Karen Titus
November 2024—David Grenache, PhD, D(ABCC), chief scientific officer of TriCore, is breaking no new ground when he considers the maternal serum alpha-fetoprotein test and says, “A screening test is designed to put people into one of two camps.”
No scientific advances there. But the reverberations of this particular screening test have landed patients and physicians in complicated situations of late.
The obvious divide is to identify increased risk of fetal abnormalities, including open neural tube defects. But the test has historically incorporated a race adjustment for Black patients. And that adjustment—seen as a way to account for reported differences in AFP values between Black and non-Black pregnant persons—in turn has come under more scrutiny recently as physicians question whether the adjustment should continue to be used, or whether it should be dropped, à la the race adjustment for eGFR/renal function.
Compared with eGFR, evaluating the race adjustment in maternal serum AFP screening is more in its infancy, Dr. Grenache says. “There are emerging differences in people’s opinions about what we should and should not do.”
Second of two parts.
Part one in October issue.
In some ways, this issue could be seen as part of the normal push-pull of practicing medicine. Data are questioned and interpreted differently; one group makes a change quickly, while another carries on with traditional approaches. But because this is a discussion about race in medicine, the conversations shoulder an added burden.
Says Dr. Grenache, who is also core laboratory director and clinical professor of pathology, University of New Mexico School of Medicine, “There’s a long history of racism in America.” And while the goal is to care for all people equally, using the word “race” in medicine is upending, he says, because it implies unequal care.
In recent interviews with CAP TODAY, a number of experts weighed in on both the immediate matter at hand—whether to keep or drop the race adjustment to the AFP calculation—and the larger forces that animate their decisions.
Often they say these things using almost the exact same words and phrases.
And yet, there’s palpable frustration in their voices, too, as they explain their decisions to drop the adjustment now, soon, or later.
As Octavia Peck Palmer, PhD, vice chair of health equity, diversity, and inclusion, Department of Pathology, University of Pittsburgh Medical Center Presbyterian and Shadyside automated testing laboratories, succinctly puts it: “Using race is not the best way to do medicine.” Dr. Peck Palmer is also associate professor, Departments of Pathology, Critical Care Medicine, Clinical and Translational Science, and division director, clinical chemistry.
Some institutions have moved fairly nimbly through the process of dropping race adjustments.
When Christopher Crutchfield, PhD, arrived at Northwestern Memorial Hospital in Chicago in 2021, as associate director of the clinical laboratory, he quickly became part of the effort to remove the race-based adjustment from eGFR, an action that was underway at many institutions nationwide.
Dr. Crutchfield, now scientific director for Northwestern Medicine clinical laboratories, which comprises 11 hospital-based labs, calls the eGFR experience “delightful.” He worked closely with Northwestern’s chief of nephrology at the time, Susan Quaggin, MD, who was also serving as president of the American Society of Nephrology. He and colleagues prepared carefully throughout that fall and in November 2021 dropped the adjustment.
Despite their relative spryness, he says, “It was actually quite a leap forward for our institution,” both in terms of dropping an entrenched practice and bringing more equitable care for renal transplant candidates.
More recently, Northwestern abolished the race adjustment for maternal serum AFP, again with relative ease. The initial push came from the then vice president for quality, Cynthia Barnard, PhD, who, after conversations with colleagues and looking at a study reconsidering use of race in AFP screening that has influenced much of the recent discourse (Burns RN, et al. Obstet Gynecol. 2023;141[3]:438–444), asked whether Northwestern was using the adjustment. The question soon fell to Dr. Crutchfield to answer.
Maternal AFP screening entails both the serum biochemical test and the calculation, Dr. Crutchfield says, which typically is part of the commercial software many laboratories use. “You need to have a special type of software to organize those medians and organize the quality statistics.”
The first step was to ask the technical supervisors and the vendor if the adjustments were indeed in the software. The next was to contact the CAP, he says, given that the language in the accreditation checklist requirement seemed to be open to interpretation regarding whether the adjustment was required. If the laboratory did drop it, they were told, they needed to document their decision-making.
