Predictors of Differential Responsivity to Pivotal Response Training and Discrete Trial Training

Background: Treatment studies indicate that substantial gains may be achieved by some children with autism when treatment is provided at an early age. However, heterogeneity of treatment response is common to all evidence-based approaches. After early intervention, many children remain considerably impaired. Investigators have hypothesized that customizing treatments based on individual child and family needs should increase the overall number of children that benefit from intervention. Improved understanding of how to match specific treatments (e.g., Discrete Trial Training/DTT, Pivotal Response Training/PRT) to children exhibiting different behavioral characteristics may enhance our ability to tailor interventions to individual children, thereby improving treatment effectiveness.  

Objectives:  (1) To evaluate the relative effectiveness of DTT and PRT for teaching children with autism under the age of 3 receptive and expressive language, play, and imitation skills, and (2) to identify variables influencing whether specific children are more likely to benefit from DTT or PRT in the tested domains.

Methods:  Preliminary data are presented for four children, under the age of three, who participated in a single-subject alternating treatments design. Language, play, and imitation targets were matched on developmental appropriateness and difficulty and then randomly assigned to treatment conditions. Children received three 45-minute sessions of in-home treatment per week in each intervention for 12 weeks. Potential predictor variables were collected at pre-treatment. Fidelity measures were collected on 33% of all procedures. Data are reported on rate of learning, spontaneous skill use, and disruptive behaviors during sessions, as well as skill acquisition and generalization during weekly probes, and maintenance of gains at 3-month follow-up.

Results:  Preliminary results indicate that participants made gains in the acquisition and generalization of the target behaviors in both treatments. However, response to treatment varied by child and curriculum area in that different children responded uniquely to DTT and PRT. Developmental level and age at intake did not influence these patterns. Similarly, a priori clinician judgment did not consistently predict responsivity patterns. Pre-treatment characteristics that did appear important included toy contact, adult avoidance, and early rates of learning. In particular, children who engaged in low levels of toy contact and were avoidant of adults were less likely to benefit from PRT. These same variables did not predict responsivity to DTT. Alternatively, children who were more interested in objects and less avoidant of adults did equally well in both interventions or had superior performance in PRT. Across most participants, PRT was superior for learning expressive language and spontaneous use of skills. Finally, patterns indicating which intervention was more effective seemed to emerge at approximately 4 weeks into treatment.

Conclusions:  These data confirm the importance of treatment individualization and begin to suggest specific methods for tailoring treatment programs to individual child needs. The strengths and weakness of DTT and PRT may vary depending on child variables, as well as curriculum area focus. Specific child behaviors, including toy contact and adult avoidance, may aid in prospective treatment planning efforts. Additionally, early rates of learning may be predictive of longer-term treatment response.