Journal Issue: Children with Disabilities Volume 22 Number 1 Spring 2012
Childhood Chronic Conditions
Figure 1 documents a marked improvement in the health of U.S. children when infant mortality is the yardstick. However, as the infant mortality rate declined during the past half-century, public attention in developed countries shifted from acute fatal health problems toward chronic problems. While there is no doubt that chronic conditions are increasing in relative importance, it is often argued that chronic childhood illnesses are increasing in absolute importance as well.
Table 1 uses the childhood retrospective module that was placed into the 2007 wave of the Panel Study of Income Dynamics (PSID) by James P. Smith.12 The PSID has been tracking incomes and working conditions of a representative sample of American families on a yearly basis since 1967. Given the quality of its economic information, the PSID is an ideal data source to track the impact of poor childhood health on adult health and socioeconomic status. The major limitation of the PSID was the absence of information on childhood health, which was addressed by the retrospective module. This module enables us to combine childhood health data with excellent adult data on health and socioeconomic status. The information in the module focuses on the presence or absence of chronic conditions rather than on functional disabilities associated with those conditions.
Using calendar life-history methods, the initial set of questions asked respondents in the module whether—in the years up to and including age sixteen—they had any of a list of important childhood illnesses or conditions. This list included asthma, diabetes, respiratory disorders (such as bronchitis, wheezing, hay fever, shortness of breath, and sinus infection), speech impairment, allergic conditions, heart trouble, chronic ear problems or infections, epilepsy or seizures, severe headaches or migraines, stomach problems, high blood pressure, difficulty seeing with eyeglasses, mumps, measles, chicken pox, and three indicators of childhood mental health problems (depression, drug or alcohol problems, and other psychological problems). In the last category, there was no mention of the specific problem, such as attention-deficit/hyperactivity disorder (ADHD) or bipolarism.
Table 1 organizes reported prevalence rates of these childhood diseases by birth years of PSID respondents. Because the most recent group in the table was born in 1986, the data do not address rates of disability among younger groups of children who have not yet reached their adult years, which for our purpose we define as beginning at age twenty-five. These data offer a valuable and consistent picture of the consequences of poor childhood health in older individuals, where these pathways can be traced.
These data show several interesting patterns. First, when effective vaccines were developed, common childhood infectious diseases almost disappeared—first measles and mumps, and more recently chicken pox, for which a vaccine was developed in 1995.13 Second, it is difficult to read conclusive evidence on the direction of secular trends with regard to rarer childhood diseases—type 1 diabetes, hypertension, and epilepsy or seizures—although there may be an increase in the most recent birth years. Third, table 1 suggests that several common childhood diseases are becoming more prevalent. This is especially the case for respiratory diseases (asthma and respiratory illness), allergies, and depression.
There are reasons why the data in table 1 should not be taken at face value. Because the data are based on recall, memory biases may play a role. Memory typically declines with time, although salient events may suffer less from this memory decay, and memories of childhood have been shown to be superior to memories of other times of life.14 A second problem is the difficulty of separating true prevalence and incidence from improved detection. For most childhood diseases, diagnosis and detection have improved over time. For some diseases, including mental illness, there may also be lower thresholds for diagnosis, reflecting both medical advances and changing social attitudes. Finally, at very old ages, mortality selection effects, whereby the least healthy die at earlier ages, may be operative because those with childhood diseases may have lower life expectancies. However, selective old age mortality is not likely to explain the increasing trends among children born in the most recent birth years. Declines in infant mortality could lead to an alternative form of selection bias if unhealthy infants become increasingly likely to survive over time.
How serious are these sources of bias? The second form of mortality selection—healthier babies surviving to older ages—cannot be playing much of a role in the rise in childhood chronic illness or childhood disability, given the low rates of infant mortality evidenced in figure 1 for people who are now less than sixty years old. For younger age groups, trends in childhood chronic disease still appear to be growing over time.
One way of assessing how important recall bias could be is to use contemporaneously reported data on childhood chronic conditions. Even then, one difficulty is that statistics on American health, unlike those related to the U.S. economy for instance, do not generally reflect consistent, comparable reporting over time. Data on health conditions over time come from two long-running U.S. health surveys, the National Health and Nutrition Examination Survey (NHANES) and the National Health Interview Survey (NHIS), and both periodically have changed definitions of what is included within a disease category.15 Using subsets of childhood diseases that can be defined more or less consistently over time, figure 2 examines trends in reported rates of asthma, bronchitis, and hay fever. Similar to trends from recall data, all three childhood chronic diseases exhibit sharply rising secular trends. The similarity between the contemporaneous record and the PSID recall data indicates that recall bias is unlikely to be the primary driver of the secular trends in table 1.
