Adaptive behavior refers to a person’s social responsibility and independent performance of daily activities. One of the first publications of intervention with a child with autism was an application of then new behavior analysis procedures to the problem of teaching a young boy to wear his glasses (Wolf et al., 1964). Since that time, behavioral interventions have been applied to building a wide variety of adaptive skills with varied populations of children and adults with developmental disabilities.
Toilet training and associated issues have been the focus of a broad range of early behavioral interventions. For example, behavioral interventions for toilet training have been based upon principles of both operant and classical conditioning (Azrin et al., 1971; Azrin and Foxx, 1971, 1974; Mahoney et al., 1971). The problem of nocturnal enuresis has been addressed with urine detection devices that serve to awaken children so they can get out of bed when wet, as well as with systematic behavioral procedures involving practice, rewards, and clean-up requirements (Hansen, 1979). Interventions have also been developed and evaluated to address encopresis (O’Brien et al., 1986).
Adaptive skills are usually taught through a process that begins with a task analysis, which breaks down a skill into its component parts (Haring and Kennedy, 1988). Instruction then proceeds through a process of teaching each component skill in small steps, and ultimately chaining the sequence of behaviors together. This approach has been evaluated through
the use of single-subject designs across many applications. A number of interventions have demonstrated that adolescents or adults with autism can be taught purchasing skills and other community living skills, such as ordering food in a restaurant (Haring et al., 1987). However, most applications of instruction in community living skills have been developed for children and adults with mental retardation. Daily living skills targeted have ranged from appropriate mealtime behaviors (O’Brien et al., 1972; Wilson et al., 1984), to eating in public places (van den Pol et al., 1981). Proactive approaches to promoting community access include instruction in clothing selection skills (Nutter and Reid, 1978), pedestrian safety (Page et al., 1976), nondisruptive bus riding (Neef et al., 1978), vending machine use (Sprague and Horner, 1984), and coin summation (Lowe and Cuvo, 1976; Miller et al., 1977; Trace et al., 1977). Additionally, procedures for teaching leisure skills have targeted independent walking (Gruber et al., 1979) and soccer (Luyben et al., 1986).
Another area of widespread application is found in investigations on the remediation of eating disorders. For example, various approaches have been documented as effective in controlling rumination, or persistent vomiting (Kohlenberg, 1970; Rast et al., 1981; Sajwaj et al., 1974), pica (Mace and Knight, 1986), and diurnal bruxism (Blount et al., 1982). Skill-based interventions have been aimed at promoting oral hygiene (Singh et al., 1982) and food acceptance by a child with a gastrointestinal feeding tube (Riordan et al., 1984). A simple procedure of requiring placement of the fork down between bites was shown to reduce the pacing and quantity of food intake by obese children (Epstein et al., 1976).
Behavioral medicine, or the application of behavioral principles to medical problems, includes an experimental case study with a child with autism, whose seizure disorder was ameliorated by a technique involving interruption early in an identified behavioral chain (Zlutnick et al., 1975). Procedures that have been developed for teaching generalized toy play skills to children with mental retardation should translate to use with children with autism (Haring, 1985). Of additional relevance to children with autism are applications of operant procedures to the assessment of hearing in persons with profound mental retardation (Woolcock and Alferink, 1982), as well as the assessment of visual acuity in nonverbal children (Macht, 1971; Newsom and Simon, 1977).
The bulk of the literature cited above was derived from research in which the subjects were described as having mental retardation, and early applications of behavior analysis were conducted primarily with adults in residential settings. However, it is likely that some of the subjects in these early applications also had undiagnosed autism. There was little attention to diagnostic precision in the early behavioral research, though the brief subject descriptions provided often mentioned behaviors com-
monly associated with autism (e.g., self-stimulatory behaviors, self-injury, echolalia, etc.).
