Problem behaviors of children with autistic spectrum disorders—and other children—are among the most challenging and stressful issues faced by schools and parents in their efforts to provide appropriate educational programs. Problem behaviors such as property destruction, physical aggression, self-injury, and tantrums are major barriers to effective social and educational development (Horner et al., 2000; Riechle, 1990). Such behaviors put young children at risk for exclusion and isolation from social, educational, family, and community activities (Sprague and Rian, 1993). In addition, problem behaviors may place an onerous burden on families, particularly as children grow from preschool into school age (Bristol et al., 1993). Concerns about school behavior problems led to new standards and procedures for discipline, student suspension, and expulsion in the 1997 amendments to the Individuals with Disabilities Education Act (IDEA, 1997; Department of Education IDEA regulations, March, 1999). Specifically, the regulations include provisions for the use of functional behavioral assessments and positive behavioral interventions and support.
The definition of problem behaviors depends on whether the behaviors are considered from the perspective of a child with an autistic spectrum disorder or from the perspective of a parent or teacher. From a child’s perspective, problem behaviors include the inability to understand demands of a classroom or a parent and to communicate his or her needs and wants, severe difficulty in initiating and maintaining social interactions and relationships, confusion about the effects and consequences of
many of his or her behaviors, and engagement in restrictive and repetitive behaviors and interests that may limit the child’s ability to learn and to fit in with peers. From a teacher’s or parent’s perspective, problem behaviors include lack of compliance with or disruption of classroom routines, tantrums, destruction of property, and aggression against self or others.
The research evidence reviewed suggests that educational interventions that do not address the development of positive and prosocial behaviors (the potential for problems from the child’s perspective) will be unsuccessful in the long-term elimination of problem behaviors from others’ perspectives. Chapters 5 through 9 discuss the essential elements (communication, social interaction, cognitive features, adaptive behaviors, and sensorimotor skills) needed for effective, appropriate educational programs for children with autistic spectrum disorders to address core problem behaviors. These elements are discussed in this chapter only as they are directly relevant, but they are essential in any consideration of problem behaviors.
Different literatures provide the empirical base for interventions for problem behaviors in young children with autistic spectrum disorders: data from comprehensive programs; single-subject design studies that address specific problem behaviors; psychopharmacological studies that assess the safety and efficacy of pharmacological interventions on both global and specific problem behaviors; the growing literature on the neurobiology of autism; and legal reviews of the 1997 IDEA provisions related to autism (Turnbull et al., 1999) and findings in due process and court cases involving children with autism (Mandlawitz, 1999).
Many studies evaluated were not designed specifically for this committee’s interest in children with autistic spectrum disorders from birth to age 8. Some degree of latitude was taken in generalizing from findings in studies of older children and children with autistic spectrum disorders if the behaviors of interest and the behavioral principles involved would be expected to apply to children with autistic spectrum disorders in the birth to eight-year-old age group. The focus of this chapter is on the most commonly reported problem behaviors of young children with autistic spectrum disorders. As reported by Horner and colleagues (2000), this focus represents only a selected subset of a large literature that primarily involves treatment of severe, dangerous, chronic cases of behavior problems, mostly in older children. An extensive review of medical studies is beyond the charge of the committee, but selected results are included here as relevant.
NATURE AND PERSISTENCE OF BEHAVIOR PROBLEMS
Most behavior problems displayed by young children with autistic spectrum disorders are “normal” behaviors in that they may be observed,
albeit at lower frequency, in typically developing children. However, in autistic spectrum disorders, the intensity, frequency, duration, or persistence of the behaviors distinguish them from similar behaviors of normally developing young children. For example, several studies have shown that self-injurious and stereotyped behaviors occur in normal infants and then decrease, although they do not necessarily disappear, as locomotion develops in these children during the first and second years of life (Thelen, 1979; Werry et al., 1983). Body-rocking occurred in 19 percent and head-banging in 5 percent of one sample of typical children ages 3 to 6 years (Sallustro and Atwell, 1978). Similar levels of body-rocking have also been reported in normal college students (Berkson et al., 1999; Rafaeli-Mor et al., 1999).
These repetitive movements and potentially self-injurious behaviors are presumed to serve some function in normal development (Berkson and Tupa, 2000). Berkson and Tupa (2000) found that about 5 percent of toddlers with developmental disabilities (including autistic spectrum disorders) engaged in head-banging, about the same percentage as reported for typically developing children. The incidence of head-banging with actual injuries in the group with developmental disabilities is presumably greater: between 1.3 and 3.3 percent, depending on the type of measurement. This rate is similar to the prevalence rates reported for older, noninstitutionalized populations of children and adults with developmental disabilities (Rojahn, 1986; Griffin et al., 1987).
