Adaptive Intervention For Communication In Minimally Verbal School Aged Children

Background: Current estimates are that 30-40% of children with autism spectrum disorder (ASD) are minimally verbal at age 5 years despite involvement in intensive early interventions. Current successful behavioral interventions, such as joint-attention and engagement with enhanced milieu training (JAE/EMT), improve social communication and spontaneous use of spoken words in children who are preverbal (preschool aged). However, these interventions have not been rigorously tested for minimally verbal school-aged children with ASD, despite the high prevalence and severity of need of this population. The IACC, NIH, and Autism Speaks all have placed a high priority on developing efficacious interventions for these children. Further, there is significant interest in the effectiveness of blending behavioral interventions with augmentative alternative communication (AAC) devices to speed the uptake of spoken communication.

Objectives: An adaptive intervention is a sequence of decision rules, which adjusts treatment over time as a function of the changing clinical status of the child. The overarching aim of this study was to construct and systematically test an adaptive interventionthat utilizes JAE/EMT and varies the addition of an AAC device with minimally verbal school aged children.  The primary specific aim of this study was to examine the effect of an adaptive intervention beginning with JAE/EMT+AAC versus with JAE/EMT alone.

Methods: Sixty children participated in a longitudinal randomized clinical trial across 3 sites (UCLA, Vanderbilt, and KKI).  At baseline, participants completed diagnostic (ADOS), and cognitive (Leiter-R) assessments.  To measure spontaneous communication, participants completed a natural language sample (NLS) with a blind assessor at entry, intervention mid-point (month 3), and exit (month 6). Children were randomly assigned to JAE/EMT+AAC or JASP/EMT without AAC.  Intervention sessions were two, hour-long sessions per week.  Intervention was adapted at mid-point based on whether the child was an early or slow responder. Early or slow response status was based on spontaneous communication during intervention sessions and the NLS.  Intervention was intensified (to three sessions per week) for JAE/EMT+AAC participants who demonstrated slow-response. JAE/EMT participants who demonstrated slow-response were had treatment intensified or had were introduced an AAC. 

Results: Fifty-nine of the children met criteria for autism, and 1 met for ASD (ADOS).  Participants were an average of 6.33 years old (SD=1.12), had an average of 16.62 (SD=14.63) unique words on the baseline NLS, and average nonverbal cognitive scores of 4.01 years (SD=1.13, Leiter-R).  At mid-point, adaptive interventions beginning with JAE/EMT+AAC had a larger total number (Cohen’s d=0.76, p<0.01) and percent (d=0.59, p=0.02) of spontaneous communicative utterances versus adaptive interventions beginning with JAE/EMT alone. The effects persist through treatment exit: total number (d=0.6, p=0.02) and percent (d=0.75, p<0.01) of spontaneous communicative utterances.

Conclusions: These data suggest that children who are minimally verbal can make significant progress in socially communicative spoken words after age 5.  Moreover, they benefit significantly more from the experimental intervention when they begin the treatment with an AAC device than without. Thus, future research should examine how AAC support can be integrated into homes and schools to further improve communicative outcomes for these children.