Characteristics of U.S. Youths With Serious Emotional Disturbance: Data From the National Health Interview Survey

The federal government invests a significant amount of resources in targeting services to youths with serious emotional disturbance. For example, in fiscal year 2006 the Substance Abuse and Mental Health Services Administration's budget for programs to serve youths with serious emotional disturbance exceeded $100 million. Youths with serious emotional disturbance are individuals younger than 18 years who currently, or at any time during the past year, have had a diagnosable mental, behavioral, or emotional disorder resulting in functional impairment that substantially interferes with or limits the child's role in family, school, or community activities ( 1 ).

National data on the characteristics of youths with serious emotional illness are critical for targeting resources to populations with the greatest need. For example, if most youths with serious emotional disturbance have private health insurance, then policies that focus only on care financed by public insurance programs may miss a large segment of the population. Similarly, if most youths with serious emotional disturbance live above the poverty line, then programs that concentrate only on communities of poor youths will also fail to reach large numbers of youths. Access and quality issues may exist for persons who are privately insured as well as for those with fewer resources.

Despite the need for national data, there has been no nationally representative survey of the prevalence and burden of child and adolescent psychiatric disorders in the United States ( 2 ). Extrapolations based on smaller epidemiological surveys indicate that there are six million to nine million youths and adolescents in the United States with serious emotional disturbance, representing 9 to 13 percent of all youths ( 3 ). Extrapolations from those surveys have been useful in providing ballpark estimates of the size of the population of youths with serious emotional disturbance. They can also be used to address questions about risk factors, such as whether poor youths are at greater risk of experiencing a serious emotional disturbance. However, epidemiological surveys that are based on convenience samples cannot address questions pertaining to the sociodemographic characteristics of the population of youths with serious emotional disturbance living in the United States.

The major task of national epidemiological surveys is to identify a large enough and representative enough sample of "true" cases, allowing national inferences to be drawn ( 2 ). In recent decades, most epidemiological studies of mental illness were conducted by translating disease taxonomies into interviews or questionnaires ( 2 ). For example, in an epidemiological study of youths in several locations in the United States, researchers developed and used the Diagnostic Interview Schedule for Children ( 4 , 5 ). However, because such diagnostic interview instruments are time intensive and expensive to administer, a large-scale nationally representative study of youths that has used diagnostic interviews has not been conducted in the United States. The United Kingdom, however, has recently carried out a national prevalence study by using the Development and Well-Being Assessment, a computer-assisted lay interview ( 6 , 7 , 8 ).

A more practical approach to gathering national data on youths with serious emotional disturbance is to use screening scales. Such screening scales may be used as a substitute for, or approximation of, psychiatric interviews ( 2 ). If such screens have adequate sensitivity in relation to the criteria for serious emotional disturbance, then they may be used to estimate population prevalence rates and identify characteristics.

In this article we examined data from a national household survey that included a screening instrument that has been correlated with psychiatric diagnoses to identify youths at high risk of serious emotional disturbance. The results provide new information on the characteristics and insurance coverage of youths with serious emotional disturbance and how they compare with all youths in the United States.

Methods

Data for this study came from the family component and the sample child component of the 2001 National Health Interview Survey (NHIS). The NHIS is a multistage probability sample survey of the health of the U.S. civilian noninstitutionalized population. It is a household survey conducted continuously throughout the year by face-to-face interviewers from the U.S. Census Bureau. Information about one randomly selected child in a household is obtained from an adult who is knowledgeable about the child's health, usually a parent. The interviewed sample for the sample child component in 2001 provided information on 13,579 youths from 0 to 17 years of age (80.6 percent response rate).

Youths with serious emotional disturbance in the survey were identified through their scores on the Strengths and Difficulties Questionnaire (SDQ). The SDQ was added to the NHIS in 2001 for youths aged four through 17 years and is a brief behavioral screening questionnaire that can be completed in five minutes by parents. The SDQ measure on the NHIS was selected for this study after a review of national household surveys that included measures of mental health status. This review indicated that although not perfect, the SDQ data from the NHIS was the best available resource for characterizing youths with serious emotional disturbance.

The SDQ consists of 25 items that are divided among five scales: emotional symptoms (five items), conduct problems (five items), hyperactivity or inattention (five items), peer relationship problems (five items), and prosocial behavior (five items). The scale items were selected on the basis of nosological concepts of the DSM-IV and ICD-10. Several studies attest to the reliability and validity of the SDQ ( 9 , 10 , 11 , 12 , 13 ). In addition, replication of psychometric properties has been achieved by investigators examining the German, Finnish, Bangladeshi, Swedish, and Dutch versions of the tool ( 14 , 15 , 16 , 17 , 18 ).

