Errorless Academic Compliance Training: Improving Generalized Cooperation With Parental Requests in Children With Autism

Errorless Academic Compliance Training: Improving Generalized Cooperation With Parental Requests in Children With Autism

Errorless Academic Compliance Training: Improving Generalized Cooperation With Parental Requests in Children With Autism JOSEPH M. DUCHARME, PH.D., A...

201KB Sizes 0 Downloads 12 Views

Errorless Academic Compliance Training: Improving Generalized Cooperation With Parental Requests in Children With Autism JOSEPH M. DUCHARME, PH.D.,

AND

TAMMY L. DRAIN, M.A.

ABSTRACT Objective: Children with autism often demonstrate distress and oppositionality when exposed to requests to complete academic or household tasks. Errorless academic compliance training is a success-focused, noncoercive intervention for improving child cooperation with such activities. In the present study, the authors evaluated treatment and generalization effects of this intervention on four children diagnosed with autism. Method: In a multiple baseline acrosssubjects design, parents delivered a range of academic and nonacademic requests to their children to determine the probability of compliance for each request. A hierarchy of academic requests ranging from those yielding high compliance (level 1) to those yielding low compliance (level 4) was then developed. Treatment began with the concentrated delivery of level 1 requests, with praise and reward for compliant responses. Over several weeks, children were gradually introduced to requests from subsequent probability levels with continued reward for compliance. Results: High compliance levels were demonstrated throughout and following treatment. Evidence of generalization to untrained academic requests and nonacademic requests emerged. Treatment gains were maintained up to 6 months after treatment. Conclusions: Errorless academic compliance training appears to be an effective intervention for enhancing generalized compliance in children with autism. J. Am. Acad. Child Adolesc. Psychiatry, 2004;43(2):163–171. Key Words: autism, parent training, generalization, compliance training, errorless approaches.

Autism is a developmental disorder that is defined by qualitative impairments in social interaction and communication as well as restricted, repetitive, and stereotyped patterns of behavior, interests, and activities (American Psychiatric Association, 1994). In a review of epidemiological surveys of autism, Fombonne (1998) concluded that the best estimate of prevalence is 5 per 10,000. Many children with autism engage in severe aberrant behaviors such as noncompliance, self-injury, aggression, screaming, and temper tantrums (Prior and

Accepted September 10, 2003. From the University of Toronto. The authors thank Leanne Baldwin and the staff of Peel Behavioural Services, Trillium Health Centre, for support of this research. All parents provided written consent for publication. Reprint requests to Dr. Ducharme, Department of Human Development and Applied Psychology, University of Toronto (OISE), 252 Bloor Street W., Toronto, Ontario, Canada M5S 1V6. 0890-8567/04/4302–0163©2004 by the American Academy of Child and Adolescent Psychiatry. DOI: 10.1097/01.chi.0000101370.03068.87

Ozonoff, 1998). These behaviors often occur when the child is exposed to unpleasant or challenging conditions, such as instructional activities or academic demands (e.g., Carr and Newsom, 1985; McComas et al., 2000; Robbins and Dunlap, 1992). Noncompliance is one of the most common problems reported by parents of children with autism (Van Bourgondien, 1993). In a study examining responses to parental prohibitions (Arbelle et al., 1994), autistic children were noncompliant significantly more than both typically developing and developmentally delayed children. Zelazo (2001) suggests that noncompliance, which often emerges in autistic children at around 18 months, precludes imitative responses in these youngsters, thereby playing a significant role in arresting language development. He notes that intervention to improve compliance should be the first step in any program to stimulate social-communicative development in children with autism. Over the past several decades, some of the most commonly used strategies for managing problem behavior in children with autism and developmental dis-