Understanding what causes these problem behaviors to emerge during the early childhood and preschool years, what maintains them, and what evokes their moment-to-moment expression holds promise of treatments to prevent them from becoming permanent and abnormal (Berkson and Tupa, 2000). Once moderate to severe problem behaviors become an established part of a child’s repertoire, unlike children with typical development, children with autistic spectrum disorders or other disabilities do not usually outgrow them. Without appropriate intervention, these behaviors persist and worsen (Schroeder et al., 1986).
With increasing research into the neurobiology and genetics of autism, the organicity of some aspects of behavior in autism is becoming clearer. For example, Lewis (1996) has attempted to explicate some of the underlying neurobiology of repetitive or stereotyped behaviors. Other researchers (Symons et al., 1999) have demonstrated that the locus of some types of self-injurious behavior might show different genetic patterns. Thompson and his colleagues (Thompson et al., 1995) argue that self-injurious behaviors have consequences other than social changes. For example, some self injurious behaviors involve the release of neurochemical transmitters and modulators that subsequently bind to specific brain receptors. By using sophisticated methods that study form, location, and intensity of self-injury, these researchers conclude that some people may
learn to self-injure in body locations that produce the greatest neurochemical release and receptor binding (Symons and Thompson, 1997; Thompson et al., 1995). Schroeder and colleagues (Schroeder et al., 1995) reviewed animal studies of neonatal dopamine depletion relevant to the prevention of self-injurious behavior and recently used an animal model to demonstrate primary prevention of self-injurious behavior using operant conditioning (Tessel et al., 1995).
Epidemiological studies indicate that a substantial minority of all young children, with or without developmental disorders, exhibit problem behaviors at some time that might benefit from intervention (McDougal and Hiralall, 1998; Emerson, 1995). Young children with poor social skills or limited communication, including children with autistic spectrum disorders, are especially at risk for such problems (Borthwick-Duffy, 1996; Koegel et al., 1992). An analysis of five reviews of intervention approaches for the general population of individuals with developmental disabilities, conducted between 1976 and 2000, found that the target behaviors most often addressed in intervention studies were aggression, destruction of property, disruption of activities, self-injury, stereotypic behavior, and inappropriate verbal behavior (Horner et al., 2000). Horner and colleagues’ review of applied behavior analysis studies that were published since 1990 and restricted to children with autism between birth and age 8 found that the behavior problems most frequently addressed were tantrums, including crying and shouting (six studies); aggression (four studies); stereotypic behavior (two studies); and self-injury (one study).
Appropriate Individualized Educational Plans
No single intervention has been shown to deal effectively with problem behaviors for all children with autism. However, there is an increasing consensus among developmental, psychosocial, applied behavior, and legal experts that prevention of such problems should be a primary focus, particularly during the early childhood and preschool years (Berkson and Tupa, 2000; Schroeder at al., 1986; Dunlap and Fox, 1999; Schopler et al., 1995). There is also a growing consensus that the most effective form of prevention of problem behaviors is the provision and implementation of an appropriate individualized education plan (IEP) based on proven interventions that have some scientific evidence supporting their value. The New York State Department of Health panel that developed The Clinical Practice Guideline for Autism/Pervasive Developmental Disorders (New York State Department of Health, 1999) went further: “The use of an ineffective assessment or intervention method [is] a type of indirect harm
if its use supplants an effective assessment or intervention method that the child might have otherwise received.”
The 1999 U.S. Department of Education Regulations for the 1997 Amendments to the Individuals with Disabilities Education Act (IDEA) provides for scientifically supported interventions (see, e.g., 20 U.S.C., 1400 © (4)). The IDEA further requires that schools must confirm, before any changes of placement due to a behavioral problem can be considered, that the IEP and placement were appropriate and that special education services, supplementary aids and services, and behavior intervention strategies were provided consistent with the IEP and placement (34 C.F.R., 300.523, 1999; Turnbull et al., 1999). In short, before assessing deficiencies in a child who is misbehaving, it is critical to assess the adequacy of the intervention program the child is receiving. IDEA requires that interventions must show demonstrable benefits to be continued. A number of different approaches to interventions for problem behaviors meet the IDEA criteria for scientific support and benefit to individual children.