There has been an assumption that behavioral interventions documented as effective in teaching adaptive skills to adults with developmental disabilities will apply equally well to child populations. For example, although written as a commercial self-help guide for toilet training normal children, the procedures in Toilet Training in Less Than a Day (Azrin and Foxx, 1974) involved rather minor modifications of the procedures previously developed for adults in residential settings (Foxx and Azrin, 1973). Similarly, faded guidance procedures that were evaluated for teaching adolescents with disabilities to brush their teeth (Horner and Keilitz, 1975) bear marked resemblance to procedures described for teaching independent daily living skills to toddlers with autism (McGee et al., 1999). In other words, many procedures for teaching self-care skills to adults with mental retardation have been extended to younger children. Yet there have been relatively few direct empirical tests of adaptations to young children with autism. This situation may partially result from the lack of emphasis on publishing systematic replications, as well as from the cost- and time-efficiency of simply using existing procedures that prove to be clinically effective.
DEVELOPMENTAL CONSTRUCTS AND THEORY
An issue of considerable relevance to understanding autism is whether associated impairments are simply developmental delays or developmental irregularities. Pertinent to this question are findings that suggest that children with autism show uneven patterns between developmental domains (Burack and Volkmar, 1992). Depending on how broadly developmental domains are defined, children with autism have also been found to show scatter within certain domains. Specifically, children with autism were found to show deviant and not just delayed development in the social and communication domains represented on the Vineland Adaptive Behavior Scales (Sparrow et al., 1984), although not necessarily in domains of daily living skills that can be more easily taught (VanMeter et al., 1997).
Several studies have queried parents on the developmental progress of their children with autism. In a survey of 100 parents of children with autism between the ages of 9 and 39 years, 48 percent of the children were still wearing diapers after the age of 3 (Dalrymple and Ruble, 1992). In addition, 25 percent of the parents surveyed reported past or present problems with their children eliminating in inappropriate places, such as outdoors or in the bedroom. Although the average reported age for urine control was 3.85 years and 3.26 for bowel continence, 22 percent of chil-
dren and adults with autism continued to wet their beds at night. Five adults, at an average age of 24, were still not toilet-trained. Health-related problems included constipation (13%) and severe diarrhea (13%). Behavioral issues included stuffing toilets with paper or items, continual flushing, smearing of feces, playing in toilets, and refusing to use a variety of toilets.
A substantially larger sample of children and adults with autism (1,442) was compared with people with mental retardation (24,048) in terms of their motor, daily living, social, and academic skills, using a database of the New York Developmental Disabilities Information System (Jacobson and Ackerman, 1990). Comparisons were made between age groupings of children (5–12 years of age), adolescents (13–21), and adults (21–35 years). Although the children with autism functioned at higher levels than did the children who had mental retardation without autism, these differences were no longer evident when examining the skill levels of adolescents. The advantage of children with autism was reversed in the groups of adults, with people with autism functioning at lower levels in academic and social skills although they continued to maintain an advantage in gross motor skills.
FORM OF ADAPTIVE BEHAVIORS
A subjective account of 25 Irish mothers of children with autism between the ages of 3 to 14 years of age presents an array of perceived difficulties in the day-to-day management of a child with autism and the consequent effects on the child’s family (O’Moore, 1978). Among the difficulties reported were parental problems in managing housework, due to the extra time needed to feed, toilet train, dress, engage, and put their children with autism to sleep. Parents often felt uncertain regarding effective behavior management techniques, and most reported the use (although not approval) of corporal punishment. Both the children with autism and the overall family had restricted levels of contact in the community, due to either the fear or reality of increased behavioral problems during community outings. Another study compared the breastfeeding patterns of children with autism with a matched group of children with more general developmental delays, and findings were that the mothers of children with autism reported no significant differences in the offering or acceptance of breastfeeding (Burd et al., 1988).
Although the range of adaptive behaviors can be defined more or less broadly, virtually all categorizations include a focus on self-care skills related to basic biological functions. In addition to issues of toileting, eating and sleep disorders are frequently reported in children with autism (Richdale and Prior, 1995). However, most research on irregularities in biological functions has been based on parental report, which can be
influenced by the behavioral characteristics of autism. For example, in a study of sleep patterns of 22 children with autism, aged 3 through 22, parental responses on a questionnaire were compared to direct measurement of ambulatory behavior with an actigraphic device (Hering et al., 1999). More than half of the parents reported that their children had sleep problems, including difficulty in getting to sleep, early morning awakening, and multiple night arousals. However, direct measures of non-sleep activity suggested fewer differences between the children with autism and a comparison group of normally developing children. Children with autism, on average, tended to awaken approximately 1 hour earlier than the typical children. The investigators speculated that parents of children with autism might be more sensitive to sleep issues with their children. Other studies have reported rates of sleep disorders that equal or exceed those of children with other developmental disorders (Dahlgren and Gillberg, 1989, Thompson et al., 1994).
