Evaluation of Behavioural Treatment

Evaluation of Behavioural Treatment

The small steps forward and the contributions made by hundreds of investigators have yielded the kind of outcome we have learned to anticipate in science. After some 40 years of behavioural treatment-research based on some 90 years of basic research in learning processes, thousands of scientific studies have been published, forming the many basic components of behavioural treatment. Cumulatively, enough data-based treatment procedures have been developed to yield a most favourable outcome: major improvements in intellectual, educational, social, and emotional functioning in children with developmental delays. This success includes average gains of approximately 10 to 20 IQ points on standardized tests of intelligence (e.g., Anderson, Avery, DiPietro, Edwards, & Christian, 1987; Harris, Handleman, Gordon, Kristoff, & Fuentes, 1991; Lovaas, 1987), similar gains on other standardized tests (e.g., Anderson et al., 1987; Hoyson, Jamieson, &. Strain, 1984; McEachin, Smith, &. Lovaas, 1993), and placement in less restrictive classrooms than those typically offered to children with developmental delays (e.g., Fenske, Zalen-ski, Krantz, & McClannahan, 1985; Lovaas, 1987). Smith (1999) provides a comprehensive review of these and similar studies.

In this chapter we illustrate how intensive and early behavioural treatment of individuals with developmental delays evolved, and we do so by focusing on data from treatment outcome studies. We also describe the criteria for appropriate treatment that resulted from the development of behavioural intervention. Note that in this discussion terms such as we and our are often used for editorial purposes. This should not be taken to imply that the major components of the treatment program were developed at the University of California, Los Angeles (UCLA). Rather, the design of the treatment was based on extensive research published in peer-reviewed journals by independent investigators. Most of the studies have been replicated. Also, note that none of these studies have reported a major breakthrough. Rather, progress in treatment developed in a gradual and stepwise fashion, such as that done in the building of a pyramid where a solid foundation (learning theory research) forms the basis for adding additional structure. Research projects at UCLA help to illustrate this development. Investigators have contended that the best way to know what aspects of a treatment make that treatment effective is by studying individual components of the treatment separately before performing an outcome study (Johnston, 1988).

The 1973 Treatment Study

The first comprehensive treatment study was started in 1964 and reflected many blind alleys in our own design (Lovaas, Koegel, Simmons, &. Long, 1973). First, we worked with the hope that if we removed the children we treated from their natural environments and placed them in an institutional setting, they would be free from distracting stimuli and we would be in a better position to treat them. Second, we hoped that 1 year of intensive one-on-one treatment (2,000-plus hours) would be enough and that treatment gains would last. Finally, we focused our major efforts on developing language because we hoped that language would be pivotal in facilitating improvement in untreated behaviours. None of these hopes were realized. However, major gains in teaching complex behaviours such as language were observed. Many have questioned the efficacy of behavioural interventions in building language (Chomsky, 1965). Nonetheless, studies have indicated that effective procedures for teaching language to children with autism and other developmental delays can be derived from laboratory research on operant conditioning involving discrimination learning, discrete trials, prompting and prompt fading, shaping, and chaining (see Chapters 10 and 16).

To illustrate, beginning steps in language training involve teaching students to understand simple instructions (Lovaas, 1977). Subsequently, students are taught to imitate simple sounds, followed by combining these sounds into syllables and then words (Lovaas, Berberich, Perloff, & Schaeffer, 1966). Once students are able to imitate words, they are taught to label items and events and then to request favorite objects and activities (Risley & Wolf, 1967). Subsequently, students are taught to combine words into simple sentences such as 'I want (item)' and 'This is an (object)' (Risley, Hart, & Doke, 1972). Students who proceed at a rapid rate may then be instructed in how to use abstract concepts, including yes-no (Hung, 1980), plurals (Baer, Guess, & Sherman, 1972), adjectives (Risley et al., 1972), prepositions, pronouns, opposites such as big-little and hot-cold, and time relations such as first-last and before-after (Lovaas, 1977). Once these concepts are mastered, students are taught to ask questions (Hung, 1977; Lovaas, 1977) and engage in simple conversations (Gaylord-Ross, Haring, Breen, & Pitts-Conway, 1984; Lovaas, 1977).