Comprehensive Treatment Programs
Various comprehensive treatment programs encompass a number of different philosophical and theoretical positions, ranging from strict operant discrimination learning (Lovaas, 1987) to broader applied behavior analysis programs (Harris et al., 1991; Fenske et al., 1985; Kohler et al., 1996), and those that highlight incidental learning (McGee et al., 1999) to more developmentally oriented programs (Schopler et al., 1995; Rogers and Lewis, 1989; Greenspan and Wieder, 1997). Comprehensive programs generally require 25 or more hours of active student engagement per week for 2 or more years and attempt to change the clinical course of an autistic spectrum disorder, including prevention of or reduction in problem behaviors. Reviews of eight model comprehensive early intervention programs (Dawson and Osterling, 1997; Harris, 1998; Rogers, 1998), taken together, identified several critical elements common to many programs that addressed problem behaviors (a more extensive review of program elements is provided in Chapter 12):
curriculum content that emphasized direct instruction in basic skill domains and abilities: attending to elements of the environment that are essential for learning, especially to social stimuli; imitating others; comprehending and using language; playing appropriately with toys; and interacting socially with others;
highly supportive teaching environments and generalization strategies;
predictability and routine;
a functional approach to problem behaviors;
plans for transition from preschool classroom;
identification of and intervention with children with autistic spectrum disorders as early as possible;
working with young children in small teacher-to-child ratios, often one to one in the early stages; and
active engagement of the child from 20–40 hours per week
Programs that do not include the above features should be reevaluated for suitability before discussing the “suitability” of the disruptive student.
Applied Behavior Analysis
Forty years of single-subject-design research testifies to the efficacy of time-limited, focused applied behavior analysis methods in reducing or eliminating specific problem behaviors and in teaching new skills to children and adults with autism or other developmental disorders. Initially, applied behavior analysis procedures were reactive, focusing on consequences of behaviors after they occurred, and interventions of this type continue to play an important role (see below). However, there has been increasing attention to intervention procedures that focus on what to do before or between bouts of problem behaviors (Carr et al., 1999a; Carr et al., 1994; Schroeder et al., 1986). Since the mid-1980s, applied behavior analysis prevention strategies have focused on antecedent conditions in the child or the environment that set the stage for or trigger the problem behaviors (Carr et al., 1999c); some of these are discussed below in the sections on positive behavioral interventions and supports and functional behavioral assessment.
Interventions that involve changing schedules, modifying curricula, rearranging the physical setting, and changing social groupings have been shown to decrease the likelihood of problem behaviors (Carr et al., 1998; Dunlap et al., 1991, 1993). This has been termed a “shift from viewing behavior support as a process by which individuals were changed to fit environments, to one in which environments are changed to fit the behavior patterns of people in the environments” (Horner et al., 2000:6). Many of these antecedent interventions have been implemented for years by some of the comprehensive, developmental programs described earlier (Mesibov et al., 2000). The broader interest in these antecedents now brings the methodological rigor of applied behavior analysis to directly test the causal relationship between these environmental changes and skill acquisition and reduction in problem behaviors.
The research evidence regarding the role that communication deficits play in the emergence, remediation, and maintenance of reduction in problem behaviors is particularly robust across researchers and methodologies (Carr et al., 1999b; Koegel et al., 1992; Schroeder et al., 1986; Wacker et al., 1998). Interventions that deal with receptive communication—for example, use of schedules, work systems, and task organization (Schopler et al., 1995) that assist students in understanding classroom routines and requirements as well as effective instruction in spontaneous, expressive communication (Schreibman et al., 2000; Wacker et al., 1996)—are needed to prevent problems and maintain reductions in those behaviors (see a more detailed discussion of functional behavioral analysis below).
AFTER THE FACT: TEACHING ALTERNATIVE BEHAVIORS
Most empirically based intervention approaches designed to reduce or eliminate specific, identified problem behaviors have an applied behavior analysis theoretical base. From this perspective, problem behavior is viewed as being composed of two environmental features and one behavior or set of behaviors that have a temporal relationship. Antecedents, the first feature, are events (e.g., mother tells child it is time to go to the store) or internal conditions (e.g., child feels pain or hunger) that occur before a problem behavior (e.g., running around the house instead of going to the door) occurs. Consequences are events that follow the behavior and that either increase the likelihood that the behavior (running) will be repeated (reinforcement, e.g., mother makes a game of chasing the child to get him into the car) or decrease the likelihood that the behavior will be repeated (e.g., mother shouts “No!” when the child runs away).
