White Doctors Kill Black Babies:

Dubious Science and Anti-Racist Medicine

This is a wooden stake post on steroids. Wooden stakes kill vampires. Wooden stake posts aim to kill the false, unjustified and misleading academic claims that suck the blood out of social science validity and credibility.

It is also the first wooden stake guest post to appear here. In it, Jake Mackey and Dave Gilbert take on the wildly influential “finding” that Black infant mortality was higher when the attending physician was White than when the physician was Black, a “finding” routinely interpreted as some manifestation of racism. Jake, a professor of classics, and co-founder of Free Black Thought has posted here before on the “virtuous lie” that the SAT is racist.

Dave Gilbert is a technologist and former visiting assistant professor of communication with interests at the intersection of technology, culture, and identity. He is a founding member of Free Black Thought. You can find him on X here.

INTRODUCTION

Jake Mackey and Dave Gilbert

In August 2020, a study published in the Proceedings of the National Academy of Sciences (PNAS) made headlines with a striking claim: When cared for by white physicians, black newborns were more than twice as likely as white newborns to die, and significantly more likely to survive if cared for by black physicians. The authors speculated that "institutional racism" or racist biases could be among the underlying causes. 

From their “significance statement”:

A large body of work highlights disparities in survival rates across Black and White newborns during childbirth. We posit that these differences may be ameliorated by racial concordance between the physician and newborn patient. Findings suggest that when Black newborns are cared for by Black physicians, the mortality penalty they suffer, as compared with White infants, is halved.

The study’s finding was widely circulated in traditional and social media and has influenced medical practices and legal discourse. Its implications for the efficacy of "racial concordance" in medical care, that is, matching patients with physicians of the same race, were quickly seized upon. According to Google Scholar, it has been cited 683 times (as of 10/10/24) and counting. It was awarded the 2020 PNAS Cozzarelli Prize for “Best Paper for Behavioral and Social Sciences at PNAS.” And it was referenced by Supreme Court Justice Ketanji Brown Jackson in her dissent in the 2023 Students for Fair Admissions v. Harvard case (more on this below). 

The study appeared in the late summer of 2020. Given the prevailing mood, one could be forgiven for feeling that its findings not only exposed the hidden racism of a “white” medical establishment but also refuted the very possibility of a racially integrated society. After all, at the time Robin DiAngelo and other "anti-racists" were pressing the case that merely being around white people was dangerous for black people. How much riskier to have racially biased white doctors attending vulnerable black babies?

Now, however, a reanalysis of the data by Harvard economist George J. Borjas and Manhattan Institute fellow Robert VerBruggen, published in PNAS in September 2024, has debunked the original study's headline-grabbing claim. The new paper reveals that the 2020 analysis suffered from a critical methodological flaw—it failed to account for very low birth weight, a key factor in neonatal mortality. When this variable was included, the supposed benefit of racial concordance between physician and patient became statistically insignificant.

The rise and fall of the 2020 PNAS paper amounts to more than just an instance of science correcting itself. It also offers an illuminating case study of how dubious findings from ostensibly objective research can swiftly propagate through traditional and social media, get adopted as justifications for novel, race-based policies and practices, influence the transformation and even the creation of institutions, and inform jurisprudence at the highest level. Here, we trace the fate of the "white doctors kill black babies" meme and discuss the "neo-segregationist" medical practice of racially concordant care that its results were taken to support.

FROM SCIENTIFIC "FACT" TO ANTI-RACIST MEDICINE: THE 2020 PNAS STUDY ON INFANT MORTALITY 

The 2020 PNAS study by Greenwood et al., "Physician–patient racial concordance and disparities in birthing mortality for newborns," concluded that black newborns experienced significantly lower mortality rates when cared for by black physicians compared to white physicians. The study ran a linear regression on data from Florida hospital discharges between 1992 and 2015 and found that black infants had a mortality rate of 0.9% when cared for by white doctors, compared to just 0.4% when cared for by black doctors. In other words, they were 2.25 times as likely to die under a white physician’s care. As the authors wrote, "Black physicians systemically outperform their colleagues when caring for Black newborns." They speculated that "stereotyping and implicit bias" on the part of white doctors as well as "institutional racism" might account for at least some of the disproportionate mortality of black newborns in their care, and that black doctors’ "[awareness] of the challenges and issues" faced by black people might account for some of the decreased mortality of their black infant patients.

