IN SEARCH OF MIRACLES: A review of interventions for children with autism

by Mitchel Fleming


Most people are aware of the numerous claims that appear in newspapers, on television and on the Internet concerning treatments that offer dramatic outcomes for children with autism. Parents are promised ‘miracles’ if they undertake a particular form of therapy, while sceptical professionals are castigated for their intolerance and shortsightedness. Profressor Patricia Howlin notes that it frequently appears that the more extravagant the promises made about a treatment, the more limited are the data on which they are based. For example, advocates of Holding Therapy claim ‘children lost their autistic symptoms completely …. became entirely normal,’ while proponents of Facilitated Communication note that it is successful even for those ‘previously assumed to be amongst the lowest intellectually functioning ….’

Such claims can put enormous pressure on parents of children with autism to pursue specific treatments. It is not that these parents are more gullible than others, but rather they don’t want to leave any stone unturned in their search for an effective treatment. However, determining whether a particular treatment is effective and worth pursuing can be complex, and parents can become easy prey to over-zealous or unscrupulous advocates of different interventions.

When considering an intervention for a child with autism, it is wise to carefully examine the various types of available evidence. The evidence for some treatments is based solely on subjective reports, e.g., testimonials, anecdotes and personal accounts. This is the most dubious type of evidence and should be viewed with caution. Anyone can make a claim that an intervention works; what is needed is proof. Next in the hierarchy of types of evidence comes the uncontrolled study or observation. Here proponents of an intervention claim that a child or group of children showed improvements over time following treatment, but didn’t compare the group with a similar group who didn’t get treatment. We know that children change and develop over time and if we are to have confidence in a treatment, then it is important to know that the improvements observed happened because of the intervention and not just as a consequence of maturation or chance. Evidence that is most impressive comes from controlled studies using objective measures, where the researcher has taken time to examine variables that influence an intervention carefully. Results from such studies provide us with good information on the effectiveness of an intervention. If an intervention truly works, then others will be able to repeat the study and get the same outcome. Also, if an intervention works and is widely practiced, then we should expect to see over time noticeable improvements in many of the children who received the intervention.

While the above comments on the importance of carefully considering the available evidence on different interventions are logical, many choose to ignore them. There are some reasons for this. In the case of new interventions there is often a dearth of quality research upon which we can make informed decisions. Unfortunately, in these situations we can only be wise after the facts have emerged. Experience, however, suggests that we should be cautious, especially if there is a possibility that the treatment might have long-term negative effects. For example, a few years ago there was a good deal of excitement about the reported benefits of secretin and many parents had their children injected with the hormone. However, since then research has shown that the treatment was ineffective and enthusiasm about secretin has waned.

Fortunately, many popular interventions have been around for several years and it is reasonable to assume that these have been researched. After all, if you thought that you had discovered an effective or miraculous treatment for autism, would you not want it to be thoroughly investigated to show that it definitely works? Accordingly, in my view, you should be suspicious if there is little or no evidence available on a particular long-established intervention. There are a number of reasons why good evidence on an intervention may not be available. Firstly, no good quality study was conducted. Secondly, quality research was conducted but the results were not made public. Why? Thirdly, the results were inconclusive and were not published. Academic journals have a strong bias towards publishing well-conducted studies that obtain significant and especially positive findings. However, findings from poorly conducted studies, particularly those that make dramatic claims, are often reported in the newspapers and on the Internet. These findings are unreliable and need to be viewed with extreme caution.

Another common reason why some people ignore evidence pertaining to the effectiveness of different interventions is that they become overwhelmed by the amount of available information or with the complexities of arguments concerning claims and counter-claims. Accordingly, they turn to seemingly knowledgeable and trustworthy ‘experts’ for direction and hope. Unfortunately, some of these experts are linked with particular interventions and their opinions may be strongly biased.

There are a number of independent individuals and agencies that have carefully reviewed the evidence pertaining to various treatments. Those that I find particularly helpful are the New York State Department of Health’s review of interventions for children with autism (available on the web) and Professor Patricia Howlin, who advises the National Autistic Society on the effectiveness of different interventions. Prof. Howlin is quite a frequent presenter at conferences in Britain and Ireland, and those interested in keeping up-to-date on the effectiveness of interventions should find her presentations helpful.