That was in February 2023. By December 2023, the adjustment had been removed. “This was actually relatively fast,” he says, given that most institutional projects of this magnitude take about 18 months. One key to their swiftness was pulling in expertise and perspectives from obstetrics, equity, and quality. Consensus was universal, as he recalls. “I don’t think there was any major dissent.”
They took a look at the current literature as well as their own data to arrive at their decision.
“When you look at the adjustments that were being made, there’s really two,” he says, based on what he calls legacy research that asserts there is elevated production of AFP in pregnant Black patients versus those who are non-Black. “So there’s a biochemical adjustment. And then there’s also a difference in the screening cutoff that’s supposed to account for prevalence.
“When we looked at our own institutional data,” he continues, “our findings projected that we would not see as big of a change as the literature asserted,” including a recent and widely discussed paper in Clinical Chemistry (Messerlian G, et al. 2024;70[7]:948–956). “So our strategy was going to be—as we have to do already for this type of test—to monitor the screen positive rate.” The typical quality threshold for positive rate on the high end is three percent, he says, adding that with the changes that the legacy parameters had established, those changes reduced the sensitivity of this test for Black patients.
Moreover, the evidence for using the calculation didn’t seem to be extremely strong, he and his colleagues felt. “A challenging component of this issue is, it’s only the designated Black patient who was receiving the correction factor.” Without comprehensive, robust data for other racial groups, he says, the reason for continuing with the correction seemed to falter further.
Colleagues had questions, of course. “It’s a challenging subject when you socialize this,” Dr. Crutchfield says. “Because the immediate response is, Well, don’t you just want the most accurate test?”
“In this case,” he says, “there’s not going to be any major clinical decision just based on AFP. It’s going to escalate access to advanced care. So to that end, it becomes an equity question.” The concern about increasing sensitivity was that it could lead to some highly abnormal frequency of false-positives, he says, referring back to the three percent cutoff. As they had with dropping race from the eGFR equation, he and his colleagues looked at the impact, asking themselves, Did this achieve what we thought it would achieve? With maternal AFP, “We did not see an increase over that three percent false-positive rate. The frequency had some nominal increase, but nothing highly concerning.”
They worked closely with their software company, Benetech (“They were a good partner,” he says) to remove the adjustment. They did an initial assessment several months after making the change, and they intend to continue to monitor “to see if somehow that potential false-positive rate gets too high,” Dr. Crutchfield says. “But at the moment we feel quite gratified that we now have an equitable test, and we are not decreasing potential access to care. And we feel, at least within our lab, that this specific test should be the last for which we still have some legacy race-based adjustment that we would have to correct.”
In dropping the race adjustment, Dr. Crutchfield and colleagues appear to have acted out a basic geometry truth: The shortest distance between two points is a straight line.
As discussions about maternal serum AFP screening unfold elsewhere, however, the road to change can look more like a Piccadilly Circus.
“This subject is going to be very emotionally charged,” Dr. Crutchfield says. “You have issues with race. You have issues with pregnant people. You have issues with fetuses.”
You also have issues with the past. As writers love to point out, the past—along with its first cousin, history—is also prologue. It’s never dead, nor even past.
In medicine, says Christina Pierre, PhD, D(ABCC), clinical assistant professor of pathology and laboratory medicine, Perelman School of Medicine at the University of Pennsylvania, reconciling the past and present is often a matter of bringing ingrained practices in line with current knowledge. The majority of research that formed the basis for race adjustment in maternal serum screening biomarkers, specifically for open neural tube defects, she says, was done before the sequencing of the human genome, when it was standard practice to look for differences in biomarker concentrations based on race.
“That’s significant,” says Dr. Pierre, who is also clinical chemist, Department of Pathology and Laboratory Medicine, Penn Medicine Lancaster General Health, “because prior to that we didn’t fully appreciate that essentially genetics did not define different races. So we didn’t appreciate that there’s not a genetically distinct pattern that we would expect in white people versus Asian people versus Black people, et cetera.” With genome sequencing having provided much more nuanced insights, it’s clear that race “doesn’t necessarily define or neatly put people into boxes as far as biomarkers are concerned.”