The study's conclusions spread rapidly through both traditional and social media. Major news outlets ran headlines like "Black newborns more likely to die when looked after by White doctors" (CNN); "Mortality rate for black babies is cut dramatically when black doctors care for them after birth, researchers say" (Washington Post); "Black babies more likely to survive when cared for by black doctors – US study" (The Guardian); and "Black babies are 3 times more likely to die when cared for by white doctors. Their mortality rate plunges with Black doctors" (Business Insider). 

Former New York City mayor Michael Bloomberg teamed up with the presidents of several historically black-serving universities and medical schools to write an opinion piece in Time that cited the study in order to campaign for more black doctors. "If we had more Black doctors, we would save more Black lives," they wrote, and they laid out a plan to achieve that result. 

Popular media typically presented the study’s finding as an established fact, with little discussion of potential methodological issues or alternative explanations. Moreover, the moral framing and policy implications that prevailed in the popular media merely reflected those of the study itself. The authors noted that their results bolster "the importance of racial concordance in addressing health care disparities." They opined sympathetically that "[f]or families giving birth to a Black baby, the desire to minimize risk and seek care from a Black physician would be understandable." And they lamented the fact that "the disproportionately White physician workforce makes this untenable because there are too few Black physicians to service the entire population."

The study’s emphasis on racial concordance aligns with a robust movement in the medical field that predates the 2020 PNAS study. For example, Culture Care, a startup founded in 2018, displays the slogan "Black Doctors for Black Women" on its homepage and offers racially concordant "telemedicine." Since 2016, Summa Health’s CenteringPregnancy program in Akron, Ohio, has offered "consultations with a Black OB/GYN," among many other services. And UCSF’s EMBRACE program has featured "race-concordant" perinatal care for black families since 2018. 

The 2020 PNAS study bolstered the legitimacy of these novel interventions by seemingly providing evidence in support of the theory that racial concordance improves outcomes for black infants. For example, Dr. William McDade, the Chief Diversity, Equity and Inclusion Officer of the Accreditation Council for Graduate Medical Education (ACGME), cited the study in a 2021 Health Resources and Services Administration (HRSA) presentation titled "Racially Concordant Care" that touted the benefits of the practice. And Dr. Alicia F. Lieberman, a  Professor of Psychiatry in the UCSF School of Medicine, cited the study in a 2024 HRSA presentation in which she discussed the success of UCSF’s racially concordant EMBRACE program. 

Racially concordant healthcare programs continue to be established, legitimized, at least in part, by the 2020 PNAS study. In July of last year, BLOOM: Black Baby Equity Clinic opened at UCSF Benioff Children’s Hospital Oakland. According to a press release, "[t]he new clinic matches Black babies, newborn to 3 years old, with a health care team of the same race." The hospital has launched a study to collect data on the outcomes of patients who use BLOOM. They express the hope that their "findings can inform refining of the intervention and the development of similar equity-promoting programs and services in California, and around the country." Under the heading "Foundational Articles," the website that announces the BLOOM study cites the 2020 PNAS article.

Indeed, the 2020 PNAS article’s second author, Dr. Rachel Hardeman, has herself recently launched a new healthcare initiative. For Hardeman, racially concordant health care is not just an academic but also a personal issue. Hardeman is a professor in the University of Minnesota’s School of Public Health and serves on the CDC’s Advisory Committee to the Director. Her faculty webpage identifies "structural racism as a fundamental cause of health inequities" and states that her work "dismantles the systems, structures, and institutions that allow inequities to persist" by "[leveraging] frameworks like critical race theory and Reproductive Justice." In an interview with Science News, Hardeman underscored the importance of racially concordant care during her own pregnancy. "I wanted a doula from the same racial background as myself, and having someone who really understood my life experience was incredibly important," she said. 

