Impairments in social communication are among the core symptoms of autism spectrum disorder (ASD), but there are currently no available, well-validated, objective measures that could be used to evaluate social communication outcomes in clinical trials. The central aim of this multisite project is to further validate two newly developed treatment outcome measures: the BOSCC (Brief Observation of Social Communication Change)¹ and ELSA (Elicitation of Language Samples for Analysis) over the course of a manualized ASD treatment program based on JASPER (Joint Attention Symbolic Play Engagement and Regulation)².

Ninety children ranging from 18 months to 5 years of age who are minimally/low verbal will be recruited at three sites (Boston University [PI: Helen Tager-Flusberg], Weill Cornell Medicine [PI: So Hyun (Sophy) Kim] and the University of California, Los Angeles [PIs: Connie Kasari and Catherine Lord]) and randomized into a clinician- (n=45) or parent-mediated (n=45) JASPER program for ten weeks, followed by an additional ten weeks of intervention if the initial response to treatment justifies this.

There are three specific aims for this project:

Aim 1 — To demonstrate that the ELSA and the clinician- and caregiver-administered BOSCC are sensitive to change and clinically meaningful. Both measures will be used at the following timepoints: pre-intervention (T1) and post-10-weeks (T2) and post-20-weeks intervention (T3).

Aim 2 — To evaluate the psychometric properties of the ELSA and BOSCC. Inter-tester reliability will be assessed on measures of the frequency of utterances and conversational turns on the ELSA (comparing two different examiners) and on the BOSCC domain and total scores (comparing examiner and parent). The convergent and discriminant validity of the measures derived from ELSA and BOSCC scores against blind Clinical Global Impressions (CGI) and therapists’ CGI of Improvement (CGI-I) will also be assessed. For ELSA, the measurement validity for each of the nine segments of the protocol will also be assessed to explore whether some could be eliminated. Finally, the changes measured by ELSA and BOSCC will be compared.

Aim 3 — To develop and validate automated measures of children’s and their social partners’ movement and speech during the ELSA and BOSCC. Such objective, quantitative measures are likely, in the future, to augment or replace at least some traditional human coding methods. These tools will be developed by Shrikanth Narayanan’s group at the University of Southern California.

The successful completion of this project will result in the validation of two novel naturalistic assessment protocols that can be administered in different settings for a wide range of children at risk for ASD from which reliable and valid objective measures of social communication can be readily derived.