Racial Discrimination Linked to Worse Pregnancy Outcomes

The Body Keeps Score Before the Baby Arrives

The numbers are brutal. A Black woman in the United States is three to four times more likely to die from pregnancy related causes than a white woman. Her baby is more than twice as likely to be born too early, too small, or not at all. For decades, the standard explanation has been a checklist of individual risk factors: income, education, access to prenatal care, smoking, obesity. Fix those, the logic went, and the gap would close.

Something else is happening inside the body, long before a woman steps into a doctor’s office. Something that a landmark meta-analysis published in BMJ Global Health has now quantified with uncomfortable precision.

Kim Robin van Daalen and her colleagues at the University of Cambridge and Harvard University combed through nearly 14,000 studies to ask a simple question: Does the experience of racial discrimination itself predict worse pregnancy outcomes? Their answer, published in 2022, is a statistical gut punch. Across 24 studies involving tens of thousands of women, the odds of giving birth preterm were 40 percent higher for women who reported experiencing racial discrimination (van Daalen et al., 2022). The odds of having a baby small for gestational age were 23 percent higher, though that finding was based on fewer studies (van Daalen et al., 2022).

No study found a protective effect. Not one.

This is not about poverty. This is not about genetics. This is about what happens when the social world gets under your skin and into your uterus.

How Do You Measure Something That Shouldn't Exist?

The tricky part of studying discrimination is that you cannot randomize people to be treated unfairly. You cannot, for ethical reasons, expose a control group to racism and see what happens. So researchers must work with what life hands them.

Van Daalen’s team did something rare in science. They performed what is called a systematic review and meta-analysis, which is essentially a forensic audit of every decent study ever done on a topic. They searched eight databases in any language, with no time cutoff, through January 2022. From 13,597 records, they whittled down to 24 studies that met rigorous quality standards.

The studies used a variety of designs. Some followed women forward in time (cohort studies). Some looked backward at women who had already given birth (case control studies). Some took a snapshot at a single moment (cross sectional studies). Twenty of the 24 studies were conducted in the United States, which makes sense given the country’s particular history of racial hierarchy, but also limits what we can say about other contexts.

The exposure they measured was self reported racial discrimination. This is not an objective count of racist incidents. It is a person’s lived experience of being treated unfairly because of their race, whether in housing, employment, healthcare, or everyday interactions. Some studies used validated questionnaires. Others used single questions like “Have you ever been treated unfairly because of your race?”

This is a limitation, but also a strength. Self reported discrimination captures something that administrative data cannot: the subjective, cumulative wear and tear of navigating a world that devalues you.

The outcomes they focused on were preterm birth (delivery before 37 weeks) and small for gestational age (birth weight below the 10th percentile for gestational age). Both are powerful predictors of infant mortality and lifelong health problems.

The quality assessment was rigorous. They used the Newcastle Ottawa Scale for individual studies and the GRADE approach for the overall body of evidence. When they excluded low quality studies, the association weakened slightly but remained significant. The odds ratio for preterm birth dropped from 1.40 to 1.31, still a 31 percent increase in risk (van Daalen et al., 2022).

The 40 Percent Number That Should Change How We Talk

Let me sit with that 40 percent for a moment.

An odds ratio of 1.40 means that if you take two women who are identical in every measurable way except one has experienced racial discrimination and the other has not, the woman who experienced discrimination has a 40 percent higher chance of delivering her baby too early.

This effect size is comparable to, and in some studies larger than, well known risk factors like smoking during pregnancy or maternal obesity. And unlike smoking, which a woman can (with support) stop, discrimination is not a behavior. It is something done to her.

The studies controlled for things like income, education, age, and medical history. The association persisted. This suggests that discrimination is not simply a marker of being poor or having less access to care. It has its own biological pathway.

The meta analysis also found something fascinating in the subgroup analyses. The association was stronger in studies that measured discrimination during pregnancy specifically, rather than lifetime experiences. This makes intuitive sense. The physiological stress of being treated unfairly while your body is building another human may be particularly potent.

The Biology of Being Told You Don't Belong

How does a social experience become a biological outcome? This is where the science gets both interesting and disturbing.

The leading hypothesis involves chronic stress. When you experience discrimination, your body activates the stress response system: the hypothalamic pituitary adrenal axis and the sympathetic nervous system. Cortisol and adrenaline surge. Heart rate and blood pressure rise. This is adaptive in the short term. It helps you flee or fight.

But when discrimination is a recurring feature of daily life, the stress response never fully turns off. The system becomes dysregulated. Cortisol levels flatten. Inflammation increases. Blood vessels stiffen.

Pregnancy is exquisitely sensitive to these changes. The placenta, that temporary organ that connects mother and baby, is packed with cortisol receptors. When maternal stress hormones are chronically elevated, they can trigger premature labor. Inflammation can damage the placental blood vessels, starving the baby of oxygen and nutrients.

This is not speculation. A growing body of research has linked experiences of discrimination to higher levels of C reactive protein, a marker of inflammation, and to shorter telomeres, the protective caps on chromosomes that shorten with stress and aging.

Van Daalen’s team did not directly test these mechanisms. Their meta analysis was about establishing whether the association exists. But they noted that the broader literature on discrimination and health supports a causal pathway mediated by stress and inflammation.

What This Research Does Not Prove

Here is where I need to be honest with you about what this study does not tell us.

First, the evidence is observational. You cannot prove causation from a meta analysis of observational studies. It is possible that women who experience discrimination also face other unmeasured risks that explain the association. The studies tried to control for confounders, but you can never control for everything.