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 43:2, FEBRUARY 2004

163

DUCHARME AND DRAIN

abilities involved the use of decelerative consequences such as timeout and manual guidance following each occurrence of problem behavior (e.g., Howlin, 1998). Such procedures have been demonstrated to be effective in reducing problem responses. However, potential limits and concerns regarding negative side effects are well documented (e.g., Masters et al., 1987; Sidman, 1989). In response to these concerns, researchers have focused on the development of new approaches to reduce the need for punitive or constraining consequences in the treatment of aberrant behaviors. For example, approaches based on operant learning principles, such as functional communication training, which involves replacing aberrant behaviors with communication responses that serve a similar function (Carr et al., 1994), are among the most effective strategies for reducing severe problem behaviors in persons with autism and developmental disabilities (Howlin, 1998). Another intervention developed from learning principles, errorless compliance training, has demonstrated promise in producing substantial reductions in child oppositionality in a variety of clinical populations without use of punitive consequences or coercion (e.g., Ducharme et al., 2000, 2001, 2002). The treatment approach focuses on child compliance, a skill that, when improved through intervention, can lead to a broad range of behavioral improvements in children with developmental and conduct difficulties (e.g, Ducharme and Popynick, 1993; Ducharme et al., 2000; Russo et al., 1981). The intervention involves an initial observational assessment of parent–child interactions to determine empirically the probability of child compliance with a range of requests and tasks. This assessment results in a hierarchical breakdown of demand situations into four compliance probability levels. At the beginning of treatment, children are initially exposed to level 1 of the hierarchy (i.e., requests that yield low levels of oppositionality). Children are rewarded (often with praise) for demonstrating prosocial responses to these easily manageable tasks. Lower probability levels involving more demanding and provocative conditions (levels 2 through 4) are gradually introduced over time at a slow enough rate to preclude noncompliant and other problem responses during and following intervention (thus the name “errorless”). Eventually, children learn to tolerate difficult demand situations (those previously associated with severe oppositionality) with substantially reduced noncompliance. Errorless com164

pliance training provides parents with effective childmanagement strategies without exposure of the children to the risk of excessive or misapplied punishments (Barkley, 2000). In recent studies, errorless compliance training was effective in reducing severe oppositional behavior in children from a variety of subgroups (e.g., child witnesses and victims of family violence, children with developmental disabilities, children of parents with brain injury). Evidence of a range of positive outcomes, such as behavioral covariation, generalization of effects, and maintenance over time has been found (e.g., Ducharme et al., 2000, 2001, 2002; Ducharme and Popynick, 1993). The purpose of the present research was to extend the findings of previous studies examining errorless compliance training in two ways. First, empirical research with this approach has never been conducted with children with autism. Given its graduated and noncoercive nature, the errorless approach could potentially diminish severe behavioral reactions associated with demand situations in these children, thereby obviating the need for constraining or coercive consequences to manage child oppositionality. Second, although earlier studies have confirmed the efficacy of errorless compliance training in improving child cooperation with general daily household requests (e.g., “put on your coat”; “turn off the TV”), they have not examined the effects of this approach on requests related to academic demands and tabletop learning tasks (e.g.,“draw a tree”; “show me the letter B”). Such an investigation would be particularly valuable for children with autistic tendencies who may fail to learn certain academic skills due to extreme distress reactions during exposure to these tasks. In the present study, we used the errorless approach with four children with autism who exhibited two areas of behavioral concern that are common with this population. First, they were unable to manage academic demands initiated by their parents without high levels of opposition. Second, their general compliance to everyday household requests by parents was low. Thus, these children presented severe management problems during most daily interactions. The study was designed to answer the following research questions: (1) Can a variation of errorless compliance training focusing on academic activities produce significant generalized increases in compliance to academic requests in children with autism? (2) Will this intervention, “errorless academic compliance train-

