Do you have to pick sides when it comes to interventions for children with autism?

Find out in this Q&A with the author of Individualized Autism Intervention for Young Children: Blending Discrete Trial and Naturalistic Strategies

About the author

Dr. Travis Thompson

Travis Thompson, Ph.D., is a licensed psychologist affiliated with the Autism Certificate Program in the Special Education Program of the Department of Educational Psychology at the University of Minnesota. He is also an adjunct professor in the Department of Applied Behavioral Science at the University of Kansas, Lawrence, and a collaborator on a multisite project on challenging behavior in developmental disabilities.

Dr. Thompson has served on several National Institutes of Health research review committees, including chairing reviews of the applicants for Collaborative Programs of Excellence in Autism awards in 2000, 2003, and 2007. He has been a member of American Psychological Association (APA) task forces concerned with the practice of psychology and psychopharmacology. He is a past president of the Behavioral Pharmacology Society, the Division of Psycho-pharmacology and Substance Abuse, and the Division of Mental Retardation and Developmental Disabilities of the APA.

Dr. Thompson has received numerous awards, including the Distinguished Research Award from the Arc of the United States; the Academy on Mental Retardation Lifetime Research Award; the APA's Don Hake Award; the Edgar A. Doll Award, for contributions to facilitate the transfer of research into practice; and the Ernest R. Hilgard Award; and the Impact of Science on Application Award of the Society for Advancement of Behavior Analysis.

He has spoken in 46 states and 15 countries about his research and clinical services and on topics related to autism and other developmental disabilities and psychopharmacology.


Customer Service:

Customer Service

Q: Can you briefly describe "discrete trial instruction" and "naturalistic, incidental teaching" and explain how they differ?

A: In discrete trial intervention (DTI) a child and therapist, parent, or teacher are seated at a child-sized table in a room or area free from distractions. Only materials essential for the task at hand are visible nearby. The therapist provides a cue, such as Do this, and demonstrates clapping hands. If the child repeats the response (or approximates it) the therapist praises the child and provides an additional reward such as tickles or an edible treat, such as a piece of apple.

Each sequence (stimulus-response-consequence) is a discrete trial. If the child doesn't respond initially, the therapist guides the child's hands through the motion and then rewards her or his attempt. This is repeated usually in groups of 10 or 20 trials, followed by a break and change of activity. Each trial focuses on learning one specific thing, like how to imitate a motor or speech sound. Cues and pacing of activities are determined by the therapist/parent.

In naturalistic or incidental intervention, therapy or instruction can occur in a variety of settings at school, community or around the home. Goals are planned in advance and opportunities are created by the therapist by making it likely the child will initiate an interaction or request, for example, by placing a toy the child would like to play with visible on a shelf but out of reach.

The therapist may say, What do you want? If the child is unable to say the name of the item, the therapist will provide a further prompt, What do you want? ... ball. If the child is in the kitchen and tries to open the refrigerator, the therapist may say, Want juice? and Say juice.

As the child's skill level increases the prompts are faded and the child initiates more independently. The responses and rewards are naturally tied to the situation rather than being determined entirely by the therapist as with DTI. The specific activity and pacing are determined to a large extent by the child's interest.

Q: Your new book outlines a blended approach to intervention for children with autism. What is blended intervention?

A: Blended intervention combines DTI and incidental intervention and also draws on techniques from some developmental interventions such as Diane Bricker's activities-based preschool interventions and Robin McWilliam's routines-based intervention.

Depending on the child's profile, more or less time may be spent in the beginning using DTI methods, with higher functioning children usually requiring less discrete trial practice before gradually transitioning to incidental strategies. Some higher functioning children may acquire the basic skills very quickly and within a few sessions may profit from transitioning to a largely incidental approach. The decision to transition from DTI to incidental is data-based.

We also make use of role-playing and thematic activities that are somewhat more structured than typical incidental intervention but less structured than DTI.

