The LENA System in Clinical Trials: Evidence from Pivotal Response Treatment Studies

Background: Treatment research in autism spectrum disorder (ASD) has been limited by the lack of objective indicators of symptom improvement. Thus, there is a critical need to develop validated, objective outcome measures for assessing treatment-mediated changes in core autism symptoms. The Language ENvironment Analysis System collects large samples of speech from the child’s natural environment and automates analysis of child utterances and conversational turns. This system is ideal to assess any changes in clinical trials involving interventions that aim at assessing language improvement in ASD. 

Objectives: This presentation will review data from two different clinical trials examining the effectiveness of two pivotal response treatment (PRT) programs in targeting language deficits in minimally verbal young children with ASD between the ages of 2 and 6 years. 

Methods: The first dataset was obtained from an uncontrolled trial examining the effectiveness of a group parent training in PRT (PRTG) in children with ASD between the ages of 2 and 6 years. The second dataset is being collected as part of a randomized controlled trial of a PRT package treatment (PRT-P), which combines parent training with clinician-delivered in-home treatment. PRT-P is being compared to a delayed treatment group (DTG) that is receiving community services. In both trials, LENA recordings were obtained at baseline and post-treatment. 

Results: Eleven children with ASD participated in the first study. Preliminary analyses revealed improvement in language abilities with increases in child vocalization during structured laboratory observation (SLO) with increase in the number of utterances between baseline (32 ± 21) and post-treatment (51 ± 26; t= -3.572, p=.009). This improvement was associated with changes in the average frequency of child vocalization per hour as assessed by in-home LENA recordings (Baseline= 205 ± 77; Post-treatment=244± 96; t= -1.882, p=.089). Data is available from 25 participants in the PRT-P trial, and preliminary analyses revealed significant improvement in language deficits as measured by the clinical global impression scale-improvement (CGI-I) (N=16; PRT-P= 9; DTG = 7; X²= 9.363; p=0.025), and the MacArthur-Bates Communicative Development Inventories with the active group showing more improvement than the control group on the words produced out of 680 (PRT-P: 10; Baseline=126.6 ± 111.2; week 24: 281.0 ± 194.3; DTG = 7; BL: 111 ± 110; Week 24: 140 ± 148; F=5.267; df (1;15); p=0.037). LENA recording were also obtained and findings from a small subset of participants, revealed a non-significant increase in the conversational turn count in the active group (N=11; Baseline 144.7 ± 94.1; Week 24: 149.0 ± 97.3).  Additional LENA data will be available and further analyses will be completed. 

Conclusions: These preliminary findings illustrate the value of including objective measures obtained in the natural environment using the LENA system in assessing improvement in language in response to two different PRT programs. Potential benefits and challenges of the LENA system will be discussed in comparisons to structured laboratory observation and standard language measures.