Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
2 The Nature and Extent of the Problem E pidemiology, the basic science of public health (Rothman and G Â reenland, 1998), provides vital information about diseases that threaten the health and well-being of the population. Epidemiology provides basic information that can be used to identify where and what kind of prevention is needed and to monitor the success (or failure) of preventive interventions. In order to be of use in the prevention of mental, emotional, and behavioral (MEB) disorders, epidemiology must provide information about which individuals are suffering from or at risk for mental, emotional, or behavioral problems, at what ages or developmental stages, and must be able to assess whether interventions have reduced the prevalence of a disorder. National surveys of adults have shown the extent of the problem. In the early 1990s, the National Comorbidity Survey (NCS) of mental illness in the United States showed that more than one in four (26.2 percent) adults had a mental disorder in the 12 months up to the time of the survey (Kessler, Anthony, et al., 1997). The NCS-Replication (NCS-R) a decade later reported this figure as close to one-third (Kessler, Chiu, et al., 2005). In these and other surveys, roughly half of all affected adults recalled that their mental disorders started by their mid-teens, and three-quarters by their mid-20s (Kessler, Berglund, et al., 2005). However, studies of young people themselves are needed to establish accurately when MEB disorders first occur and what their consequences are in terms of chronicity, impaired functioning, and impact on their ability to reach developmental milestones, such as graduating from school, finding work, and forming adult relation- 35
36 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS ships. The NCS-R includes a sample of over 10,000 adolescents ages 13 and older, but the findings are not yet available. MEB disorders in young people are a public health concern for sev- eral reasons: (1) they cause suffering to individuals and their families; (2) they limit the ability to reach normal goals for social and educational achievement; (3) they increase the risk of further psychopathology, func- tional impairment, and suboptimal functioning throughout life; and (4) they impose heavy costs to society because of the resultant need for extra care, the social disruption that they can cause, and the risk that affected young people will underperform as adults. The significant economic costs of treating disorders warrant an increased focus on preventing them (Smit, Cuijpers, et al., 2006). However, support for prevention programs depends on knowing the size of the problem and its societal burden and on being able to monitor reductions in that burden when prevention programs are put in place. The United States is significantly behind other countries in supporting the necessary information-gathering programs. In this chapter, we review the evidence available from epidemiological studies to answer the following questions: â¢ What kind of research methods and data are needed to answer questions about areas of high priority for prevention? â¢ How prevalent are MEB disorders of major public health concern? â¢ Is prevalence increasing or decreasing? â¢ How many new cases are there (incidence)? â¢ Is incidence increasing or decreasing? â¢ At what age do diagnosable disorders first occur (onset)? â¢ What is known about factors affecting prevalence, incidence, and age of onset? â¢ Are rates of these factors increasing or decreasing? â¢ Are some groups at particularly high risk for specific disorders? Chapters 4 and 5 provide additional information related to the fac- tors that affect the prevalence of disorders and define high-risk groups. A closely related set of questions deals with the cost to society of the harm caused by MEB disorders and the cost-effectiveness of prevention. These are addressed in Chapter 9. research methods and data The prevention of disease is a challenge for the whole community, not just for clinicians and their patients. Prevention is, by definition, an intervention that occurs before it is known who will develop a disorder and who will not. It follows that epidemiological information about whole
THE NATURE AND EXTENT OF THE PROBLEM 37 communities (or representative samples of whole communities) is usually needed to answer questions about prevalence (the total number of cases in a given period of time) and incidence (the number of new cases in a popula- tion). In addition, many young people have more than one MEB disorder (Angold, Costello, and Erkanli, 1999). This comorbidity can increase the severity of a disorder (Kessler, Chiu, et al., 2005). Rates of comorbidity can- not be determined using clinic-based data, because cases seen in treatment settings are different in many ways from untreated cases (Berkson, 1946). Population-level information is needed to determine which diseases are of public health concern. It needs to encompass a wide range of disorders, including their rates of occurrence and co-occurrence and the burden they cause to individuals, their families, and the social organizations and agen- cies in which individuals live their lives. The standard method of finding out how many cases of a disease exist in the community is to carry out a randomized survey of the general popu- lation. The size of a sample needed to provide precise answers to questions about the prevalence of an emotional or behavioral disorder depends on how common or rare it is. The less common the disorder, the larger the sample needed to provide a reliable prevalence estimate. For example, if a disorder occurs in 1 child in 10,000, researchers would need a population sample of at least 1 million children to find approximately 100 cases. If a disorder produces such a high level of disability that every case comes to the attention of doctors, schools, or other agencies, then agency records can sometimes be used to estimate prevalence and even incidence. This method has been used by the Centers for Disease Control and Pre- vention (CDC) to estimate the prevalence of autism. In some countries, databases of inpatient and outpatient treatment are maintained and can be used to estimate treated prevalence. But many MEB disorders rarely come to the attention of doctors or teachers. Studies in the United States show that fewer than one in eight children with an MEB disorder is currently receiving treatment in the mental health or substance abuse systems, and only about one in four has ever received treatment (Burns, Costello, et al., 1995; Farmer, Burns, et al., 2003; Kataoka, Zhang, and Wells, 2002). To estimate the full burden of MEB disorders among children and adolescents, it is usually necessary to interview large community-based samples of par- ents and their children. As mentioned earlier, there have been two recent surveys of mental illness in representative samples of the U.S. adult population: the NCS (Kessler, 1994), a follow-up of the same participants (NCS-2) (Kessler, Gruber, et al., 2007), and a second sample (NCS-R) a decade later (Kessler, Chiu, et al., 2005). The NCS included no one younger than 15. The NCS-R includes a sample of 10,000 adolescents (ages 13-17), but the data on this sample are not yet published. Although the United States supports several