One approach rewards behaviors that are incompatible with the problem behavior: for example, rewarding a child for taking his mother’s hand to go to the car so the child cannot engage in running away at the same time (differential reinforcement of alternative behavior). Another approach removes the consequences of the behavior that are thought to be reinforcing (extinction-based procedures). For example, when adult attention is thought to be a reinforcer for the child’s running away from his mother, an extinction-based strategy would be for the mother to demonstrate no attention to the running, provided the child is safe. In the example above, for some children, the parent’s shouted “No!” functions as a punisher and reduces running behavior. For others, the parent’s
attention is rewarding and increases the likelihood that the child will “play” the running away game.
Pivotal Response Training
Interventions that enable children to have some control over their environments, such as task preferences, choice-making, reinforcement selection, and self-monitoring, can all contribute to reductions in problem behaviors (Fisher et al., 1992; Koegel et al., 1987; Koegel et al., 1992; McGee and Daly, 1999; Newman et al., 1997). Teaching of pivotal skills, such as increasing motivation or self-management, can produce improvement in wide areas of functioning that might otherwise require hundreds or even thousands of discrete trials for the child to master individually (Koegel et al., 1999).
Functional Behavioral Assessment
Functional assessment is the process of identifying the variables that reliably predict and maintain problem behaviors (Horner and Carr, 1997). Although such an approach is implied in much of the research described above, a more formal approach to functional behavioral assessment has evolved in the literature and is required in certain cases of discipline under IDEA (see, for example 34 C.F.R., 300.520, 1999). The functional behavioral assessment process typically involves:
identifying the problem behavior(s);
developing hypotheses about the antecedents and consequences likely to trigger or support the problem behavior;
testing the hypotheses; and
designing an intervention, based on the conclusions of the assessment, in which antecedents or consequences are altered and the child’s behavior is monitored.
Initial identification of the problem behavior and development of ideas of why it occurs often involve interviews with people in the child’s classroom or family and direct observation of the behavior in its usual context. Testing hypotheses may occur through additional observation or, less frequently, through systematic functional analysis in which the environment is manipulated to test the hypotheses (Carr et al., 1994; Dunlap et al., 1993; Iwata et al., 1982; Repp and Horner, 1999). Such analyses are expected to lead to the identification and training of alternative, appropriate behaviors that can give the child the same “payoff” he or she received from the previous problem behavior. In several reviews, as many as 16 different motives for problem behavior were identified (Reiss
and Havercamp, 1997; Carr et al., 1999c), and more may exist, including multiple functions for some behaviors. Prominent among these functions or motives for problem behaviors are: a means of communicating needs and wants effectively; social attention; social avoidance; escape from difficult or boring tasks or other aversive situations; access to desirable tangible items and preferred activities; and generation of sensory reinforcement in the form of auditory, visual, tactile, olfactory, and gustatory stimulation.
For example, problem behaviors such as self-injury or destructive behavior often produce reliable changes in the child’s environment. A child bites his hand and learns over time that the parent or teacher may approach and soothe the child, provide a favorite toy, or “rescue” the child from a difficult situation. For young children with autistic spectrum disorders, who often have little or no ability to communicate using conventional words or even gestures, hand-biting, tantrums, or other disruptive behaviors become effective ways for the child to convey a message. Wacker and colleagues (Wacker et al., 1998) trained parents, in their homes, to conduct functional analyses of the problem behaviors of their young children (between 1 and 6 years of age) with autism or other severe developmental disorders. Parents were then trained to use a functional communication system with their children, based on their own child’s existing communication skills. Children learned from their parents verbal or nonverbal appropriate means (such as signing “please” to gain parent attention, “break” to get a brief break from tasks, or “help” to obtain parent assistance in completing a task) to obtain what they wanted. After treatment, aberrant behavior had decreased an average 87 percent across the range of children, and appropriate social behavior had increased an average 69 percent. The intervention took approximately 10 minutes per day. On a parent-rating measure of acceptability (from 1= not acceptable to 7=very acceptable), the average overall acceptability was 6.35.
Three findings on functional behavioral assessment emerge from 10 reviews of research from 1988 to 2000 on problem behaviors in persons with developmental disabilities including autistic spectrum disorders (Horner et al., 2000): (1) functional behavioral assessment results more frequently in the choice of positive rather than punishment procedures than do problem reduction strategies not starting from functional behavioral assessment; (2) interventions developed from functional behavioral assessment information are more likely to result in significant reductions in problem behaviors than those that do not systematically assess the function of the problem behavior; and (3) in some cases in which functional behavioral assessments were conducted, interventions were designed that were not consistent with or may even have been contraindicated by the assessment information (Scotti et al., 1996). Thus, additional
training of how to implement results of a functional behavioral assessment in home and school interventions is often needed to link the assessments with interventions.
