I started noticing a pattern. Every Black woman I know who has given birth in the past decade delivered by cesarean. Some of those surgeries were scheduled in advance. Some followed hours of labor. Some were described as unavoidable emergencies. But the outcome was the same.
At first, I treated it as coincidence. Childbirth is unpredictable. Cesarean delivery can be lifesaving. Obstetric care is complex. No two births unfold the same way. But the pattern unsettled me, and I began to ask a different kind of question. Not just what was happening to these women, but what was shaping the information they were given when decisions were being made. What did these women actually know about how their risk was being calculated? And who, or what, had shaped that calculation before they ever walked into a consultation room?
These questions led me to vaginal birth after cesarean—commonly called VBAC. VBAC refers to attempting a vaginal delivery in a pregnancy after a prior cesarean section. National clinical guidelines state that many patients who have had one prior cesarean are candidates for a trial of labor. The primary concern is uterine rupture at the previous scar site (uncommon, but serious). But repeat cesarean is not a neutral default. It is major abdominal surgery, with increased risks of infection, hemorrhage, and dangerous complications in future pregnancies. Each additional cesarean compounds that exposure. For many patients, the decision extends far beyond a single birth.
But across the United States, VBAC became difficult to pursue even for patients who wanted it. Cesarean rates rose dramatically through the late 1990s and 2000s, reaching roughly one in three births today. During the same period, VBAC rates fell sharply, from around twenty-eight percent in 1996 to approximately eight percent by 2006. Liability pressures and institutional caution narrowed access to trial of labor.
Into that landscape came a prediction tool. In 2007, researchers from the Maternal-Fetal Medicine Units Network published a VBAC prediction model in Obstetrics & Gynecology. The model, commonly known as the MFMU VBAC calculator, estimated the individualized likelihood that a patient attempting labor after a prior cesarean would deliver vaginally. Clinicians could enter information available early in pregnancy: a patient’s maternal age, body mass index, prior vaginal delivery, reason for the first cesarean, race, and ethnicity. The calculator returned a percentage.
That percentage entered real conversations. An obstetrician might say: your predicted chance of successful VBAC is sixty percent. Or forty. Those numbers shape how risk is framed, how possibilities are imagined, how patients weigh their options. They feel like facts.
Race correction was built into the original calculator. Race correction refers to the practice of adjusting clinical estimates based on a patient’s racial classification. In the original MFMU model, patients categorized as Black or Hispanic were assigned lower predicted chances of successful VBAC than white patients with otherwise similar clinical profiles.
The justification was methodological. In the dataset used to build the model, VBAC success rates differed across racialized groups. Race was a statistically significant predictor. Including it, the reasoning went, improved accuracy. That reasoning sounds neutral. But it collapses a distinction that matters enormously.
To describe disparity is to observe a difference in outcomes. To interpret disparity is to decide what that difference means, and to let that interpretation shape clinical decisions. By placing race inside the calculator, the model treated race as something that should govern how a patient’s body is assessed and counseled. The question of why outcomes differed (whether because of differential hospital quality, access to continuous labor support, exposure to chronic stress, or provider bias) was not answered. It was bypassed. The association was attached directly to the patient’s racial category and converted into a number.
Here is what that means in practice. A Black woman arrives at a prenatal appointment. She is thirty-two years old, healthy, with one prior cesarean. She tells her provider that she wants to try for a vaginal birth. Her provider enters her information into the calculator. Before she has said another word about her body, her history, her wishes, before her testimony has been heard at all, the tool has already adjusted her predicted odds downward based on her race. Her provider receives a VBAC success probability percentage that has been shaped by who she is before it reflects anything she has said or anything clinically specific to her.
The information on which the patient is then counseled is not neutral. It has been pre-formed by a racial classification that stands in for a constellation of social conditions the calculator makes no attempt to name or address. When she says, I feel ready for this, I want to try: that testimony lands in a clinical encounter whose epistemic frame has already been set. Her self-knowledge meets a prejudged picture of what her body is likely to do.
This is not a matter of individual clinician bias. The clinician may be attentive, well-intentioned, and genuinely committed to patient-centered care. The distortion is not in their attitude. It is upstream, built into the inferential tool before the conversation begins. The wrong is structural, not interpersonal. And that is precisely what makes it so difficult to see and to contest.
This pattern of encoding race into medical inference has precedent. Nineteenth- and early twentieth-century obstetrics classified pelvic shapes into racialized typologies (“gynecoid,” “android,” “anthropoid,” “platypelloid”) with the so-called anthropoid pelvis frequently associated with Black women and described as predisposed to particular labor patterns. These classifications were presented as anatomical science. They were not. They were racialized assumptions laundered through measurement. Enslaved Black women were subjected to gynecological experimentation without anesthesia. Beliefs about racial differences in pain perception persisted in clinical education well into the twenty-first century.