38 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS national surveys of health and drug abuse, these include very little on child and adolescent mental illness, and so there are almost no national preva- lence and incidence estimates. Table 2-1 is a summary of various nationally representative studies, spon- sored by federal agencies, that have made some effort to produce estimates of the prevalence of MEB disorders of youth and, in some cases, the need for or use of mental health services. There is a dramatic contrast between the richness of the data on drug use and abuse from the National Health and Nutrition Examination Survey (NHANES), the National Survey on Drug Use and Health (NSDUH), and Monitoring the Future (MTF), and the paucity and lack of continuity of measures of MEB disorders. MTF has been collect- ing information on drug use and abuse since 1975, and NSDUH since 1988. However, the latter added some mental health questions only in 1994, and the results have not yet been published. NHANES used selected modules of a diagnostic interview for about five years, but since 2004 has limited its rel- evant data collection to a screener for depression for two years (2005, 2006) and some questions about conduct disorder since 1999. For three years, the National Health Interview Survey (NHIS) included the Strengths and Diffi- culties Questionnaire (Goodman and Gotlib, 1999), a 25-item parent report that produces symptom scales but not diagnoses. The current NHIS includes only three to five mental health questions. The new National Childrenâs Study, which will begin recruiting participants in 2009, offers a wonderful opportunity for nationally representative, longitudinal data collection on the development of MEB disorders, the need for services, and the role of preven- tion and treatment in their course. No plans have been published for the data to be collected beyond the first few months, so it is unknown whether this opportunity will be realized. Given the limitations of national surveys, conclusions about prevalence and incidence of MEB disorders among young people have to be drawn from (1) national surveys from other countries and (2) local population surveys in the United States. Despite being the best available data, both of these also have limitations. In the first case, rates can be very different in different countries, so that extrapolation to the United States is difficult. For example, using the same diagnostic interview (Development and Well-Being Assessment) with 8- to 10-year-olds in three different countries produced rates of conduct disorder in Norway that were much lower than those found in the United Kingdom (Heiervang, Stormark, et al., 2007) or the United States (see below). Within the United States, local surveys also show variation in rates. For example, in a set of studies using identical methods, the prevalence of disruptive behavior disorders was lowest in Puerto Rican youth living in Puerto Rico, higher in mainland Hispanic and white youth, and highest in mainland African Americans, even after controlling for a range of risk factors (Bird, Canino, et al., 2001).
TABLE 2-1â Review of National Surveys Providing Information About Emotional and Behavioral Disorders in Youth Survey and Agency Relevant Information Collected Design and Comments National Health Interview Adult respondent is asked whether a doctor has ever told the Cross-sectional household interview survey. Survey (NHIS) respondents that the child has mental retardation, Sampling and interviewing continuous Agency: National Center developmental delay, ADHD, or autism, or if a school or throughout each year. Multistage area for Health Statistics (CDC) health professional has said that the child had a learning probability design. Oversampling of both disability. blacks and Hispanics. Sample size ~43,000 In 2001, 2002, and 2003, ~10,000 adults completed 25-item households (~106,000 persons) annually. For Strengths and Difficulties Questionnaire (SDQ) for children, information provided by a children ages 4-17. responsible adult family member. From 2007 on, 3-5 mental health questions asked (depending on childâs age and sex). No diagnostic information. National Health and Age/topic/method/dates: Examines a nationally representative sample of Nutrition Examination 12+/Depression screener/(CAPI)/2005, 2006 about 5,000 persons each year. 15 counties Survey (NHANES) 12-19/Alcohol use/(ACASI)/1999 on are visited each year. Agency: National Center 12-19/Conduct disorders/(ACASI)/1999 on for Health Statistics (CDC) 12+/Drug use/(ACASI)/1999 on 12-19/Tobacco use/(ACASI)/1999 on 8-19/Eating disorders/(CDISC)/2000-2004 8-19/Depression/(CDISC)/2000-2004 8-19/Panic and anxiety/(CDISC)/1999-2004 8-15/ADHD/(parent CDISC)/2000-2004 8-15/Conduct disorders/(parent CDISC)/2000-2004 8-15/Depression/(parent CDISC)/2000-2004 8-15/Eating disorders/(parent CDISC)/2000-2004 8-11/Elimination disorders/(parent CDISC)/2000-2004 No data yet published. 39 continued
TABLE 2-1â Continued 40 Survey and Agency Relevant Information Collected Design and Comments CAPI = computer-assisted personal interview ACASI = audio computer-assisted self-interview CDISC = computerized diagnostic interview schedule for children Youth Risk Behavior Covers tobacco use, unhealthy dietary behaviors, inadequate Since 1990, monitors health risk behaviors using Surveillance System physical activity, alcohol and other drug use, sexual self-report questionnaires administered in (YRBSS) behaviors that contribute to unintended pregnancy and school. Agency: National Center sexually transmitted diseases, including HIV infection, for Chronic Disease behaviors that contribute to unintentional injuries and Prevention and Health violence. No MEB disorders included. Promotion (CDC) National Survey on Drug Designed to produce drug and alcohol use incidence and Running since 1988 (formerly National Use and Health (NSDUH) prevalence estimates and report the consequences and Household Survey on Drug Abuse). Extensive Agency: Office of Applied patterns of use and abuse in the general U.S. civilian data on drug use, including age at first use, Studies, Substance Abuse population ages 12 and older. Since 1994, questions added lifetime, annual, and past-month usage for and Mental Health Services on mental health and access to care. Treatment for youth alcohol, marijuana, cocaine (including crack), Administration ages 12-17 is defined as receiving treatment or counseling hallucinogens, heroin, inhalants, tobacco, for problems with behaviors or emotions from specific pain relievers, tranquilizers, stimulants, and mental health or other health professionals in school, sedatives; substance abuse treatment history home, or from other outpatient or inpatient settings in the and DSM-IV diagnoses. past year. A module on lifetime and past year prevalence of major depressive episode (MDE), severity of the MDE as measured by role impairments, and treatment for depression was administered to adults ages 18 or older and youth ages 12-17, from 2004 to 2006; 8.5% of youth had an episode of MDE in the past 12 months (see http://oas.samhsa.gov/2k8/youthDepress/youthDepress. pdf).