In February 2021, six months after the PNAS study appeared, Dr. Hardeman founded and began serving as Founding Director of the Center for Antiracism Research for Health Equity at the University of Minnesota. The Center’s "Research Library" webpage, which bills itself as "A curated catalog of trusted antiracist scholarship," links to Hardeman’s 2020 PNAS study and provides the abstract. Another webpage details the Center’s origin and links to a 2022 New York Times article about Hardeman and her work. The Times article cites her 2020 PNAS study and acknowledges that the "results of her research…sometimes garner controversy."

DO THE MATH: THE SUPREME COURT MISREADS THE 2020 PNAS STUDY

Perhaps most strikingly, the 2020 PNAS study by Greenwood, Hardeman et al. made an appearance in the Supreme Court. In 2023, Justice Ketanji Brown Jackson cited its conclusions in her dissent in Students for Fair Admissions v. Harvard. Arguing that the University of North Carolina’s promotion of diversity in medicine through race-conscious admissions "saves lives" (p. 22), Justice Jackson wrote:

For high-risk Black newborns, having a Black physician more than doubles the likelihood that the baby will live.

Justice Jackson found this claim in an amicus brief, which she cites in a footnote. The authors of the brief, Heather Alarcon and Frank Trinity, both with the Association of American Medical Colleges, later informed the Court, through their counsel, Jonathan Franklin, that they had overstated the study’s findings. 

This admission came on July 7, 2023, two days after Ted Frank, Director of Litigation at Hamilton Lincoln Law Institute (which had filed its own amicus brief on behalf of the plaintiffs in the case), pointed out Justice Jackson’s mistake in a Wall Street Journal op-ed. Frank noted that according to the study, the actual improvement in survival rates for black newborns attended by black physicians was between 0.13% and 0.2%, not the dramatic more-than-doubling that Justice Jackson had cited. 

Despite having been admonished in the pages of the Wall Street Journal, Franklin was careful to affirm that even if the amicus brief had overstated the 2020 PNAS study’s finding, its claim still conformed with the larger facts:

But however it is summarized, the [2020 PNAS] study strongly supports the statement in Justice Jackson’s dissent that "the diversity that UNC pursues for the betterment of its students and society . . . saves lives."

Thus, even as Franklin refuted the mistake of the amicus brief and Justice Jackson, he reaffirmed the deeper truth that the mistaken claim bore witness to. What matters is not whether the details of the brief and Justice Jackson’s dissent are literally true. What matters is the larger truth of racism and racial disparities that even overstated or mistaken claims rightly direct our attention toward.

The trouble is that while there may well be racism in American medicine, the study, even when summarized accurately, provides no evidence of it, because its authors omitted a crucial control that would have changed its finding dramatically. To this we now turn.

DEBUNKING THE KILLING BLACK BABIES MYTH: THE 2024 PNAS STUDY AND BEYOND

A 2024 reanalysis by George Borjas and Robert VerBruggen revisited the data from the original 2020 study and demonstrated that its conclusions were in error. The original study had controlled for numerous variables, including hospital and physician fixed effects, as well as the 65 most common comorbidities affecting newborns. However, it failed to account for one crucial variable: very low birth weight (defined as less than 1,500 grams), a key determinant of neonatal mortality. Black newborns are disproportionately more likely to be born with very low birth weights compared to their white counterparts.

Borjas and VerBruggen used the same dataset but introduced their one, crucial adjustment. As they point out, "the finding of racial concordance in newborn mortality is very sensitive to subjective decisions about which comorbidities to include as control variables." By omitting this factor, the 2020 study had created the appearance of a racial concordance effect that was actually driven by differences in the newborns' health.