The combination of youth, parent, and teacher reports has been shown to have good specificity and moderate sensitivity in predicting the presence of a psychiatric disorder ( 19 ). Although the SDQ was not intended to measure or screen for specific DSM diagnoses, three of the content areas measured by SDQ items correspond to mood, conduct, and attention problems. Disorders associated with these problems are among the most prevalent conditions observed among youths with serious emotional disturbance. In a validation study of British youths aged five through 15 years, parent-rated SDQ scores above the 90th percentile increased the odds of having a DSM disorder 15-fold, and self-reported SDQ scores above the 90th percentile predicted a sixfold increase in risk of a DSM disorder ( 10 ).

In the study presented here, youths with an SDQ total difficulty score at the 90th percentile or above were considered to have serious emotional disturbance. Syntheses of epidemiological studies indicate that approximately 9 to 13 percent of youths have serious emotional disturbance ( 3 ). This fact, coupled with the research showing that youths in the 90th percentile of SDQ scores are at much higher risk of psychiatric disorders, led us to use the 90th percentile as the cutoff point for serious emotional disturbance. Analysis of the SDQ data from the NHIS has found that scores around the 90th percentile correlate strongly with contact with or use of mental health services ( 20 ).

The prevalence of serious emotional disturbance was examined by sociodemographic characteristics that are often used to direct resources to specific populations. Specifically, the characteristics of the youths examined included age, gender, race, ethnicity, poverty status, and insurance status. We also examined insurance status for users of mental health services among youths with serious emotional disturbance and in the general child population. Respondents were coded as receiving mental health care if they answered the following question affirmatively: "During the past 12 months, have you seen or talked to any of the following health care providers about [child's name]'s health? A mental health professional such as a psychiatrist, psychologist, psychiatric nurse, or clinical social worker."

Information on insurance status was based on respondent self-report. Respondents were offered a list of insurance types—for example, Medicare, Medicaid, or private insurance—and were asked whether they had that type of coverage. They were also asked to present an insurance card to confirm the coverage.

The frequencies of sociodemographic characteristics of youths identified as having serious emotional disturbance were calculated. They were compared with the characteristics of youths without serious emotional disturbance and the characteristics of all youths. The frequencies of the characteristics of the seriously emotionally ill population and of the remaining population were compared by using chi square statistics. All estimates were weighted by using the NHIS person-level weights. The weight was the product of the inverse probability of selection at each probability stage. The weights were also adjusted for participant nonresponse. We used the program STATA (SVY: tabulate) to account for the complex sampling design.

Results

Sociodemographic characteristics

The average age of youths with serious emotional disturbance in the sample was 10.4 years, essentially the same as that of all youths in the United States (10.5 years). As shown in Table 1 , compared with those without serious emotional disturbance and with the general population, youths with serious emotional disturbance were somewhat less concentrated at either end of the age range and slightly more concentrated in the 12- to 13-year age group.

Enlarge table

Most youths in the United States were Caucasian, and most youths with serious emotional disturbance were Caucasian. However, serious emotional disturbance was overrepresented among African-American youths. Overall, 14.9 percent of youths were African American, and 21.2 percent of youths with serious emotional disturbance were African American. Youths with serious emotional disturbance were also somewhat more likely to be Hispanic: 19.1 percent of youths with serious emotional disturbance were Hispanic, whereas only 16 percent of all youths were identified as being Hispanic.

The risk for serious emotional disturbance was more common at lower income levels, but most children with serious emotional disturbance were living at 200 percent of the poverty level or higher. Among youths living below 200 percent of the poverty level, the percentage of youths with serious emotional disturbance was 1.6 times as large as that of all youths (46.1 percent compared with 28.6 percent). A majority of youths with serious emotional disturbance, however, were living at 200 percent of the poverty level or higher, as were a majority of all youths.

Insurance coverage

As shown in Table 2 , private insurance was the most common type of insurance coverage among youths with serious emotional disturbance (43.6 percent). The next most common health insurance program was Medicaid and Children's Health Insurance Program (CHIP) (30.9 percent). About 13.6 percent of youths with serious emotional disturbance were uninsured at the time of the interview. The distribution of health insurance among youths with serious emotional disturbance differed from those without serious emotional disturbance and from the population as a whole. Nationwide, 67.2 percent of the population received health insurance from private sources, compared with only 43.6 percent for youths with serious emotional disturbance. Across the United States, 15.1 percent of youths received Medicaid and CHIP, compared with 30.9 percent of youths with serious emotional disturbance. Overall, 9.6 percent of American youths were uninsured, compared with 13.6 percent of youths with serious emotional disturbance.

Enlarge table

In order to understand funding sources for treatment of youths with serious emotional disturbance who are poor and those who are not poor, Table 2 describes the distribution of insurance coverage among youths from families with income less than 200 percent of the poverty level and among those with income at 200 percent of the poverty level or higher. Not surprisingly, for youths with serious emotional disturbance with family income less than 200 percent of the poverty level, the most common insurance coverage was Medicaid and CHIP (50.2 percent); next likely was being uninsured (19.1 percent), followed by having private insurance (17.2 percent).