Second, the vast majority of studies were conducted in the United States. We do not know if the same pattern holds in countries with different racial histories, different healthcare systems, or different forms of discrimination. A Black woman in Brazil, a Dalit woman in India, a Maori woman in New Zealand, they may have similar or different experiences. We need more global data.

Third, the measurement of discrimination is crude. Self reported discrimination captures only the incidents that people recognize and are willing to report. It misses subtle forms of bias, internalized racism, and structural discrimination that operates without any single identifiable perpetrator. If anything, this measurement error would tend to underestimate the true association. But it remains a limitation.

Fourth, the meta analysis could not fully disentangle the effects of discrimination from the effects of racism more broadly. Discrimination is one dimension. But racism also operates through segregated neighborhoods, under resourced schools, environmental toxins, and biased medical systems. These structural factors may be even more important than individual experiences of unfair treatment, but they are harder to measure and harder to isolate.

Finally, the number of studies on small for gestational age was only three. The pooled odds ratio of 1.23 was not statistically significant, meaning we cannot be confident that discrimination causes smaller babies. The trend was in the same direction, but the evidence is weaker.

Why This Changes the Conversation

If you have been following the literature on racial health disparities, this meta analysis may not surprise you. But it should sharpen your thinking.

For years, the dominant narrative has been that Black women have worse pregnancy outcomes because they are poorer, less educated, or less healthy. This meta analysis adds to a growing body of evidence that those explanations are incomplete. Even when you account for all the standard risk factors, the experience of being Black in a racist society still predicts worse outcomes.

This shifts the locus of responsibility. If the problem is discrimination, then the solution cannot be simply more prenatal vitamins or better health education for individual women. The solution must involve changing the social environment that produces the discrimination.

Consider what this means for clinical practice. A doctor who asks a pregnant patient about smoking, alcohol, and diet but never asks about experiences of discrimination is missing a major risk factor. Some researchers have begun developing screening tools for discrimination exposure. But screening is useless without a response. What do you do if a patient reports frequent discrimination? You cannot prescribe a medication for it. You cannot tell her to avoid it. You can acknowledge it. You can validate her experience. You can connect her with support. And you can advocate for systemic change.

What the Studies Actually Measured

Let me give you a sense of how the studies worked.

One study followed 1,035 Black women in Detroit. Researchers asked about everyday discrimination: being treated with less courtesy, receiving poorer service, being called names. Women who reported high levels of discrimination had twice the odds of preterm birth compared to women who reported low levels.

Another study used data from the Pregnancy Risk Assessment Monitoring System, a national survey. They asked women whether they had experienced racism during their pregnancy. The association with preterm birth was significant, even after controlling for income, education, and prenatal care.

A third study measured discrimination using the Experiences of Discrimination scale, which asks about nine specific situations: at school, getting a job, at work, getting housing, getting medical care, getting service in a store, getting credit, on the street, and from the police. Each additional domain of discrimination increased the risk of preterm birth.

The meta analysis pooled these studies using random effects models, which account for the fact that different studies have different populations and methods. The heterogeneity across studies was moderate, meaning the results were reasonably consistent.

The Gap in the Evidence

Here is what frustrates me as a science journalist. The paper calls for higher quality evidence from large, ethnographically diverse cohorts. That is science speak for “we need better data.” But who will fund it?

Research on discrimination and health has been chronically underfunded compared to research on individual behavior and genetics. It is politically uncomfortable. It implicates systems, not just individuals. It requires collecting data on race and racism, which some policymakers would prefer to ignore.

The few studies that exist are mostly cross sectional, meaning they measure discrimination and outcomes at the same time. This makes it hard to know whether discrimination preceded the outcome or vice versa. Longitudinal studies that follow women from before pregnancy through postpartum are rare and expensive.

There is also a need for studies that measure discrimination more precisely. Not just whether it happened, but when, how often, how severe, and in what contexts. Does discrimination in healthcare matter more than discrimination in housing? Does childhood discrimination have a different effect than discrimination during pregnancy alone? We do not know.

And we need studies that include objective measures of stress physiology. Cortisol patterns, inflammatory markers, epigenetic changes. These would help establish the biological plausibility and identify points of intervention.

What This Actually Means

Here is the bottom line, stripped of academic caution.

• ▸Discrimination is a biological event, not just a social one. The experience of being treated unfairly because of your race changes your physiology in ways that can harm a developing pregnancy. Clinicians should assess discrimination exposure the same way they assess smoking or nutrition.

• ▸The effect size is clinically meaningful. A 40 percent increase in preterm birth risk is comparable to many established risk factors. Ignoring discrimination means ignoring a major driver of racial disparities in birth outcomes.

• ▸Individual level interventions will not solve this. Teaching Black women to “manage stress” or “build resilience” places the burden on the victims of discrimination, not the perpetrators. The intervention that would have the largest effect is reducing discrimination itself.

• ▸Healthcare systems need to address their own racism. Many of the studies measured discrimination in healthcare settings. Black women report being dismissed, disrespected, and undertreated. This is not just a social injustice. It is a medical risk factor.

• ▸This is not about Black women being biologically different. The idea that racial disparities in pregnancy are genetic or inherent has been used to justify inaction. This meta analysis shows that the disparity is driven, at least in part, by a modifiable social exposure. That is good news, because it means change is possible.

The body keeps score. And the score, for too many women, is a baby born too soon. The question is whether we are willing to change the game.