Some professionals insist that discrete trial methods be used consistently throughout and some prefer to use only incidental intervention methods. In the last analysis, which is used should be based on what is effective in helping the child learn important social, communication, and other skills.

Q: In your book, you mention that a difficulty faced in the field of autism is that a child is expected to fit within a service model, rather than the model needing to fit the child. Can you elaborate what you mean by that?

A: Most community-based service programs specialize in one type of intervention, such as Discrete Trial Teaching, Relation Developmental Intervention, or Floortime. When parents visit with intake staff members, they quickly learn their child will receive that particular type of intervention and are unlikely to be exposed to other types of intervention. In short, their child must fit the intervention, instead of the intervention being adjusted to fit the child.

Some programs only teach skills sequentially based on a standardized assessment—such as the ABLLS, which is a very useful guide in designing interventions—but it is not necessary to teach skills only in the order listed on such an instrument. Other programs do not try to teach skills explicitly but focus on the quality of the child–adult relationship.

Such choices are confusing to parents who have no way of evaluating which would be the best fit for their child and which would be likely to yield significant gains. In my work with the Minnesota Early Autism Project and other endeavors, we attempt to draw on a range of approaches for which there is evidence of efficacy and combine them to meet individual children's needs.

Q: How do you individualize treatment?

A: The process is actually quite involved. We study all previous testing by psychologists and school programs and medical assessments the child has received. For some children that can be a large number of documents.

We want to determine whether the types of services we are able to provide are likely to be appropriate for the child. Some children with severe sensory, physical, or intellectual disabilities—or who may have serious health problems that could be incompatible with regular intervention sessions—may not be able to participate in many of the types of interventions employed in intensive early behavioral intervention.

Next, we schedule a face-to-face visit in our clinic with the child and at least one parent, often both. While a therapist is seated on the floor attempting to probe the child's skills in critical domains through play, I interview the parents around family and child history, and current presenting issues while simultaneously observing the child. We describe to parents our general approach to intervention but indicate we will discuss with them in detail our thoughts after we have had a chance to discuss our observations among ourselves.

Following the interview and observation, one of our clinic staff and I meet to review our observations in the context of all of the prior information received. I complete a review of primary autism symptoms and moderating symptoms (the AIRS® can be used for this purpose) that predict how likely it is the child will profit from one or another type of intervention strategy.

At that point we indicate to the parent our tentative conclusions about how best to proceed. If the parents continue to be interested, we schedule a home visit at which time an initial treatment plan is developed based on parent priorities and all of the information gathered to that point.

Q: How does understanding a child's autism profile help determine what type of intervention is appropriate?

A: The decision about the type or blend of interventions is rarely either/or, black or white, with the exception of individuals toward either end of the continuum.

For children with severe autistic disorder, very limited attention, lack of social interest, or considerable developmental delay, it is usually obvious that beginning with a discrete trial intervention is most likely to be effective. Conversely, for some high functioning children, especially those with Asperger disorder or PDD-NOS, it may be equally clear that a naturalistic intervention approach is most appropriate.

For about half of the children we have served, a blended approach has been most effective, depending on both core autism features as well as such moderating factors as severity of ADHD symptoms, anxiety, and compulsive symptoms.

Q: Can you give an example of a blended approach in practice?

A: Lettie is 4 years old, and says just a few words, (hi, bye, more, juice, play), and engages in repetitive behavior such as hand flapping and "edging." She is interested in other children but doesn't know how to play with them. She is anxious around groups of people. She has tantrums when preferred routines are interrupted.

We might begin by using a discrete trial intervention approach with object identification to teach her how this learning situation works. We want her to understand the nature of her relationship to the therapist or parent/instructor.

Once she catches on, we move to incidental teaching by asking her similar questions in the context of her daily routines throughout her home, and later outside her home (Point to car, Point to boy, What is this? while holding up an apple).