Other adaptive behaviors pertain to home and community living skills, with applicable areas for young children including dressing, grooming, and safety-related behaviors. A broader perspective on adaptive behaviors may expand to school-related skills, such as academic behaviors (McGee et al., 1986), play skills (Haring, 1985), or overall engagement with work materials or the social environment (McGee et al., 1997). For example, children with autism often need to be directly taught how to request help when facing challenging tasks (Carr and Durand, 1985). Finally, most views of adaptive behaviors also cover domains of language, social, and motor skills, which are reviewed in other sections of this report.
Not surprisingly, there are correlations between levels of adaptive skills and intellectual ability (Carter et al., 1996). For example, lower cognitive and verbal levels are highly correlated with age of accomplishment of bowel and urine training (Dalrymple and Ruble, 1992). However, successful use of toileting intervention procedures based on operant and classical conditioning may be more related to physical maturity and social responsiveness than to cognitive level (Azrin and Foxx, 1971). There is some evidence that levels of adaptive behavior predict future independent functioning more accurately than measures of cognitive or academic functioning (Carter et al., 1996).
ASSESSING ADAPTIVE BEHAVIOR AND PLANNING FOR INTERVENTION
The aim of assessment of adaptive skills is to obtain a measure of the child’s typical functioning in familiar environments such as the home and the school. Such measures provide clinicians with an estimate of the degree to which the child can meet the demands of daily life and respond
appropriately to environmental demands. A significant discrepancy between IQ and the level of adaptive skills or between observed performance in a highly structured situation and in more typical situations indicates that an explicit focus on acquisition and generalization of adaptive skills is important. For a diagnosis of mental retardation, assessment of adaptive level is required.
Assessment of adaptive functioning is particularly important for children with autism for several reasons. First, measures of a child’s typical patterns of functioning in familiar and representative environments, such as the home and the school, can be obtained. Assessment of adaptive skills provides a measure of a child’s ability to generalize teaching across settings; given the nature of the cognitive difficulties in generalization in autistic spectrum disorders, such assessments are especially important. As with other children with developmental difficulties, acquisition of basic capacities for communication, socialization, and daily living skills are important determinants of outcome. Significant discrepancies, for example, between performance in a highly structured setting and in less structured settings, or between intellectual skills and adaptive abilities, indicate the importance of including an explicit focus on teaching such skills and encouraging their generalization across settings. Adaptive skills may be in marked contrast to a child’s higher ability to perform in one-on-one teaching situations or in highly structured behavioral programs.
Second, assessment of adaptive behaviors can be used to target areas for skills acquisition. Third, there is some suggestion that relatively typical patterns of performance in autistic spectrum disorders can be identified and that some aspects of adaptive assessment (e.g., of social skills) can contribute to a diagnostic evaluation (Carter et al., 1998; Loveland and Kelley, 1991). This can be especially important in high-functioning children, in whom IQ scores may not reflect the ability to function independently in natural environments. Fourth, assessment of adaptive skills, as well as of intellectual ability, is essential in documenting the prevalence of associated mental retardation and, thus, eligibility for some services (Sparrow, 1997).