Researchers have also described procedures for facilitating conversational speech between children with autism and normally developing peers (Charlop & Mil-stein, 1989; Gaylord-Ross et al., 1984; Haring, Roger, Lee, Breen, &. Gaylord-Ross, 1986). Thus, experimentally validated procedures exist for helping students progress from being mute or echolalia to possessing some or all of the language skills displayed by typical children and adults. Children can also be helped to develop learning strategies such as nonverbal imitation and to reduce tantrums and self-injurious behaviours.

Despite all of our successes, we may have made the most progress by recognizing our mistakes. Our assumption that increases in language would be associated with concurrent improvements in other areas of functioning was not supported. This was a major disappointment because we had hoped that, once the children learned to talk, they would develop the kind of response generalization that would 'push them over' into normalcy. Instead, the children revealed themselves to be without much prior knowledge. A 'little child' did not seem to be hiding inside, waiting for the opportunity to come out from the autistic shell and talk with us, as many theoretical formulations had postulated (and still do). Nonetheless, the acquisition of language gave the clients access to future educational environments where additional appropriate behaviours could be built (a kind of 'successive' response interaction).

The second lesson we learned during this time concerned the lack of generalization across environments, including post treatment environments (see Stokes & Baer, 1977, for an instructive paper on ways to increase generalization across behaviours and environments). When we discharged the clients to the state hospital from which they came, they inevitably regressed. It was heartbreaking

to observe Pam and Rick, children who had gained so much with us, slowly but surely lose the skills they had acquired. When we brought the children back for treatment a second time, they recovered many of the gains they had made during the first treatment period, only to lose them again after their second discharge. The data we secured before, during, and after treatment served as the most important guide for development of the project as it stands today. It may have been possible for us to fool ourselves without such data; others apparently had. By that time we knew there were no shortcuts. Instead, we realized there was a great deal of hard work ahead.

At about this same time, we learned another bitter lesson: Colleagues from other orientations considered it extremely difficult to adopt and test treatment programs developed from the behavioural perspective. Our pleas to continue working with the children placed in the hospital were met with immediate rejection by the hospital staff. It was their opinion that the children were to be regressed in order to recover the childhood experiences they had been deprived of. Only then, the hospital staff asserted, could the children develop into normal individuals. The staff could not imagine how this behaviour modification program, the same one that trained dogs and pigeons into becoming robots, could ever actually facilitate the development of human beings.

Much in contrast to the children placed in the state hospital, the children who were discharged to parents eager to be informed about our treatment did much better at maintaining the skills gained through the intervention. The important role parents play as colleagues in treatment is discussed later in this manual.

The 1987 UCLA Young Autism Project

Six observations made during the 1973 treatment-research study played a major role in the design of our next effort (Lovaas, 1987). First, we made the serendipitous discovery that the youngest children in the 1973 study made the greatest progress. Second, we learned that treatment effects were situation specific. Thus, we moved treatment away from a hospital or clinic setting and into the children's homes and other everyday environments. Third, we found limited evidence for response generalization and therefore designed treatments for most or all of the children's behaviours. Fourth, we learned that parents could become skilled teachers, and they were the best allies one could ask for in accelerating and maintaining treatment gains. Fifth, we offered treatment for most of the children's waking hours for 2 or more years, and taught the children to develop friendships with typical peers in an attempt to continue treatment at that level. This arrangement more closely resembled that available to typical children who learn from their environment (parents, peers, etc.) from morning to night, vacations included. Finally, and most important, by the 1970s we possessed a large range of data-based procedures that could be amalgamated so as to enrich our treatment program. Through these procedures, our treatment expanded to consist of hundreds of separate teaching programs.

The 1987 UCLA Young Autism Project generated major and lasting increases in intellectual, educational, social, emotional, and other aspects of behaviour (McEachin et al., 1993). The question is often raised as to why the younger children did so much better than the older ones. There could be several reasons for this. One reason may be that the intensive treatment (40 one-on-one hours per week) was started early enough so that a sizable minority (47%) could 'catch up' and acquire an adequate amount of language, social, play, and self-help behaviours so as to be successfully mainstreamed among typical children in regular preschools. Once kindergarten was successfully passed, the children went on to successfully pass the first grade and subsequent classes in public schools. The development of friendships by these children with average children may have helped to build further pro-social behaviour and protect against relapse.

