American parents may be more terrified than ever before that they will make errors that will have long-term negative impacts on their children. In this article, I argue that some of that terror is actually warranted. We know more than ever before about how relatively subtle factors previously thought to be harmless can hurt children, especially during the fetal period. But the bigger picture is that child health has improved enormously over the past 20 to 30 years in the United States, and that inequalities in children’s health have declined.
Some of this improvement no doubt reflects changes in parenting and parents’ behavior, and some of it reflects the enormous expansion of programs like Medicaid. While there are challenges that remain, such as increases in childhood obesity, it is important to put the bad news in the context of the great gains in children’s health and well-being over the past two decades. If we want to, we as a society can produce healthier children. The proof is that we already have.
The Fetal Origins Hypothesis:
The thalidomide episode was one of the first events to alert the general public to dangers that children face while still in utero. Thalidomide was licensed in 1957 and was widely prescribed to pregnant women for morning sickness until 1961, when it was identified as the cause of an epidemic of severe birth defects such as missing arms and legs. Researchers began to understand that there were other in utero shocks that could lead to severe and permanent abnormalities at birth. Another example is congenital rubella syndrome, which was first recognized in 1941 by Norman Gregg. Mothers infected with rubella in the first trimester of pregnancy have a 50 percent chance of delivering a baby with a severe problem such as cataracts, deafness, or heart disease. More recently researchers have realized that common influenza, at the wrong time in pregnancy, can also damage the fetus. Similarly, in 1973, Kenneth Jones and David Smith described “fetal alcohol syndrome,” a cluster of anomalies found in children of alcoholic mothers. Previously, the distinctive features and behaviors of this syndrome had been attributed to heredity.
Many researchers began to explore the so-called “fetal origins” hypothesis, which was named and propounded by David Barker, a British physician and epidemiologist. Barker focused on inadequate nutrition in utero and argued that it can “program” the fetus in a way that leads to future disease. For example, he argued that individuals starved in utero become more efficient in using calories, which makes them more likely to be overweight as adults, as well as more likely to suffer heart disease and diabetes.
Studies of the Dutch “Hunger Winter” of 1944 provide some of the sharpest tests of Barker’s hypothesis about prenatal nutrition. Food supplies in the Netherlands had been adequate up to October of 1944 when the Nazi regime halted food shipments to much of the country. By April of 1945, official food rations had fallen to 500 calories per day and starvation was common. Work by Aryeh Stein, L.H. Lumey, and others examined the cohort that was affected by famine in utero using military records for over 400,000 18 year-old men. Obesity rates were double among those who had been exposed to famine in the first trimester of pregnancy. Other studies associated with Ezra Susser using national Dutch psychiatric registries discovered increases in the prevalence of schizophrenia in the famine-affected cohorts, something that has also been seen among survivors of the great Chinese famine of 1959.
These studies clearly show that catastrophes such as the Dutch “Hunger Winter” have a long-run toll on a wide range of outcomes. But what of the kind of lesser, “everyday” health insults that pregnant women can suffer? Such insults include cigarette smoke, poor nutrition, stress, and environmental pollution. These things are quite common, and are also more likely to impact disadvantaged women; hence, if factors like these have important effects on fetal outcomes, then they are a possible source of health inequality that starts even before birth.
It is now well known that cigarette smoking during pregnancy is a leading preventable cause of low birth weight in the United States, and that smoking rates are much higher in disadvantaged communities. In turn, low birth weight is linked to a higher risk of infant death as well as to lower test scores and a higher incidence of asthma and ADHD in surviving infants.
Poor nutrition is also common. Although starvation became thankfully rare in the United States with the advent of food and nutrition programs like food stamps (now called the Supplemental Nutrition Assistance Program or SNAP) and WIC (the Supplemental Nutrition Program for Women, Infants, and Children), there are still many people who experience food insecurity, as well as people who consume enough calories in total (or too many calories) but still lack important micronutrients. A fascinating series of studies by Douglas Almond, Bhashkar Mazumder, and others focus on the effects of mild calorie restriction during Ramadan, when observant Muslims fast during daylight hours but eat after sundown. Since Ramadan cycles through the year, it is possible to identify cohorts of children who would have been affected in utero, and to distinguish Ramadan effects from seasonal fluctuations in nutrition. Cohorts subjected to calorie restriction in utero suffer effects on adult outcomes in a wide variety of settings (for example, Uganda, Iraq, Indonesia, and Denmark), and on a range of outcomes including educational attainment, test scores, adult height and weight, mental disability, and wealth.