Positive Behavioral Interventions and Supports
IDEA requires that if a child’s behavior impedes his or her learning or the learning of others, the IEP team must consider, if appropriate, strategies, including positive behavioral interventions, strategies, and supports to address that behavior (20 U.S.C., 1414 (d)(3)(B)(i), 1999; 34 C.F.R., 300.346(a)(2)(i), 1999. “Positive behavioral interventions and supports” describes an approach to deal with a child’s impeding behavior that focuses on the remediation of deficient contexts (i.e., environmental conditions and behavioral repertoires) that are confirmed by functional behavioral assessment to be the source of the problem (Carr et al., 1999a). An expanded definition of this proactive rather than reactive process brings together four interrelated components that draw on aspects of many of the interventions described above. Positive behavioral interventions and supports include (Turnbull et al., 1999):
systems change (e.g., the process of considering, modifying, or substantially changing an agency’s policies, procedures, practices, personnel, organization, environment, or funding);
environmental alterations (including building on a child’s strengths and preferences, connecting the child with community supports, increasing the quality of the student’s physical environment, making environmental alterations, such as changing when or for how long an activity occurs or introducing a schedule for the student, and making instructional accommodations for the student);
skill instruction, consisting of instruction for both the student and those who interact with him or her on appropriate academic, independent living, or other skills; teaching the student alternative behaviors and adaptive behaviors that reduce or ameliorate the impeding behaviors; and teaching skills to those involved with the student regarding communication with the student, development of social relationships, problem solving, and appropriate responses to the student’s impeding behaviors; and
behavioral consequences (so that impeding behaviors are eliminated or minimized and appropriate behaviors are established and increased).
The expected outcomes from positive behavioral interventions and supports are increases in positive behavior, decreases in problem behavior, and improvements in life-style (Horner et al., 1990). This includes the expectation of systems change, including changes in the behaviors of oth-
ers in the environment and broad environmental reorganization and restructuring. Many of these features are implemented as standard practice in the comprehensive or focused behavioral programs reviewed above and in Chapter 12. The concept of positive behavioral interventions and supports represents a theoretical, scientific, and legal attempt to bring all aspects of these successful, positive interventions to bear on resolving behavior problems in children with autism or other disorders.
A total of 366 outcomes of positive behavioral interventions and supports were examined in a detailed review of applied behavioral analysis studies of persons with autistic spectrum disorders (10%), mental retardation (about 50%), or combined diagnoses of retardation and autism, frequently accompanied by additional diagnoses (e.g., seizure disorder, brain damage; about 40%) (Carr et al., 1999a). These outcomes included 168 outcomes for children from birth to age 12; they addressed problems of aggression, self-injurious behavior, property destruction, tantrums, and combinations of problem behaviors. The success rate (90% or greater reduction in problem behavior from baseline levels) across pooled outcomes was generally within 5 points of 50 percent of individuals, regardless of the type of intervention. Good maintenance rates were observed for a substantial majority of outcomes (68.7%, 63.6%, and 71.4% for 1–5 months, 6–12 months, and 13–24 months, respectively). Males and females scored equivalent successes.
A similar review of a differently defined, overlapping data set (Horner et al., 2000) concluded that the available interventions are reasonably effective at reducing problem behaviors of persons with developmental disabilities, including autistic spectrum disorders. Reductions of 80 percent or greater were reported in one-half to two-thirds of the comparisons. Some reductions of 90 percent or greater were reported for individuals with all diagnostic labels and all classes of problem behaviors. The lowest success rate (23.5 percent) was for interventions that targeted sensory functions, compared with approximately 60 percent success rates for interventions based on functional behavioral assessments that identified other functions of problem behaviors (e.g., attention, escape, tangible, or multiple types).
A review of applied behavioral analysis interventions specifically for children with autistic spectrum disorders from birth to age 8 (Horner et al., 2000) addressed problems of tantrums, aggression, stereotypy, and self-injury. This targeted review found, for 37 comparisons, mean rates of reduction in problem behaviors of 85 percent (with a median reduction level of 93.2% and a mode of 100%). Fifty-nine percent of the comparisons recorded problem behavior reductions of 90 percent or greater, and 68 percent of the comparisons reported reductions of 80 percent or greater. The mean length of maintenance (assessed in 57% of studies; rates main-
tained within 15% of initial outcome levels) was 12 weeks, with the longest assessment occurring at one year after intervention (Koegel et al., 1998).