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 43:2, FEBRUARY 2004

ERRORLESS ACADEMIC COMPLIANCE

ing,” produce generalization to general household requests that were never included in treatment? METHOD Participants and Setting Participant children were referred to a community health care agency in Ontario that provided treatment for children with developmental disabilities. Five families initiated intervention. One family withdrew shortly after beginning baseline observations due to demands unrelated to the project. The four remaining children engaged in destructive, oppositional, or disruptive behavior in both academic and general demand situations. Child 1 was a 42-month-old boy who was diagnosed with autism by a psychologist using the Childhood Autism Rating Scale (Schopler et al., 1988) and the DSM-IV (American Psychiatric Association, 1994). His developmental level was estimated at around 18 months. He did not use words or gestures in a meaningful way. Child 2 was a 52-month-old boy who was diagnosed with autism by a pediatrician using the Autism Diagnostic Observation Schedule (Lord et al., 1989). He could communicate relatively effectively. Child 3 was an 80-month-old girl who was diagnosed with autism by a psychologist using the Gilliam Autism Rating Scale (Gilliam, 1995) and DSM-IV criteria. Her developmental level was estimated at 36 months and she could communicate relatively effectively. Child 4 was a 76-month-old boy who was diagnosed with autism by a psychiatrist using DSM-IV criteria. He had severe expressive language deficits and delayed receptive language. None of the children was taking medications or had comorbid diagnoses. Three mothers and one father served the role of primary intervention agent for the families. Participant parents ranged from 32 to 40 years of age. Education levels for the four parents were as follows: high school, some university education, an undergraduate degree, and a master’s degree. Three parents were employed outside of the home; the other was an in-home family manager. All parent-training workshops took place at the health care agency. Parents conducted all academic compliance sessions in the family or living room of their homes, where study materials and activities were accessible. All other requests were delivered in the area of the home where the request was most relevant. Experimental Design The design was a multiple baseline across subjects (Barlow and Hersen, 1984). Such time-series designs are ideal experimental strategies for developing and evaluating novel treatment approaches, as they allow systematic evaluation of both process and outcome of intervention (e.g., Morgan and Morgan, 2001). As required in multiple baseline designs, baseline evaluation for all participant children was initiated simultaneously and implementation of treatment procedures was time-lagged across children. Throughout baseline and treatment, child compliance was assessed observationally in time-series fashion. Dependent Measures Observations of Child Compliance. Parents collected compliance data in the home after training in the clinic (see below) throughout baseline, treatment, and follow-up. Child compliance was defined as initiation of the requested activity within 10 seconds of the request and completion within approximately 40 seconds. Portions of the home-based sessions in all phases of intervention were si-

multaneously observed and coded by a research therapist to obtain a measure of interobserver agreement. Interobserver agreement was collected on 27% of baseline requests, 36% of treatment requests, 21% of generalization requests, and 75% of follow-up requests. Interobserver agreement scores averaged 85% during baseline (range 72%–97%), 91% during treatment (range 69%–100%), 84% during generalization (range 53%–100%), and 94% during follow-up (range 85%–100%). Compliance Probability Checklists. We administered Compliance Probability Checklists similar to those used in previous errorless compliance training studies (e.g., Ducharme et al., 2000, 2001, 2002) to parents during the initial parent-training workshop. These checklists were used to assist with selection of requests for development of the compliance probability levels necessary to activate treatment with each child. Participant parents completed two checklists, one comprising 45 requests associated with academic and tabletop tasks (Academic Compliance Probability Checklist) and the other comprising 60 requests focusing on general areas of compliance such as dressing, hygiene, cleanup, and mealtime (General Compliance Probability Checklist). For each request, we asked parents to rate the likelihood of child compliance as “almost always” (76%–100%), “usually” (51%–75%), “occasionally” (26%–50%), or “rarely” (0%–25%). Parents also completed the Academic Compliance Probability Checklist following treatment. Parent’s Consumer Satisfaction Questionnaire. We adminstered a consumer satisfaction questionnaire (e.g., Ducharme et al., 2000, 2001, 2002) to the participant parents after treatment. The questionnaire comprised 15 items rated on a 5-point Likert scale. Items focused on satisfaction with the overall program and the therapist. Group Training Workshop 1: Introduction to Errorless Compliance Training Participant parents attended four clinic-based workshops interspersed throughout baseline and treatment. During the first workshop, parents provided background information about their children and the difficulties they were experiencing with behavior management. The trainer then provided parents with an overview of errorless compliance training. Finally, parents completed the two compliance checklists and signed consent forms. Group Training Workshop 2: Request Delivery and Data Collection Before the second group training session, the research therapist selected 24 requests from each of the completed Academic Compliance Probability Checklists (approximately 6 requests from each of the four probability levels) and 10 requests from each of the completed General Compliance Probability Checklists (approximately 5 requests from each of probability levels 3 and 4) as rated by parents during the first workshop. Only levels 3 and 4 general requests were selected, as general requests were used solely as measures of generalization (see below). The selected requests were printed on individualized baseline data sheets. During this workshop, parents learned to deliver requests effectively and to record their child’s response to requests on the baseline data sheets provided to them. For the purpose of this study, proper request delivery involved gaining the child’s attention before issuing a request, using single component requests, using the imperative rather than the interrogative, and issuing the request only once in a firm but polite voice. Parent training in data collection and request delivery was facilitated through use of modeling, role-playing, and performance feedback procedures.