The Vineland Adaptive Behavior Scales (Sparrow et al., 1984) are the most widely used instruments to assess adaptive skills (Harris and Handleman, 1994). The Vineland assesses capacities for self-sufficiency in various domains such as communication (receptive, expressive and written language), daily living skills (personal, domestic and community skills), socialization (interpersonal relationships, play and leisure time and coping skills), and motor skills (gross and fine). A semistructured interview is administered to a parent or other primary caregiver; the Vineland is available in four editions: a survey form to be used primarily as a diagnostic and classification tool for normal to low-functioning children or adults, an expanded form for use in the development of indi-
vidual education or rehabilitative planning, a classroom edition to be used by teachers, and a preschool form. Particularly for children with autistic spectrum disorders, the expanded or preschool form may be most helpful since it can be used to derive goals that can be directly translated in an individualized education plan (IEP) (Volkmar et al., 1993). In addition, several research studies have delineated Vineland profiles that are relatively specific to autism (Loveland and Kelley, 1991; Volkmar et al., 1987). This unique pattern consists of relative strengths in the areas of daily living and motor skills and significant deficits in the areas of socialization and, to a lesser extent, communication. Supplementary Vineland norms for autistic individuals are also now available (Carter et al., 1998).
Other instruments with subtests for assessing adaptive behaviors of very young children include the Brigance Inventory of Early Development (Brigance, 1978), the Early Learning Accomplishment Profiles (ELAP) (Glover et al., 1988), the Scales of Independent Behavior-Revised (Bruininks et al., 1996), the AAMD Adaptive Behavior Scales (Lambert et al., 1993) and the Learning Accomplishments Profile (LAP) (Sanford and Zelman, 1981). The Developmental Play Assessment Instrument (DAP) (Lifter et al., 1993) provides an evaluation of the quality of a child’s toy play skills in relation to those of typically developing children, which can help to target the play level and actions that a child with autism needs to learn.
A primary consideration in selection of adaptive living goals should be the skills needed to promote age-appropriate independence in community living, so that a child can have access to the larger social community. For example, children who are not toilet trained are not likely to have access to classrooms with normally developing peers, and parents of children who present safety risks will be less likely to take them on community outings. Classrooms and home programs may begin with an early focus on independent daily living skills early in a child’s intervention program, because progress in these areas is more easily achieved than in the more challenging domains that are diagnostic descriptors of autism (i.e., social, communication, and behavior). Thus, parents and teachers are pleased when their child makes tangible early progress, and they may be motivated to collaborate on more challenging tasks.
There are a number of published manuals that provide practical guidance on the design of instructional programs, along with detailed task analyses of various daily living and self-help skills. For example, Steps to Independence (Baker and Brightman, 1997) provides easy-to-follow guidelines for teaching skills such as shoe tying or hand washing. Behavioral intervention techniques can readily be used to teach adaptive skills (e.g., Ando, 1977; Azrin and Foxx, 1974; McGee et al., 1994), and self-care and other skills can be systematically taught (McClannahan et al., 1990), although it is critical that generalization of teaching be accomplished. Other
methods for teaching adaptive skills include peer tutoring for teaching community living skills (Blew et al., 1985). Additional resources for commonly encountered difficulties include books written for parents on eating disorders (Kedesdy and Budd, 1998) and sleep problems (Durand, 1998).
The books listed above, and similar resources, include suggestions for data collection during baseline planning, implementation and follow-up. The complexity of the data collection procedure will vary according to the challenge of the skills being taught (e.g., bladder control training usually requires very detailed records on successes and failures, while teaching a child to throw away their paper towels may be monitored with one probe). Ongoing assessment typically requires at least some baseline measurement, as well as periodic measures of skill performance during and after intervention. In order to assess the level of independence achieved for a given skill, it is necessary to evaluate the performance of the new skill in conditions of decreasing prompting.
With the exception of research in communication and socialization, there are surprisingly few studies that directly evaluate the use of behavioral interventions to teach adaptive skills to young children with autism. However, there is a body of research on reinforcer potency that is directly relevant to efforts to use behavioral techniques for skill instruction with children with autism. Thus, constant versus varied reinforcement procedures were compared in a study of the learning patterns of three boys with autism, aged 6 to 8 years (Egel, 1981). Using a reversal design, it was shown that correct responding and on-task behavior during a receptive picture identification task increased using varied reinforcers. Satiation for food reinforcers was problematic in conditions in which constant reinforcers were used. Similar results were found in comparing sensory versus edible reinforcers; rewards having sensory properties were found to be less vulnerable to satiation (Rincover and Newsom, 1985). The importance of systematic reinforcer assessment has been demonstrated to improve learning and attention to task, and the use of highly potent rewards on learning tasks has also been shown to yield positive side-effects in terms of substantial drops in levels of maladaptive behaviors (Mason et al., 1989).