Furthermore, the 2020 study did not adequately account for the fact that low-birth-weight black infants were disproportionately cared for by white physicians. These newborns accounted for 3.4% of white physicians’ black patients, compared to 1.4% for black physicians. This skew in patient-physician demographics falsely amplified the apparent advantage of racial concordance, making it seem as if white physicians were less competent at keeping black newborns alive when, in fact, they were simply more likely to attend at-risk black newborns. When this crucial factor was included in the analysis, the racial concordance effect became statistically insignificant. Black infants were more likely to die under the care of white physicians than of black physicians, but it was not because of racism or lack of "racial concordance" between physician and baby. It was because white physicians cared for disproportionately more very low birthweight black babies and such very low birthweight babies have much higher mortality. The 2020 study had thus created the illusion of a racial concordance effect driven by the uneven distribution of unhealthy black newborns among doctors.

This correction highlights the limitations of the racial concordance narrative. While the idea that black physicians were inherently better suited to care for black infants may have had intuitive moral appeal, especially in the summer of George Floyd, Borjas and VerBruggen’s analysis demonstrated that focusing on physician race was misguided. Racial concordance turned out to be perfectly irrelevant to infant survival rates. The inference to be drawn from the study, for anyone who cared about black lives, was that an excess of low birth weights among black babies, not a dearth of black doctors, urgently needed to be addressed.

This result cannot be emphasized enough: the obsession with race as a causal factor in health disparities was shown, once again, to mislead us away from real causes that we could potentially address. In this case, tackling the root causes of very low birth weight in black newborns—such as socioeconomic inequalities and inadequate access to prenatal care—would be far more effective in reducing neonatal mortality than attempting to match patients with physicians of the same race in the hopes that a mysterious racial sympathy between them will improve health outcomes.

There is, however, more to the story of the 2020 PNAS study, beyond its debunking by the 2024 study. A year ago, in the Wall Street Journal op-ed referenced above, Ted Frank noted that "hidden deep in an appendix" of the 2020 PNAS study was data revealing that the black doctors considered in the study have "a statistically significant higher mortality rate for white newborns, and a higher mortality rate overall, all else being equal." If we take these results seriously, and given the fatal methodological flaws in the study we surely should be cautious about doing so, we would then have to question not only the racial concordance narrative, but also the "we need more black doctors" narrative. For why would we wish to increase the share of black doctors, if it’s true that their patients experience higher mortality rates? 

The fact that these results of this flawed study were nowhere explicitly pointed out and discussed in the body of the PNAS article is consistent with an observation of Cory Clark and her coauthors earlier this year. They discovered that scientists are increasingly using "harm-based criteria to evaluate research" and that they are willing to censor one another and themselves for "prosocial" reasons, out of "concerns for the well-being of human social groups," including "historically marginalized groups." In light of this trend, it is not surprising that Greenwood and his team ran with the finding that inculpates white doctors for a mortality disparity that affects black babies, and buried their less congenial results in an appendix.    

In these buried findings, too, we see another danger of a hyperfocus on race. As the economist Glenn Loury has noted more than once, the game of seeking out racial disparities that imply black disadvantage and white culpability is "a double-edged sword" that can be picked up and turned against the person who wields it. If we make a practice of looking to race as the determinative factor in the outcomes we care about, we should not be surprised if others, too, seize with glee upon their own set of race-related disparities, such as those in the 2020 study’s appendix, that are unflattering to historically marginalized groups.

THE 2020 PNAS STUDY: A "VIRTUOUS LIE"?

The episode we have laid out serves as a case study in the way activists, knowledge-economy elites, and the institutions they control can harness spurious scientific research to justify ideologically-driven policies and practices and to bolster legal arguments. This pattern, whereby incomplete or flawed, and in any case highly contestable, claims or findings become the basis for policy innovations and jurisprudence, is a hallmark of social justice movements. Activists adopt these dubious findings—these "virtuous lies"—and present them as the incontrovertible "facts on the ground" that necessitate radical change. 

The term "virtuous lie" was coined by Jake Mackey. Virtuous lies often go hand-in-hand with and serve as the "factual" foundations of what Rob Henderson has dubbed "luxury beliefs." Luxury beliefs are prescriptions for novel policies and social practices that confer status on those who propose them while entailing costs for the people subject to them.

