Consequences of Childhood Mental Health Disorders

Featured in print Bulletin on Aging & Health

In recent years, there has been increasing recognition of the prevalence of childhood mental health disorders. One well-regarded study found that one in five children and adolescents in the U.S. exhibit some functional impairment from a mental or behavioral disorder, with one in nine experiencing significant impairment and one in twenty experiencing extreme impairment.

Despite the fact that millions of children in the U.S. suffer from mental health problems, relatively little is known about the long-term effects of childhood mental disorders. In "Mental Health in Childhood and Human Capital" (NBER Working Paper 13217), researchers Janet Currie and Mark Stabile examine the relationship between several common mental health conditions and future educational outcomes.

The authors use data from two large data sets, the American National Longitudinal Survey of Youth and the Canadian Na-tional Longitudinal Survey of Children and Youth. The comparison between the U.S. and Canada is of interest since children from the two countries would be expected to have a similar underlying propensity for mental disorders, yet reported prevalence and treatment rates for mental disorders are much lower in Canada.

The authors focus on several of the most common mental disorders, including Attention Deficit Hyperactivity Disorder (ADHD), depression and anxiety, and aggression. They examine a set of outcomes designed to capture the child's human capital accumulation, including grade repetition (a predictor of eventual school attainment), math and reading scores, enrollment in spe-cial education, and young adult delinquency.

This study offers several innovations relative to the existing literature. First, the authors use "screener" questions asked of all children in the survey to identify mental disorders rather than relying on diagnosed cases. The latter approach is problematic be-cause mental illness may be over-diagnosed or under-diagnosed relative to its true prevalence, creating the potential for bias in the estimation. Screener questionnaires are routinely administered to parents of troubled children as the first step in evaluating their child's mental health, so a high score on a screener questionnaire is a good indication of the existence of a mental disorder. The authors use the screener questions to determine a hyperactivity score, an emotional behavior score, and an aggressive be-havior score, as well as a measure of overall behavioral problems.

Second, while past studies have found that children who have mental disorders have worse outcomes, this association may reflect other problems suffered by these children. For example, prenatal exposure to alcohol, tobacco, or drugs is a known risk factor for childhood mental health disorders. Thus, the correlation between mental disorders and child outcomes could be captur-ing the long-term effects of prenatal factors rather than the effect of mental disorders per se. To address this, the authors focus on sibling comparisons, a strategy which allows them to control for omitted factors that are common to all children within a family.

Turning to the results, the authors find that children with higher hyperactivity scores have worse academic outcomes. A one-unit change in the hyperactivity score (where the average score is 5 and the maximum score is 16) increases the probability of grade repetition by about 1 percentage point, or roughly 10 percent relative to the mean. A one-unit change in the score also low-ers math and reading scores by about 5 percent of a standard deviation and raises the probability of being enrolled in special edu-cation by about 1 percentage point or 15 percent. Interestingly, there is no significant effect of hyperactivity on delinquency.

Aggression and emotional problems have adverse effects on child outcomes as well, though they are not as large. A higher aggression score raises the probability of delinquency and grade repetition and lowers math and reading scores. A higher depres-sion score raises the probability of grade repetition but has no significant effect on test scores, suggesting that cognitive impair-ment is not the primary mechanism for the grade repetition effect.

Where they can be compared, the results are quite similar in the U.S. and Canada. The effects of mental health problems are large relative to that of family income, which is known to be an important predictor of child outcomes. They are also large relative to the effect of chronic physical health problems.

The results raise an interesting question about why early childhood mental health problems affect educational outcomes - is it simply because they predict later mental health problems, or do they have an independent effect? The authors find that for grade repetition and special education, early mental health problems matter mostly because they predict later mental health prob-lems. However, in the case of test scores, these problems have an independent effect. This could be because even temporary childhood mental health problems affect test scores, or alternatively, because earlier onset of a mental health problem is associ-ated with a more adverse effect.

Perhaps surprisingly, the authors find little evidence that income protects against the negative effects of mental health prob-lems. They suggest that this may be because treatment for childhood mental health problem is still a relatively new phenomenon, making it difficult for even wealthy parents to identify and purchase effective care.

The authors acknowledge funding from the Annie E. Casey Foundation and the Social Science and Humanities Research Council of Canada.