Monitoring the Future Has been collecting self-report anonymous data on drug use Ongoing study of the behaviors, attitudes, and (MTF) since 1975. Is able to show rise and fall of use of different values of U.S. secondary school students, Agency: National Institute drugs. college students, and young adults. Each year, on Drug Abuse ~50,000 8th, 10th, and 12th grade students are surveyed (12th graders since 1975 and 8th and 10th graders since 1991). Annual follow-up questionnaires mailed to a sample of each graduating class for several years. National Survey of Questions asked for ADHD, depression, anxiety, oppositional One-time survey (2007-2008) of ~86,000 Childrenâs Health defiant disorder, behavioral or conduct problems, autism, children ages 0-17. Data collected from Agency: Maternal and developmental delay, Tourette syndrome: responsible adult by telephone. Child Health Bureau of the Has a doctor or other health care provider ever told you Health Resources and that selected child (SC) hadâ¦? Services Administration Does SC currently haveâ¦? Would you describe his/her â¦ as mild, moderate, or severe? In case of ADHD, a fourth question is asked: Is SC currently taking medication for ADD or ADHD? Results not yet published. National Childrenâs Study No information yet collected. So far, there is no planned Will examine the effects of environmental Agency: National Institute collection of information on emotional or behavioral influences on the health and development of for Child Health and disorders. ~100,000 children across the United States Human Development from before birth until age 21. Congress authorized the National Childrenâs Study with the Childrenâs Health Act of 2000. Will take place in 105 representative counties around the United States. 1,000 mothers and their children will be recruited from each site and followed for 20 years. 41
42 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS Precise estimates of the size of the problem of MEB disorders of youth in the United States, or changes in the problem over time, require nationally representative population surveys that make valid and reliable diagnoses. However, as discussed below, the consensus from a large number of recent studies with smaller samples or from other countries provides a ballpark estimate. prevalenCE of MENTAL, Emotional, AND Behavioral DISORDERS Clinical psychiatry has mapped out a range of MEB disorders and related problems seen in children and adolescents. These are listed in the two main taxonomies of disease, the section on mental and behavioral disorders in the International Statistical Classification of Diseases and Related Health Problems (ICD) (World Health Organization, 1993) and the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM-IV) (American Psychiatric Association, 1994). Some other major public health problems, like crime and violence, are subsumed within the diagnostic criteria for conduct disorder. The disorders examined in this chapter are those in the American Psychiatric Associationâs DSM-IV. The DSM-IV includes abuse of and dependence on alcohol and illicit drugs, as well as dependence on tobacco. This section reviews current epidemiological information about the more common MEB disorders up to age 25: conduct disorder and oppo- sitional defiant disorder, often combined as disruptive behavior disorders; attention deficit hyperactivity disorder (ADHD); anxiety disorders, including posttraumatic stress disorder; depression; and drug abuse and dependence. Disorders of low population frequency, with little reliable epidemiological data but considerable societal burdenâsuch as autism spectrum disorders and pervasive developmental disorders, schizophrenia, bipolar disorder, eating disorders, and obsessive compulsive disorderâare discussed when information is available. More specific information may be available when the adolescent version of the NCS is published. Table 2-2 presents the results of a meta-analysis of data on the preva- lence of MEB disorders in young people from more than 50 community sur- veys from around the world, published in the past 15 years (updated from Costello, Mustillo, et al., 2004). The analysis controlled for sample size, number of prior months that subjects were asked about in reporting their symptoms, and age of participants. Not all studies report on all diagnoses. The table includes the 16 diagnoses or diagnostic groupings that were reported by at least 8 studies (number of studies shown in parentheses). Figure 2-1 illustrates with a box-and-whisker plot the range of estimates from these surveys for each diagnosis. The ends of the âwhiskersâ for each
THE NATURE AND EXTENT OF THE PROBLEM 43 TABLE 2-2â Prevalence Estimates of Mental, Emotional, and Behavioral Disorders in Young People Standard Diagnosis or Diagnostic Group Prevalence Error Lower Upper (N of studies contributing to estimate) (%) (%) 95% 95% One or more disorders (44) 17.0 1.3 14.4 19.6 Unipolar depression (31) 5.2 0.7 4.0 7.0 Any anxiety disorder (29) 8.0 1.0 6.2 10.3 Generalized anxiety disorder (17) 1.3 0.3 0.9 2.0 Separation anxiety disorder (17) 4.1 0.9 2.6 9.4 Social phobia (15) 4.2 1.1 2.4 7.3 Specific phobia (13) 3.7 1.3 1.7 7.7 Panic (12) 0.7 0.2 0.3 1.5 Posttraumatic stress disorder (7) 0.6 0.2 0.3 1.1 Attention deficit hyperactivity disorder (34) 4.5 0.7 3.3 6.2 Any disruptive behavior disorder (23) 6.1 0.5 5.4 7.3 Conduct disorder (28) 3.5 0.5 2.7 4.7 Oppositional defiant disorder (21) 2.8 0.4 2.1 3.7 Substance use disorder (12) 10.3 2.2 6.3 16.2 Alcohol use disorder (9) 4.3 1.4 2.1 8.9 NOTE: The prevalence estimates from each study were transformed to logit scale and their standard errors computed using the available information about the sample size and preva- lences. Using weights inversely proportional to estimated variances, weighted linear regression models were fit in SAS, using PROC GENMOD with study as a fixed effect (class variable). The overall estimate (on the logit scale) and its standard error were then used to recompute the overall prevalence and its standard error using the delta method. SOURCE: Based on a meta-analysis for the committee by Alaattin Erkanli, Department of Biostatistics, Duke University. A list of the data sets used in the meta-analysis is in Appendix D, which is available online. diagnosis show the highest and lowest estimates, and the upper and lower bounds of the box show the interquartile range of the Âestimatesâthat is, the 75th and 25th percentiles of the range of estimates. It shows estimates only for diagnoses reported by at least eight studies (number of studies shown in parentheses). The mean estimate for any diagnosis was 17.0 percent (standard error, SE, 1.3 percent) and the median 17.5 percent. The most common diagnostic group was substance abuse or dependence, including nicotine dependence (10.3 percent, SE 2.2 percent). Anxiety disorders were common (8.0 percent, SE 0.1 percent), followed by depressive disorders (5.2 percent, SE 0.07 percent) and ADHD (4.5 percent, SE 0.07 percent). Some disorders, notably anxiety disorders, have a much wider range of estimates than others. The range of estimates for specific phobias was particularly broad. It is also noticeable that the top 25 percent of the range
44 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS 45 Percentage Prevalence 40 35 30 25 20 15 10 5 0 1 or More Disorders (44) Major Depressive Disorder (27) Any Anxiety Disorder (29) Generalized Anxiety (17) Separation Anxiety Disorder (17) Social Phobia (15) Specific Phobia (13) Panic (12) Posttraumatic Stress Disorder (7) Any Disruptive Behavior Disorder (23) Conduct Disorder (28) Oppositional Defiant Disorder (21) Attention Deficit/Hyperactivity Disorder (34) Substance Use Disorder (12) Alcohol Use Disorder (9) FIGURE 2-1â Ranges in data on the prevalence of mental, emotional, and behavioral Fig2-1.eps disorders among young people. NOTE: Lines represent the range of estimates from different studies. Boxes repre- sent the interquartile range. SOURCE: Based on a meta-analysis for the committee by Alaattin Erkanli, DepartÂ ment of Biostatistics, Duke University. A list of the data sets used in the meta- a Â nalysis is in Appendix B, which is available online. of estimates is generally much wider than the lowest 25 percent range, indicating that a few studies tend to generate much higher estimates than do the majority. Several factors contribute to the variability in prevalence estimates: (1) changes in the taxonomy or definitions and criteria used for disorders in different versions of the DSM and the ICD, (2) the evolution of assessment tools over the past few decades, and (3) differences in the populations sampled and the inclusion and exclusion criteria used. For example, since different disorders have different onset ages (see the section on incidence below), samples with different age ranges will show different prevalence rates for many disorders. A fourth factor is that, in surveys of young people (but rarely in surveys of adults), it is normal to collect infor-
THE NATURE AND EXTENT OF THE PROBLEM 45 mation from several informants: mothers, fathers, teachers, and children themselves. Each informant brings a unique view of the child, so the num- ber and nature of informants affect the prevalence estimate. Missing from both Table 2-2 and Figure 2-1 are some rare but often severe disorders; for example, schizophrenia, bipolar disorder, and perva- sive developmental disorders. The reason is that studies to date have not been large or numerous enough to capture these rare disorders with any hope of accuracy. For example, the two studies that included schizophrenia had rates of 6 per 1,000 and 7 per 1,000, respectively (Wittchen, Essau, et al., 1992; Costello, Angold, et al., 1996). The three available estimates for adolescent bipolar disorder (two from the same study) fell between 1 and 3 per 1,000 (Lewinsohn, Rohde, et al., 1998; Costello, Angold, et al., 1996), although prevalence increases in young adulthood (Wittchen, N Â elson, and Lachner, 1998). No population-based estimates are available for prepubertal bipolar disorder. Despite the variability across studies, it is possible to draw some gen- eral observations about prevalence. The mean (17 percent) and median (17.5 percent) estimates for one or more MEB disorders were very close, with 50 percent of studies producing estimates between 12 and 22 percent, suggesting that this estimate is fairly reliable. The rank ordering of preva- lence estimates for the different disorders was remarkably consistent across the individual studies. Of the diagnoses included in Figure 2-1, the lowest prevalence rates came from studies of younger children, especially those from Scandinavia, while the highest rates were reported from studies of young adults (ages 19-24). However, from the point of view of prevention, it should be noted that a review of studies of preschool children concluded that almost 20 percent of 2- to 5-year-olds had at least one DSM-IV dis- order in the past three months (Egger and Angold, 2006), the same rate as seen in older children, adolescents, and young adults. Within studies, after controlling for risk exposures that are often con- founded with race/ethnicity, such as poverty (Costello, Compton, et al., 2003), parental incarceration (Phillips, Erkanli, et al., 2006), or migrant s Â tatus (Bengi-Arslan, Verhulst, et al., 1997), similarities across differ- ent racial/ethnic groups are much more noticeable than are differences (Costello, Keeler, and Angold, 2001; Loeber, Farrington, et al., 2003). Dis- ruptive behavior disorders (conduct disorder, oppositional defiant disorder), ADHD (Rutter, Caspi, and Moffitt, 2003), and substance use disorders (Wittchen, Nelson, and Lachner, 1998) tend to be more common in boys than girls, while the opposite is true of emotional disorders (depression, anxiety disorders). About half of the children with a diagnosis have a disorder that causes significant functional impairmentâthat is, a disorder that impedes their ability to function and develop appropriately in human
46 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS relationships or in cognitive, social, or emotional development (Angold, Erkanli, et al., 2002; Costello, Angold, et al., 1996). As noted earlier, no representative population surveys of rates of the full range of MEB disorders in children in the United States have been pub- lished, although results from a survey of 13- to 17-year-olds in the NCS will be published in 2009. The NSDUH, a household survey from the Substance Abuse and Â Mental Health Services Administration, includes adolescents ages 12 and over. In 2005 and 2006 it included a module on major depres- sive episodes and found that 8.8 percent (2005) and 7.9 percent (2006) of youth reported such an episode in the past 12 months. NHANES has also begun to include selected modules addressing MEB disorders in young people, but no data have yet been published. In addition, more work is needed to expand epidemiological studies to include representative samples of all racial/ethnic groups in the United States, to control for socioeconomic confounds in such studies, and to develop international collaborations that provide comparisons among nations using comparable measures (see Heiervang, Goodman, and Goodman, 2008). Cumulative Prevalence Several longitudinal studies have calculated the proportion of the popu- lation that has received at least one diagnosis of a MEB disorder across repeated assessments, from childhood through adolescence and into early adulthood (Costello, Mustillo, et al., 2004). Jaffee and colleagues compared three such studies and found that between 37 and 39 percent of youth in the three studies had received one or more diagnoses between ages 9 and 16 (Jaffee, Harrington, et al., 2005). In later follow-ups of these Â studies, the cumulative prevalence rose to between 40 and 50 percent by age 21 ( Â Arseneault, Moffitt, et al., 2000; Costello, Angold, et al., 1996). This is similar to a 46.4 percent lifetime prevalence rate based on retrospective data from the NCS of adults (Kessler, Berglund, et al., 2005). In the one study for which cumulative data are available by diagnosis (Costello, Angold, et al., 1996), rates of reporting one or more episodes of a disorder by age 21 were 16.4 percent for disruptive behavior disÂorders, 14.5 percent for anxiety disorders, and 10.4 percent for depressive disorders. Comorbidity Many children have more than one MEB disorder. Figure 2-2 summa- rizes the data from a meta-analysis of comorbidity among the major classes of disorder, after controlling for comorbidity between the comorbid condi- â See http://oas.samhsa.gov/NSDUH/2k6nsduh/tabs/Sect6peTabs1to41.htm#Tab6.27B.
16 14 12 10 8 Age 6 4 2 0 ADHD-DBD Dep-Anx SUD-DBD SUD-ADHD DBD-Dep ADHD-Dep SUD-Dep DBD-Anx ADHD-Anx SUD-Anx FIGURE 2-2â Comorbidity between major disorders (controlling for other comorbidities). Fig2-2.eps NOTE: ADHD = attention deficit hyperactivity disorder; Anx = anxiety; DBD = conduct disorder/oppositional defiant disorder; Dep = depression; SUD = substance use disorder. Bars represent odds ratios; lines represent 95 percent confidence intervals. broadside SOURCE: Adapted from Angold, Costello, and Erkanli (1999) and Armstrong and Costello (2002). 47
48 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS tion and other disorders. For example, it adjusts the comorbidity between anxiety and depression for comorbidity between anxiety and ADHD and depression and ADHD. As the figure demonstrates, comorbidity is wide- spread, and there are clear patterns; there is greater comorbidity among disruptive behavior disorders, ADHD, and substance abuse disorders, on one hand, and among the emotional disorders (anxiety and depression), than between emotional and disruptive behavioral disorders, on the other. Comorbidity remains high from early childhood (Egger, Erkanli, et al., 2006) through adolescence (Roberts, Roberts, and Xing, 2007) and into adulthood (Kessler, Chiu, et al., 2005). In summary, there is consistent evidence from multiple recent studies that early MEB disorders should be considered as commonplace as a frac- tured limb: not inevitable but not at all unusual. The prevalence of these disorders is the same in young people as it is in adults. An implication for prevention is that universal programs will not be wasted on large numbers of risk-free children. Is prevalence increasing or decreasing? Repeated surveys are needed to tell whether rates of any disorder are going up or down. For adults, a second NCS has recently been com- pleted, and should provide some information for the population ages 18 and older. The one area of problem behavior in which data on trends in young people are available is alcohol and other drug use and abuse. Three national surveysâNSDUH, the Youth Risk Behavior Surveillance System, and MTFâregularly measure alcohol and drug use and abuse in young people. All restrict their data collection to adolescents (12 and over for NSDUH, 8th, 10th, and 12th grade students for MTF). MTF tends to pro- duce slightly higher estimates than NSDUH; however, they are remarkably consistent in their reports of trends, which show a clear reduction in use across nearly all categories between 2002 and 2007 (see http://oas.samhsa. gov/NSDUH/2k6NSDUH/2k6results.cfm#Tab9-1). Reviews or meta-analyses have used cross-sectional studies conducted at different periods, together with the small longitudinal data sets available, to put together a picture over time (Collishaw, Maughan, et al., 2004; Costello, Foley, and Angold, 2006). Evidence of this sort has produced two fairly clear conclusions: there has been an increase in disruptive behavior symptoms over the past few decades (Collishaw, Maughan, et al., 2004), whereas there is no evidence for a similar increase in child or adolescent depression (Costello, Erkanli, and Angold, 2006). The question of whether the prevalence of autism has increased (Fombonne, 2005) is fraught with problems of broad- ening of the diagnostic category, heightened public awareness, and more attention from clinicians (Schechter and Grether, 2008). The same is true of
THE NATURE AND EXTENT OF THE PROBLEM 49 ADHD and juvenile onset bipolar disorder (Moreno, Laje, et al., 2007). It is certainly the case that more young people are being given these diagnoses. incidence OF MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS To estimate the incidence, or number of new cases, in a given period of time, it is necessary to make repeated estimates in the same representative population sample, excluding those who had the disorder at the previ- ous assessment. The same lack of basic data from repeated, representa- tive sampling hampers the ability to answer questions about incidence. However, in this case, some of the small community-based longitudinal studies can provide data about incidence of the more common disorders. For example, data on 1,420 youth ages 9-21, over a 14-year period, from the Great Smoky Mountains Study (GSMS), a community study from the southeastern United States, shows a mean annual incidence rate of any dis- order of around 3.5 percent in this age group. Of the 55 percent of youth in this community sample who had MEB disorders in one or more years of assessment, more than half (57.2 percent) had a diagnosis at two or more assessments, indicating that, in the majority of cases, the disorder was not confined to a single episode (Costello, Angold, et al., 1996). A related issue relevant to prevention is the age at onset of child and adolescent emotional or behavioral disorders. In the NCS and NCS-R s Â tudies of adults, which ask people with a lifetime history of mental illness to remember their age at the first episode, half of all adults report onset in childhood or adolescence; the NCS-R found that in a population sample ages 18 and older, âhalf of all lifetime cases start by age 14 years and three fourths by age 24 yearsâ (p. 593). Similarly, as noted earlier, in the GSMS, 55 percent of participants had been diagnosed with at least one MEB dis- order by age 21 (see also Kim-Cohen, Caspi, et al., 2003). Age at Onset Figure 2-3 shows the age at onset of the first symptom in youth from the GSMS sample who would eventually receive a diagnosis by age 21, as well as the age at onset of the full-blown disorder. Disruptive behavioral disorders and ADHD had the earliest onset, followed by emotional disor- ders (anxiety and depressive disorders). Although many adolescents began using alcohol and other illicit drugs in their early teens, they tended not to meet criteria for abuse or dependence until their late teens. Epidemiological findings like these raise questions of the utmost impor- tance for prevention. If at least half of those who will have an MEB disorder during their lives have onset in childhood, then prevention resources need
50 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS 25 Age at first symptom 20 Age at first diagnosis Age 15 10 5 0 sis ce y e HD D CD on i et us OD en no ssi AD nx Ab nd i ag pre yA ce pe yD De An tan De An bs ce Su tan bs Su FIGURE 2-3â Age at onset of first symptom and of full psychiatric disorder, by Fig 2-3.eps age 21: Data from Great Smoky Mountains Study. NOTE: First symptom = age at first symptom in youth who at some point received this diagnosis. First diagnosis = age when subject reported the minimum number of symptoms for this diagnosis. SOURCE: Costello, Angold, et al. (1996). to be focused on this period of life. In addition to universal prevention programs, Figure 2-3 suggests that there may be a window of opportunity lasting two to four years between the first symptom and the full-blown disorder, when preventive programs might be able to reduce the rate of onset of specific disorders. Recently developed measures (Egger and Angold, 2006) now make it possible to identify children with symptoms of several disorders at an early stage. In addition, developmentally informed interven- tions that aim at known antecedent risk factors during childhood and early adolescence can provide important opportunities for prevention. Is Incidence Increasing or Decreasing? To determine whether the number of new cases is rising or falling over time, it is important to distinguish between incident (new) cases and newly referred or treated cases. For example, according to one survey of clini- cal referrals, the number of children and adolescents in the United States treated for bipolar disorder increased 40-fold from 1994 to 2003, to about
THE NATURE AND EXTENT OF THE PROBLEM 51 1 percent of the population under age 20 (Moreno, Laje, et al., 2007). In contrast, the three studies that have assessed rates of mental illness across time in the general population found a prevalence of bipolar disorder of between 1 and 3 per 1,000 children, with no increase over the past two decades (Lewinsohn, Rohde, et al., 1998; Costello, Angold, et al., 1996). The reason for this discrepancy between epidemiological and clinical data may arise from the increased use of psychopharmacological treatments for children. The availability of a treatment may encourage clinicians to make a diagnosis and parents to seek professional help. Thus, the advent of a new drug or greater willingness of parents to bring their children for treatment can greatly increase the number of children seen by professionals, while the baseline prevalence of the disease in the population may remain unchanged. In order to find out whether population incidence and prevalence are changing we need several longitudinal studies covering different time peri- ods, so that new case rates can be calculated for different historical periods. National surveys like MTF make it possible to chart, for example, the rise and fall of alcohol and cocaine use by adolescents (Banken, 2004). Data like these are not available for other MEB disorders. Although a variety of federal agencies are making efforts to monitor mental, emotional, or behav- ioral problems, with the exception of substance use disorders, these efforts have not yet produced the repeated estimates over time necessary to plot the rise and fall of disease prevalence and the effects of interventions. factors affecting prevalence and incidence In the language of infectious disease epidemiology, it is possible to talk about various pathogens as âcausesâ of disease. Epidemiology invented the term ârisk factorsâ in the 1950s when the Framingham Heart Study showed that cardiovascular disease did not have a single cause but many different factors contributing to increased risk, no single factor being either necessary or sufficient. MEB disorders seem to have more in common with chronic diseases like cardiovascular disease than with infectious diseases, in having multiple risk factors. A mountain of research on environmental risk and protective factors for MEB disorders in young people has identified a large number of predic- tors, from internal (e.g., intellectual ability, brain development) to familial, educational, communal, and national (see also Chapter 4). Several theorists have developed multilevel risk models that predict complex interactions among the various levels of risk and protection. As with the prevalence and incidence of disorders, the prevalence and incidence of risk factors vary across the nation and at different developmental stages. To take a single example, data from the 2000 decennial census show that the proportion of
52 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS families living in poverty in 2004 varied from 5.3 percent in Minnesota to 17.6 percent in Mississippi. In order to focus prevention efforts most effectively, it is essential to know when vulnerability to an emotional or behavioral disorder increases simply with an increasing number of risk factors, irrespective of their nature, and when increased risk follows specific risk exposures. (Of course, both may occur at the same time.) We illustrate how both aspects of risk come into play with data from over 6,000 assessments of 1,420 youth from the GSMS. On one hand, there was a clear relationship between total risk exposure, using a list of over 80 risk factors, and MEB disorders. Rates of nearly all of these disorders were three or more times higher in the highest risk group than in the lowest risk group, irrespective of the type of risk. On the other hand, when the question of specific risk factors for specific disorders was examined in the same data set, both general and Â disease- s Â pecific risk factors emerged (Shanahan and Hofer, 2005). Parental unem- ployment and maternal depression were associated with increased risk for most MEB disorders, but the analyses revealed âsignature setsâ of factors associated only with certain diagnoses. For example, while sexual abuse, poor parental supervision, and deviant peers were risk factors for both conduct disorder and oppositional defiant disorder, parental depression and loss of close relations and friends were specific to conduct disorder in these analyses. In the emotional disorders, parental depression was a spe- cific risk for depression but was not associated with any anxiety disorders, whereas parental drug use and unemployment were associated with anxiety disÂorders but not with depression (see also Chapter 4). The role of individual differences in genetic makeup has been the focus of intensive study in recent decades (see Chapter 5). Twin and adoption studies have identified a genetic component of risk for most child and ado- lescent psychiatric disorders (Rutter, Silberg, et al., 1999a, 1999b), and genetic research in psychiatry began with the hypothesis that genes âcauseâ mental illness (Kendler, 2005). However, with the exception of a number of rare disorders, such as Williams syndrome, Turner syndrome, fragile X syndrome, and velocardiofacial syndrome (Davies, Isles, and Wilkinson, 2001; Inoue and Lupski, 2003; Thapar and Stergiakouli, 2008) so far no unequivocal candidate genes for specific mental, emotional, or behavioral disorders in children or adults have survived the test of replication in mul- tiple studies (Joober, Sengupta, and Boksa, 2005; Thapar and Stergiakouli, 2008). There are some indications that variations in specific genes may contribute to such disorders as depression (Levinson, 2006; Lopez-Leon, Janssens, et al., 2008). â The term âpsychiatricâ rather than âmental, emotional, or behavioralâ is used here as that is the term used by the authors.
THE NATURE AND EXTENT OF THE PROBLEM 53 Current efforts focus on the search for genes that influence underlying processes, such as threat appraisal or risk aversion, that may be common to more than one mental, emotional, or behavioral disorder. More recently, genetic approaches are also being used to map out the role of environmental factors in the etiology of MEB disorders in people with different genetic profiles; that is, the extent to which (1) a disorder occurs in the presence of a given risk factor only in those with a specific genetic trait or (2) genetic effects on environmental exposure increase risk of a disorder. As discussed in Chapter 5, continued research may make it possible to identify and target the most genetically vulnerable children for prevention interventions. Also, identifying gene variants that are associated with MEB disorders may eventually lead to prevention approaches based on modify- ing components of the pathways from genes to behaviors. However, the focus of prevention for the foreseeable future will still be on psychosocial interventions that change environmental risk factors. Research on signature sets of risk factors suggests that it may also be possible to target prevention efforts for some disorders to youth with high levels of signature risk for that disorder, potentially including both environmental and genetic factors. There is also an argument to be made for paying attention to risk factors, like maternal depression or family disruption, that affect multiple types of MEB disorders (see Chapter 4). Are Rates of Causal Factors Increasing or Decreasing? There is, of course, no simple answer to this question. National surveys and databases can be helpful in monitoring some of the epidemiological factors thought to be associated with emotional or behavioral disorders. For example: â¢ Low birth weight and other perinatal hazards may be increasing in the United States because of the increasing number of births from in vitro fertilization, the increasing age of women at first birth, and other factors. The proportion of newborns under 2,500 grams rose by more than 20 percent between 1980 and 2005. â¢ Family poverty fell in the 1990s but has been level since then (according to the 2000 U.S. census). â¢ Divorce rates have fallen since their peak in the 1980s (U.S. census). â¢ Single-parent households have risen steadily, especially since the 1970s (U.S. census). â See http://www.cdc.gov/media/pressrel/r061121.htm?s_cid=mediarel_r061121_x.
54 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS However, unless these changes can be linked with outcomes in spe- cific data sets, the causal links remain very weak. Countries that maintain national databases on illness, crime, and household structure are beginning to use record linkage to monitor changes in risk exposure, but this is not possible in the United States. High Risk of Some Sociodemographic Groups for Specific Disorders It appears that boys are more vulnerable to disorders with early onset, such as developmental disabilities, autism, disruptive behavior disorders, and ADHD (Rutter, Caspi, and Moffit, 2003). After puberty, several diver- gences appear. Depression and anxiety increase markedly in girls but not in boys (Rutter, Caspi, and Moffitt, 2003). Substance abuse develops faster in boys than girls, and behavioral disorders remain higher in boys (Rutter, Caspi, and Moffitt, 2003). However, sex differences can vary depending on how a disorder or its consequences are defined. For example, the DSM-IV diagnosis âconduct disorderâ is not much more common in boys than girls, but boys are increasingly more likely than girls to be arrested, charged with an offense, convicted, and incarcerated (Copeland, Miller-Johnson, et al., 2007). Similarly, conduct disorder is equally common in African American and Hispanic youth, controlling for socioeconomic status and rural/urban residence (Angold, Erkanli, et al., 2002), but arrests, criminal charges, and convictions are more common in African American youth (U.S. Public Health Service, 2001c). Even in urban settings, after controlling for socio- economic status, delinquency rates were similar in three urban and African American samples (Loeber, Wei, et al., 1999), perhaps due to the tendency for poor African American youth to be concentrated in urban ghettos (Sampson, ÂRaudenbush, and Earls, 1997). CONCLUSIONS AND RECOMMENDATIONS Epidemiology provides the basic information needed to establish the size and community burden of MEB disorders and to track the effective- ness (and cost-effectiveness) of large-scale preventive interventions. To carry out this task, a nation needs to be able to monitor the changing rates of risk exposure and illness in the population as a whole, at different developÂ mental stages, and also in minority groups that may have different patterns of risk. Based on an amalgam of small surveys, about one in five or six young people has one or more recent MEB disorders. Retrospective studies of adults show that half or more had their first episode as a child, adoles- cent, or young adult. The first symptoms of most disorders precede onset of the full-blown condition by several years, so the opportunity exists for preventive intervention.
THE NATURE AND EXTENT OF THE PROBLEM 55 Conclusion: Mental, emotional, and behavioral disorders are as com- mon among young people as among adults. The majority of adults with a mental, emotional, or behavioral disorder first experienced a disorder while young, and first symptoms precede the full-blown disorder, pro- viding an opportunity for prevention and early intervention. As discussed in more detail in Chapter 9, MEB disorders impose a heavy national burden of disability. Early emotional and behavioral prob- lems predict school failure, unplanned pregnancy, and crime. MEB disor- ders are not well tracked by the mortality statistics that are among the few monitoring tools available in the United States. Other tools are needed, including regular household surveys and surveys of institutions, such as hospitals and prisons, where rates of mental illness are high. The United States supports several household and school-based surveys suitable for this purpose. Although these provide very detailed coverage of drug use and abuse, they have many limitations in the area of mental illness, particularly for younger populations, and they are sketchy in their measurement of risk. Data specific to the United States come from a patchwork of small, local studies. Conclusion: Although the United States collects rich data related to drug use and abuse, systematic data related to the prevalence and inci- dence of mental, emotional, and behavioral disorders in young people are sparse. It is notable that the Foundation for Child Developmentâs annual Child Well-Being Index, which has been charting trends in child well-being since 1975, because data are not available, includes only one measure related to MEB disorders: the teenage suicide rate. Similarly, given the limitations of available data, the only national indicators related to MEB disorders reported by the federal Forum on Child and Family Statistics are alcohol and drug use and the percentage of children ages 4-17 reported by their parent as having serious emotional or behavioral difficulties. The forum is planning to add an indicator related to adolescent depression using data collected in NSDUH. Recommendation 2-1: The U.S. Department of Health and Human Ser- vices should be required to provide (1) annual data on the prevalence of â See http://www.soc.duke.edu/~cwi/. â See http://www.childstats.gov/americaschildren/index.asp. â The indicator is based on a parental response to one question from the Strengths and Dif- ficulties Questionnaire and does not provide information about any diagnosis.
56 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS MEB disorders in young people, using an accepted current taxonomy (e.g., the Diagnostic and Statistical Manual of Mental Disorders, the International Statistical Classification of Diseases) and (2) data that can provide indicators and trends for key risk and protective factors that serve as significant predictors for such disorders. Methods for collecting such data should: â¢ be capable of providing reliable prevalence estimates for minority populations and high-risk groups (e.g., incarcerated youth, foster children, immigrant children, youth with chronic diseases, children with developmental delays); â¢ be capable of providing accurate estimates at the level of individual states, ideally with unique identifiers that would facilitate the use of data by local communities and potential linkage with other state databases, such as those created as part of the No Child Left Behind Act of 2001; and â¢ include measurement of identified risk and protective factors, either directly or by building links to appropriate databases (e.g., parental death, foster care placement, divorce, incarceration). As illustrated in Table 2-1, multiple agencies of the U.S. Department of Health and Human Services (HHS) administer surveys that collect data related to MEB disorders. The Centers for Disease Control and Prevention, which has public health surveillance and prevention within its mandate and administers several major surveys potentially relevant to this task, is one possible lead agency for the collection of prevalence and incidence data. Similarly, the Substance Abuse and Mental Health Services Administration is the lead federal agency charged with âbuilding resilience and facilitat- ing recoveryâ in relation to substance abuse and mental disorders. It has recently expanded its population survey, NSDUH, beyond substance abuse, making it another potential option. However, while a specific agency may need to be identified to provide data on the prevalence and incidence of disorders, inclusion of data related to risk and protective factors is likely to require the involvement and input of multiple HHS agencies, making this a departmental responsibility. The Office of Disease Prevention and Health Promotion and the Office of the Assistant Secretary for Planning and Evalu- ation, both in the Office of the Secretary, would potentially be able to serve a coordinating function. Young people with MEB disorders tend to receive care from a wide range of service providers and agencies, including the child welfare, educa- tion, and juvenile justice systems, as well as primary medical and specialty mental health care providers. Very little is known about the adequacy of
THE NATURE AND EXTENT OF THE PROBLEM 57 this patchwork of care. Under its statutory mandate, the Substance Abuse and Mental Health Services Administration (SAMHSA) must provide national data on mental health and substance abuse treatment services and on persons with mental and substance use disorders. This mandate includes the determination of the national incidence and prevalence of the various forms of mental disorder and substance abuse, as well as characteristics of treatment programs. SAMHSA has focused much of its efforts on specialty providers and services supported through state substance abuse and mental health agen- cies. However, nontraditional settings, such as jails, prisons, schools, and general hospitals, are becoming increasingly important as sites of care for youth with MEB disorders. Exclusion of other settings in which young people often receive care provides a misleading and incomplete picture of service use. Recommendation 2-2: The Substance Abuse and Mental Health Services Administration should expand its current data collection to include measures of service use across multiple agencies that work with vulner- able populations of young people. The Centers for Medicare and Medicaid Services (CMS) and programs funded by CMS collect information on use of Medicaid-funded services for prevention and treatment. These data could provide a rich set of informa- tion on trends in utilization of services across various health care providers. Analysis of these data in conjunction with the above prevalence and service use data, with appropriate privacy protections, could provide additional insights.