A Japanese study reported the first early application of operant conditioning procedures to the toilet training of five 6- to 9-year-old boys with profound mental retardation and “clear signs of autism” (Ando, 1977). Like other early behavioral interventions, aversive consequences (i.e., intense spankings) were prominent, and results of an evaluation using an ABAB reversal design showed inconsistent effectiveness. In
contrast, a toilet-training manual written for parents of typical children (Azrin and Foxx, 1974) can be adapted relatively easily for use with children with autism (McGee et al., 1994) and may be more successful.
Peer tutoring was shown to be effective in teaching community living skills to two boys with autism, aged 5 and 8 years, who lived in a residential school (Blew et al., 1985). Single-subject, multiple baseline designs were used to evaluate the effectiveness of treatment components across intervention settings. Skills targeted included buying ice cream at a restaurant, checking out a book at the library, buying an item at the store, and crossing the street. Modeling by typical peers was not sufficient to produce acquisition, but both boys learned all target skills when the peers provided direct instruction.
Physical exercise has been found to decrease self-stimulatory behavior in children with autism, as well as to yield collateral changes in appropriate ball play, academic responding, on-task behavior, and ratings of general interest in school activities (Kern et al., 1982). With physician approval for each of three children (ranging in age from 4 to 7 years, plus three older children/adolescents), mildly strenuous jogging sessions were begun at about 5 minutes per day and gradually increased to 20 minutes per day. In a follow-up study with three children with autism, one of whom was age 7 (and two who were 9), it was shown that mild exercise (e.g., playing ball) had virtually no impact on self-stimulatory behavior (Kern et al., 1984), but positive benefits were replicated in conditions of vigorous physical exercise.
An assessment of the grooming of children and adolescents with autism may have some application for either skill assessment or for measuring the quality of care provided to children with autism. Quality of care was the central focus of a multiple baseline study of 12 children with autism, and the single-subject multiple baseline was nested within multiple baselines across residential settings (McClannahan et al., 1990). Thus, a grooming checklist (e.g., fingernails clean, hands washed, clothing unstained, etc.) was administered to children residing in one large residential placement, and major improvements were documented when these children were transitioned into community-based group homes that provided more individualized care. Further, when feedback on grooming details was regularly provided to group home teaching parents, the children’s appearance improved to a level similar to that of children with autism who lived at home with their families. Applications to young children with autism would likely involve both skill building and assistance to parents in managing the responsibilities of caring for their children with autism. Appearance becomes a practical concern as more and more children with autism are gaining access to inclusion with typical peers, and attractiveness may influence their receipt of social bids.
When adaptive skills are broadly defined, there are a number of ap-
plications reported for young children with autism. Thus, a variety of approaches have been used to increase engagement both with adult-directed tasks and in general attending to the environment; these include delayed contingency management (Dunlap et al., 1987), self-management techniques (Callahan and Rademacher, 1998), and strategies for environmental arrangement (McGee et al., 1991). Inclusion and interaction with typically developing peers (Kohler et al., 1997) have been used as a medium for increasing engagement and play skills (Strain et al., 1994; Wolfberg and Schuler, 1993). Now that children with autism are beginning to gain access to regular preschool and elementary school settings, there has developed a need for teaching them to transition smoothly across educational activities (Venn et al., in press).
As discussed earlier, there have been demonstrations that young children with autism can be taught to increase the frequency and variety of their play skills. Such interventions are expedited by pivotal response training and by targeting the skills displayed by typical children at similar developmental levels (Lifter et al., 1993, Stahmer, 1995). Young children with autism have been taught peer imitation abilities in the course of Follow the Leader games (Carr and Darcy, 1990).
