A Developmental Framework for Promoting Joint Attention in Toddlers with ASD: Formative Analysis of An Intervention

Background:  

Single Case Design (SCD) research is typically used to study intervention effects.  We review an alternative use – formative analysis of an intervention’s design features.  Joint Attention Mediated Learning (JAML) is a parent-mediated intervention for toddlers with autism spectrum disorders (ASD) that is designed to support joint attention development through promotion of theoretical developmental precursors.  An earlier pilot study showed JAML’s effects for three toddlers.  A needed next step was to implement JAML with multiple interventionists in different sites and to use what was learned to design intervention refinements. 

Objectives:  

The study’s purpose was to explore individual responses to JAML enhancements and to identify patterns of individual response relative to precursor-focused intervention phases.

Methods:  

Seventeen toddlers with high risk for ASD and their parents participated in JAML to promote preverbal social communication.  A SCD was used to track child progress in four targeted outcomes: focusing on faces (FF), turn-taking (TT), responding to joint attention (RJA), and initiating joint attention (IJA) through intervention phases corresponding with these outcomes.  Participants progressed through phases as outcomes for previous phases were achieved.  Earlier participants received the original version of JAML.  Procedural changes introduced for later participants included merging the final two intervention phases and adding: guided reflection on video-recorded parent-child interaction, enhanced intervention materials, and video clips to illustrate targeted outcomes and mediated learning principles.  Coders, blind to intervention condition, assessed weekly 10-minute recorded video sessions of parent-child interaction for the occurrence of each targeted outcome in each of 60 ten-second intervals.  Data were plotted for visual analysis of child movement through the phases (i.e., their achievement of targeted outcomes) and the strength of individual responses to the intervention. 

Results:  

All except one child demonstrated joint attention in parent-child interaction over the course of the intervention.  For half of the participants, joint attention appeared before it was introduced as a phase of intervention.  Seven of the 10 participants who received the enhanced version of JAML showed more than 10 instances of joint attention over at least two sessions.  Only one of the seven participants who received the original version of JAML achieved this level of joint attention engagement.     

Conclusions:  

Before an intervention design is tested and disseminated for field-based implementation as an evidence-based practice, it is useful to consider features of the intervention that may affect its outcomes.  Enhancements to JAML’s design appeared to result in stronger toddler responses to the intervention. Analyses of individual responses to JAML and half of the participants’ attainment of joint attention before it was introduced in the intervention support JAML’s inclusion of the theoretical precursors FF and TT as a foundation for JA.  Joint attention, in turn, is an important foundation for verbal language, as demonstrated in well-replicated research.  Because of its depiction of individual patterns of response to an intervention over time, SCD is a useful tool for micro-analysis of an intervention’s conceptual design and procedural features.