The MFMU calculator did not invoke pelvic typology. Its authors used contemporary statistical methods. But the underlying interpretive move (when unequal outcomes appear in data, treat the racial category as explanatory) belongs to a longer history of treating social inequality as biological fact. This is not an isolated failure in obstetrics. It is a recurring tendency in American medical epistemology: to reach for race as an explanation when the deeper explanatory work, attending to structural conditions, institutional failures, and the physiological toll of discrimination, remains undone.
There is a further dimension. The woman in the consultation room typically does not know that race was a variable in the calculator. She has no access to the model’s construction, no language for what has been done to her risk estimate before it reached her. She experiences the VBAC success probability percentage as objective, as a fact about her body derived from medical science. She cannot contest what she cannot see.
If she feels that the consultation was somehow foreclosed, that her own sense of her body and her readiness was not quite heard, she has no framework to explain why. The very tool shaping her care is invisible to her.
The result is that she is doubly disadvantaged: first by the way her risk has been framed, and then by her inability to see how that framing was constructed. The distortion propagates silently through a conversation that appears, on its surface, to be evidence-based and patient-centered.
Maternal health outcomes in the United States make this urgency concrete. Black women are approximately three times more likely to die from pregnancy-related causes than white women. They experience higher rates of severe maternal morbidity, including hemorrhage and organ failure. These disparities persist across income and education levels. Vaginal birth carries documented protective benefits that repeat cesarean forecloses: shorter recovery, lower rates of postpartum hemorrhage, and reduced risk of placenta accreta in subsequent pregnancies.
For Black women who already face disproportionate mortality and morbidity from surgical complications, being systematically counseled away from VBAC does not neutralize risk—it redirects it toward the very outcomes they are already most vulnerable to. A tool that lowers predicted VBAC success for Black patients is not offering a conservative option; it may be pointing toward the more dangerous one. In that context, the systematic deflation of predicted VBAC success for Black patients is not just technical imprecision. It shapes the horizon of possibility in a domain where the stakes include life, future fertility, and surgical risk accumulating across pregnancies.
In 2021, after sustained criticism from clinicians and researchers, the authors of the MFMU model published a revised calculator that removed race and ethnicity as variables. The revision acknowledged that including race could perpetuate inequity. A commentator in the AMA Journal of Ethics argued that equitable access to vaginal birth requires abandoning race-based medical reasoning in favor of models grounded in clinical factors and structural conditions.
This was the right move. But removal alone does not resolve the deeper question. Race correction persists elsewhere in American medicine: in kidney function estimation, in pulmonary function testing, in cardiac risk models. Each instance poses the same problem. When disparity appears in data, do we encode the category as cause, or do we investigate the conditions producing it?
The answer matters because the category cannot do the explanatory work assigned to it. Race, as it appears in American medical records, does not reliably capture a physiological mechanism relevant to labor success. What it captures, at best, is a social position, shaped by histories of exclusion, differential access to care, the accumulated physiological effects of discrimination. To encode it as risk is to take the residue of injustice and use it to further constrain the options of those who have already borne its costs.
The pattern that first unsettled me does not prove that the VBAC calculator caused the cesarean deliveries I observed. Emergencies are real. Not every surgical delivery reflects a failure of justice. But learning that race had been built into the risk tool shaping those counseling conversations reframed what I thought I understood about how medical decisions are made.
We tend to think of a clinical encounter as a meeting between a patient’s knowledge of herself and a clinician’s knowledge of medicine. What the VBAC calculator revealed is that something else can enter that room first: a mathematical representation of who the patient is, constructed before she speaks, carrying the weight of her racial classification as though it were a fact about her body. By the time she says I want to try a VBAC, the conversation has already been shaped by an answer she was never asked to give.
The wrong that interests me begins before the conversation. It lives in the infrastructure of inference itself, in the tools that determine what counts as evidence about a patient’s body before she has a chance to speak. And when that infrastructure redirects Black women toward compounding surgical risk—foreclosing the protective benefits of vaginal birth for those already most vulnerable to its absence—the harm is not merely epistemic. It is bodily, cumulative, and borne most heavily by those who have already paid the highest price. Recognizing this kind of wrong, and taking it seriously, is not only a step toward better obstetric care. It is a step toward understanding what genuine epistemic justice in medicine would actually require: not just fairer conversations, but fairer calculations.
The Women in Philosophy series publishes posts on those excluded in the history of philosophy on the basis of gender injustice, issues of gender injustice in the field of philosophy, and issues of gender injustice in the wider world that philosophy can be useful in addressing. If you are interested in writing for the series, please contact the Series Editor Elisabeth Paquette or the Associate Editor Shadi “Soph” Heidarifar.
Zinhle Mncube
Zinhle Mncube is an assistant professor of Philosophy at the University of Massachusetts, Amherst. She is a philosopher of science with an interest in race and personalizing medicine. Zinhle’s current research concerns the epistemology and ethics of personalizing medicine, as well as the use of race in clinical and predictive diagnostic equations, guidelines, and tools.