Another strand of this literature looks at the effect of stress. One pathbreaking study by Anna Aizer, Laura Stroud, and Stephen Buka measured pregnant women’s cortisol, sometimes referred to as the stress hormone, and found that infants exposed to higher cortisol levels prenatally have up to one year less school at age seven than their siblings, indicating that they have been delayed in starting school or held back. Given rich background information on the mothers, this study was able to ask which children were most affected. The study found that children born to less educated mothers suffer larger impacts of exposure to a given cortisol level, suggesting that there may be fewer resources available to buffer the impact.
Other studies examine stressful events, such as the death of a close relative. One problem with such a study is that a relative’s death is likely to bring many unwelcome changes to the household, such as lower-income. Petra Persson and Maya Rossin-Slater use administrative data on a very large sample from Sweden and compare children affected in utero to those affected in early childhood. They find larger effects of deaths in the in utero period. In adulthood, children affected in utero were 23 percent more likely to use ADHD medication, and 9 percent more likely to use antidepressants compared to children affected in early childhood.
A large and growing body of work looks at the effects of in utero exposure to pollution on birth outcomes and on long-run outcomes. This work shows that even low levels of pollution exposure can have measurable negative effects on pre-maturity and low birth weight, placing children at risk. One innovative recent study by Adam Isen, Maya Rossin-Slater, and Reed Walker looks at the long-run effects of reductions in pollution caused by the Clean Air Act of 1970. The Clean Air Act required the implementation of pollution reduction measures in counties that were above thresholds set by the act for target pollutants. Counties just below the thresholds were not required to clean up. By merging data from several large administrative data sets, such as unemployment insurance records, the authors are able to link air pollution changes around the time of birth to the adult earnings and employment of people in their 30s. They find an estimated gain of US$4,300 in earnings per person for a total of US$6.5 billion (in 2008 dollars) in gains.
Interventions that Work
Evidence about the fetal origins hypothesis suggests that, in many instances, the best thing we could do for children would be to improve their mother’s health and circumstances prior to conception — so that good pro-mother policies are actually good pro-child policies.
Several US policy initiatives have this favor, including some that do not explicitly target pregnant women, such as anti-tobacco legislation. Through a combination of education, cigarette taxes, and bans on smoking, governments at all levels have succeeded in reducing smoking during pregnancy from 50 percent of all mothers in 1960, to 9 percent of mothers in 2013. This radical change in behavior is likely responsible for some of the dramatic reduction in infant mortality over the same period in the United States, from 26 deaths per 1,000 live births in 1960 to 5.8 deaths per 1,000 births in 2014.
In the late 1980s and early 1990s, a pathbreaking bipartisan effort on the part of federal and state governments led to an expansion of public health insurance for poor pregnant women, and then for their children. Initial evaluations by me and Jonathan Gruber showed that these changes improved access to care and reduced infant mortality. Recent studies follow surviving children to quantify the long-term effects of being eligible for public health insurance under the Medicaid program from before birth. These studies use the fact that the Medicaid expansions were phased in at different rates across the states (although the expansions eventually became mandatory in every state). Researchers have found lower childhood mortality, lower hospitalization rates for chronic conditions, and higher levels of college attendance, employment, and earnings among children who became eligible for insurance. The estimated effects are generally strongest for African-Americans, who were most strongly affected by the expansions, given lower average incomes.
Another program that has been shown to be effective in terms of improving short and long-term outcomes is the Nurse Family Partnership (NFP) program developed by David Olds, a modified version of which is being expanded through the Affordable Care Act. NFP programs provide nurse home visits to poor, unmarried, young women who are pregnant for the first time. The visits occur monthly during the pregnancy and during the first two years of the child’s life. Nurses provide guidance to pregnant women and new mothers regarding healthy behaviors, child care, and maternal development. The program has been shown to reduce child abuse and adolescent crime, and to improve children’s academic achievement.