Notwithstanding the scientific underpinnings of genome sequencing, Dr. Pierre says, researchers have observed, historically and currently, differences in maternal serum AFP that appear to be linked to race, though various studies have yielded conflicting results. That led to recommendations—including from the CAP and American College of Medical Genetics—to adjust for those differences based on race, without understanding why those differences occurred.
The reasons, of course, are likely multifactorial, as Dr. Pierre and others point out.
Dr. Pierre notes that self-reported race can, to some degree, correlate with genetic ancestry; thus, some gene variants might be observed at higher rates in some self-reported races than others. But, she cautions, “That is not to say it is true that every single person in some self-reported race has that specific trait or genetic variants.” The sickle cell allele is a prominent example. It’s most prevalent in areas endemic for malaria—sub-Saharan Africa and South Asia. “But you can’t look at the color of someone’s skin and tell them they have that allele.”
Dr. Pierre, who co-wrote with Dr. Peck Palmer and others a clinical opinion published late last year (Pierre CC, et al. Am J Obstet Gynecol. 2023;229[5]:522–525), suggests focusing on how using the race adjustment affects patient care.
First, by saying that Black pregnant people need to have a higher serum AFP concentration to be classified as screen positive, “Is that impacting the rate at which we’re detecting these defects, specifically in Black women?” she asks.
The second question gets back to Dr. Pierre’s earlier point: What is the reason for the differences some studies have observed? Whatever the answer(s), future risk assessments should take that, not race, into account.
But the bigger picture of race is more troubling. “I think about these things often, because I have mixed ancestry—Black, white, South Asian,” says Dr. Pierre, who identifies as Black. “So where do you correct for my AFP, or do you not correct for my AFP?”
She’s not sure herself how she would answer the question, showing how dubious even self-identification can be. Before she entered the medical field, she says, “I would tick ‘Black,’ without thinking twice on any of these clinical questionnaires: ‘Race: Black’—done. Now, essentially, what I’m understanding is depending on what box I tick, that can change the way I’m managed clinically.”
Not only is this near and dear to her, as she puts it, but it also raises huge questions for health care providers. The medical field doesn’t know how to accommodate people who are multiethnic. Dr. Pierre points to the commonly used category of “Hispanic,” which comprises people who identify as Black as well as white. Would the race adjustment be used for someone considered to be Hispanic Black, Dr. Pierre asks, or would they be considered Hispanic? “You could see how this gets tricky really fast.”
UPMC’s Dr. Peck Palmer propels the conversation further.
“We are crudely using race to try to deliver care quickly in a manner that’s not appropriate,” she says. “To group populations by race and say a whole population is different from another population is inaccurate.”
Laboratories, long encouraged to look beyond their walls and into their institution, might want to look outside the building itself, into the communities they serve, Dr. Peck Palmer suggests. In that sense, fixing the problems created by the adjustment in question is also a matter of making larger adjustments to the practice of medicine.
“There are ways in which we can look at environmental factors, dietary factors, social factors, even access to care,” she says. “Is the test menu even appropriate? Did they get an ultrasound? Have they been seen throughout their pregnancy to make sure they have a healthy pregnancy?”
Dr. Pierre suggests laboratories talk with their Ob-gyn colleagues to dig into how they follow up with a positive AFP result. “Is that even their primary mode of screening?” she asks, noting that many have adopted high-resolution ultrasound instead. Other questions worth asking: “How do you counsel a patient? What do you do next? How do you confirm this result? Do you confirm this result? In what populations do you use AFP as your primary screen? And then how do you tackle those patients in terms of follow-up, to make sure we don’t cause unnecessary harm?”
As Dr. Peck Palmer and Dr. Pierre make clear, the problem is not only about the adjustment in particular (though they make clear it is indeed problematic); rather, it’s about medicine coming to terms with its past.
“We in medicine, just as a field, are starting to grapple with and untangle some of the implications of things we’ve done historically using race,” Dr. Pierre says. As with any topic having to do with race, this is relatively controversial, she acknowledges. Some are quick to respond defensively, she says, when they hear the words “race,” “racist,” and “racism,” often dichotomizing that into good versus bad. More useful, she says, is “recognizing that racism has to some extent been built into our systems, and something we’ve all been taught, regardless of what race we are.”
