These policies were supposed to stop pregnant women from drinking. A new study says they’re hurting babies.

Few debates in maternal health are as heated as the one around drinking during pregnancy.

Experts agree that heavy drinking can harm a developing fetus, but there’s widespread disagreement on whether smaller amounts are safe. Meanwhile, efforts to protect future children by getting women to stop drinking — sometimes even before they become pregnant — have been criticized as paternalistic and unnecessary.

Now, new research suggests that those efforts could even be harmful.

In a study published on Wednesday in the journal PLOS One and provided to Vox exclusively ahead of publication, Meenakshi S. Subbaraman, a biostatistician at the Public Health Institute, and Sarah C.M. Roberts, an associate professor of obstetrics and gynecology in UCSF’s Advancing New Standards in Reproductive Health (ANSIRH) research group, looked at state policies designed to stop pregnant women from drinking. They found that several of these policies, including posting warning signs in bars and restaurants and defining drinking while pregnant as child abuse or neglect, are actually associated with worse health outcomes for babies, specifically low birth weight and premature birth. One reason, the researchers say, is that the policies can actually discourage women from seeking prenatal care.

“Usually, when we talk about harms from substance use during pregnancy, we talk about the harms from the use itself,” Roberts told Vox. But in the study, she and Subbaraman also found “harms from the policies that we adopt in response.”

The results have implications for how doctors treat pregnant women and how state governments approach maternal health. And the issue goes beyond alcohol — as states legalize marijuana and consider strategies to combat opioid addiction, they’re struggling to figure out how to approach the issue of maternal drug use. The study by Subbaraman and Roberts suggests that when it comes to protecting children’s health, scaring moms could do more harm than good.

Drinking during pregnancy can cause miscarriage, stillbirth, or a range of disabilities known as fetal alcohol spectrum disorders (FASDs), according to the Centers for Disease Control and Prevention. People born with FASDs can have problems with learning and memory, speech and language delays, and problems with organs like the heart and kidneys.

After an influential paper on the effects of maternal alcohol consumption was published in 1973, policymakers started trying to discourage pregnant women from drinking, OB-GYN Dr. Jen Gunter wrote at the New York Times earlier this year. The number of states with policies around alcohol and pregnancy — like requiring warning signs about the effects of maternal drinking in places where alcohol is sold — went from one in 1974 to 43 in 2013.

Roberts told Vox she started thinking about studying these policies after working with pregnant women in Northern California who used alcohol and drugs. In studying the barriers keeping some of these women from getting prenatal care, she found “that they were really scared of being reported to Child Protective Services and having their children removed.” Some of the women she worked with were also “really worried that they had already irreversibly damaged their babies, and that if they went to prenatal care, that they would get some confirmation of that” — and perhaps be judged by a doctor or have their use publicized in some way.

Roberts wanted to find out if what she saw in California held true nationwide, and how state policies might affect women’s fears. For the study published Wednesday, she and Subbaraman decided to look at the effects of a variety of policies, including mandating warning posters in restaurants and other places that sell alcohol and classifying drinking during pregnancy as child abuse or neglect. The researchers also looked at policies aimed at helping pregnant women who drink or use drugs, like giving them priority for substance abuse treatment or limiting criminal prosecution for drinking during pregnancy (for example, by prohibiting the use of prenatal medical test results in criminal trials).

Somewhat surprisingly, the researchers found that each of these four policies — including those designed to help women — had significant negative effects. They led to an increase in the number of babies born prematurely or at low birth weight.

The reasons for the effects probably vary, according to the researchers. In the 23 states that classify drinking while pregnant as child abuse, pregnant women who drink may be afraid to visit the doctor, and a lack of adequate prenatal care can lead to complications like premature birth.

Warning signs may seem harmless compared to threatening pregnant women with child abuse charges, but Roberts said the signs may also contribute to the kind of anxiety she saw among women in California. Many of them, she said, believed that by drinking during pregnancy, they had already done irreversible harm to the fetus. In addition to keeping women from visiting an obstetrician, this fear might even keep them from trying to quit drinking. Women may think “that they’ve already used earlier in their pregnancy, and so it doesn’t matter what they do now,” even though stopping drinking later in pregnancy could still benefit the fetus, Roberts said.

As she and Subbaraman write in their paper, it’s more surprising that limiting prosecution for pregnant women who drink, and giving them priority for treatment, can have ill effects. But, Roberts explained, it’s possible that states that give priority to pregnant women are the ones with a shortage of treatment opportunities to begin with — “it might be an indication of something else going on in the treatment delivery system more broadly,” which could affect the health of women across their reproductive lives.

State policies limiting prosecution for drinking during pregnancy, meanwhile, may have arisen in places where lots of pregnant women were prosecuted for drinking previously, Roberts said. The ill effects may be less about the limits and more about the ongoing fear of prosecution in those states, which in turn can cause women to fear seeking prenatal care.

Nationwide, state lawmakers have typically enacted policies like limiting criminal prosecution in reaction to more punitive policies, Roberts said, rather than by working with doctors and pregnant women to develop really effective ways to reduce drinking.

The research by Subbaraman and Roberts could help draw attention to the potential drawbacks of some state alcohol policies, said David Jernigan, a professor of health law, policy, and management at Boston University who was not involved in the study.

“It’s really important that they’re looking at this,” he told Vox, “because the last thing we need is not only ineffective policies but ineffective policies that target a vulnerable population” — that is, pregnant women.

Researchers already know that things like warning signs don’t actually work to reduce drinking, he said. What does work is reducing the physical availability of alcohol by limiting the hours and places where it can be sold, and increasing taxes on alcoholic beverages.

Raising alcohol taxes leads to fewer homicides, fewer drunk driving crashes, and fewer sexually transmitted infections, Jernigan said, and it would likely lead to less drinking during pregnancy as well. “At the end of the day, we’re all influenced by price,” he said.

In her earlier work, meanwhile, Roberts has found that policies that aim to reduce alcohol consumption among the general population, like restricting where alcohol can be sold, are associated with reductions in low birth weight babies and premature births.

The study published Wednesday, she said, suggests that when it comes to state alcohol policies, “a pause is in order” while policymakers figure out what really works. “What we need is to step back and start again and center the voices of people who are using during pregnancy and the people who take care of them,” she said.

That applies to drug policies too, Roberts said. Many of the policies she and Subbaraman studied, like classifying substance use during pregnancy as child abuse, apply to drugs like opioids too. And as states figure out how to handle legalized marijuana, they should take note of the research around alcohol warning signs before mandating similar signs for weed.

The findings also have implications for the way doctors talk to pregnant women, Roberts said. They may need to communicate more clearly, for example, that just because someone drank alcohol earlier in pregnancy, that doesn’t mean there’s no benefit to stopping later on.

“We need to be more careful about the health messages related to potential harms from use during pregnancy,” Roberts said. “Scaring people is not necessarily having the intended effect.”