Insurance coverage among mental health users

Table 3 presents insurance coverage data among mental health service users with and without serious emotional disturbance and for both groups together. For all groups, private insurance was the most common source of coverage among users of mental health services. The distribution of coverage for mental health service users with serious emotional disturbance was very similar to that for youths without serious emotional disturbance ( Table 2 ); however, this was not the case for the overall child group with mental health service use. For these individuals, Medicaid coverage was much more common among youths with serious emotional disturbance who used mental health services than among mental health service users without serious emotional disturbance (35.9 percent compared with 16.5 percent). Among youths without serious emotional disturbance who used mental health services, 72.2 percent had private insurance. This may indicate that parents with private insurance have a lower threshold before seeking help for their youths. It is also interesting that most youths who use mental health services would not be classified as having serious emotional disturbance. This is similar to epidemiological findings in the adult population ( 21 ).

Enlarge table

Discussion and conclusions

The goal of this study was to characterize the population of youths with serious emotional disturbance by using a nationally representative household survey. The results must be understood in light of their limitations. The most important of these is the use of the SDQ to identify youths with serious emotional disturbance. Although the SDQ has been widely used in the United States and other countries, it does not indicate a specific diagnosis and it does not directly measure functional impairment. Therefore, it is unlikely that the tool perfectly identifies youths with serious emotional disturbance who would be identified by studies using more extensive interview and diagnostic procedures administered by trained clinical personnel. Another limitation of the study is that it reports solely on youths with serious emotional disturbance living in households. It provides no information on the numbers or characteristics of youths who are homeless or living in institutional settings. Finally, the data are from 2001, and characteristics of this group may have changed since then. Despite these limitations, use of an SDQ cutoff of the 90th percentile as a method of identifying such youths allows for the examination of nationally representative information on their characteristics and insurance status.

Other questions from the SDQ have been recently used to examine behavioral problems among youths. Specifically, Simpson and colleagues ( 22 ), from the Centers for Disease Control and the National Institute of Mental Health, published a study that used data from the SDQ from the 2001, 2002, and 2003 NHIS to report on the prevalence of youths aged four to 14 years with difficulties in emotions, concentration, behavior, or ability to get along with others. In contrast to our study, the study by Simpson and colleagues used a section of the SDQ in which a respondent most knowledgeable about the child's health, usually the parent, was asked, "Overall do you think that [child's name] has difficulty in any of the following areas: emotions, concentration, behavior, or being able to get along with others?" They found that 4.8 to 5.5 percent of youths met this criterion. The prevalence rates from questions that ask parents to identify behavioral problems of their youths may be lower than that derived from a symptom-specific report because of lack of awareness or stigma.

The NHIS data presented here indicate that compared with the general population of youths, youths with serious emotional disturbance are overrepresented among low-income families, those with public insurance coverage or who are uninsured, and those who are African American or Hispanic. These findings are consistent with other studies that suggest that the prevalence of serious emotional disturbance is greater among certain sociodemographic populations ( 20 , 22 , 23 ).

Although the sociodemographic risk factors for serious emotional disturbance are important for public policy, they are not the only consideration for policy planning. Another important factor is the distribution of sociodemographic characteristics among youths with serious emotional disturbance. From this perspective, youths with serious emotional disturbance share many characteristics with youths in general. They are most likely to be white, live in families above the poverty level, and have health coverage through private insurance. Furthermore, their distribution across age groups does not differ much from that for the general child population.

A recent study by Warfield and Gulley ( 24 ) found relatively high levels of barriers to accessing mental health care among youths with special care needs whose parents reported them as needing mental health services. Specifically, 15 percent had trouble obtaining referrals, 27 percent had trouble finding providers with skills or experience, 11 percent had trouble getting enough visits, and 12 percent had access trouble because of the amount that the parents had to pay.

Studies of unmet mental health need by insurance status have found that youths who are uninsured are at greater risk of not receiving services ( 25 , 26 ). However, approaches to providing services to youths with serious emotional disturbance that focus solely on those with low incomes who are uninsured or who rely upon Medicaid for their insurance coverage will likely miss large segments of youths who experience access or quality shortfalls.

Medicaid coverage for youths with special needs has traditionally been more comprehensive than the coverage of most private insurance plans, including the provision of case management, rehabilitation services, personal care, psychological counseling, long-term residential care, and many other services. Private insurance has typically been geared toward acute treatment, with limits on the number of hospitalizations and relatively high coinsurance rates. Studies comparing youths with Medicaid to those with private insurance have found that publicly insured youths have access to mental health services that is greater or equal to that of privately insured youths ( 26 ). This study highlights the importance of efforts to learn more about the service utilization patterns of youths who are privately insured and have serious emotional disturbance. Such research could aid in the design of programs or policies directed to this population.

Acknowledgments

This study was funded by the Substance Abuse and Mental Health Services Administration (SAMHSA). The views expressed here do not necessarily represent those of SAMHSA or the Department of Health and Human Services. The authors thank Peter Jensen, M.D., Sherry Glied, Ph.D., and William Narrow, M.D., for their assistance in reviewing data sources for understanding youths with serious emotional disturbance. They also thank Derek Fugh, B.A., and Xue Song, Ph.D., for their programming assistance.