Though these are very positive findings, evaluating studies, and their results, requires cognizance of the prevailing scientific trend, adopted by many journal editors, that favors publication of studies that report successful, rather than unsuccessful, interventions. Thus, the results summarized above, represented as percentages of published comparisons, represent possible outcomes when these procedures are carefully implemented and progress monitored; they do not reflect the number of unsuccessful interventions, which are not reported. As described above, research concerning problem behaviors in individuals with developmental disabilities has generally been strong and plentiful. However, there are relatively few studies directly addressing issues for young children with autistic spectrum disorders. In many cases, interventions that were successful with other populations may be appropriate for young children with autistic spectrum disorders (Wolery and Garfinkle, 2000). Studies testing this assumption with appropriately described and diagnosed children are crucial before it can be accepted. Using the guidelines established by this committee, published research concerning positive behavior approaches to young children was relatively strong in measurement of generalizability and in internal and external validity (see Figures 1–1, 1–2, and 1–3 in Chapter 1). Limitations in the existing studies are not due to a generally poor quality of research, but to changes (and differences) in standards of reporting and research designs in applied behavior analysis and those of the more general, educational and clinical guidelines for treatment evaluation (see Chapter 1). These limitations in these studies were particularly apparent in the selection and description of subjects, random assignment to treatment conditions, and independence of evaluation. As for other areas, these limitations also related to differences in the contexts in which methods were developed. For behavioral interventions that addressed such targets as dangerous self-injury in institutionalized adolescents with profound mental retardation, random assignment, accurate diagnosis, and independence of evaluation may have been of less concern than developing an immediately implementable effective individualized program. However, in order to evaluate treatments for milder difficulties in young children with autistic spectrum disorders, provision of standard, descriptive information about subject selection, subject characteristics and other aspects of research design is crucial in determining what approaches will be most effective for which children.
With these caveats in mind, consistent findings across reviews of published studies indicate several conclusions about current positive behavioral interventions and supports:
effectiveness in significantly reducing (and maintaining reduction of) problem behaviors in at least one-half, if not more, of all applied behavioral analysis studies of problem behaviors in children with autistic spectrum and related disorders;
doubled effectiveness when functional behavioral assessments were used to determine what reliably predicted and maintained the behavior before undertaking the intervention, that is, what function that behavior served for the child;
ability to be effectively carried out in community settings by the children’s usual caregivers, although effective treatment data for the most difficult cases generally involved specialized personnel in less typical settings;
minimal effectiveness (fewer than one-fourth of the cases) in reducing behavior problems that prior functional behavioral analysis indicated were maintained by sensory input; and
increased problem behavior for a small percentage of outcomes studied (6–8%).
These conclusions are particularly important because such interventions must be considered, under IDEA, if a child’s behavior impedes his or her learning or the learning of others.
IDEA contains what has been termed a “rebuttable presumption” (Turnbull et al., 1999) in favor of using positive behavioral interventions and supports in cases of “impeding behavior.” This presumption (having legal weight) can be refuted by evidence to the contrary, but positive behavioral interventions and supports is the only intervention strategy specifically required for consideration by IDEA; other strategies may be considered. If positive behavioral interventions and supports is seen as a rebuttable assumption, it means that an IEP team can consider other intervention strategies only in comparison with positive behavioral interventions and supports and must have adequate cause for adopting a different strategy. Evidence for the efficacy of positive behavioral interventions and supports (presented above), although encouraging, also indicates that current positive behavioral interventions and supports strategies, as presently implemented, may be ineffective or only minimally effective for up to one-third of all problem behaviors and for up to three-quarters of those problem behaviors maintained by sensory input. In these cases, different or additional strategies may be required, after first considering positive behavioral interventions and supports.
Physically Intrusive or Physically Aversive Procedures
In an analysis of 102 interventions included in an overall review of positive behavioral interventions and supports (Carr et al., 1999a), about one-half were associated with the use of extinction (removing or preventing the occurrence of whatever has been found to reinforce and increase the problem behavior) and one-half with the use of seven different punishment procedures: verbal reprimand, forced compliance, response cost, overcorrection, timeout, brief restraint (physically interrupting the response and preventing its recurrence), and water mist (one case, considered highly intrusive and aversive). Although research indicates that reinforcement-based procedures are often not as effective in eliminating severe problem behaviors as quickly as are punishment-based procedures (Iwata et al., 1982), punishment-based procedures can also cause undesirable side effects, such as the child avoiding the punisher.
