Research debunks ‘racial concordance’ study cited by Supreme Court Justice Ketanji Brown Jackson
New research from a Harvard University professor refutes a widely cited study that concluded black newborns had better survival rates when cared for by black doctors.
When factoring in infants’ birth weights, Harvard Kennedy School Professor George Borjas (pictured left) and Manhattan Institute fellow Robert VerBruggen (pictured right) found the conclusion of the 2020 study fell apart.
“Robert and I are both interested in the fragility of empirical results in social science,” Borjas told The College Fix in a recent email. “We both found that by doing the statistical analysis and adding a bunch of controls, the result in the original paper disappeared.”
Their research, published in the “Proceedings of the National Academy of Sciences” journal Sept. 16, used the same data as the 2020 study to replicate the findings, but also controlled for low birth weights.
“The estimated racial concordance effect is substantially weakened, and often becomes statistically insignificant, after controlling for the impact of very low birth weights on mortality,” the researchers found.
Racial concordance is the idea that doctors of the same race as their patients provide better care.
The 2020 study was cited as evidence to support the theory. It concluded that “disparities in survival rates across Black and White newborns during childbirth … may be ameliorated by racial concordance between the physician and newborn patient.”
It received widespread media coverage, and attention from other universities and medical associations. But watchdog groups like Do No Harm medicine have warned that the concept is “flawed” and medical schools should stop embracing it.
Borjas, when asked if he expected to find different results than the 2020 study, said he was not surprised.
“I can say that in my case I am not at all surprised by the fact that by looking at the data from a different angle, you get the results to change,” he said.
“However, I never suspected that the reason the result would change was because there was a left out variable which was obviously the kind of variable that should never have been left out,” he told The Fix.
“I did my job and I think our paper is pretty conclusive. People are going to interpret the evidence whichever way they want,” he said.
MORE: Former med school dean debunks push to match patients and doctors by race
Borjas mentioned that he was in contact with Brad Greenwood, the lead author of the original study, who was helpful.
“He helped guide us through some of what they did in the original paper,” Borjas said.
“[H]e responded within a couple of days after we sent him the original draft of our paper with an acknowledgement that we were making an important point and encouraged us to send [it] to PNAS,” he told The Fix.
When contacted about the new study, Greenwood told The Fix he was “glad that there is a continued and robust dialogue in the scholarly community” about the subject of racial concordance.
“It is important work that needs to be continued,” he said in a recent email.
VerBruggen, a policy researcher at the institute, told The Fix over email that he had been interested in the 2020 study since it came out.
“At the time, I was skeptical — but the models had a lot of control variables that did seem to address some of the obvious criticisms, so I wasn’t entirely sure what to think,” he said.
VerBruggen said he started looking into the data again after U.S. Supreme Court Justice Ketanji Brown Jackson mentioned the research in her dissent to Students for Fair Admissions v. Harvard. The 2023 decision overturned race-based college admissions policies.
“That’s when I started looking into getting the data and asked Prof. Borjas if he’d be interested in working on the issue,” VerBruggen said.
He and Borjas wrote an additional report, published Tuesday at the Manhattan Institute, that explains their research in more detail:
We discovered an important repercussion of the way the authors controlled for the infants’ health conditions. The analysis accounts for the 65 health conditions that are most common in the data set, using “diagnosis codes” that are reported for each patient—effectively comparing how the race of the doctor and the race of the newborn interact among newborns who are “equally healthy” in terms of those top 65 “comorbidities.” However, the top 65 comorbidities do not include a single indicator for whether the newborn has a very low birth weight, which is a crucial determinant of infant mortality.[3]
Controlling for very low birth weight—i.e., comparing newborns within the same weight class—eliminates the racial concordance result in the most detailed statistical models.
VerBruggen also said the 2020 study’s authors were “extremely helpful” as they conducted their research. The authors included Greenwood, from George Mason University, Laura Huang from Harvard, and Aaron Sojourner and Rachel Hardeman, both from University of Minnesota.
“They gave us the data they’d collected on doctor race, and answered numerous questions about how they handled technical matters. They also provided feedback on a draft of our paper and encouraged us to publish the findings,” he said.
“It was a good example of how research can move forward and improve when scientists are open to replication and challenge,” he told The Fix.
MORE: Med schools push flawed race-matching, watchdog warns
IMAGE: Harvard University, Manhattan Institute
Please join the conversation about our stories on Facebook, Twitter, Instagram, Reddit, MeWe, Rumble, Gab, Minds and Gettr.