The Cost of Fad Treatments in Autism.

When considering interventions for autism, there is agreement on the importance of proof that a treatment is actually effective; that is, it actually produces positive gains in skills of the person with autism. Most behavior analysts and treatment providers adhere to a standard of "effectiveness" that incorporates objective measurement using an experimental design that is implemented with adequate control over validity threats and other potential confounding variables. Thus, consumers should expect that treatment providers have some objective evidence to support claims of treatment effectiveness. This "empirical frame of reference" for judging effectiveness is supported by many committed to autism treatment. For example, the Organization for Autism Research (2008) advocates dissemination of the current state of the "Science" in autism research. The Autism Society of America (2008) lists several guidelines, one of which is, "has the treatment been validated scientifically?" Even the federal education law requires that teachers use "scientifically-based practice" when working with children, both typical and those with special needs. Specifically, what are criteria for valid evidence of effectiveness? An important publication addressing these criteria was the New York State Department of Health (DOH) Clinical Practice Guidelines (1999). The DOH formed a panel of professionals and parents that developed criteria for what constituted quality research evidence for treatment effectiveness. Included in these criteria were: (a) use of experimental design, (b) controls for bias, and (c) multiple studies done by multiple investigators. The guidelines exerted a major influence on the shaping of evidenced-based practice in the early intervention of autism. In addition, Newsom and Hovanitz (2005) presented a compelling list of characteristics that would be part of any criteria. They argued that any test of treatment effectiveness must meet several standards of quality, including that: (a) terms must be operationally defined, (b) reliability of measurement must be assured, and (c) the treatment in question must be tested using experimental procedures (e.g., identification of independent and dependent variables, controlling for internal validity threats, etc.). Similar criteria were identified by Chambless, Baker, Baucom, Beutler, Calhoun, Crits-Christoph, et al. (1998) who proposed criteria that must be met by treatments used by clinical psychologist for those treatments to be considered effective. These criteria included: (a) a number of within-subject design experiments with more than nine subjects, (b) treatment manuals must exist specifying the details of the treatment methodology, (c) same effects demonstrated by at least two different researchers, and (d) subject characteristics must be detailed. Therefore, there is substantial body of criteria for research that can be considered well controlled and whose results then can be judged to be most believable (e.g., Kasari, 2002; Green, 1996). The importance of using effective treatment is underscored when considering the cost of caring for individuals with autism. It is likely that children who do not receive effective early intervention services will require long-term special and custodial care throughout their lives, which for 1996 was estimated to cost over $13 billion a year (FEAT, 1996). More recent studies suggest that the US spends $90 billion per year (Autism Society of America, 2008) to care for the 1.5 million children and adults with autism. This cost could skyrocket to between $200 billion and $400 billion by 2013 (The Autistic Society, 2008). Given the enormous cost of caring for these individuals over their lifetime, efforts are being focused on effective early intervention strategies in the hopes of offsetting some of the long-term costs. Applied Behavior Analysis (ABA) has been shown to produce substantial benefits for many children with autism (Anderson, Avery, DiPietro, Edwards, & Christian, 1987; Birnbrauer & Leach, 1993; Fenske, Zalenski, Krantz, M Lovaas, 1987; McEachin, Smith, & Lovaas, 1993). …