There are other potential explanations for why younger children did better than older children. Laboratory studies on animals have shown that alterations in neurological structure are quite possible as a result of changes in the environment in the first years of life (Sirevaag & Greenough, 1988). There is reason to believe that alterations are also possible in young children. For example, children under 3 years of age overproduce neurons, dendrites, axons, and synapses. Huttenlocher (1984) hypothesized that, with appropriate stimulation from the environment, this overproduction might allow infants and preschoolers to compensate for neurological anomalies much more completely than older children. Caution is needed in generalizing these findings from studies on average children to early intervention with children with autism, particularly because the exact nature of the neurological anomalies in children with autism is unclear at present (e.g., Rutter & Schopler, 1987). Nevertheless, the findings suggest that intensive early intervention could help compensate for neurological anomalies in children with autism. Finding evidence for this type of compensation would help explain why the treatment in the UCLA study was effective. More generally, it might contribute to a better understanding of brain-behaviour relationships in young children.

It is a compliment to the field of Applied Behaviour Analysis that literally hundreds of investigators have been able to generate thousands of replicable studies that add in a cumulative manner to a vast array of useful knowledge. No other area within clinical psychology, special education, or other helping professions within the social sciences has accomplished this. Along the same lines, there is every reason to believe that progress in other fields (e.g., special education, psychotherapy) will not occur until such a strategy of replicable and cumulative findings is established.

Adversity

It was not always smooth sailing during our investigations. We would have been in a better position to protect ourselves from disappointments had we been forewarned of the adversity we would experience. Parents who have sought financial assistance from state agencies, such as school districts, to help pay for their children's treatment have too often been subjected to major distortions about the nature of behavioural treatment (see Chapter 40). Protection from disappointment for a service provider is also important when it appears that he or she must stay in it for the long run, as when trying to help severely disadvantaged children and their parents.

An area of disappointment we experienced pertains to dissemination. There is at least a 25-year delay between what is now known about how to teach children with developmental delays and what has been adopted. Paradoxically, special education teachers appear eager to receive training in behavioural treatment and are in a good position to deliver such treatment. It is difficult to know where the obstacles are. When treatment is adopted, it is often a watered-down version. Watering down of treatment is nothing new; it has occurred since the inception of attempts to treat individuals with developmental delays (cf. Lane's, 1976, description of Itard's work).

Another area of disappointment centres on pronouncements by colleagues representing other areas of investigation. These take on numerous forms. There has been an alarming tendency in psychological treatment-research to attribute failure to the client, as in invoking organic limitations when treatment fails. For example, in their review of research of autistic children, DeMeyer, Hingtgen, and Jackson (1981) concluded that 'infantile autism is accompanied by permanent intellectual/behavioural deficits' (p. 432), adding that no one would even give lip service to changing such deficits. Zigler and Seitz (1980) suggested that one would fail in one's efforts to alter IQ scores to any substantial degree. Some also attribute failures to defects of the investigator. Spitz (1986) characterized those who have reported increases in intellectual functioning accompanying educational enrichments as 'fools, frauds, and charlatans.' Still other colleagues propose that behavioural treatment is harmful. Bettelheim (1967) attacked behavioural treatment as follows: 'Perhaps we may say of the operant conditioning procedures what has been said of lobotomy: that lobotomy changes a functional disorder that is potentially recoverable into an organic one for which there is no treatment' (p. 411). More recently, Greenspan (1992) presented 'behavioural schools of thought' as an example of a 'common unhelpful approach' that 'ignores the delayed child's many needs' and allows 'disordered patterns to become more stereotyped and more perseverative as [the children] grow' (p. 5). These comments are obviously inconsistent with findings from scientifically sound research.  Functioning appear to represent the only areas that develop spontaneously (as side-effects) during treatment.

Second, little or no data exist to support the notion that the changes in behaviours learned in one environment or taught by one or two teachers transfer to other environments or to other persons. Rather, there is evidence for 'situation specificity' in treatment effects. This implies that the individual must be treated in all significant environments (home, community, and school) and by all significant persons (family, teachers, and friends).

Third, there is strong evidence of relapse if treatment is discontinued. The only data-based exception to this can be observed in very young children with autism and developmental delays who are treated with intensive behavioural intervention through which a significant minority can be successfully mainstreamed and achieve normal functioning (McEachin et al., 1993). Other than such children, data from research implies that clients must be in special educational environments all their lives. In short, optimal treatment effects require a much more comprehensive intervention than previously considered.

