Studying Racial and Ethnic Disparities in Leukemia: Q&A with Richard Aplenc, MD

Sep 18 2018

Studying Racial and Ethnic Disparities in Leukemia: Q&A with Richard Aplenc, MD

Acute myeloid leukemia (AML), the second most common leukemia in children, affects different populations of pediatric patients in different ways. With the support of a new Epidemiology Grant from Alex’s Lemonade Stand Foundation, Richard Aplenc, MD, PhD, assistant vice president and chief clinical research officer at Children’s Hospital of Philadelphia, is leading a powerful research project based on the automated extraction of data to learn more about racial and ethnic disparities observed in African American children with AML.

“We’ve known for a long time that African American children do worse with treatment for acute myeloid leukemia than non-African American children,” Dr. Aplenc said. “But we’ve never really understood why that is, and there are a lot of different possibilities.”

In order to truly explore and investigate these possibilities, Dr. Aplenc and his team are taking a new approach to building a pediatric oncology cohort by mining data from geographically diverse pools of electronic health records (EHR) across four leading U.S. children’s hospitals. The resulting data set — which encompasses factors like the neighborhoods patients live in, their laboratory tests, and a plethora of information in between — allows the team to analyze what might underlie disparities in AML outcomes for African American children. Ultimately, their findings could not only identify methods to decrease these disparities, but also advance the way investigators build patients cohorts for research. We sat down with Dr. Aplenc to learn more about this unique and critical project.

This research project actually had its start nearly 10 years ago. Tell us the story behind that.

Almost 10 years ago, we identified that African American children were more likely to have kidney damage during their AML chemotherapy, and we identified that in a data set that’s considered ‘administrative and billing’ data. There can be limitations in administrative and billing data, but it was one of the very first observations we made. So over the last eight years, with help from experts in the Department of Biomedical Health and Informatics here at CHOP, we’ve been able to develop a way to get data out of the electronic medical record in an automated and scalable fashion for patients with relatively rare diseases.

What does this kind of data extraction reveal, and how might it advance what you initially observed from administrative and billing information?

With data from the electronic medical record, we can get out not only the laboratory data that patients have, and what medications they’ve prescribed, but we can also get out their residential addresses, and we can link that address to an incredible array of information about the neighborhood in which they live. So, when we looked in the laboratory data, we saw again that African American children with AML at CHOP were more likely to come in with kidney damage and were more likely to develop kidney damage during therapy. That confirmed our findings in the data set from eight years ago. So, then we thought, there must be really something here because we’re seeing kidney injury in two different data sets using two different data types. We put together the ALSF grant application, where we really want to look into that and look at other disparities in African American kids, relative to their AML treatment and outcome.

What makes this research project unique?

I think what makes the grant unique is that we’re going to implement this automatic data extraction tool at four major hospitals in the U.S. — so us, Seattle Children’s, Children’s Hospital of Atlanta, and Texas Children’s. We’re going to be able to create a data set that represents a geographically diverse sample in the United States from four of the leading pediatric hospitals. For these AML patients, we’ll have every lab value for their renal function that was ever recorded. And every blood pressure that was ever recorded in the electronic medical record. And every medication that was ever ordered and administered — as well as a lot of granular information about their neighborhoods. I think that’s going to help us hopefully begin to understand why African American kids are more likely to come in with renal dysfunction and develop it during the course of their treatment.

We know you’re still at the beginning of this research project, but do you have any theories as to what factors might cause these disparities?

We don’t know at this point. We know that in general, African Americans are more likely to have high blood pressure and renal dysfunction generally, so there may be sort of an underlying increased risk that happens. We are very interested in looking to see whether there are neighborhood-level nutritional and food availability and activity factors that could contribute to this, because you could imagine that if someone has access to food that is predominantly high-salt, high-fat, and they live in an environment where it’s difficult to have physical exercise, their chance for renal dysfunction would be higher. Their kidneys would be less healthy to start with, so that’s one of the unique aspects of the data. We can understand something about the neighborhoods people come from. So those are the two things that we’re thinking of right now.

In terms of timeline, what are the steps you’re going to take with this project?

We’ve built the initial cohort here at CHOP, and we’ll be adding patients to bring it up to date. Our colleagues in Atlanta will hopefully have data ready for us by the fall, and we’re now working on getting things set up in Seattle and in Houston. My hope is that in the next six to nine months, we’ll have data from both of those places, so by the end of this fiscal year, we’ll have a complete data set, and we’ll have some initial analyses done. This time next year, we’ll be focused on doing all of the analyses with the full data.

Beyond kidney damage, have you seen any of these racial/ethnic disparities within AML?

American Journal of Hematology. What [Dr. Winestone] found was that African American children are almost twice as likely to die during their first course of chemotherapy than non-African American children with AML. Approximately 60 percent of that increased risk comes from the fact that they are more ill when they get to the hospital. So, not all of that increased risk, but over half, comes from the fact that they’re just more ill when they get to us. That paper uses administrative and billing data, but [Dr. Winestone] is now writing a follow-up study that looks at laboratory results that come from our new cohort.
Blood, I’ve also published a paper about clinical trials run by the Children’s Oncology Group (COG)where we found that in four consecutive trials, African American patients with AML did worse. And last year at the American Society of Hematology, we presented an abstract that describes how African American children who get AML chemotherapy are more likely to have heart damage than non-African American kids. That manuscript has been accepted at the Journal of Clinical Oncology and will be published shortly.

Would the cohort be able to be used for these other disparities, outside of kidney damage?

Yes, we are working on expanding the cohort to include cardiac complications.

Is there anything else you’d like to add about the new grant?

We are very appreciative of the support from Alex’s Lemonade Stand Foundation. We would not be able to do this work without their invaluable support.

Learn more about AML treatment and research at CHOP