Other studies using contemporaneously reported statistics also show increased rates of chronic illnesses among Americans. James Perrin and others documented substantial increases in childhood chronic illnesses such as obesity, asthma, and ADHD in the United States.16 Jeanne Van Cleave and others, using data from three National Longitudinal Survey of Youth groups aged two to eight, reported that the prevalence of any chronic health condition was 12.8 percent for a group in 1988 that was followed to 1994, 25.1 percent for a group in 1994 followed to 2000, and 26.6 percent for a group in 2000 followed to 2006.17 Using data from the Centers for Disease Control and Prevention, Lara Akinbami and others showed an increase in the prevalence of childhood asthma from approximately 3.6 percent in 1980 to about 9.7 percent in 2007.18 Some 14 percent were reported as either currently having or having once been diagnosed with asthma during their lifetimes, based on the 2009 National Health Interview Survey.
Rising rates of chronic diseases among children present a puzzle in light of rapidly declining infant mortality rates. And because many indicators of adult health have been improving over this period, questions arise about the extent to which childhood health contributes to adult health, and more basically the extent to which chronic childhood conditions are actually increasing.
Some of the major factors thought to contribute to better childhood health have been improving rather than worsening. Table 2 focuses on some determinants of child health and shows that the proportion of children who grew up in a home where neither parent smoked has been rising and that the proportion of PSID respondents who thought that they grew up in a poor family, as they self-defined what poverty meant, has been declining over time. While this may seem surprising in light of today's headlines about rising levels of child poverty, the period when these PSID adults were children was a time of significant declines in U.S. poverty, including among children. Although older mothers (those age thirty-five or older) are a risk factor for poor childhood health, once again we see declining trends in table 2. Many environmental problems (like air pollution) related to children's health are being corrected, although it is possible that new environmental toxins are being introduced.
While childhood obesity rates have risen rapidly in recent years, figure 3 demonstrates that most of that rise in childhood obesity affected the youngest age groups in table 1 and hence cannot be responsible for the table 1 trends. Figure 4 indicates that there has been only a small rise in low-birth-weight babies over time.
Although there is almost universal agreement that reported rates of childhood chronic conditions are rising, we believe that any conclusion about rapidly rising rates of childhood chronic physical health conditions over time are premature at best, especially concerning the magnitude of trends. More work is needed to separate out the impacts of improved detection and diagnosis. The real trends in health may be nowhere near as dramatic as suggested by simple time-series of reported prevalence rates of childhood disease. Documenting real changes in prevalence of specific diseases is a high-priority research topic.
One area of greater confidence about deteriorating trends over time concerns childhood mental health issues. As table 2 shows, one of the drivers of depression in childhood—being raised in the absence of both parents—has been worsening over time.19 Similarly, in a careful study, Marissa King and Peter Bearman showed that changing diagnosis alone does not explain the increasing rates of childhood autism, even though up to half of the reported increase might be due to changes in reporting and definitions.20 The PSID childhood retrospective module can also display trends based on family income. Table 3 documents trends for children who lived in households with family incomes above and below the median. Historically, the reported prevalence of the three childhood infectious diseases was greater among better-off American families, as was the effectiveness of vaccines for measles and mumps, reflecting the greater availability of vaccinations to the above-median income group. The pattern of rising prevalence of childhood disease is true both above and below the median income, but with few exceptions (respiratory illnesses, allergies, ear problems) in the most recent birth groups, prevalence rates are higher for children who grew up in below-median income families. Increases in reported rates of diabetes and hypertension in the below-median income group are especially striking. Our findings are consistent with a large body of research showing that children from low-income families experience higher prevalence rates for the main childhood health problems.
Prevalence rates also vary by race. For example, Akinbami and others, using one measure of social-economic differences, racial differences in asthma prevalence, show Asian American children having lower prevalence than whites, black children having 1.6 times the prevalence rate of whites, and Puerto Rican children having 2.4 times the prevalence rate of whites.21 For more on socioeconomic disparities, see the article in this volume by Neal Halfon and others.22