To add a new learning task—identifying people by their facial features—we revert to DTI for the first several days, practicing naming familiar people shown in pictures (such as Mom, Dad, sister Betty, brother James, Grandma Ettie and so on). Once Lettie is identifying people from their pictures we ask her to identify and name them "in real life."

In preparation for applying this skill at school, we obtain pictures of four children in her preschool classroom in whom she shows interest. We use a similar DTI method at home to teach her to name each child by their picture. Then we teach her to practice greeting each child, Hi, Penny, using a role-playing format. Next we teach her to practice the same skill with each of the four children in school. Then we reinforce her for establishing eye contact with the child as she greets them.

In all likelihood there would be intermediate steps along the way for each skill, but this gives a general idea of the approach. Start with DTI to overcome the initial difficulty of the task so the child understands what is the nature of the problem at hand and then transition to incidental teaching as performance warrants.

Q: What has research shown so far about the impact of blended intervention?

A: We have completed one clinical study that is currently under review for publication which shows about 64% of children are integrated into regular education after completing 1+ year of blended intervention.

Similar approaches have been used by Robert Koegel, Lynn Koegel, Laura Schreibman and colleagues, Gail McGee, Lynn McClannahan, Pat Krantz, Sam Odom, and Phil Strain. They have not done comprehensive evaluations of their effectiveness with a wide range of autism symptoms but have clearly demonstrated efficacy with some important communication, social, and academic skills.

Recently Geraldine Dawson and Sally Rogers have used a somewhat similar approach (The Early Start Denver Model) and conducted a randomized clinical trial yielding positive results. We are not aware of previous attempts to match individual child profiles with autism early interventions, with the exception of two Pivotal Response Training studies by Laura Schreibman and her colleagues.

Q: What barriers exist to a more widespread adoption of child-centered, individualized autism interventions?

A: Professionals who have been successful in adopting an evidence-based approach, such as a largely discrete trial intervention strategy, are understandably loath to stop what they are doing and consider adopting alternative strategies. Nothing succeeds like success is an understandable argument.

The issue is whether they might actually have even greater success if they were able to individualize interventions incorporating some naturalistic methods that also draw upon developmental psychology principles. It is also possible—and our data strongly suggest—that similar or even greater gains may be obtained with somewhat lesser intensity of a more naturalistic strategy for some children on the autism spectrum.

In the long run, professionals are convinced by objective evidence. It's our job to help provide that evidence.

Q: What steps need to be taken to overcome those barriers?

A: Two steps would help. Obtaining more data from controlled studies that examine the relationship between child profiles and intervention types and outcomes could be very helpful. Resources are not unlimited, and it is likely service budgets will be cut, so we need to be mindful of selecting the most efficient methods.

Secondly, we need to make it clear that we are not questioning the value of Ivar Lovaas's original discrete trial approach for many children with autism. Lovaas's contribution to the field of autism services was monumental. There would be no field of ABA autism early intervention without Ivar Lovaas. Like every advance in science, we build on past successes as we understand more about the phenomena we are studying, modifying our strategies based on the best evidence.

Q: Do you have any specific advice or insight you would like to give/share with practitioners who have been trained in one "school" of intervention versus another?

A: Evidence-based blended approaches do not mean a little of this and a little of that depending on the practitioner's preference, regardless of evidence that it is effective. I urge practitioners to study the evidence regarding what methods have been demonstrated to be effective for particular types of children, and judiciously integrate them when it make sense to do so.

By keeping your eye on the ball, what leads to the best outcomes for the children with whom you work, you will tend to adopt methods not based on one or another theory of child development or learning, but based on objective evidence of what works.

Our children's futures depend on setting aside old quarrels about one theory versus another theory and focusing our attention on what works best for each child. I am not suggesting abandoning solid science, but there is often more than one way to do good science.

Individualized Autism Intervention for Young Children: Blending Discrete Trial and Naturalistic Strategies

Be sure to see these other titles: