Monday, February 22, 2010


ASSESSING EMOTIONAL AND BEHAVIORAL DISORDERS IN PRESCHOOL CHILDREN Clinical Presentation, Classification, and Epidemiology C. Keith Conners, Ph.D. Professor Emeritus Duke University Medical Center Presented at the annual meeting of the National Association of School Psychologists Chicago, Illinois, March 3, 2010 Introduction Once again it is a privilege to address this audience of committed professionals who work on the front lines, who bring their knowledge and skills to children who manifest the entire spectrum of mental, emotional, and behavioral handicaps. It is a particular pleasure to be able to introduce some new tools to aid in the understanding of the most vulnerable children of all, those just beginning their journey into the hazards of this world, the preschoolers. I am fortunate to have two experienced and brilliant young colleagues to help present our latest work--Elizabeth Sparrow and Jenny Pitkanen--who will provide you with an in-depth briefing on a somewhat ambitious and comprehensive assessment package, our Early Childhood symptom and developmental scales. [SLIDE 1] My task in this workshop is to provide the background of research and the conceptual basis for the way the scales were constructed. We started with the premise that mental illness is intimately related to the stages of normal development, so that assessment must establish both the features of illness as well as the developmental proficiencies and limitations of the child. Aims of this Talk In line with my Zen approach to workshops, I have made a vow to avoid that dread affliction known as “Death from Powerpoint.” So I will keep slides to a minimum. Much of what I will say is extensively reviewed in a paper by Helen Egger and Adrian Angold [1 ] with over 300 references. I am indebted to Professor Angold, head of the Child Epidemiology Program at Duke, for supplying some of the key slides from his paper. Many years ago George Miller published one of the most famous articles in psychology, which became known as Miller’s Law: “The Magical Number Seven, Plus or Minus Two.” Most people can remember about 7 items without practice, with a confidence band of about 2 items; so 5 to 9 items are reliably retained by the average adult. But being mindful that the attention span of the average adult may be greatly exaggerated, let’s be conservative; I only want you to remember 3 things from this talk. [SLIDE ] But these claims are not without controversy. Imagine that you are sitting in a meeting with your fellow psychologists, or a principal, or some administrators from downtown. (I imagine you have many of those). Or perhaps you are in one of the many IEP meetings, which might include parents, teachers, lawyers or other mental health professionals. Suppose you have decided to do an assessment of the preschooler in question, but there is an outcry against you: “How can you think of doing such a thing? How dare you presume to apply psychiatric syndromes or labels to preschoolers, as if they were the same as older children or adolescents or adults?” After all, look what happened to the last guy who claimed that preschoolers could be diagnosed:] [SLIDE Now, I am convinced that evidence supports those three claims I have listed above; but it is important that you become convinced, in order that you can defend your role as an expert who is assessing a preschooler, without harm to your person or reputation. It is not enough for you to say that you heard me say it at a workshop; you should become familiar with the evidence supporting those claims of preschooler psychopathology. Generally there are only five arguments against labeling, diagnosing, classifying, “medicalizing” or alleging some mental illness in preschoolers: [SLIDE ] (1) Individual differences in temperament in normal development will be inappropriately identified as psychiatric symptoms or disorders. The argument of normal variations in temperament (2) Preschool age involves such rapid physical, neural, behavioral, emotional and cognitive development that it is not possible to identify valid symptoms or clusters of symptoms that can be reliably measured. The argument of rapid, variable, unstable developmental changes. (3) That the dominant psychiatric classification system, DSM, does not take into account developmental variation. The argument that DSM is based solely on adults and older children and there is no continuity of illness from early childhood to adolescence or adulthood. (4) That a young child will be inappropriately ‘labeled’ with ‘diseases’ that will adversely affect self-perceptions or perceptions by others of the child. The argument that labeling creates harmful stereotypes. (5) That problem emotions or behavior are not located in the child but rather in the relationships between parents and children and the wider environment. The argument that mental illness is caused by the context in which the child grows. Temperament and Symptom Cutoff Scores Let’s examine the first claim, that problems might just be normal variations of temperament. Over several decades now, numerous studies have used checklists and cut-points on symptom scales showing many symptoms with high prevalence (7% to 25%) in preschoolers. Evidence shows these to be stable characteristics that can be reliably measured. Many studies consistently validate a major class of ‘externalizing’ as well as ‘internalizing’ syndromes that map onto the broad and specific DSM diagnostic categories. Twin studies demonstrate the heritability of many of these syndromes. Temperament studies do in fact reveal that certain extreme temperament types such as behavioral inhibition and behavioral disinhibition are risk factors for the later development of psychiatric disorders, as well as being associated with problem behaviors in preschoolers. For example, preschool children who respond to negative mood induction by overly-expressive or under-expressive facial response have more anxiety and depression at follow-up in the first grade [2]. Negative emotions of fear, anger, frustration, poor adaptability and high emotional intensity are predictors of later internalizing and externalizing symptoms, as well as antisocial behavior in adulthood. Some studies also link negative affectivity to later anxiety and depression [2 ]. It is true, then, that temperament extremes among preschoolers are linked to significant later psychopathology. However, there are many syndromes that temperament concepts do not cover, and so cannot serve as a sufficient conceptual framework. There is no contradiction in recognizing that there are normal dimensions of behavior, such as activity level, or social withdrawal, or impulsive action, and that one can establish cut-points at which these normal behaviors shade into pathological extremes. In the same way that we set certain extremes of blood pressure as causing pathology, we can establish a category of temperament beyond which bad things result. Thus, a category such as extreme activity level is no longer a harmless normal variation; shyness is a normal temperamental trait, but extreme social withdrawal leads to significant impairment and maps on to the DSM categories we recognize in older children and adults. The key finding in the temperament research has been that extremes at either end of a dimension of temperament, such as activity level or social involvement, are associated with recognizable syndromes at the preschool level, and are predictive of later symptom categories and significant level of impairment. CATEGORICAL DIAGNOSIS IN PRESCHOOLERS: RELIABILITY AND VALIDITY. A carefully developed DSM psychiatric interview showed excellent reliability in a large epidemiologic study [3]. Now the second argument against early diagnosis, that DSM included little direct mention of preschoolers is true; but the American Academy of Child and Adolescent Psychiatry formed a Task Force on preschoolers which produced a modified version of DSM. It is available to you online at: The modifications adhere as closely as possible to the original DSM-4 criteria but with modifications that take into account the developmental status of the child. For example, consider the category of Conduct Disorder as presented by the Task Force: Conduct Disorder Diagnostic Features Conduct disorder is characterized by a repetitive and persistent pattern of behavior that violates norms and rules and the basic rights of others. The diagnosis of CD rests on the assumption that a child knowingly violates rules, a supposition that requires both knowledge of the rules and intent to break them. Most preschool children are generally able to understand the concept of rules and can control their behavior accordingly. The duration requirement was shortened from 12 months to 6 months. This decision was made because 12 months is a disproportionate amount of a child’s life span in this population relative to older children. Because preschoolers are not skilled in verbal negotiation, they may make threats (e.g., I won’t be your friend) as a means of resolving disputes. Bullying and threatening should be endorsed positively only when threats and intimidation are persistent patterns of behavior and involve threats of aggression or cruelty (e.g., social ostracism) Infrequent, reactive mild aggression towards peers or objects is common during this period. Atypical aggression is more frequent and may be severe (e.g., kicking, biting, and choking). The effect of physical constraints on the manifestation of symptoms during this period must be considered. For example, most preschool children are not likely to have access to firearms or knives, but may use rocks or sticks to hurt someone. Six of the 15 DSM-IV symptoms were modified and 5 symptoms were not. Four DSM-IV symptoms were deleted because they were inappropriate in relation to the developmental capacities of this age group (A10 “broken into someone else’s house, building, or car”; A13 “stays out at night”; A14 “run away”; and A15 “truant”). Since only 1 new symptom was added, this makes fewer possible symptoms available for children to meet the diagnosis. In the task force document a complete table of the empirical studies supporting the categories of psychopathology in preschoolers and infants is presented, including Reliability, Face validity, Descriptive validity, Predictive validity, and Construct validity. The figures for all of the diagnostic