Think, for example, of virtuous lies such as "police hunt and kill black people in preference to white people" and "standardized tests are racist." Neither of these claims is true. The virtuous lie about the police motivates the luxury belief of "defunding" or "abolishing" them, while the virtuous lie about standardized tests underwrites the luxury belief that they should be discontinued in college admissions for the sake of "equity." These luxury beliefs have the ring of moral clarity but in practice they entail costs for the very people they purport to serve. Depolicing results in massive increases in crime and murder in the most disadvantaged neighborhoods while, as Mackey explained here at Unsafe Science, dropping the SAT removes a metric by which gifted but under-resourced young people can be identified by colleges, even as it makes it easier for the wealthy to game admissions.

The untrue meme that "white doctors kill Black babies" was, as we saw in the previous section, already following this pattern of passing from "fact" to institutionalization in policy and practice before it was debunked. The spurious finding of the 2020 PNAS study was thus a virtuous lie in the making. It aligned with the broader goal of addressing racial disparities in healthcare, making it appealing during a time of heightened concern about systemic racism. And it was, as we have seen, rapidly adopted to justify racially-concordant perinatal practices by hospitals and medical schools. In this particular case, the imperative to promote racial concordance between physicians and newborns was the associated luxury belief, itself also nascent but growing. (We ask whether all racial concordance practices are luxury beliefs in the next section.) 

The uncritical acceptance of the finding and the adoption of the associated luxury belief came at a potential cost. By focusing on physician race instead of the substantive factors driving neonatal mortality, the fledgling virtuous lie risked validating luxury beliefs about racially concordant perinatal care that could have diverted resources and attention away from interventions that might have had a more meaningful impact on neonatal outcomes. The finding and its associated policies could also have led to other unintended consequences, such as stigmatizing white physicians with unblemished track records of caring for at-risk black neonates, or reinforcing overly tendentious narratives of racial conflict in healthcare. 

In general, virtuous lies tend to thrive in environments where there are strong conformity pressures to adopt a "liberatory" narrative, and where questioning that narrative is equated with supporting oppression or injustice. They spread during "Awokenings" and within social justice movements, where certain claims—such as the ubiquity of systemic racism, the dangers of implicit bias, or the supposed efficacy of race-based policies—are treated as moral axioms. Challenging these claims, even when based on flawed or incomplete data, risks social and professional ostracism. As a result, virtuous lies can become entrenched, not because of the overwhelming evidence of their truth, but because they are both morally appealing and socially and economically advantageous for those who endorse them. All these conditions were present when the 2020 PNAS study was published. We were at the peak of an Awokening in which the Black Lives Matter movement had captured the country's attention.

This brings us to one of the most striking features of social justice campaigns based on virtuous lies and luxury beliefs: their teleological, yet non-agential, nature. Through the uncoordinated actions of numerous well-intentioned individuals and organizations, virtuous lies and their concomitant luxury beliefs become embedded in public discourse and institutional practices. In the case of the 2020 PNAS study and its wide dissemination, there was no grand conspiracy, no Marxist "long march through the institutions" of the sort that "anti-wokes" love to rail against, to fabricate data or push a false narrative. Instead, individual actors—researchers, journalists, activists, and even an ex-mayor of New York City and a Supreme Court Justice—spontaneously, each for his or her own reasons, aligned around an overarching common goal: addressing racial disparities in healthcare. The virtuous lie that white doctors are fatal to black babies and the luxury belief in racial concordance were, in a sense, emergent properties of this alignment, driven by the moral imperative to correct perceived injustices. 

The non-agential nature of this process makes it particularly difficult to counter. There is no central authority or orchestrated conspiracy to unmask. Rather, virtuous lies and luxury beliefs become deeply ingrained in the discourse, institutional practices, and public policy through distributed channels. This decentralized propagation contributes to the resilience of virtuous lies, as there is often no single point of origination to trace the lie back to. Who, for example, originated the virtuous lie that police hunt and kill black people? Hell, it used to be true! As a result, such narratives can persist and influence decision-making long after their validity has been successfully challenged.

