The Hawthorne Effect
Why parents swear by ineffective treatments for autism.
Also in Slate: Arthur Allen on how the vaccine/autism theory may be dead, but the treatments live on.
Autism can present in many ways—hence "autism spectrum disorders"—but that range is nothing compared with the diverse techniques that parents use in their attempts to cure, ameliorate, or disrupt the progress of the disease. In the 60-plus years since autism was first described, many methods to treat it have been proposed—one research paper identified 111 recognized treatments or strategies. Studies have found that parents try an average of between 4.3 and seven interventions simultaneously; one family reported using 47 different treatments at one time.
Alas, almost none of these treatments are evidence-based, and some have been clearly demonstrated to be worthless. In dealing with other medical problems, like the common cold, I've always annoyed medication-seeking parents by pointing out the obvious: If there is any illness for which 100 treatments are available, you can be sure that none of them works. But with autism, the stakes are much higher.
It is especially difficult to know where to look for treatments when a condition is poorly defined and characterized. There are no laboratory tests or gross anatomical findings that establish the diagnosis, but experienced clinicians often "know it when they see it" almost instantly, especially when patients are severely affected. I once made the diagnosis from a dog-eared snapshot. Since most of the ways we diagnose autism are based on behavior, we can't rely on biological, structural, or chemical findings to determine if a treatment is working. We primarily measure success based on a patient's change, or lack thereof, in behavior.
Medications, new styles of teaching, classical psychological conditioning, physical manipulation, vitamins, diets, special eyeglasses—many kinds of treatments have been proposed and tried, but few have been tested in a rigorous way. Fewer still—some behavioral conditioning methods, a few anti-psychotic medications—have demonstrated some degree of efficacy. Some autistic patients exhibit very difficult patterns of behavior, ranging from simple stubbornness to compulsiveness to screaming to destructiveness to explosive violence. The behavioral changes produced by the few effective treatments make life in social settings (including the home) possible, but we have no idea whether they have any effect on the underlying cause (or causes) of autism or whether they even make severely affected patients feel better. The people who work with autistic clients often come to depend on their own sensitivity and empathy to judge whether a treatment has had a positive or negative impact.
Other treatments are iffier in their ability to cause behavioral change; some are utterly worthless. For instance, patients with autism frequently have huge difficulties in communication, so there has always been the hope that addressing that problem would have great benefit, both in improving quality of life and perhaps even in fixing the underlying problem. One method intended to help, "facilitated communication," is based on the idea that a sensitive facilitator will hold the hand of a patient over a kind of Ouija board. She will then help the patient respond to questions by sensing his intention and helping guide his hand to spell out answers. Rigorous studies have shown that the spelled-out answers come from the unconscious (or, worse, the conscious) mind of the facilitator. Nonetheless, the practice is still in use, and I know parents who are utterly convinced that it is valid and useful. Frankly, something important did happen when facilitated communication was introduced to my patients: They improved, they brightened, they became more social and more interactive, and they seemed, somehow, happier, even though facilitated communication didn't actually translate their thoughts into words. I'll come back to "why" in a minute.
Sensory integration treatment is another method in very wide use for autistic patients. The technique, developed by occupational therapist/clinical psychologist Dr. A. Jean Ayres, is based on the observation that some children, particularly in autistic, learning disabled, or developmentally delayed populations, show an excessive sensitivity to a variety of external stimuli—touch, position in space, sound. She posited that this was the result of a poor ability to process sensory messages received by the brain—for example, skin contact or signals from the balance organ in the inner ear. Ayres and her followers suggested that occupational therapists could help repair and reintegrate improperly processed sensory inputs. In doing so, they hoped to address and improve the underlying conditions that led to (or perhaps were caused by) dysfunctions in sensory integration. The techniques of sensory integrative treatment include rubbing or brushing skin (using graded and tactile stimulation), balance exercises, exposure to soft music, and the use of weighted clothes, among other things. Does it work? Most of the research has been of very poor quality, but, in virtually all of the recent studies, sensory integration doesn't seem to be any more beneficial than any other treatment.
The problem is this: When it comes to human behavior, almost any (positive) attention or intervention is likely to be somewhat beneficial. Between 1924 and 1932, some industrial psychologists and efficiency experts studied the Western Electric manufacturing plant in Cicero, Ill., to determine what interventions might lead to an increase in productivity. * Increase the lighting, even a little bit? Definite improvement for a while. Shorten the workday? Definite improvement for a while. Lengthen it? Definite improvement for a while. Dim the lighting? Definite improvement for a while. It looks as if environmental alteration, especially if coupled with increased attention and perhaps expectation, often leads to change in human behavior. It's called the "Hawthorne effect," after the Hawthorne Works of Western Electric.
People respond—mostly favorably—to positive attention and interaction. The question we need to ask about all the treatments available for autism is whether they actively shape and change brain development and thus treat the underlying condition, as many proponents believe, or whether the benefits (if they are present at all) are simply another example of the Hawthorne effect.
Perhaps my patients who became more alive and more interactive after facilitated communication was introduced changed because their families and caretakers were taking them more seriously as people who might have an inner life—people worthy of attention and interaction.
Correction, April 27, 2009: This article originally misidentified the Illinois town in which the studied manufacturing plant was located. (Return to the corrected sentence.)
Sydney Spiesel is a pediatrician in Woodbridge, Conn., and clinical professor of pediatrics at Yale University's School of Medicine.
Illustration by Mark Alan Stamaty.