categories are very similar to data in older children and adults. So despite the rapid developmental changes taking place in preschoolers, there is good evidence that assessment is both reliable and valid for most of the types of mental illness found in older children and adults. RESEARCH ON DIAGNOSIS IN PRESCHOOLERS: EPIDEMIOLOGY There are four large scale studies documenting the reliability and validity of preschool diagnostic categories. For example, here is the method used in the Egger and Angold study, which used a DSM-based structured psychiatric interview (the PAPA, or Preschool Age Parent Assessment): [SLIDE ] The prevalence of disorders meeting DSM diagnostic criteria in these studies is similar to the prevalence found in older children and adults: [SLIDE ] It is important to note that pediatricians in private practice have been shown to considerably underestimate the percentage of their patients with emotional and mental disorders [5] The comorbidity with other disorders is also approximately the same as data from older children and adults: [SLIDE ] ADHD, its comorbidities, and levels of impairment were found to be highly similar in preschool and school age children by Wilens, et al. [4] An interesting finding in the Egger and Angold study, since replicated in other studies, is that although there appears to be associations among all of the comorbid factors, when the effects of each disorder on all other disorders are controlled for simultaneously, it turns out that the apparent associations between some pairs of disorders are mediated by another disorder. In the following slide, a preschooler with an anxiety disorder, but not ODD, is no more likely than a child without a disorder to have CD. Notice how ODD seems to be a central mediator between the relationships of other disorders to each other. [SLIDE ] It almost appears that ODD is a gateway to other disorders. Of course oppositional behavior is normal to some degree in two and three-year-olders. But the question is whether any individual child is more oppositional than expected for age and gender. Here, the use of extensive age and gender specific norms becomes crucial. By using a cutoff, say the 90th percentile, we can have some confidence that an intervention is warranted. In this case, parent training has been shown to be an effective intervention for young ODD clients. IMPAIRMENT A finding that a preschooler has more symptoms than expected for age and gender is not enough to demonstrate the need for intervention. It is important that your assessment also shows that the symptoms lead to impairment (in social, educational, or developmental status such as play). Percent Impairment of Preschoolers with Diagnoses [SLIDE ] Data from the preschool studies show a strong relationship between a diagnosis and the number of symptoms and impairment. However, even for many children who do not meet the threshold for diagnosis (“sub-syndromal” patterns), there may be significant impairment. This supports the notion of a continuum of impairment as a function of the number of symptoms. When you examine a child and they have many symptoms, but not enough to make a diagnosis, you still have to consider whether there is some impairment present. ARE THERE HARMFUL EFFECTS OF PSYCHIATRIC LABELS ON PRESCHOOLERS? If you type that question into Google you will get literally thousands of “articles” berating the use of psychiatric labels in children. After reading a few dozen of these articles you may become impatient for any evidence or data to support the conclusions. Many of the articles come from Scientologists or people who can quote Tom Cruise, but cannot quote a single empirical fact to support their argument. A few will quote Peter Breggin or other well-known advocates against psychiatry, particularly those against the use of medications. Most however, are simple expressions of outrage based on no data whatsoever or an anecdote from personal experience. Anecdotes are not data. I adhere to the philosopher David Hume’s method of evaluation: “if it has no tables or data, commit it to the flames.” But we mustn’t stop there; the next step is to use the vast resources of electronic searches in Ovid’s Medline, or other large databases. There it is equally frustrating to find that there are literally hundreds of thousands of articles on children, mental health, stereotyping, self-esteem, side-effects, diagnosis, and various combinations thereof; but not a single study that appears to demonstrate the harmful effect of psychiatric stereotyping or labeling on children. Am I missing something? Or have there been some acceptable empirical studies of this common attribution that simply have not been captured in my database searches? I confess that in my own experience, weighing the benefits of good assessment and treatment against the fear of labeling effects leaves little doubt that the benefits outweigh the risks. After all, finding a correct label means there is some guide to action. There is benefit to knowing what the onset, course, and response to treatment is likely to be for a known disorder. A diagnosis is not a harmful label; it is an injunction to informed action, unless 2000 years of medical wisdom