pediatrics and emergency medicine at UPMC, spearheaded a study to determine whether the use of diagnostic imaging in pediatric emergency departments differed based on a child’s race and ethnicity. She said choosing to investigate this topic was actually accidental.

“We had done a prior study that looked at overall trends in imaging — how frequently different imaging tests were being used,” Marin said. “As more of a secondary analysis, we found that white children appeared to have a higher likelihood of having imaging. We continued to discuss it and decided it was something we needed to dive a little deeper into.”

The group of researchers on this study wasn’t a coincidence. The study’s senior author, Dr. Mark Neuman, an associate professor of pediatrics and emergency medicine at Harvard Medical School, and co-author Dr. Samir Shah, director of the division of hospital medicine at Cincinnati Children’s Hospital, have a research group dedicated to pediatric emergency medicine. This group has published numerous studies on pediatric emergency medicine.

Neuman said this research fascinated him because he’s interested in looking at variation in care and trying to identify the optimal way to care for children in the emergency department. His goal is to minimize the differences that exist in the way health care providers impart care to all children.

“The findings from the study demonstrate differences in the rates of imaging based upon race and ethnicity,” Neuman, a pediatric emergency medicine physician at Boston Children’s Hospital, said. “I am hopeful that this study will lead to other research to explore why these differences exist, and future work to minimize or eliminate these differences.”

While the study found that racial disparities in pediatric medical imaging exist, it doesn’t conclude why they exist or why the imaging was performed. Marin said with data sets, the “granularity” is missing.

“One of the challenges of working with data sets where you’re really just getting computer-generated data points is I can only tell you the age and sex of the patient, when they left the hospital, whether they got a CT scan and what their diagnosis was,” Marin said. “I can’t tell you why that CT was ordered, I can’t tell you whether that CT was appropriate and I can’t tell you whether the patient was sent to the [emergency department] ED by their pediatrician.”

Even after adjusting their analyses for some of the preexisting differences between populations — if, for example, white children have been determined to be more sick when they arrive at the emergency department — Marin said the disparity and differences remained.

Co-author Dr. Alon Peltz’s research focuses on the impacts of health and social policies on health disparities. He said some factors such as structural and implicit biases contribute to health inequities.

“Often health inequities are caused by the cumulative effects of structural factors and implicit bias, which are important issues that our health care system needs to address,” Peltz, an instructor in population medicine for Harvard Medical School and Harvard Pilgrim Health Care Institute, said. “We hope that these results encourage providers and clinical organizations to closely examine equity in how they care for children in the emergency department.”

But the researchers feel hopeful that there are methods providers can take to approach a more equal emergency department experience for all children. Peltz said a “standardized” pathway to determine what imaging is necessary can provide a more objective tool for providers to use.

“I think one potential way to improve equity in care is to use standardized, evidence-based, clinical treatment pathways to help guide providers on when it is appropriate to offer imaging and when it is less appropriate to offer imaging,” Peltz said.

Neuman also said determining the conditions for which imaging is appropriate can ensure that all children receive the appropriate imaging when indicated.

A more hidden problem, Marin said, is something only health care providers can control — their implicit bias.

“As physicians, we’re human beings, and we don’t like to admit it, but we’re all susceptible to implicit bias,” Marin said. “Interestingly, those biases tend to come out in times of stress. And the emergency department, as you can imagine, is generally a very stressful place.”

Marin said other research in the field has shown that physicians and clinicians have a “preference” for white patients, so unless providers are taught to recognize biases, they may just continue letting it happen.

“Believing that there’s a problem and diving deeper into local data is what physicians and departments in hospitals need to do,” Marin said. “It’s hard to change behavior because the data is so large and so distant from that one physician — so that’s where you dive into your own local data.”

Marin added the timing of this study being released was “lucky” because people are more receptive to the message, with the past year shining a greater focus on disparities faced by communities of color.

“Now is the time to promote implicit bias training and awareness. It can start as early in the undergraduate years, certainly in medical school,” Marin said. “You know, it’s not too late for us old dogs, because as I think the word implies, implicit, it’s not something that we are deliberately doing.”

Because of the service nature of the health care profession, Marin said she thinks physicians will be up to the task of actively changing the way they approach this implicit problem.

“Physicians go into medicine because we want to help people,” Marin said. “I think physicians would want to know when these implicit biases are rising to the surface and potentially impacting the care that they’re delivering to their patients.”