I have attempted to summarise, below, the evidence from the above sources concerning some of the most frequently cited treatments. Treatments are divided into those that tested negatively, those where little or no evidence exists to support them, those that show promise (but where the evidence is limited), and finally those that are supported by good evidence. It is beyond the scope of this article to describe each of the interventions mentioned.

Those interventions that have tested negatively include: Facilitated Communication, Secretin, Auditory Integration, and Fenfluramine. Facilitated Communication (mentioned above) was heralded by some a few years ago as a means of unlocking the hidden genius of people with autism. A facilitator would hold a person’s hand, wrist or arm while that individual used a keyboard or letter-board to spell out words or sentences. Results from over 50 studies that involved over 300 people on whom the approach was used showed no evidence of independent communication.

Little or no convincing evidence has been found to support the following as effective treatments for autism: Holding Therapy, Music therapy, Psychoanalysis, Physical Therapy, Sensory Stimulation, Sensory Integration, vitamins (unless a person has a vitamin deficiency), diets, intravenous immune globulin, anti-yeast, Son Rise Programme, Walden Approach and Pet Therapy. This is not to say that people with autism don’t enjoy some of the above activities or that for some there are not any benefits, but rather the existing evidence does not indicate that these approaches on their own are effective treatments for autism. Many of the above are often included with other approaches.

Interventions that show promise, but where the evidence is still limited, include TEACCH, the Picture Exchange Communication System (PECS), Hanen, the Early Bird Programme, and some drugs, such as the new SSRIs (to reduce certain maladaptive behaviours).

The intervention with the best evidence concerning its effectiveness as a treatment for autism is Applied Behaviour Analysis (ABA). Indeed some of the promising approaches mentioned above use elements of, or have their origins in, ABA. ABA, formally known as Behaviour Modification, is a structured approach to learning that is based on scientific principles and which uses techniques such as reinforcement, discrete trials, modeling, etc. The approach first rose to prominence in the area of autism in the late 1980s and early 1990s through the work of Dr Ivar Lovaas at UCLA. He compared a small group of 19 preschool children with autism who received an intensive ABA programme with another group who received a less intense intervention programme, over two years. Those who received the intensive programme made much better progress than the other group. Nine children (or 47% of the intensively treated group) went on to mainstream school. Inn a seven-year follow-up study, 8 were reported to be ‘indistinguishable from their normal peers’, and were said to have made a ‘complete recovery’.

In my view there seems to be little doubt that behavioural approaches that focus on education, development and the reduction of problem behaviours are effective for many children with autism, but I doubt whether the extent of the claims made by Lovaas will be replicated in other studies. ABA programmes for children with autism have become increasingly popular, especially in the USA. If Lovaas’ findings were replicable, then I would have expected to see a large decrease in the number of children with autism following treatment. Instead, survey after survey shows a dramatic increase in the reported incidence of autism in the USA, but so far I have not seen any surveys to indicate a substantial decrease in autism among those who have received intensive ABA intervention, especially in California, where the approach is used extensively.

Today there are many types of ABA programmes available—Precision Teaching, Discrete Trial Training, Pivotal Response Training, Incidental Teaching, Embedded Learning, etc—and very few programmes limit themselves to the type of ABA approach used by Lovaas. Also, most now regard aversive techniques as used by Lovaas as unwarranted. For these reasons it is unlikely that Lovaas’ research will be replicated in its original format. However, it we are to determine the effectiveness of ABA as an intervention for autism we will need more, especially large-scale, independently conducted studies that use ABA techniques.

On a different but related topic, I recall a presentation given by Prof Jerome Kagan from Harvard University some years ago in Dublin. His research team investigated a large number of different educational approaches used with ‘normal’ children to see which were the most effective. These ranged from highly structured approaches where every aspect of the teaching programme was carefully pre-designed, to very flexible approaches that encouraged learning through discovery. Prof Kagan and his team found that the effectiveness of an approach was largely determined by the competence, commitment and belief that the teacher had in the approach, and to a much lesser extent by the specific nature of the approach. It is my experience that proponents of ABA often have such qualities in abundance, I wonder if these apparently essential qualities, together with the use of a structured and individually tailored approach based on sound learning principles, are the real ingredients of success.


Please enter your comment!
Please enter your name here