In 2007, the Maternal Fetal Medicine Units Network (MFMU) created an algorithm, called a vaginal birth after cesarean (VBAC) calculator, to determine the probability of whether someone could have a successful VBAC. The algorithm incorporated race and ethnicity as a variable, along with factors like age, body mass index, and the reason for the previous cesarean. Due to the way the VBAC calculator factored race/ethnicity, it assessed Black and Hispanic patients as having lower chances for a VBAC, on average 5-15 points lower than White patients.

More than a decade later, people started to raise concerns about how the use of race and ethnicity in the VBAC calculator furthers structural inequities. Scholars criticized algorithms that used race more broadly for perpetuating the false idea that race is biological in nature, thus obscuring structural root causes of health disparities.

New research led by Nicholas Rubashkin, MD, and colleagues, used interviews with clinicians and patients, along with recorded prenatal visits, to analyze how the use of the VBAC calculator helped to “automate” racism by coding race into institutional practices and care interactions.

The researchers identified 4 processes that facilitated automation of inequities:

• Adhering to strict cutoffs. When clinicians used cutoff scores—often set around a score of 60% or less—below which they discouraged or banned VBAC, they facilitated the calculator’s unequal assessment. Clinicians reported varying rationales for using cutoff scores to determine who would be offered a vaginal birth. Some were required by institutional policies to use cutoffs, and clinicians struggled with these policies, arguing that cutoff scores shouldn’t override a patient’s preference. Thus, some clinicians continued to support VBAC-interested and low-scoring patients despite official policies forbidding low-scoring patients from attempting VBACs. Other clinicians believed the cutoffs kept patients safe.
• Routine adoption of calculators. Even in institutions that didn’t require the use of a calculator, it was adopted into routine counseling in many sites and became a standard for educating ob-gyn trainees. One Latinx maternal-fetal medicine specialist told researchers, reflecting on the automatic adoption of the calculator in their training: “I feel like I was indoctrinated to the calculator.. It’s like this script that we’ve passed on but at no point were we really talking about why are each of these steps so important.”
• Obscuring the use of the calculator and its factors. Clinicians often subjectively interpreted calculator scores as more or less favorable for the patient. Some shared their interpretations with patients without explicitly referencing the calculator or the factors that went into producing the score. When clinicians, often unintentionally, obscured the existence or the inner workings of the calculator, they made it more challenging for Black and Hispanic patients to identify the bias that went into their VBAC scores.
• Reflexively categorizing race and ethnicity. The calculator required clinicians to categorize patients into mutually exclusive racial and ethnic categories of Black, Hispanic, and white. Clinicians sometimes encountered challenges when entering Asian American, Indigenous, multiracial, multi-ethnic, or Afro-Latinx patients. One patient, who identified as Mixed-Latinx and -white ethnicity, was encouraged by her clinician to identify as white for the purposes of the calculator, That the calculator could have such flexible results based on her having a mixed-ethnic heritage, led her to an analysis of the calculator’s racism, stating “I was infuriated at that time…I was just thinking like, this just shows how arbitrary this entire process is. I’m like, if this calculator is based on science, then it shouldn’t be like that flexible.”

While some have expressed concern that the VBAC calculator systematically disadvantaged Black and Hispanic patients, this is the first qualitative study to document how that disadvantage worked out in practice. By naturalizing false biological notions of racial and ethnic differences contributing to “failure” rates, the VBAC calculator contributed to obstetric racism in the US.

Algorithms have the potential to inscribe historical racism into new information architectures, institutional practices, and human interactions. When clinicians used cutoff scores, they facilitated the calculator’s inequitable assessment. When they did not transparently show the calculator in their counseling, Black and Latinx patients had to do extra work to expose the inequitable assessment. However, some clinicians and some patients disrupted the calculator’s inequitable assessment by rejecting the role of race/ethnicity in calculating VBAC probability.

In 2021, amidst growing calls for the abolition of race-based medicine, the MFMU developed a new calculator that excludes race and ethnicity. While removing race and ethnicity helped mitigate the most negative consequences, racism might continue to operate implicitly. Racism may explain in part why Black and Latinx people undergo more unnecessary first-time cesarean births. Because the new VBAC calculator treats every prior cesarean as if it were clinically necessary, Black and Latinx birthing people may be more likely to be processed by the new VBAC calculator. If a more fair and just calculator exists, it would have to pay attention to the explicit and implicit ways that racism structures the risk for primary and repeat cesareans. Otherwise, the VBAC prediction model could perpetuate historical structural racism.