Implications of Data from Behavioural Treatment Research

Data from research that supports the need for persons with developmental delays to receive the comprehensive treatment described in this manual are summarized in this section. First, there is no evidence that the changing of one behaviour changes any of the individual's other behaviours to a significant degree. There is no evidence for the existence of a 'pivotal' or 'critical' behaviour which, when altered, suddenly brings about large-scale progress in overall functioning. For example, with increased language skills, there is little to no evidence that play behaviour and self-help skills show a concurrent change. On the other hand, an increase in language skills should facilitate the teaching of later skills such as peer play and academic achievements. Similarly, helping an individual reduce tantrums and other interfering behaviours should facilitate the teacher's effectiveness and help the individual enter a less restrictive environment where alternate behaviours may further be taught. If scores on tests of intelligence represent what an individual has learned from his environment up to the time of testing, then one may reasonably expect such scores to increase with increased exposure to effective educational environments. Increased attention, higher IQs, and improvement in emotional

Criteria for Appropriate Treatment

It has been possible to formulate a consensus among many scientific researchers and practitioners that appropriate treatment contains the following elements (Simeonnson, Olley, & Rosenthal, 1987).

1. A behavioural emphasis. This involves not only imposing structure and rewarding appropriate behaviours when they occur, but also applying some more technical interventions. These

interventions include conducting discrete trials, shaping by successive approximations, producing shifts in stimulus control, establishing stimulus discriminations, and teaching imitation (R. L. Koegel & Koegel, 1988).

2. Family participation. Parents and other family members should participate actively in teaching the person who is developmentally delayed. Without such participation, gains made in professional settings such as special education programs, clinics, or hospitals rarely lead to improved functioning in the home or community (Bartak, 1978; Lovaas et al., 1973).

^ 3. One-to-one instruction. For approximately the first 6 to 12 months of treatment, instruction should be individualized rather than in groups because persons with autism and other developmental delays learn more readily in one-on-one situations (R. L. Koegel, Rincover, & Egel, 1982). This training needs to be supervised by degree professionals educated in Applied Behaviour Analysis and trained in one-on-one treatment. Treatment may be administered by people who have been thoroughly trained in behavioural treatment, including undergraduate students and family members (O. I. Lovaas & Smith, 1988).

4. Integration. Prior to integration in a group setting, the individual should be taught as many socially appropriate behaviours as possible. When an individual is ready to enter a group situation, the group should be as typical (normal or average) as possible. Persons with autism perform better when integrated with typical persons than when placed with other autistic individuals (Strain, 1983). In the presence of other persons with autism, any social and language skills the individual may have developed usually disappear within minutes, presumably because these behaviours are not reciprocated (Smith, Lovaas, & Watthen-Lovaas, 2002). Mere exposure to typical persons, however, is not sufficient to facilitate appropriate behaviour. Persons with autism require explicit instruction on how to interact with their peers (Strain, 1983).

5. Comprehensiveness. Persons with autism initially need to be taught virtually everything. They have little appropriate behaviour, and new behaviours have to be taught one by one. As mentioned earlier, this is because teaching one behaviour rarely leads to the emergence of other behaviours not directly taught (Lovaas & Smith, 1988). For example, teaching language skills does not immediately lead to the emergence of social skills, and teaching one language skill, such as prepositions, does not lead directly to the emergence of other language skills, such as mastery of pronouns.

6. Intensity. Perhaps as a corollary for the need for comprehensiveness, an effective intervention requires a very large number of hours, about 40 hours per week (Lovaas & Smith, 1988). Ten hours per week is inadequate (Lovaas & Smith, 1988), as is 20 hours (Anderson et al., 1987). Although increases in cognitive function (as reflected in IQ scores) may be observed, this should be understood to mean not that the student will be successfully integrated among typical peers, but rather that the student may regress unless treatment is continued. The majority of the 40 hours, at least during the first 6 to 12 months of intervention, should place a major emphasis on remedying language deficits (Lovaas, 1977). Later, this time may be divided between promoting peer integration and continued remedying of language deficits.

