Racial segregation is returning to health care, driven by a new generation of woke activists. The seminal study advocating this backward policy was released in June 2018 and revised three years ago this month. Titled “Does Diversity Matter For Health: Experimental Evidence from Oakland” and published in the American Economic Review, it argued that matching black patients with black physicians would save lives. The medical establishment immediately took up the banner of “racial concordance” between physicians and patients.
But the Oakland study is fatally flawed. In a new analysis, we show the shortcuts and un-scientific methods its authors used. No sound evidence supports resegregating health care, and taking this dangerous road will surely lead to worse health outcomes.
It’s important to realize how the Oakland study is reshaping health care, starting with the push for diversity in medical education and training. There aren’t enough minority physicians to make racial concordance a reality—so to get more nonwhite physicians, medical schools are ditching the MCAT for some minority students, while trainee assessments are being changed to minimize the documented differences in performance between white and so-called underrepresented minorities. These policies require lower standards, a direct threat to patient health.
Medical providers are also moving toward racial concordance. A February 2021 leak showed that UnitedHealth Group, the largest health-insurance company in America, supports matching patients with physicians based on their race. Activists are pushing hospitals and policymakers to follow suit. Yet these measures inherently undermine trust—a key part of health care—between patients and physicians with different skin colors.
The Oakland study does not justify this campaign. The authors conclude that racial concordance would lead to better health screening, leading to “a 19 percent reduction in the black-white male cardiovascular mortality gap and an 8 percent decline in the black-white male life expectancy gap.” The authors admit, however, that these are “back of the envelope calculations.” They should have designed a better study.
The first problem is that the Oakland study contains no meaningful control group. The authors recruited more than 1,300 black men from local barbershops and flea markets. The patients were randomly paired with a black or nonblack physician who attempted to persuade them to take preventative checkups. The patients generally agreed to more services with the black physicians, leading to the study’s conclusions.
Yet without a substantial control group of non-black patients, there was no way to test for many possible variables, such as the effectiveness or persuasiveness of these particular physicians unrelated to their race. The researchers belatedly tried to create a quasi-control group of 12 nonblack patients who were accidentally recruited. But this sample almost certainly differs in both observable and unobservable characteristics from the sample group, which means the two aren’t comparable.
The second problem is an abundance of unproven logical leaps. The authors believe that five one-time interventions—such as blood pressure measurement and diabetes testing—will transform individual health. Yet preventative screening is effective only if used in a patient with a high likelihood of being at risk for a particular clinical condition, something the study made no attempt to discern. These procedures also generally need to be combined with lifestyle changes and medication adherence. The researchers accounted for none of this, discounting the study’s astounding findings.
The third problem is that both the patients and physicians were unrepresentative. The patients who came to the checkups were, on average, more likely to be unemployed, less likely to have a high school diploma, and older than the general black male population. As for the physicians, only 14 were used—eight nonblack, six black. The authors use percentages to imply a far larger sample, noting in one case that 67 percent of black physicians in the study specialized in internal medicine, but that simply means four out of six. Such a small and non-random sample of physicians is not generalizable to the medical profession.
Why has such a poorly designed study seen no sustained pushback? Simple: medical journals are overrun by the same ideology that surely motivated the Oakland study itself. Academics are tacitly encouraged to submit papers that support this agenda, while those who raise questions are ignored or denied publication. The Oakland study has been uncritically referenced or relied upon by hundreds of subsequent studies and papers. Only one academic paper has raised concerns; it was subsequently retracted and its author disciplined by his medical school. The author maintains that his punishment constitutes retribution for opposing affirmative-action programs in cardiology.
On the basis of faulty research, and with no meaningful debate, health care is at risk of being resegregated. Yet more credible assessments, including a prominent 2011 study of a database of 22,000 patients, show that racial concordance between physicians and patients doesn’t produce meaningful improvements in health outcomes. Of course, patients should be free to choose their own physicians, including those who look like them. But they shouldn’t be pushed to take this path; nor should medical educators, providers, and policymakers deliberately undermine trust and lower the standards of physicians and medicals care. Segregation was horrible in the American past; it will do terrible harm now, too.
Stanley Goldfarb, a physician and former associate dean of curriculum at the University of Pennsylvania’s Perelman School of Medicine, is chairman of Do No Harm, where Alexander Raikin is a research fellow.
Get up to speed with the threats facing healthcare – and how we’re protecting patients and physicians.
"*" indicates required fields