is to be ignored in the face of fear and ignorance. ADVERSE FAMILY CIRCUMSTANCES, PARENTING AND PSYCHOPATHOLOGY IN PRESCHOOLERS Finally, let us consider the question of whether classifying or diagnosing a preschooler erroneously places an emphasis on innate, biological or genetic problems rather than environmental problems. Does this turn us away from thinking about the relationships between parents and children, or from the dangers inherent in the wider environment? Isn’t it true that much of the advance in our understanding of childhood mental illness comes from a long tradition of uncovering the kind of trauma, abuse, and neglect in the family environment, or the well-documented effects of poverty, poor housing, isolation, and what Emile Zola called, “The short and simple life of the poor?” The answer of course, is that there is a false dichotomy here. We now know that the unfolding of the somatic development of the body and brain, the role of the genes, is markedly subject to the environment in which genetic expression must take place. The simple fact of the matter is this: if you hold environmental variables constant, then emotion, behavior and mental development are governed by the genes. But if you hold the genetic variables constant, any variation of the phenotype is due to the environment. All diseases and disorders appear to fluctuate as a function of both environmental and genetic variation. Let us abandon this false dichotomy between genes and environment, nature and nurture, and recognize that a valid clinical assessment requires a careful investigation of both sorts of influence on a particular child. We believe that the innate genetic expression of many of the kinds of mental illness will be found in the very young child, even those blessed with an optimal environment. On the other hand, many children with normal genetic endowments may suffer from the slings and arrows of a dangerous world in which they live. Remarkably, the little available data on the role of family relationships and parenting in preschoolers appears to show that early adverse family circumstances and parenting characteristics do not contribute to the prediction of later psychopathology once child characteristics are accounted for. In a longitudinal study of 420 two-to-three years olds followed up at 10 to 11 years, of the environmental risk factors, only stressful life events contributed independently to the prediction of later externalizing behavior problems [6]. CONCLUSIONS In summary, I argue that even though there is rapid and variable developmental changes, most of the broad patterns of illness found in older children and adults will be present in a certain percentage of preschoolers. Extremes of temperament patterns do predict later mental illness, but temperament alone is not sufficient to explain the many types of individual disorders that will appear in preschoolers. With all of its faults, DSM nevertheless is a helpful guide to the variety of disorders to be found in very young children. Appropriate modifications of the DSM for developmental stage in preschoolers have good supporting data from the available epidemiologic trials, and from our own normative studies. We believe that the benefits of early classification leads to appropriate intervention strategies, and that whatever the downside might be in terms of stereotyping or labeling, the benefits are indisputable. In all of these arguments, the presence of assessment tools for disorders and development in the preschool age relies upon careful census-based age and gender-based norms. It is only from a good foundation of normal behavior and development that we can venture into the world of the preschool child. REFERENCES 1. Egger, Helen Link & Angold, Adrian (2006). Common emotional and behavioral disorders in preschool children: presentation, nosology, and epidemiology. J Child Psychology & Psychiatry, 47(3), 313-337. 2. Cole, Pamela M; Zahn-Waxler, Carolyn; Fox, Nathan A; Usher, Barbara A; Welsh, Jean D. (1996). Individual differences in emotion regulation and behavior problems in preschool children. J of Abnormal Psychology, 105(4),518-529. 3. Egger, Helen Link, Erkanli, A., Keeler, Gordon, Potts, Edward;Walter, Barbara K. & Angold, Adrian (2006). Test-retest reliability of the preschool age psychiatric assessment (PAPA). J of the American Academy of Child & Adolescent Psychiatry, 45, 538-549. 4. Wilens, T.E., Biederman, J., Brown, Sarah, Tanguay, Sarah, Monuteaux, M.C., Blake, Christie, B.S., & Spencer, T.J. (2002. Psychiatric comorbidity and functioning in clinically referred preschool children and school-age youths with ADHD. J American Academy of Child & Adolescent Psychiatry. 41, 262-268. 5. Lavigne, J.V., Binns, Helen J., Christoffel, Katherine K., Rosenbaum, Diane, Arend, R., Smith, Karen, Hayford, Jennifer R., McGuire, P.A. (1993). Behavioral and emotional problems among preschool children in pediatric primary care: prevalence and pediatricians’ recognition. Pediatrics, 91, 649-655. 6. Mesman, Judi; Koot, Hans M. (2001). Early preschool predictors of preadolescent internalizing and externalizing DSM-IV diagnoses. J of the American Academy of Child & Adolescent Psychiatry. 40, 1029-1036.