By looking at data on millions of births, a researcher shows how violence against unarmed African American men might cause acute stress on pregnancies.

In 2018 Joscha Legewie was reading The New York Times when he came across an article on health disparities between white and black babies. Black infants are twice as likely to die before their first birthday. Black women are three to four times more likely to die from complications during childbirth than white women. To explain the disparities, medical researchers were increasingly looking to the stresses of structural racism, the article explained. Legewie, a sociologist at Harvard, couldn’t help wondering if violence by cops might be one way those stresses entered mothers’ and babies’ lives.

His research, published Wednesday in the journal Science Advances, supports his early hunch. According to an analysis of police shootings of unarmed blacks between 2007 and 2016, such violent incidents can cause acute stress to pregnant black women in the area near an incident. The killings can even affect the health of their infants in utero.

Legewie gathered data on 3.9 million births in California, which are available to researchers, in anonymized form, through the state’s Department of Public Health. He focused on two markers of infant health—gestational age and birth weight—and compared them with the data on police shootings in California over the same nine-year period. Legewie’s findings suggest that police killings of unarmed black victims have an effect on black women in those communities, namely during the first and second trimester of their pregnancies. Later, when those babies were born, they had a lower birth weight on average than is typical for black babies. The infants also were more likely to be born several weeks early. The closer geographically mothers were to the shootings, the more severe the impact.

“It’s distressing, but I’m not surprised,” says Paula Braveman, director of the Center on Social Disparities in Health at UC San Francisco. Birth weight and gestational age are important predictors of both infant health and lifetime health. Low birth weight and preterm births are associated with physical and mental developmental problems. They can lead to the early onset of chronic illnesses like diabetes and heart disease. “If you start off as preterm or low birthweight, you have so many strikes against you,” says Braveman. “There’s only so much catching up that you can do.” Braveman’s work has produced similar findings, which suggest that black mothers’ chronic stress over racism could contribute to health disparities between black and white infants.

Last year, the American Public Health Association designated law enforcement violence a public health concern. One 2018 study found that shootings hurt black men’s mental health. “Gun violence, segregation, these things have not traditionally been public health issues. But there’s been a growing call for people to understand that everyday life affects your health,” says Margaret Hicken, a public health researcher at the University of Michigan.

Getting shot is undoubtedly bad for your health. Connecting police shootings with community health, however, is more complicated. When Legewie examined his data set, he tried to account for other factors that might affect the birth metrics, such as neighborhood violence and race. He compared nearly 250,000 black births with more than a million white and 2 million Hispanic births. White and Hispanic infants didn’t seem to be affected, and police shootings of unarmed victims of other races didn’t produce a strong effect either. He also didn’t find the same impacts when the black victims were armed.

Legewie even examined siblings, comparing women who had two pregnancies, one that coincided with a nearby police shooting and one that did not. He found that even in the same families, babies in utero close to a police shooting had worse outcomes than their siblings.

Quantifying the far-reaching effects of police violence is a complicated endeavor, in part because it’s difficult to get police data about shootings. Legewie used Fatal Encounters, a crowdsourced database, but spent years cleaning and coding the data. “I think he did a really good job,” says Melody Goodman, a biostatistician at New York University who wasn’t involved in the study. “It’s a really unique and interesting and good way for us to repurpose data that wasn’t collected for this purpose.”

It’s not clear what exactly is causing such a specific effect among exclusively black women. Legewie’s analysis didn’t turn up changes in infant health when mothers lived in the vicinity of generalized violence or police shootings of armed individuals. Legewie speculates that killings of unarmed victims are reminders of discrimination, which can exacerbate “stress and anxiety related to perceived injustice.”

Hicken suggests the issue might be more complicated and related to a concept called vigilance. “Black men and women need to continually anticipate and worry about how they are perceived in society,” she says. “They constantly need to prove that they are worthy of their humanity because our institutions don’t view it that way.” Police shootings of unarmed people could cause a spike in chronic stress and affect infants in utero.

These pressures and stresses may also be racially and culturally specific. Legewie was surprised he didn’t find any effect on Hispanic mothers. A previous study found that after an ICE raid in Iowa, Hispanic infants were also more likely to have lower birth rates. It may be that those events generate less media coverage, or that the history of unjust policing is represented in different ways: fears of deportation instead of fears about shootings. What hits one community might not hold the same emotional significance for another.

Legewie sees this research as offering a bird’s-eye view of the problem: It links police violence, stress, and infant health, but doesn’t puzzle out all the relationships that might be at play. The hope is that as more studies emerge, public discourse on the effects of police violence will evolve. “Doing this kind of work helps us think more broadly about the impact policing has,” Legewie says.