The claim that white doctors put their black infant patients at risk did, of course, have a determinate point of origin. We hasten to note that we are not asserting that the authors of the 2020 PNAS study acted insincerely or cynically concealed facts that would have debunked their own claim. Rather, in the summer of 2020, all the incentives—epistemic, moral, economic, and reputational—for academics, journalists, and other knowledge-economy professionals tended toward finding examples of anti-black racism that could be used to promote emancipatory projects. The fact that the study's second author, Dr. Hardeman, is a proponent of Critical Race Theory and of the "dismantling" approach to social justice may have made the team especially sensitive to these incentives.

That said, the claim never fully attained virtuous lie status. It was merely a spurious research finding that had been latched onto as proof of the need for race-conscious medical interventions before it was shown to be false. From the moment the 2020 PNAS paper appeared, it was apparent to many researchers that it was deeply flawed. It was disparaged by experts on Twitter and Dr. Vinay Prasad, a professor in the Department of Epidemiology and Biostatistics at UCSF, devoted a podcast to tearing it apart days after its publication. The 2024 PNAS paper should clear up the "misinformation" propagated by the flawed 2020 PNAS study, ideally reducing its impact on medical practice. However, if the past is any guide, the 2024 findings are likely to be quietly ignored. The institutional and bureaucratic structures supporting perinatal racial concordance will likely remain in place, even without the crucial support of PNAS 2020.

Even if the specific claim never does attain full social acceptance as an unassailable virtuous lie, the pursuit of racial concordance will surely persist in other domains of medicine and therefore warrants examination. It is worth asking whether racial concordance per se is a luxury belief—a concept that, despite a lack of robust evidence, continues to influence medical policy and practice due to its moral appeal.

RACIAL CONCORDANCE: A LUXURY BELIEF?

While the untrue claim of the 2020 PNAS study regarding perinatal racial concordance has been debunked, belief in the efficacy of racial concordance practices in general existed well before it ever appeared and will no doubt endure in some quarters. What should we make of racial concordance? Is it a mere luxury belief, or does it have a firm foundation? 

Faith in racial concordance is high in some medical circles. Some proponents, such as Dr. William McDade (mentioned above), argue that given current disparities in physician race, racial concordance alone is not enough. Even as steps are taken to diversify medicine, physicians must learn "cultural competence" and even "cultural humility." Other proponents, however, such as Dr. Gina Guillaume (North by Northeast Community Health Center, Portland, OR) Dr. Juan Robles (Albert Einstein College of Medicine, Bronx, NY), and Dr. José E. Rodríguez (University of Utah Health) argue that mere cultural competence is inadequate: the goal must be full racial concordance, which will require policies that transform the racial and ethnic composition of medicine. 

However, because it seems to attribute quasi-causal properties to race, which is a fiction, we would caution that racial concordance may distract researchers and practitioners from seeking out the actual causal factors that truly do influence health. The fate of the racial concordance hypothesis per se is of course not tied to that of the 2020 PNAS study; it must stand or fall on its own merits. Still, much like luxury beliefs, it risks diverting attention from more significant factors affecting health outcomes, thus entailing potential costs for racial and ethnic minority patients.

Writing at the height of the COVID-19 pandemic and inveighing against ubiquitous talk of the virus's disproportionate impact on Black Americans, political scientist Adolph Reed Jr. asked what exactly do people

"mean when they refer to 'race' as a factor contributing to vulnerability to Covid-19, or to anything else for that matter?" 

He decried the "resurgent racialist determinism" underlying the "popular hype about supposedly race-targeted pharmaceutical interventions," and expressed the concern

that nowadays the inclination to treat "race" as a natural category, one capable of exerting independent effects, is as likely to come from those who presume to be advocates for the concerns of what were described a generation or two ago as "historically underprivileged minorities" as from open racists. 