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 43:2, FEBRUARY 2004

165

DUCHARME AND DRAIN

Baseline Sessions

Phase 1

Following workshop 2, parents collected observational data in the home to empirically determine the probability of their child’s compliance to the selected requests. During baseline assessment sessions, we asked parents to deliver the 34 requests (i.e., 24 academic and 10 general requests) in a natural manner throughout the day during regular tabletop or household activities. In addition, we asked parents to respond to noncompliance as they typically did but to avoid any physical interventions, such as spanking. After each request, parents recorded the child’s response (compliance or noncompliance) on the baseline assessment sheet. The number of baseline sessions per child ranged from 10 to 22 (session frequency was varied for research design purposes).

During phase 1, parents delivered only level 1 academic requests to their children. Request delivery procedures were identical to baseline. Following each compliant response, parents rewarded their children with social (e.g., praise, hugs, tickles) and occasionally other rewards. Early in treatment, it appeared that praise and social interactions were not potent reinforcers for children 2 and 3. For this reason, simple token systems that could be used to access more tangible rewards (e.g., computer time, treats, board games) were initiated to supplement social rewards for these two children. When the child was noncompliant, we asked parents to ignore the behavior by continuing with ongoing activities and providing no indications of disapproval. As in baseline, the parent recorded all child responses.

Hierarchy Development After all baseline sessions were completed, we calculated the probability of child compliance for each of the 34 requests by dividing the number of compliant responses by the total number of compliant and noncompliant responses and multiplying by 100. For each child, the requests were then arranged in order from highest to lowest compliance probability. The 24 academic requests were divided into four compliance probability levels (level 1, 76%– 100%; level 2, 51%–75%; level 3, 26%–50%; level 4, 0%–25%), with approximately 6 requests in each level. We selected four of the six academic requests from each level for use during treatment. For each compliance level, a separate treatment data sheet that included four repetitions of these four requests was constructed to be used in each phase of treatment (i.e., level 1 academic requests on the phase 1 data sheet, etc). Thus, four individualized treatment data sheets were developed for each child. Examples of academic requests that made up the hierarchy for a sample child included “count the _____,” “point to the _____” (level 1); “write your name,” “trace the _____” (level 2); “tell me what this is,” “put away the _____” (level 3); and “cut out the shape,” “draw me a picture” (level 4). The remaining two academic requests from each of the four levels were not included on these treatment data sheets as they were designated for posttreatment assessment of generalization. Additionally, the 10 requests selected from levels 3 and 4 of the General Compliance Probability Checklist (i.e., general requests) were not included on the treatment data sheets as these requests were used as a second posttreatment measure of generalization (see below). Examples of general requests for the sample child included “turn off the television,” “wash your face” (level 3) and “put your schoolbag away,” “pick up your _____(e.g., toys)” (level 4). Individual Training Workshop 3: Treatment Procedures We conducted workshop 3 individually for each family to ensure that parents were trained sequentially (for research design purposes) in the use of treatment procedures. Untrained parents continued baseline procedures until their turn to initiate treatment. At the beginning of the session, the trainer reviewed compliance definitions and request delivery procedures. Parents learned three primary treatment procedures, including reinforcing compliance (i.e., immediate, consistent, enthusiastic, and effective praise and/or other reinforcement), ignoring noncompliance (i.e., no punitive consequences, responding as though the request had never been delivered), and avoiding requests from subsequent levels (i.e., using gentle prompts to complete essential tasks that involve requests from succeeding levels). The trainer used modeling, role-playing, and performance feedback procedures to teach the required skills. The individualized phase 1 data sheet was provided to parents during this workshop.

