By Aaron E. Carroll, New York Times

Many more babies die in the United States than you might think. In 2014, more than 23,000 infants died in their first year of life, or about six for every 1,000 born. According to the U.S. Centers for Disease Control and Prevention, 25 other industrialized nations do better than the United States at keeping babies alive.

This is hard for some to comprehend. Some try to argue that the disparity isn’t real. They assert that the United States counts very premature babies as infants because we have better technology and work harder to save young lives. Therefore, our increased rate of infant death is not due to deficiencies, but differences in classification. These differences are not as common, nor as great, as many people think. Even when very premature births are excluded from analyses, the United States ranks poorly.

Even among those people who accept the statistic, most assume that high infant mortality is because of poor prenatal care. But new evidence is coming to light that contradicts that conclusion. This could change our thinking about the problem.

Infant mortality is not distributed equally in the United States. In 2013, the infant mortality rate among non-Hispanic whites was five per 1,000 births, as was the rate among Hispanics. The rate among non-Hispanic blacks was more than 11 per 1,000 births.

Mothers younger than 20 or older than 40 have children with a higher rate of mortality. Firstborn babies have a higher chance of death than later siblings. The children of unmarried mothers also have a higher rate of death, more than 70 percent higher than that for children of married mothers.

The No. 1 cause of infant mortality among newborns is premature birth, which has traditionally been linked with inferior prenatal care. That explanation may not hold true in the United States.

A 2006 study published in the journal Epidemiology looked at preterm delivery among women in active duty at a large military installation. Such women receive the same prenatal care regardless of race or socioeconomic status. Because they were guaranteed care, their overall risk of premature delivery was low, just over 8 percent. But even among these women, black women were more than two times as likely as white women to deliver prematurely, regardless of military rank.

A Cochrane systematic review of the additional support women received during at-risk pregnancies included 17 studies and more than 12,000 women. Additional care was not associated with any improvements in perinatal outcomes. Cesarean sections were less common, as was hospital admission after birth, but infant mortality was not affected.

Another such review examined how the number of prenatal visits affected infant mortality. Seven studies involved more than 60,000 women in countries of varying income. There was no difference in high-income countries in the number of deaths of those who had more or fewer visits. In low- and middle-income countries, perinatal mortality was higher in groups with reduced visits, but the overall difference was small.

The authors concluded that in places where the number of visits was already low, reducing the number of visits further was a bad idea. This does not necessarily apply to standard care in the United States, though.

A recently published paper in American Economic Journal: Economic Policy adds to this discussion. Alice Chen, Emily Oster and Heidi Williams combined data from the United States with data from Finland, Austria, Belgium and Britain. As in earlier studies, they adjusted for differences in coding of very premature births. And as other studies found before, the United States had a significant infant mortality disadvantage.

This study was different, however. It used microdata, or individual birth and death records, as opposed to the aggregate data usually employed for comparisons across nations.

First, the researchers differentiated between neonatal mortality (death before 1 month of age) and postneonatal mortality (death from 1 to 12 months of age). The results showed that when it comes to neonatal mortality, the United States and other countries were pretty similar.

Differences in postneonatal mortality, or from 1 month to 1 year, however, were much more stark.

This difference does not appear to be because of race. A subanalysis that excluded blacks from the sample found a similar postneonatal mortality disadvantage in the United States. Racial differences may be more applicable to neonatal mortality.

Deaths in the postneonatal period are due, in large part, to sudden infant death syndrome (SIDS), sudden death and accident — and seem to occur disproportionately to children of poor women.

It’s not clear that “health care” might reduce deaths in this group. That does not mean there is nothing we can do.

The authors of this paper estimated how much we might consider spending. They calculated that decreasing postneonatal mortality to that of comparable European countries might lower the death rate by one in 1,000. Assuming a standard value of $7 million per life, it might make sense to spend $7,000 per infant. That is not out of the realm of what we spend on many other medical interventions.

What exactly we might do with that money is up for debate.

But some things do seem evident. The first is that improved and more prenatal care may not significantly improve our disadvantage in infant mortality. The second is that spending a significant amount on poor women to improve the health of their 1-month-olds to 1-year-olds might not only save lives; it might be cost-effective, too.

To Learn More:

U.S. Has Highest Mortality Rate for Newborns of any Industrialized Country (by Noel Brinkerhoff and Danny Biederman, AllGov)

                Why Does the U.S. Have Such a High Infant Mortality Rate? (by Matt Bewig, AllGov)

Babies in U.S. More Likely to Die than Those in Cuba or Europe (by Noel Brinkerhoff, AllGov)