Applied Behavior Analysis and Autism

Applied Behavior Analysis and Autism Applied Behavior Analysis & Autism Reichow, B. And Wolery, M. (2008, June). Comprehensive synthesis of early intensive behavioral interventions for young children with autism based on the UCLA Young Autism Project model. Journal of Autism and Developmental Disorders, 39, 23-41. DOI 10.1007/s10803-008-0596-0 The use of applied behavior analysis as a treatment intervention for young children with autism is a frequently used and applauded model. The researcher most commonly associated with applied behavior analysis is Ivar Lovaas of the University of California at Los Angeles.

Lovaas work with applied behavior analysis, commonly referred to as ABA, began in the 1960s. Eventually, Lovaas work culminated in the establishment of the early intensive behavioral intervention (EIBI) programs for young children with autism independent of UCLA, but which were based on Lovaas' Young Autism Project (YAP). A comprehensive review of 10 intervention programs by the Committee on Educational Interventions for Children with Autism of the National Research Council (NRC) did not result in recommendations of any of the programs.

Rather, the committee cited a need for more research and issued a set of general program guidelines. Methodology The purpose of this study by Reichow and Wolery was to provide a comprehensive synthesis of research conducted on EIBI. The authors utilized three methods for analysis of the EIBI programs: Descriptive analysis, effect size analysis, and meta analysis. The study methodology is robust in its examination of experimental methods, selection and assessment of participants, and intervention program. Study selection.

Articles were selected for the analysis based on criteria that associated them with the YAP program, with Lovaas' methods, or with funding from the National Institute of Mental Health Multi-Site Young Autism Program (MYAP). A second set of criteria was used to ensure that the research included in the analysis showed content validity with regard to participant characteristics, participant diagnoses, duration of programming, and experimental design. Finally, the research must have been published in a peer-reviewed journal in the English language. Coding of data & reliability.

All data was manually coded and inter-observer agreement, assessed in 4 out of 14 samples by two independent recorders, was found to be 91.6% -- sufficiently high to consider the data coding reliable. Outcome data and three levels of study characteristics -- research methods, participant characteristics, and intervention characteristics -- were defined and coded. Results Descriptive analysis. In order to evaluate the influence of experimental methods on study outcomes, the following five methodological areas were analyzed: (1) Experimental rigor; (2) study design; (3) method of group assignment; and (4) procedural fidelity.

The overall rigor of the studies analyzed was acceptably high, however the studies had important methodological limitations. Data collection often occurred after the intervention had concluded or, in other instances, intervention was begun before data collection started. Random group assignment -- a tenet of empirical studies -- was used in only 2 of the 13 studies. Little data could be ascertained about the comparison treatment conditions and uniformity did not exist across the studies.

Further, standardization across the comparison groups was lacking, there were no measures of procedural fidelity, and data on whether subjects in the groups received supplemental intervention was absent. Assessment of procedural fidelity was mixed across the studies. Other studies of IEBI have shown that parents and therapists may find it difficult to achieve good levels of procedural fidelity. The authors suggest that parametric studies of the level of precision needed for EIBI to be effective should be established in order for these measures to be empirically meaningful.

Perhaps the most important finding in this part of the analysis was that the constructs of academic placement and diagnostic reclassification -- commonly referred to as recovery -- were questioned as measures of outcome. Academic placement was found to be a flawed outcome measure that should not be used as an indication of the effectiveness of interventions.

And diagnostic reclassification must be determined through the use of valid instruments designed specifically for the diagnosis of autism and such testing must be conducted by qualified professionals who are unaware of either earlier diagnosis or membership in an experimental group. Effect size analysis. While effect sizes were large enough to suggest that "children receiving EIBI made more gains than children who received minimal behavior intervention, eclectic treatment, or treatment as usual," conclusions must be limited with regard to stating the superiority of EIBI over other treatments.

As mentioned in the descriptive analysis, this is due to inadequate comparison groups and non-random assignment of subjects into groups. Meta-analysis. The comparison group data could not be included in the meta-analysis because the groups were not similar across the 12 studies that were part of this evaluation. The result of using only sample data is that the threats to internal validity (maturation, history, instrumentation issues, lack of procedural fidelity) cannot be eliminated. A random effects model indicated that the effect size was 0.69, which was statistically significant at p < 0.001.

The only variables that were statistically significant were duration and total hours of therapy. The authors suggest that therapist behaviors may be more important than the number of hours of therapy, and note that the findings that indicate supervisory personnel trained according to the UCLA model were better able to produce.