166

Transition Sessions Between phases, parents conducted transition sessions to facilitate shifts between phases. These sessions occurred after the child demonstrated compliance levels of 75% or more for several consecutive days. Transition sessions involved a combination of requests from the two adjoining levels. Transition sessions were conducted until the child demonstrated compliance of 75% or higher for approximately two consecutive sessions. Phases 2, 3, and 4 The treatment procedures used in phases 2, 3, and 4 were identical to those in phase 1 with the exception of the probability level of the requests used (level 2 requests were used in phase 2, level 3 requests in phase 3, and level 4 requests in phase 4). At the beginning of each of these phases, we provided parents with the data sheet designated for that phase. Parent Training Workshop 4: Problem-Solving Midway through treatment, parents reconvened to discuss their children’s progress and any concerns. In addition, the trainer reviewed treatment procedures. Generalization Sessions Parents conducted generalization sessions after phase 4 of treatment. During generalization sessions, parents delivered level 3 and 4 academic requests that were not included in treatment sessions, as well as level 3 and 4 general requests. These sessions provided the opportunity to examine two types of generalization. With academic generalization requests, we examined whether the intervention produced broad generalization to other academic requests not included in treatment. With general requests, we investigated whether compliance to general household directives that were not the focus of treatment improved following treatment that included only academic requests. After the hierarchical breakdown of the 24 requests, there were not enough academic requests in levels 3 and 4 to separate treatment requests from generalization requests for child 3. Consequently, generalization to academic requests was not assessed for this child. Follow-up Sessions Academic Requests. Parents conducted academic request followup sessions 1, 2, and 6 months after treatment to evaluate maintenance of treatment effects. Level 4 treatment and generalization academic requests were used during these sessions. During aca-

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 43:2, FEBRUARY 2004

ERRORLESS ACADEMIC COMPLIANCE

demic request follow-up sessions, parents delivered two repetitions of the level 4 requests. General Requests. Parents conducted general request follow-up sessions at approximately 3 and 6 months after treatment to examine whether potential generalization effects for general household requests were maintained after treatment completion. During each general request follow-up session, parents delivered all level 3 and 4 general requests once to the child.

RESULTS Baseline

Figure 1 depicts time-series observational data for child compliance to academic requests across baseline and treatment phases for each child. Baseline data points on this graph represent percentage of child com-

Fig. 1 Percentage of compliance across all phases of intervention. Shaded areas represent mean baseline compliance levels to requests used in that phase.

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 43:2, FEBRUARY 2004

167

DUCHARME AND DRAIN

pliance to level 4 academic requests only, as these requests were the primary focus of the intervention (see also Ducharme et al., 2000, 2001, 2002). The mean probability of compliance across the four children during baseline was 88% for level 1 requests, 71% for level 2 requests, 57% for level 3 requests, and 35% for level 4 requests. Level 4 mean compliance levels ranged from a high of 46% for child 3 to a low of 18% for child 2. Treatment

The percentage of compliance to academic requests during each phase of treatment for each child is graphed in Figure 1. The shaded areas in each treatment phase indicate the baseline compliance level for the requests being trained in that phase. This shading permits a direct comparison between baseline and treatment levels of requests from the same probability level for each child. The mean percentage of compliance across the four children to level 1 academic requests in phase 1 of treatment was 89%, a level that was similar to baseline levels for these requests. During phase 2 of treatment, mean compliance to level 2 academic requests increased 10 percentage points over mean baseline levels to 81%. Substantial treatment effects began to surface in phase 3. Mean compliance to level 3 academic requests improved by 29 percentage points over baseline to 86%. Phase 4 included academic requests that yielded the lowest level of compliance during baseline. As can be seen in Figure 1 (comparing data points to the shading in phase 4), significant treatment effects emerged for all four children in this phase. The mean percentage of compliance across all children to level 4 academic requests was 86%, a mean improvement of 50 percentage points over baseline levels. Children 2 and 4 demonstrated the most pronounced treatment gains to level 4 requests, with mean increases of 70 and 60 percentage points respectively. By the end of phase 4, all four participants were exhibiting consistently high levels of compliance to level 4 requests, in comparison to typically low and variable compliance levels in baseline. During all transition phases, compliance levels were comparable to the four treatment phases.

of academic requests evaluated in baseline). Mean generalization across the three children to academic requests that were not included in treatment was 83% for level 3 requests (ranging from 78% for child 1 to 86% for child 4) and 85% for level 4 requests (ranging from 77% for child 1 to 96% for child 3). These figures represented mean gains over baseline levels of 26 percentage points for level 3 generalization requests and 52 percentage points for level 4 generalization requests. Mean generalization across all four children to general requests was 79% for level 3 requests (ranging from 67% for child 2 to 93% for child 3) and 78% for level 4 requests (ranging from 62% for child 2 to 94% for child 3). These results represent a mean improvement of 28 percentage points for level 3 requests and 51 percentage points for level 4 requests over baseline levels. Even child 2, who showed the lowest level of improvement to level 4 generalization requests, demonstrated an increase of 30 percentage points over baseline levels to these requests. Thus, errorless academic compliance training was associated with broad generalization to requests not included in treatment, including general household requests that occurred outside of academic tasks. Follow-up