We share Reed’s concerns and believe they apply to the racial concordance hypothesis. Focusing on race as a determinative factor in health is at best a proxy for and can distract from the material factors that cause adverse health outcomes, as the failure of the 2020 PNAS study showed. In this sense, whether we should view them as reflecting "luxury beliefs" or not, race-focused practices do carry potential costs for the people they purport to benefit.

The primary mechanisms by which racially concordant medicine is supposed to exert its putative beneficial effects are somewhat mysterious but are often attributed to better communication and trust between patients and doctors who "look like them." On the one hand, a large-scale study published in 2011 found "[l]ittle evidence of clinical benefit resulting from sex or race/ethnicity concordance." A systematic review of studies on racial concordance from 2018 found that racial concordance generally led to improved communication for black patients but could "not conclusively support an overarching hypothesis that patient-physician communication is worse for black patients than white patients." On the other hand, however, there are striking examples in which concordance seems to decrease mortality, presumably because better communication and trust lead to "improved medication adherence." 

Moreover, it is true that patients do tend to select caregivers of their own racial or ethnic background when they can, perhaps as a result of homophily or affinity bias. Recall Dr. Hardeman’s preference for a black doula. We are all for freedom of association, but is racial matching a "best practice"? 

As our discussion thus far suggests, the literature appears ambiguous to us. According to a 2023 study, "concordance has only modest associations with health care satisfaction or perceived respect, and it is not significantly associated with the number of medical visits or other outcomes," such as "having enough time in care, [and] ease of understanding the clinician." Instead, the study’s authors suggest, cogently in our view, that more material factors such as prior "[p]oor health status, being uninsured, and lacking a usual source of care are more strongly associated with patient experience." In other words, even though people may tend to prefer doctors of their own background, concordance is not indisputably linked either with subjective metrics such as greater satisfaction or with objective metrics such as time in care. 

It thus appears to us that the best we can do is to conclude that in some cases concordance is a luxury belief, as it appears to be for neonates, for college athletes with concussions, and for Hispanics (especially men), whereas in other cases there is evidence of its efficacy.

NEO-SEGREGATION?

Whatever we should conclude about the benefits or lack thereof of racial concordance, we note that the 2020 PNAS study, despite its fatal flaws, fulfilled its function where racial matching is concerned. The efficacy of racial concordance in perinatal contexts received reinforcement, UCSF’s racially concordant BLOOM program could appeal to the study to rationalize its practice, and Dr. Hardeman’s Center for Antiracism Research for Health Equity was established and can display the article in its "curated catalog of trusted antiracist scholarship." 

In that sense, the fate of the study’s findings no longer matters. A kind of neo-segregationist movement is underway in medicine, whether white doctors really do kill black babies or not. This impulse to resegregate, perhaps stemming from perceptions of integration’s failures after more than half a century, is to be lamented. Racial concordance—much like sorting children by race and affinity groups in K-12, as well as POC-only dorms and separate "cultural" graduations in colleges, not to mention laws and policies that establish separate packages of norms and standards for racially defined identity groups—is on the rise. This is in large part a result of the practical application of Critical Race Theory (of which, recall, Dr. Hardeman is a proponent) and its project of dismantling the liberal order, replacing equality with "equity." Debunking one poorly designed study will not stop the inevitable.

And here we arrive at our primary complaint. Whatever the merits of research showing benefits of racial concordance for black patients, we are concerned about the way that research will be used to substantiate normative and ethical proposals, some of which will be luxury beliefs. Famously, one cannot derive an ought from an is. We should therefore think hard about what we do next. For what we are being asked to accept really is a re-segregation of medicine. 

Indeed, studies can be found that show benefits of concordance for other non-white and even white patients (see also here and here) as well as black. If the proponents of concordance were consistent, the logic of their position would militate for the application of "separate but equal" medical regimes for every federally recognized racial and ethnic group. However, we suspect these proponents would tend to eschew consistency. What they are actually asking Americans to accept is a double standard, according to which favored groups get race-conscious, racially concordant medicine, while everyone else continues to get racially integrated medicine, whose comparative efficacy is a question sure to be politely ignored. Call it "separate but equitable." 