All of this emphasis on the prenatal period may leave readers feeling that there is little that can be done after birth. However, several later interventions have been shown to be effective in improving children’s outcomes. Quality early childhood education programs have been shown to improve long-term child outcomes in low-income families. Evidence suggests that intervening during the pre-school years is often more cost-effective, although some interventions with adolescents have also been shown to work. Many evaluations of interventions later in life suffer from short follow-up periods. While only programs that start before the age of three seem to have long-lasting effects on IQ, a moment’s refection will confirm that non-cognitive skills are also important for success in life, and these skills can be improved into adulthood.
Means-tested income transfer programs have also been found to improve maternal health and child outcomes. Studies in Canada and the United States use changes in transfers to low-income working families to measure the effect of transfers on children’s outcomes. These studies show improvements in mother’s mental health and children’s test scores.
An intriguing recent study by Stephen Billings and Kevin Schnepel that is especially topical in light of the tragic epidemic of lead poisoning in Flint, Michigan, offers evidence that some of the negative effects of lead can be reversed with a multi-pronged intervention. In Charlotte, North Carolina, lead poisoned children who received lead remediation, nutritional and medical assessments, WIC, and special training for their caregivers, experienced reductions in antisocial behavior, and improvements in school performance.
The Big Picture
Altogether, the United States has invested considerably more resources in its children since the late 1980s. In addition to the Medicaid expansions, income transfers under the Earned Income Tax Credit (EITC) program have grown to make the EITC the largest “welfare” program in the country. Other programs such as WIC, Head Start, and SNAP have also grown. Public attitudes increasingly find smoking and drinking during pregnancy to be unacceptable, and the environment has become much cleaner. It should not come as a surprise then to find that children are healthier now than they have ever been by almost any measure.
For example, dramatic declines in infant mortality were discussed above. But in recent work with Hannes Schwandt, I have shown that mortality has also fallen for children and young adults between 1990 and 2010, and has fallen more quickly in poor areas than in richer ones. The declines in mortality have been especially steep for African-American children, an unheralded but extremely heartening development. This finding of reduced inequality in mortality among children contrasts with recent work showing that there have been increases in inequality in mortality among middle-aged and older adults. One obvious interpretation is that our greater investments in children are paying off, reducing inequality in mortality among vulnerable children at the same time that factors such as increasing economic inequality might have been expected to increase it.
While mortality is perhaps the least controversial measure of well-being, one sees similar trends in a broad array of other measures. For example, teen pregnancy has fallen continuously, as have smoking and drinking among teenagers. The share of people over 24 who have at least a high school degree has increased. While the diagnosis of some chronic conditions such as asthma, ADHD, and autism has been increasing over time, at least some of this increase is driven by improvements in diagnosis and by the availability of new treatments which have made diagnosis and treatment more beneficial.
Obesity rates are one measure that has been moving unambiguously in the wrong direction, at least until very recently. Obesity in children mimicked the climb among adults by rising steeply between the late 1980s and 2012. The most recent data suggest, however, that childhood obesity rates have begun to level off, and that obesity may even be declining among children under 11. While it is too early to tell for sure, it is possible that the increasing attention paid to childhood, and the range of measures aimed at reducing obesity — such as Michelle Obama’s “Let’s Move!” campaign, reforms of the National School Lunch Program, banishing vending machines from schools, and revising calorie labels — are starting to have an impact.
We now know that the fetus is uniquely vulnerable to myriad health insults, but we also know much more about how to protect infants from these shocks, and about how to remediate them after birth if necessary. Over the last several decades, we have seen a strengthening of the safety net for poor mothers and children, with the expansion and improvement of effective programs such as Medicaid, the EITC, NFP, Head Start, SNAP, and WIC. It is unfortunate that none of these anti-poverty programs are counted in official poverty measures, so that, as a society, most of what we do for poor children is “below the radar” and cannot (by construction) budge official poverty numbers. It is also unfortunate that the potential role fathers could have, especially when the parents are unmarried, remains largely unexplored.
Still, there is evidence that our investments in children are paying off, most dramatically in falling mortality rates and falling inequality in mortality among children, but also in a range of other indicators of child well-being. Obviously problems remain. But we have shown that the situation of poor children can be improved through thoughtful interventions targeting both children and their mothers. While our fears for our children are unfortunately often well-founded, so are our hopes.