The Life and Death of a Medical Study

A 2015 study found that black newborns attended by white doctors die at twice the rate of those in the care of black doctors. The study’s refutation last year has not altered the progressive narrative of systemic racism in medicine.

By Stewart Justman. From Quillete. Excerpt:

"Working with the same data set (provided by Greenwood, it should be noted), Borjas and VerBruggen discovered that when the risk factor of very low birth weight is taken into account, the disparity between infant deaths associated with white and black doctors approaches zero and loses statistical significance. Black newborns in the care of white doctors do die disproportionately, but not because the doctors are racist. They die disproportionately because black newborns of dangerously low weight—specifically, under 1,500 grammes—are disproportionately attended by white doctors (presumably in intensive care units). 

In the tradition of Unequal Treatment, Greenwood et al. supplemented inadequate data with social-psychological theorising and reached a foregone, if dramatic, conclusion. Borjas and VerBruggen, publishing their findings in the same journal (PNAS), corrected this error in September 2024. And yet what changed as a result of the refutation of the most important medical study of racial concordance ever to see the light of day?

“Prior work,” Greenwood et al. conclude, “suggests stereotyping and implicit bias contribute to racial and ethnic disparities in health. Taken with this work [that is, the Greenwood study itself], it gives warrant to hospitals and other care organizations to invest in efforts to reduce such biases.” Following the refutation of their findings, did Greenwood et al. infer that the crusade “to reduce such biases” had lost some of its “warrant”? Were they led to question the evidence base of the “prior work” on unconscious bias in medicine? Did those who cited the Greenwood findings as proof of racism moderate their rhetoric when the study lost its evidentiary value? In all cases, no."