But it’s more than time to “take a step back from thinking about it being taboo and reconsider the way we use race,” Dr. Pierre adds. For those who are hesitant to label race adjustments as being racist, she has a matter-of-fact response: “You know what? The practice is racist,” she says, laughing. “But that doesn’t mean the people who made the equations are racist. People need to separate those two things—just treat it as something we have to actively unlearn. We saw a difference, and we thought that was significant, so we needed to correct for it. Now we need to think about it more closely and where the pitfalls and advantages lie.”
“Either way you look at it,” Dr. Pierre says, “when you start to stratify people and manage people differently, essentially you’re exposing one group to different conditions than another group. That’s the reality of the situation. And for something as imprecise as race, I think there are a lot of harms.”
Dr. Pierre also would like to scrutinize the literature more closely. If, as has long been maintained, the race adjustment makes the tool better, then how does it accommodate people with mixed ancestry? Furthermore, she asks, what is the evidence that these disorders are not being underdetected in patients who identify as Black?
“I’ve read the papers. I understand the arguments,” she says. “Just like I read all the papers around eGFR, and why people thought it should continue to be used. And I remain unconvinced.”
But others, including TriCore’s Dr. Grenache, say the data for using it are persuasive—that AFP is about 10 percent higher in pregnant people who are labeled or who self-identify as Black.
“We don’t know those reasons, and we should,” he says. But dropping it before the underlying reason(s) for the difference is found might lead to other inequities.
Dr. Grenache acknowledges the problems with earlier research, but he finds the analysis published this year in Clinical Chemistry, by Messerlian, et al., convincing, even as it troubles many. “It has caused some consternation among individuals who believe there is no place in medicine for race adjustments, for reasons that I don’t disagree with,” he says.
“Everyone involved in this conversation wants to do the right thing,” Dr. Grenache says. But in his view, “There’s simply not enough data yet to know exactly what the right thing is. My position is that we should stay the status quo, because the data are compelling enough to continue doing that. We have to understand the ‘why’—that will continue to be a goal. But we can’t ignore” differences between groups of patients, he says.
And for now, he says, the test is here to stay. “I don’t see any blood test that is going to replace AFP anytime soon.” He considers it to be a good test, and “absolutely imperfect” because of the race adjustment. “I acknowledge that. I hate to say it, but it’s the best we can probably do right now.”
“But I’m certainly not opposed to changing it,” he says. “I want to be able to answer the ultimate question: ‘Am I doing the right thing?’ I have to rely more on published data than my own data because I just don’t have enough of my own. I just don’t think I have the clearest answer.”
Geralyn Messerlian, PhD, and coauthors looked deeply into the data, and for now she thinks it makes a credible argument to keep using the adjustment. Dr. Messerlian is director of chemistry, Care New England Health System; professor in the Department of Pathology and Laboratory Medicine and the Department of Obstetrics and Gynecology, Warren Alpert Medical School of Brown University; and vice chair of pathology.
Her study emerged from her own interest in the topic, she says. Like many others, Dr. Messerlian was familiar with the growing literature focused on removing race from medical algorithms. Like many other institutions, hers had dropped race as a variable in the eGFR equation.
She asked the same question others have: We use race in neural tube defect screening. What would happen if we didn’t? “It’s an interesting and important question to ask,” she says. “I started with AFP rather than some of the more complex prenatal screening tests. Not everyone gets an ultrasound to look for neural tube defects, so AFP is likely to persist into the future.”
Her study is one study, and further research is needed, she says. “As scientists, it’s important to be open-minded and to be willing to have difficult conversations. We should examine data and talk about it. That’s going to be the best way forward, by having conversations.”
She continues: “The question I feel would be beneficial to answer is, What is the underlying basis of differences in AFP?” Since race is a social construct, she says, “we need to drill down to what the real cause of the differences in AFP levels are.” It’s complicated, she acknowledges, “but we could start to try to tease it apart.”