The mean percentage of compliance to level 4 academic requests (which were a combination of both academic treatment and generalization requests) across the four children for the three follow-ups was generally well above baseline levels (Fig. 2). The mean compliance level was 83% at 1 month, 78% at 2 months, and 83% at 6 months. Compliance to general requests in follow-up was similarly encouraging. Mean compliance to a combination of level 3 and 4 general requests across the four children was 86% for the 3-month follow-up and 71% for the 6-month follow-up. Thus, out of 20 follow-up sessions for both academic and general requests across all children and all follow-up periods, compliance levels were more than 70% for all but 2 sessions, and more than 80% for 13 sessions. These results suggest that generalized compliance improvements were maintained after treatment.

Generalization

Generalization data for both academic and general requests were collected in sessions immediately following phase 4 of treatment (note that academic generalization could not be assessed for child 3 due to a dearth 168

Academic Compliance Probability Checklist

Parent responses to this checklist after intervention showed a substantial change in their perception of their child’s compliance. On average, parents rated only

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 43:2, FEBRUARY 2004

ERRORLESS ACADEMIC COMPLIANCE

Fig. 2 Mean percentage of compliance in baseline (BL) and follow-up (1, 2, 3, and 6 months) for academic (black bars) and general requests (striped bars).

40% of the requests as level 1 or 2 (higher compliance probability levels) before treatment. After intervention, they ranked 92% of the requests as level 1 or 2. Parent’s Consumer Satisfaction Questionnaire

The results of the Consumer Satisfaction Questionnaire indicated that the four parents were satisfied with the program and the therapists. The mean satisfaction score was 4.0 (out of 5) for the program and 4.5 (out of 5) for the therapists. DISCUSSION

In the present study, we examined whether a variation of errorless compliance training that focused on academic/tabletop requests would produce significant increases in compliance to these requests for children with autism. In addition, we evaluated generalization to nontreatment academic requests and to nonacademic general requests following intervention. The findings indicated that errorless academic compliance training was associated with substantial improvements in child compliance to low-probability academic requests. These improvements were encour-

aging, especially considering that these gains did not require the use of punitive or constraining consequences to reduce problem responses. Notwithstanding the high levels of resistance and protest behavior demonstrated in the face of parental demands before treatment, the children exhibited few aberrant responses under such conditions throughout and following intervention. These improvements appeared to be related to the graduated introduction of demands that greatly increased the probability of successful task completion by the child and ensured opportunities for acknowledgment and reinforcement from the parent. Thus, coercion or constraint of the child was not required to achieve task completion. Intervention effects extended to untrained level 3 and 4 academic requests, suggesting that errorless academic compliance training produced generalized cooperation to academic/tabletop activities. In addition, the intervention appeared to occasion a much more comprehensive form of generalization—that is, generalization to nonacademic household requests that represent the essence of cooperation between parent and child. Requests for this evaluation included such important

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 43:2, FEBRUARY 2004

169

DUCHARME AND DRAIN

tasks as “brush your teeth,” “turn off the TV,” and “wash your hands” that had been associated with serious pretreatment difficulties. These findings suggest that errorless academic compliance training may be a highly efficient strategy for enhancing a child’s willingness to cooperate with most tasks typically viewed as important by the parents of youngsters with autism. One-, 2-, and 6-month follow-up results suggested that improvements in academic compliance were maintained after treatment completion for all four children. Similarly, 3- and 6-month follow-ups for general requests demonstrated similar levels of maintenance. These findings indicate that errorless academic compliance training has the potential to produce clinically significant and durable improvements in the compliant responses of children with autism to both the teaching and everyday household requests of their parents. Limitations