Virtually all of the well-known programs for young children with autism provide instruction in adaptive daily living skills, which often form the basis for development of communication, social, and even motor skills. Several published program outcome evaluations have specifically examined progress in adaptive skills as measured on the Vineland. For example, 20 children with autism enrolled in the Douglass Developmental Disabilities Center showed better-than-average progress in all four domains assessed on the Vineland, but the most marked progress was in communication skills (Harris et al., 1995). Similarly, the Walden family program component was shown to yield developmental gains that were larger than those expected in typical development (i.e., greater than one month gain per month), and children’s progress at home corresponded closely to the intervention priorities selected by parents (McGee et al., 1993).
The Vineland results were less robust for children treated in the Young Autism Project at the University of California at Los Angeles, although the children were described as “indistinguishable from average children in adaptive behavior” (McEachin et al., 1993). The nine children with best outcomes in the 1987 treatment outcome study (Lovaas, 1987) were reassessed at an average age of 11.5 years. Although their overall composite scores were within the normal range, five of the nine had marginal or clinically significant scores in one more domain. Results were
also mixed in two systematic replications. The May Institute’s home-based program reported that 31 percent of the children receiving intervention made at least one month gain in social age per month of intervention, and 12 of 13 showed progress on another measure of adaptive behavior (Anderson et al., 1987). A more recent systematic replication compared intensive and nonintensive interventions (Smith et al., 2000). A randomly assigned group of children with autism and pervasive developmental disorder-not otherwise specified (PDD-NOS) received intervention for approximately 25 hours per week for at least 1 year, while a similar group of children received 5 hours per week of parent training over a period of 3 to 9 months. Vineland results showed no significant differences between the two intervention groups. Chapter 12 presents more information on various model programs’ approaches to intervention in the area of adaptive behavior.
Unless a specific focus on generalization of skills is included in the intervention program, it is possible for children with autistic spectrum disorders to learn skills in a highly context-dependent way. That is, even though a child is capable of some particular behavior, it occurs only in highly familiar and structured contexts. Thus, results of adaptive behavior assessments have been less robust in some cases (McEachin et al., 1993; Smith et al., 2000). However, inclusion of an explicit home-based program has been reported to be associated with progress on measures of adaptive behavior (Anderson et al., 1987).
FROM RESEARCH TO PRACTICE
Because there is a substantial literature about teaching adaptive skills to children with developmental disabilities, one question is how often and when strategies that are effective with other populations of young children are applicable to children with autism. Skills requiring specific adaptations peculiar to autism may benefit from direct investigation (e.g., severely restricted eating patterns, toileting rituals, etc.). A major adaptation that is often required is the improved assessment and selection of reinforcers so that the child with autism will be motivated to develop new adaptive skills (Mason et al., 1989). Questions of generalization are important but need to be considered for a particular behavior and child. For example, a child might learn a very structured tooth brushing routine that is tied to a specific kind of toothbrush—which may be very helpful even if not very generalizable. However, only using a particular kind of toilet would be much more problematic.
Overall, results are encouraging regarding the potential for teaching a range of adaptive behaviors to young children with autism. However, the variability of results in this area are of crucial importance in considering each individual child’s preparation for independent functioning in
everyday situations. Interpretations of positive findings on one measure cannot be used to make blanket declarations of “recovery” from autism.
There is substantial data, particularly with older children and adolescents, that behavioral interventions, particularly those with attention to generalization, can result in improved adaptive behavior in children with autism. Adaptive goals are a significant part of both home and school programs for young children. Although general measures of adaptive behavior may indicate increasing discrepancies from normal development with age, the potential to make practical changes in the lives of children with autistic spectrum disorders through teaching specific skills that have value in the community (e.g., toilet training, pedestrian safety) or for the child (e.g., dressing) is very clear, not only for their own sakes, but also because of the increased opportunities they offer. Teaching adaptive skills, with specific plans for generalization across settings, is an important educational objective for every young child with autism. This objective includes teaching behaviors that can be accomplished within a year and that are anticipated to affect a child’s participation in education, the community, and family life.
At this time, the greatest challenge is one of translation from research to practice. Often teachers do not know what is available in the research literature. User-friendly manuals and training resources are needed to ensure the availability of effective instruction in adaptive skills for young children with autistic spectrum disorders to teachers and parents.