Finally, we speculate with some dread that as racially concordant care becomes the norm for favored groups, white people will begin to demand services that match them with white doctors. Once fully out in the open, the principle of race-matching will be impossible to limit. For if we are really being asked to believe that re-segregated medicine provides benefits for patients and that integrated medicine entails costs for them, then what concordance advocates are in effect asking of all but their favored groups is to accept those costs. 

Dystopian futures aside, what is often posited to drive the benefits of concordance is not so much special skills on the part of concordant physicians but rather increased trust as a result of cultural commonality. The beneficial effects lie not in something the doctor does, but in the patient’s greater affiliation with the doctor and greater adherence to the doctor’s prescriptions. 

Whence this trust differential? Obviously, racism is real. Moreover, there are people alive today who still bear the scars of Jim Crow. Still, could disempowering social justice narratives about the inescapability of debilitating racism, 60 years on into the post-Jim Crow era, drive at least some of this lack of trust? Pervasive virtuous lies, like the misrepresentation of the Tuskegee Study of Untreated Syphilis as a racist medical experiment, clearly play a role in fostering mistrust of a "white" medical establishment. It has been documented that the false narrative about Tuskegee contributed to lower COVID vaccine uptake among black Americans. 

Whatever its causes, should we resign ourselves to the idea that the problem of racially differential trust is intractable and simply surrender to it by mandating concordance? The question underlying all of this is what are we willing to settle for as a society? A retreat into racial and ethnic separatism, or an ever more perfect union? 

Research that shows genuine benefits of concordance should be the beginning of an investigation into the possible causal mechanisms at work, not a stopping place where we resegregate medicine. We should be asking, Why is it that black patients are more willing to adhere to the advice of black physicians, if indeed they are? We shouldn’t be satisfied merely to pair black patients with black doctors and ascribe the benefit to race. If the causal mechanism really is trust, for example, what can we do to repair and increase it? (Here the stock anti-racist answer, diversity training, is inadmissible: such training has no beneficial and occasionally harmful effects.) The story of the misleading 2020 PNAS study leads us to conclude that in all cases where we find racial disparities in health outcomes associated with concordances and discordances, we should look further, beyond mere racial matching and mismatching, for the deeper variables associated with causality and remedies.  

To sum up, there are some cases, such as pairing black infants with black doctors, in which concordance has no effect. But even in cases where its effect is robust, we would still urge that race-level analysis and race-level policy are insufficient. There are likely many health issues for which resegregation makes no difference, and yet it is being touted widely, determining entire research programs, and being put into practice with increasing frequency, all to the possible detriment of research into causes of health disparities deeper than skin color.

CONCLUSION

The 2020 PNAS paper is a case study in the way "social justice" campaigners in the knowledge economy harness dubious scientific research to justify questionable new policies and institutional transformation. The finding that white doctors are a lethal danger to black babies was well on its way to becoming a virtuous lie before it was debunked by Borjas and VerBruggen. Its normative corollary, the luxury belief that black newborns require racially concordant medical care, still appears to be going strong, if the new programs offering this service are any indication. 

The danger in any campaign that exhibits this pattern is that if its factual basis is a "virtuous lie" or its normative "luxury belief" agenda entails costs for those it purports to help, then the campaign may do more harm than good. In the current case, focusing on the irrelevant variable of race and worrying about the biases of white doctors risks obscuring the actual determinants of black infant mortality. In addition, devoting institutional and individual energy to implementing racial concordance risks misallocating resources that would be better directed elsewhere.

Finally, glancing at the bigger picture, we close by predicting that the narrative that was spun from the flawed PNAS study risks deepening ideological divides—reinforcing growing distrust in scientists and contributing to skepticism among those disinclined to consider a role for racism in health outcomes, while encouraging increasingly uncritical doubling down among those already committed to "race-first" practices. All our babies, black, white, or otherwise, deserve better.

Thor’s Fight with the Giants, 1874, by Martin Eskel Winge, via Wikimedia.