This study represents the first demonstration of the effects of errorless academic compliance training on the cooperative responding of children with autism. As such, the sample size was necessarily limited to provide the opportunity to systematically examine the process and outcomes of this new intervention. Replications of this approach with a larger sample would help to determine whether the findings of the present study can be generalized to the diverse range of children that exhibit the broad spectrum of behavioral characteristics defining autism and other pervasive development disorders. A second limitation of the study was the length of time required to collect assessment data and carry out treatment. The intervention required a mean of 16 weeks of parent-conducted sessions, from the initiation of baseline to the completion of treatment (ranging from 13 weeks for child 2 to 19 weeks for child 1). Although this time span may appear longer than interventions that focus on more coercive approaches to inhibit problem responses, the extended compliance assessment that is required as part of the intervention provides clinicians with intensive knowledge about the conditions associated with prosocial and antisocial behavior for each child. This makes possible the development of an individually tailored intervention that is completely proactive, enabling abundant opportunities for acknowledged child success and greatly reducing child distress, thus obviating the need for confrontation, coercion, and constraining consequences (Ducharme et al., 2000, 2001). 170

Clinical Implications

The findings from the present study extend the existing literature on intervention approaches for children with autism and related disorders. Given that many children with autism display severe disruptive behavior in relation to academic tasks and other parental requests, errorless academic compliance training may provide an effective and efficient strategy to improve parent–child cooperation. Through initial exposure of these children to demands that are easily tolerated, hierarchical introduction of more demanding requests, and reinforcement of prosocial responses to requests during this graduated process, children are taught to tolerate difficult demands and to cease viewing parental requests as aversive events that must be escaped by means of disruptive behavior. Throughout this process, children learn that cooperative responses are easy and lead to pleasant and reinforcing outcomes (Ducharme, 2000). The fact that the intervention was successful with children from widely divergent developmental levels suggests potential efficacy with the broad range of youngsters with autistic characteristics. These findings support previous investigations in which errorless compliance training was effective with children representing a wide range of intellectual levels, from severe mental retardation to typical development (e.g., Ducharme and Popynick, 1993; Ducharme et al., 2000, 2002). Although errorless academic compliance training appears to be highly effective as a discrete intervention, it may prove advantageous as an adjunct to other treatment strategies that have been demonstrated to be effective for children with autism. For example, evidence suggests that one of the most potent treatments for achieving more normative cognitive and adaptive outcomes for children with autism is early intensive behavioral intervention using discrete trial training (Lovaas, 1987; McEachin et al., 1993), especially when conducted at a high level of intensity (e.g., 40 hours per week during early child development). This form of intervention requires child cooperation with numerous daily tasks and instructions, creating a learning environment that may be much more arduous than that to which these children are accustomed before treatment. As a result of these high demand conditions, children with autism may initially resist the demands and expectations through protest and aberrant behaviors (e.g., Lovaas, 1981). Errorless academic compliance training may serve as a useful precursor to discrete trial training programs by

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 43:2, FEBRUARY 2004

ERRORLESS ACADEMIC COMPLIANCE

reducing the aversiveness of the teaching atmosphere and enhancing the willingness of children to participate in challenging activities. In fact, shortly after the termination of the present treatment, child 1 was accepted into an early intensive behavioral intervention program. Consumer satisfaction comments revealed that the errorless intervention was viewed by the mother and the treatment therapists conducting the subsequent intensive intervention as highly beneficial in preparing child 1 for his excellent progress in his teaching program. A useful focus for future investigation would be the effects of the present intervention on child behavior in the school, where compliance to academic tasks is particularly relevant. Assessments of generalization in the school after home-based treatment would determine whether the broad generalization effects observed in the home extended to requests by teachers in the classroom. Of similar importance would be an evaluation of the effects of the intervention when implemented by school personnel, as teachers of children with autism would benefit from additional strategies for building the cooperation of these youngsters to classroom demands. In summary, the present investigation demonstrated that errorless academic compliance training may be a parsimonious strategy for assisting parents in noncoercive management of two of the most common complaints of children with autism: inability to cooperate with teaching activities and unwillingness to complete everyday household tasks requested by parents. The intervention appeared to engender substantial generalized improvements in child cooperation in both of these request domains without parent–child confrontation. The intervention appears to be useful as a standalone strategy or as an antecedent to more intensive behavioral teaching approaches for children with autism. REFERENCES American Psychiatric Association (1994), Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV). Washington, DC: American Psychiatric Association Arbelle S, Sigman MD, Kasari C (1994), Compliance with parental prohibition in autism. J Autism Dev Disord 24:693–702 Barkley RA (2000), Commentary: issues in training parents to manage children with behavior problems. J Am Acad Child Adolesc Psychiatry 39:1004–1007

