Time to take race out of medical studies

Nothing symbolises a fall from grace like having your statue toppled. Such was the case for Dr J. Marion Sims in April, 2018, when workers unceremoniously loaded his bronze lookalike, which had peered down on New Yorkers in Central Park since 1934, into the back of a 4×4 pick-up truck ready to be shunted off to the nineteenth century surgeon’s grave site in Brooklyn.

The extraordinary reason for Sims’ demotion is the focus of an article by Princeton University historian Keith Wailoo, one of a quartet of pieces published in the Journal of the American Medical Association (JAMA) that examines the uncomfortable nexus between race and medicine in America.

In the 1840s Sims undertook a set of pioneering surgical experiments on women with fistulas – connections between the bladder and vagina often caused by traumatic childbirth. His aim was to relieve the appalling symptoms, which included urine leaking from the vagina, infections, and, in some cases, pariah status.

All of which sounds like an impeccable public service on Sims’ part. The problem was that his subjects were Alabama slaves, surgery was carried out with no anaesthetic (ether was new to the medical kit and not widely available) and the women did not give consent.

Sims, instead, reached out to the slave owners:

“I made this proposition to the owners of the negroes: ‘If you will give me Anarcha and Betsey, I agree to perform no experiment or operation on either of them to endanger their lives, and will not charge a cent for keeping them, but you must pay their taxes and clothe them.’”

While Sims praised the women’s courage, Wailoo describes how the surgeries took place in an era when many physicians believed black people were less sensitive to pain than whites. Perhaps this was salve for Sims’ conscience, but it was merely one example of entrenched false beliefs about race and biology.

In the early twentieth century, for example, lower rates of cancer in black people – probably just under-diagnosed in rural black communities with limited health care access – were put down to “negro immunity”.

This flawed racial determinism fed the mill of Jim Crow segregation, where biological difference could be invoked to assign black people separate, and mostly inferior, options in everything from buses to public toilets and hospitals.

The removal of Sims’ monument, writes Wailoo, “signals how far society has come in terms of respect for research participants, concern for research ethics, and appreciation for informed consent”.

When it comes to racial stereotypes, however, medicine is still learning. Wailoo cites a study that found some young hospital doctors mistakenly thought the blood of black patients clotted more quickly than that of whites. It was published in 2016.

This wrongheadedness speaks to deeply conflicted views in the research community about, how, or indeed whether, race ought to be included in the design of studies.

A trio of authors, led by Richard Cooper from the Loyola University Medical School in Illinois, US, writes that without studies singling out race we would never have learnt that black people with a broken bone wait longer for pain relief in the emergency department. Nor would we know that people of colour are less likely to get a flu shot or receive guideline-based screening for glaucoma or bowel cancer.

On the other hand, those authors point out, race as a demographic variable is often deduced to do some kind of work in a study’s results; a typical summing up might be “after controlling for other variables, race was still significant”.

What does that even mean? Do members of that race share some physiological property that altered an outcome, a disease trajectory for example? And what might that property be? 

One possibility is genetics. The authors point to the APOL1 genotype, possessed by 13% of people of West African descent, that confers a greater risk of kidney disease.

Such hard connections are the exception. More often, results simply fail to tease out inexorable links between race and the social determinants of health – socioeconomic disadvantage that can predict, for example, high risk health behaviours such as smoking and poor diet – that are actually responsible.

Even if race can be linked to a genotype or common geographic ancestry, the US is predicted to be “minority white” by 2045. Ethnic and genetic variation will make it even more difficult to defend race as a meaningful demographic in studies.

The upshot, the authors say, is that when analysis by race reveals neither social inequity nor mechanism of disease, “investigators are at risk of simply reinforcing the normative view of race as the great social divide”.

“This mode of analysis is wholly unjustified and should be avoided,” they conclude.

At least part of the antidote, according to a third article lead authored by Vence Bonham from the National Institutes of Health, is to ensure that race is never misused as a proxy for socioeconomic status. These authors point to an exemplary study that found education beyond high school, but not African ancestry, correlated with lower blood pressure.

It is a message reinforced by Phil Fontanarosa and Howard Bauchner, JAMA’s executive editor and editor-in-chief respectively, in an accompanying editorial:

“[I]nvestigators should ensure that the design of the study considers factors such as racial and ethnic background, ancestry [and] socioeconomic status. … [to] contribute to scientific understanding and generalisability rather than reinforcing stereotypes and bias.”

Wailoo, however, thinks it’s going to be an uphill battle. Nearly 200 years after Sims’ experiments, that 2016 study of hospital doctors found a significant minority believed black people have higher pain tolerance.

“In medicine and society,” he concludes, “the ghosts of bygone racism are always present.”