“We’re not the first study by a long shot to show that the levels are different when women self-report their race,” she says. “But why? What’s the underlying basis of those differences that we see in the marker levels?” She would like to replace the question of race with more specific questions that reflect possible contributing factors to serum hormone levels in pregnancy, including socioeconomic and environmental contributors. She’s confident the work will be done, though it will take large populations and time.
And in the meantime? “In my opinion, it would be better to know more before we make any changes to the prenatal program.”
It’s a difficult choice, as Dr. Messerlian acknowledges. Her study refers repeatedly and deliberately to “self-reported Black race.” It doesn’t say more than that, she says, but there’s something common in that self-report of race that is linked to altered levels. Among other possibilities, she says, it could be the experience of anti-Black racism, with the stresses related to that affecting placental physiology.
Laboratories routinely ask patients plenty of questions, she says, and in her view, patients can (and often do) choose to answer or decline to answer any of them, not only about race but also those related to smoking, weight, insurance, address, etc. But the lab’s work is far from done, she says. “I’m asking myself, What do I do next? How do I go forward? Prospectively, we could start to ask different questions that may help enlighten us.”
“But the broader context clearly is race and medicine, and the disadvantages that women who identify as Black are experiencing,” Dr. Messerlian adds. “Obstetrical outcomes are getting worse, and getting worse especially for women who are Black. That’s sounding alarm bells in our country.” In her own study, she says, it was apparent that Black women received prenatal care later in their pregnancies, at a later gestational age. “That’s the broader conversation—access to care and outcomes.”
“So in my mind, when we’re doing prenatal testing,” she continues, “I don’t want to contribute or to do anything that would make their experience of prenatal care even more difficult.”
She’s familiar with the suggestion that race should be removed from prenatal screening under any and all circumstances. “But in my mind, it’s important to know if we choose that route, what the impact would be.” Her concern is that doing so would further disadvantage some individuals who may already be struggling to get prenatal care in a timely manner. “The big issue is health equity.”
Like so many other things that seemed straightforward 50 years ago, maternal serum AFP screening is rooted in something quite complex.
“These algorithms were set up for quick triage,” Dr. Peck Palmer says. “But people don’t realize that when you’re using race-based algorithms, you’re not treating them as their own person; you’re treating them as a population. And we’re not trying to do population medicine; we’re trying to do personalized medicine. Why would we want to use something that has been so crudely used to disenfranchise people?” she asks.
It’s possible that looking at race has induced a different type of color blindness, one that obscures patients’ experiences, including how environment and geography can affect risk factors and access to care. Dr. Peck Palmer prefers looking through that wider lens. It’s not simply a matter of deciding to keep or drop the adjustment. AFP concentrations can fluctuate for a host of reasons, she says, “but race is not where we should be making adjustments. Pull your own data, and remove race, and look at weight, diet, past pregnancies, and other risks, like preeclampsia or other abnormalities from past pregnancies and miscarriages. That’s where we should be starting from.”
Rather than separate cohorts by race, researchers could use social determinants to design clinical studies. “That’s when people really start to peel back the layers and say, Oh, why did I design it this way? Why did I need to know Black versus white?” Dr. Peck Palmer says. A more encompassing approach benefits everyone; as she points out, the race adjustment to eGFR affected patients who weren’t Black, too.
Moreover, she says, “The lab is the owner of so much data. There are ways we could be really forward thinking, of saying we’re beyond race. It empowers us to say, We are the owners of this data. We have demonstrated that we are measuring what is going on with the patient. There is no need for any other modification or interpretation of our own work to demonstrate there would be a reason to change a value.”
“I think people get hung up on, We have to have race,” Dr. Peck Palmer says. “We don’t have to have race, actually. We just need to deliver care to that person in a personalized manner.” But thinking through its use more thoroughly “allows us to think about our own biases and stereotypes. Because we all hold them.” Indeed, she says, “It’s a daily reawakening for myself to make sure that those types of biases and stereotypes don’t actually affect how I interact with other people.”
Karen Titus is CAP TODAY contributing editor and co-managing editor.