Barlow DH, Hersen M (1984), Single Case Experimental Designs. New York: Pergamon Carr EG, Levin L, McConnachie G, Carlson JI, Kemp DC, Smith CE (1994), Communication-Based Intervention for Problem Behavior. Baltimore: Brookes Carr EG, Newsom CD (1985), Demand-related tantrums: conceptualization and treatment. Behav Modif 9:403–426 Ducharme JM (2000), Treatment of maladaptive behavior in acquired brain injury: remedial approaches in post-acute settings. Clin Psychol Rev 20:405–426 Ducharme JM, Atkinson L, Poulton L (2000), Success-based, noncoercive treatment of oppositional behavior in children from violent homes. J Am Acad Child Adolesc Psychiatry 39:995–1004 Ducharme JM, Atkinson L, Poulton L (2001), Errorless compliance training with physically abusive mothers: a single-case approach. Child Abuse Negl 25:855–868 Ducharme JM, Popynick M (1993), Errorless compliance to parental requests: treatment effects and generalization. Behav Ther 24:209–226 Ducharme JM, Spencer T, Davidson A, Rushford N (2002), Errorless compliance training: building a cooperative relationship between braininjured parents at risk for maltreatment and their oppositional children. Am J Orthopsychiatry 72:585–595 Fombonne E (1998), Epidemiological surveys of autism. In: Autism and Pervasive Developmental Disorders, Volkmar FR, ed. New York: Cambridge University Press, pp 32–63 Gilliam JE (1995), Gilliam Autism Rating Scale Manual. Austin, TX: Pro-Ed Howlin P (1998), Children With Autism and Asperger Syndrome: A Guide for Practitioners and Carers. New York: Wiley Lord C, Rutter ML, Goode S (1989), Autism Diagnostic Observation Schedule: a standardized observation of communicative and social behavior. J Autism Dev Disord 19:185–212 Lovaas OI (1981), Teaching Developmentally Disabled Children. Austin, TX: Pro-Ed Lovaas OI (1987), Behavioral treatment and normal educational and intellectual functioning in young autistic children. J Consult Clin Psychol 55:3–9 Masters JC, Burish TG, Hollon SD, Rimm DC (1987), Behavior Therapy. San Diego, CA: Harcourt Brace Jovanovich McComas J, Hoch H, Paone D, El-Roy D (2000), Escape behavior during academic tasks: a preliminary analysis of idiosyncratic establishing operations. J Appl Behav Anal 33:479–493 McEachin JJ, Smith T, Lovaas OI (1993), Long-term outcome for children with autism who received early intensive behavioral treatment. Am J Ment Retard 97:359–372 Morgan DL, Morgan RK (2001), Single-participant research design: bringing science to managed care. Am Psychol 56:119–127 Prior M, Ozonoff S (1998), Psychological factors in autism. In: Autism and Pervasive Developmental Disorders, Volkmar FR, ed. New York: Cambridge University Press, pp 64–108 Robbins FR, Dunlap G (1992), Effects of task difficulty on parent teaching skills and behavior problems of young children with autism. Am J Ment Retard 96:631–643 Russo DC, Cataldo M, Cushing PJ (1981), Compliance training and behavioral covariation in the treatment of multiple behavior problems. J Appl Behav Anal 14:209–222 Schopler E, Reichler RJ, Renner BR (1988), The Childhood Autism Rating Scale (CARS). Los Angeles: Western Psychological Sciences Sidman M (1989), Coercion and Its Fallout. Boston: Authors Cooperative Van Bourgondien ME (1993), Behavior management in the preschool years. In: Preschool Issues in Autism, Schopler E, Van Bourgondien ME, Bristol MM, eds. New York: Plenum, pp 129–145 Zelazo P (2001), A developmental perspective on early autism: affective, behavioral, and cognitive factors. In The Development of Autism, Burack JA, Charman T, Yirmiya N, Zelazo PR, eds. Mahwah, NJ: Erlbaum, pp 39–60

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 43:2, FEBRUARY 2004

171