^ 7. Individual differences. Large individual differences exist in students' responses to behavioural treatment. Under optimal conditions, a sizable minority of children gain and maintain so-called normal functioning (McEachin et al., 1993). These are children who can be labelled auditory learners. The remaining children, visual learners, do not reach normal functioning with behavioural treatment at this time and are likely to require intensive one-on-one treatment for the remainder of their lives to maintain their existing skills and continue to develop new skills. Programs designed to facilitate communication for visual learners appear promising but are in need of further research and outcome data (see Chapters 29 and 30).

8. Duration. Treatment must last for the lifetime of the person with autism because termination of treatment is likely to lead to loss of treatment gains (Lovaas et al., 1973). As previously discussed, the only data-based exception to this is for that proportion of young children who reach normal functioning with intensive and early behavioural intervention by the time they are 7 years of age.

9. Quality control. Given the visibility of the UCLA data, it is important to specify as many dimensions of the treatment as possible so that it can be replicated by others. This becomes particularly important because almost anyone can falsely present oneself as qualified to deliver such treatment by, for example, having attended a 1-day or 1-week seminar and having read The ME Book (Lovaas, 1981) or the present manual. To help protect against potential misunderstandings, we strongly suggest that the reader refer to Chapter 34. To further protect against misunderstandings, a system of certifying aides needs to be created, helping to ensure that the quality of behavioural treatment remains high. The UCLA certification project may help guide such an effort. There are currently two levels of certification through the project: Level I (Staff Aide) and Level II (Supervising Aide).

Criteria for Certification at Level I, Staff Aide

1. Take the UCLA course Psychology 170A (Behaviour Modification) or a similar course in learning theory that provides the theoretical and empirical bases for the experimental analysis of behaviour and their application to autism and other developmental delays.

2. If Requirement 1 is passed with a grade of B or better, enter Psychology 170B (Fieldwork in Behaviour Modification), which includes lectures and supervised practicum experience in one-on-one treatment for a minimum of 6 months (not less than 60 hours of one-on-one treatment) supervised by a Level II aide.

3. After completing Requirements 1 and 2, submit a 15-minute videotape of the applicant conducting one-on-one treatment with a client with whom the applicant has worked a minimum of 10 hours. The Level II aide who supervises the client selects one program that is currently on acquisition for the child in each of the following areas: (a) verbal imitation (if the child has not yet begun the Verbal Imitation Program, a component of the Nonverbal Imitation Program is selected; if the child has mastered all components of the Verbal Imitation Program, a conversation-related program is selected); (b) receptive language; and (c) expressive language (if the child has not yet begun expressive language programs, a component of the Nonverbal Imitation Program is selected [if a Nonverbal Imitation Program component was used in , this category should consist of a second component of the Nonverbal Imitation Program; for example, if was imitation of facial expressions, this category may involve imitation of drawing or block building]). The applicant is videotaped as he or she conducts each program for 5 minutes. Only the applicant, the client, and the cameraperson are present during these sessions, and the cameraperson refrains from making any comments on the applicant's performance until after the completion of the videotaping. Scoring is based on a measure developed by R. L. Koegel, Russo, and Rincover (1977).

Criteria for Certification at Level II, Supervising Aide

1. Obtain Level I certification.

2. Complete an intensive 9-month, full-time internship or 1 year of full-time service as a staff member supervised by a Level II aide and a doctoral-level director of clinical services. For at least 3 of these months, the applicant must be engaged in training novice aides (i.e., aides who have not yet obtained Level II certification).

3. Complete assigned readings on the application of learning theory and obtain a satisfactory grade on a test based on these readings.

4- Obtain satisfactory ratings from the Level II supervisor and from the novice aides trained by the applicant.

5. After completing Requirements 1 through 4, submit a 20-minute videotape of the applicant conducting one-on-one treatment with a client with whom the aide has never worked. Based on a review of the client's records, the applicant teaches novel items to the client in each of the following areas: (a) verbal imitation (if the child has not yet begun the Verbal Imitation Program, a component of the Nonverbal Imitation Program is selected; if the child has mastered all components of the Verbal Imitation Program, a conversation-related program is selected); (b) receptive language; (c) expressive language (if the child has not yet begun expressive language programs, a component of the Nonverbal Imitation Program is selected [if a Nonverbal Imitation Program component was used in , this category should consist of a second component of the Nonverbal Imitation Program; for example, if was imitation of facial expressions, this category may involve imitation of drawing or

block building]); and (d) interactive play. Only the applicant, the client, and the cameraperson are present during these sessions, and the cameraperson refrains from making any comments on the applicant's performance until after completion of the videotaping. The completed videotape is scored by a reviewer in the Multi-site Young Autism Project.