The increase in efficacy of positive interventions, when based on functional behavioral analysis, reduces the need for punishment-based procedures (Neef and Iwata, 1994). When a behavior is not maintained by social reinforcement, however, it may be difficult to treat effectively with reinforcement-based procedures only (Iwata et al., 1994). Suppression of competing problem behaviors may sometimes be needed before reinforcement of functional alternative behaviors can be effective (Pelios et al., 1999). In any case, there is agreement (New York State Department of Health, 1999) that physically intrusive measures (e.g., response interruption, holding, or physical redirection) should only be used after positive measures have failed and only as a temporary part of a broader intervention plan to teach appropriate behaviors. The use of physical aversives such as hitting, spanking, or slapping is not recommended.
Medications to Reduce Behavior Problems
Although a comprehensive review of medications and medical interventions is beyond the scope of this report, because of the widespread use of psychoactive medications, they are addressed briefly as they relate to problem behaviors in young children with autistic spectrum disorders. Psychoactive medications alter the chemical make-up of the central nervous system and affect mental functioning or behavior. Most were developed to treat a variety of psychiatric and neurological conditions other than autistic spectrum disorders; all may have benefits, side effects, and toxicity (Aman and Langworthy, 2000; Gordon, 2000; King, 2000; and McDougle et al., 2000). There are currently no medications that effectively treat the core symptoms of autism, but there are medications that can reduce problematic symptoms and some that play critical roles in severe, even life-threatening situations, such as self-injurious behavior.
Just as autism coexists with mental retardation, autistic spectrum disorders may coexist with treatable psychiatric and neurological disorders (Tuchman, 2000). Treatment of such diagnosed disorders will not cure autism, but can, in some cases, enable a child to remain in less restrictive community placements and enhance the child’s ability to benefit from educational interventions (Cohen and Volkmar, 1997). Medications have been shown in some instances to enhance and to be enhanced by systematic, individualized behavioral intervention programs (Durand, 1982; Symons and Thompson, 1997).
More than 100 articles have been published on the use of psychoactive medications for autistic spectrum disorders. A more limited number of published reports include double-blind, placebo-controlled studies with young children with autism. Double-blind studies of haloperidol (Cohen et al., 1980; Campbell et al., 1982; Anderson et al., 1984, 1989), naltrexone (Campbell et al., 1990; Bouvard et al., 1995; Kolmen et al., 1997; Willemsen-Swinkels et al., 1995, 1996), clonidine (Fankhauser et al., 1992; Jaselskis et al., 1992), and fenfluramine (Stern et al., 1990) included young children with autistic spectrum disorders, some as young as 2 years of age. In addition, newer medications, including selective serotonin uptake inhibitors, atypical neuroleptics, other antidepressants, and stimulant medications such as methylphenidate, have been studied, although most not yet in double-blind studies. A double-blind, placebo-controlled study of Risperidone™ was completed in adults (McDougle et al., 2000), and a study in children is presently under way in the National Institute of Mental Health Research Units for Pediatric Psychopharmacology (NIMH RUPPs).
The key findings from the published studies include:
Haloperidol was effective in reducing aggression and agitation and had mixed results for improving learning with long-term users, but it carries significant risk of involuntary muscular movements (dyskinesias).
Naltrexone-treated groups showed less irritability and hyperactivity than placebo groups on some measures, particularly global ratings, did not differ from placebo groups on others, and showed increases in particular problem behaviors in some instances.
Clonidine-treated subjects showed improvements in hyperarousal but reported increased drowsiness, decreased activity, they showed increasing tolerance when used to treat attention deficit disorders.
Risperidone shows promise in treating aggression and agitation with less concern about the development of dyskinesias than for the older neuroleptics.
Open trials of serotonin selective uptake inhibitors have shown promise in treating stereotypic or perseverative behavior, possibly because of effects on anxiety.
Stimulants may affect sleep and growth in developing children and, in some cases, may worsen autistic symptoms, especially self-stimulation behaviors.
Secretin-treated children did not differ from placebo-treated children.
Functional behavioral assessment to determine the function(s) of the problem behaviors increases the likelihood of choosing the correct medication and behavioral interventions.
Research is under way to predict responders and nonresponders to medication and to determine which children will benefit from behavioral treatment alone and what combinations of medication and behavioral treatment are most effective. Many parents also treat their children with nonprescription drugs and nutritional supplements. These agents have received even less study than prescription drugs, and their assessment is beyond the scope of this report. Several psychotropic medications have appeared to result in improvements for some patients but make others worse. Since medication is often instituted in a crisis, the possibility that it is actively contributing to deterioration is often not considered. Children’s responses to medication must be monitored very carefully.