6. Obtain recertification every 2 years by submitting a videotape like the one in Requirement 5. Only Level II aides are authorized to supervise treatment or conduct workshops. All applications for Level I and Level II aides are reviewed by doctoral-level persons who are key personnel in the Multi-site Young Autism Project (NIMH 1 R01 MH48863-01A4).

Are Other Treatments Effective?

In deciding how to treat individuals with developmental delays, it is crucial to obtain information about the effectiveness of the various treatments that claim to be of help. A detailed review is not provided in this manual because there are several recent and comprehensive discussions and evaluations of the many treatments proposed to help persons with autism and other developmental delays. For example, G. Green (1996a) has provided certain guidelines by which one can evaluate whether a given treatment has proven to be effective. Facilitated Communication and Sensory Integration are used as examples to illustrate how such an evaluation may be conducted. G. Green (1996b) has also provided an assessment of early behavioural intervention for autism, describing the strengths and weaknesses of this particular treatment approach.

Smith (1996) wrote a critical review of a large number of alternative interventions, such as Project TEACCH; the Higashi School; sensory-motor therapies such as Sensory Integration, Auditory Integration, and Facilitated Communication; and psychotherapies such as Floor-Time, Holding Therapy, Option Method, and Gentle Teaching. It is distressing to note that, despite the wide use of these therapies, either outcome data are missing altogether or available data fail to support the effectiveness claimed by the originators of these programs. Likewise, many individuals with autism are provided with speech and language therapy, yet there appear to be no scientific studies demonstrating that speech and language therapy is effective with this population. The Natural Language Paradigm, Incidental Teaching, Pivotal Response Training, and Fast For Word have similarly not been supported by objective data pertaining to long-term evaluation and comprehensive treatment outcomes. While many treatments lack supporting scientific data, some treatments can actually cause harm.

Smith (1996) concluded his review of treatments by noting the large scientific support for behavioural treatment and advised parents that the best initial course of action may be to concentrate exclusively on carrying out behavioural treatment as well as possible rather than looking for ways to supplement it with other treatments. On the other hand, for individuals who are found to progress slowly in behavioural treatment, alternative interventions should be supplementary, although no guidelines exist at this time for deciding which ones.

When evaluating whether or not to apply a treatment, the cost of treatment should be considered. This issue relates to whether funds are distributed equally across those who need help or absorbed by a small group of individuals. Similar issues deal with whether the cost of treatment will eventually result in savings to the state. That consideration involves whether the client at some point will be able to seek employment, pay taxes to repay the state's expense, and no longer be in need of services.

A conservative estimate of the incidence of autism can be placed at 1 in 250, which means that in the United States approximately 800,000 people are diagnosed with autism. Autism is the third most common childhood disorder after cerebral palsy and mental retarddation, and occurs at a higher rate than childhood cancer, cystic fibrosis, and Down syndrome. Autism has been estimated to cost the United States in excess of $10 billion per year. Intensive early behavioural intervention is estimated at $5,000 per month (slightly over $30 per hour), or $60,000 per year. The average length of treatment for the 47% of children who reach normal functioning is 2 years, resulting in a total expense of $120,000 per child. Such persons will in all likelihood be employed and pay taxes at some time after treatment. The estimated average cost per individual who needs lifelong protective and institutional care is $40,000 per year or $2.4 million for the 60 or more years that such services will be rendered. The savings accrued from intensive behavioural treatment for those who do not recover is also substantial (Jacobson, Mulick, & Green, 1996). A recent study by Smith, Groen, and Wynn (2000) compares the effectiveness of early and intensive behavioural intervention for two groups of children, one group diagnosed with autism and the other with pervasive developmental disorder-not otherwise specified (PDD-NOS). Data show that the children diagnosed with PDD-NOS may have gained even more from the treatment than the children diagnosed with autism. Besides showing support for the continuity model (see Chapter 2), these data support the inference that it is

economically wise to invest in behavioural treatment.