Children with autistic spectrum disorders are also at increased risk for certain medical conditions, notably seizure disorders. From one-fourth to one-third of people with autism are expected to develop seizure disorders sometime in their lifetime (Bristol et al., 1996; Kanner, 2000; Tuchman, 2000). School personnel need training in recognizing the symptoms of seizures and other medical problems and in monitoring the effects of medications over time. Although technically outside the scope of the educational program, it is important that educators and parents be informed of the importance of quality medical care and both the potential value and the possible problems of pharmacological intervention.
Except in unusual medical circumstances, medications are usually not considered first-line interventions for behavior problems in young children, but an exception, for example, would be behavioral manifestations of seizure disorder. Because young children are developing and learning, it is essential that both positive outcomes and unintended side effects of medications for behavioral problems are considered and that cognitive as well as behavioral effects are monitored if a decision is made to use medication. In addition to a functional behavioral analysis of the problem behavior, medication for behavioral intervention should be based on knowledge of medical pathology, psychosocial and environmental conditions, health status, current medications, history, previous intervention, and parental concerns and desires (New York State Department of Health, 1999).
A new generation of rational drug design will be based on emerging
findings concerning the neurobiology of behaviors in autism. Future research offers the possibility of developing or refining medications based on the specific mechanisms that maintain problem behaviors in children with autism and related disorders. The new field of pharmacogenomics goes farther by hoping ultimately to match medications to genetic profiles for individual patients.
Other Interventions to Reduce Behavior Problems
Although there are some effective interventions to address sensory aspects of behavior problems, e.g., substituting appropriate sources of kinesthetic, visual, auditory, or olfactory stimulation for aberrant ones (Favell et al., 1982), there is a pressing need for more basic and applied research in this area. Interventions such as relaxation training (Groden et al., 1998) and physical exercise (Quill et al., 1989; Kern et al., 1984), appropriately adapted for young children, are also promising avenues for stress reduction and concomitant decreases in problem behaviors.
FROM RESEARCH TO PRACTICE
Problem behaviors such as property destruction, physical aggression, self-injury, stereotypy, and tantrums put young children at risk for exclusion from social, educational, family, and community activities. Serious behavior problems occur in a minority of young children with autistic spectrum disorders, but they are costly financially, socially, and academically to children, their families, and their classmates. The concept of problem behaviors in autism varies depending on whether the problem is defined in terms of the child’s needs or the effect of the behavior on the home or classroom environment. Both research evidence and clinical judgment agree that the primary approach to problem behaviors in young children should be prevention by providing the child with the skills needed to effectively deal with the physical, academic, social, and sensory aspects of his family’s school, preschool, early childhood, or community environment.
The foundation of prevention of problem behaviors is an appropriate and fully implemented IEP. Critical elements of effective, comprehensive educational programs and of focused, problem-specific applied behavior analysis programs, identified for young children with autistic spectrum disorders, need further independent replication, direct comparison of different treatment approaches, and clinical trials of methods that have proven effective in what Rogers (1998) notes as the clinical equivalent of “open trials.” Broader implementation and evaluation of functional behavioral assessment and positive behavioral interventions and supports should lead to an expanded array of effective strategies for the majority of
problem behaviors. More rigorous evaluation of existing medications and development of new medications based on burgeoning knowledge of the neuroscience of autism is likely to add new tools to the armamentarium.
IDEA requires benefits from interventions, presumes in favor of positive interventions, disallows those that do not produce benefits, and authorizes a wide range of beneficial interventions without preferring any particular ones. Although the interventions discussed in this chapter have shown evidence of accelerating the child’s development and reducing behavior problems, none attains the strict research standards for replicated, randomly assigned, controlled, long-term comparison studies (Bristol et al., 1996).
Education is at heart an enterprise that must be informed by science, and it should stimulate hypotheses, case studies, and descriptive research to identify promising approaches for further rigorous study. As Greenspan (1999) points out, researchers, clinical practitioners, and consumers will need to work together to refine existing methods and develop new approaches. Joint efforts of federal, state, and local agencies will be needed to stimulate and fund longitudinal sites sharing common measures and a common database to address the daunting questions of which treatments are most effective for reducing behavior problems for which children. In the meantime, researchers, educators, and parents should not ignore testable, not yet fully assessed methods or measures that hold promise for reducing problem behaviors in children with autistic spectrum disorders. IDEA sets up perhaps the most practical and in some ways the most difficult challenge—that of generating a functional analysis of each child’s behavior to fashion an individualized program that will enable the child to progress and to participate in the academic and social life of family, school, and community.