A Randomized Control Trial to Evaluate the Efficacy of a Mobile Application to Treat Prosodic Deficits

Background: For those individuals with ASD who acquire spoken language, 50 – 80% will present with prosodic deficits (Baltaxe & Simmons, 1985). Emerging research suggests that mobile technology may be useful in treating these types of impairments; however, no randomized control trial has been conducted to validate their efficacy (Simmons, Paul & Shic, 2015). 

Objectives: The purpose of this study was to assess the preliminary efficacy from a randomized control trial of an application, SpeechPrompts, for treating impairments in prosodic functioning.

Methods: Fifty-three students with prosodic deficits (Mean age = 8.64 years, SD = 3.06; Expressive language SS = 84.1, SD = 19.1) were recruited from school speech-language pathologists (SLPs; N = 22) and randomly assigned to either the experimental treatment condition (SpeechPrompts; n = 32) or treatment as usual (TAU; n = 21). For the experimental condition, SLPs used the SpeechPrompts software with their students 1x/week for 12 weeks. Those assigned to the control group continued receiving their speech and language intervention as usual. SpeechPrompts is an iOS application with four components that each target a prosodic construct (Intensity, Stress, Rate/Rhythm, Pitch). Students’ prosody was rated pre- and post-treatment. Time spent using the application compared to time spent engaging in non-prosody related speech-language activities was measured.

Results: Global intonation ratings suggest there was no significant difference in prosody between groups at pre-treatment (p=0.42). The SpeechPrompts sessions had a mean length of 8.66 minutes (SD = 5) whereas non-SpeechPrompts sessions (e.g., sessions pertaining to non-prosody speech and language goals) had a mean length of 14.72 minutes (SD = 8.3).  Following intervention, there was a trend for those assigned to the experimental group to be rated as more “typical” sounding in post-treatment prosody ratings of rhythm and intensity with moderate effect sizes; however, these results were not significant. Within-group analyses suggest that students in the SpeechPrompts condition were significantly more likely to be rated as more “typical” in Rate, Stress, Rhythm, Articulation, and Intonation (p’s<.05) at study conclusion with moderate to large effect sizes. Conversely, students in the TAU condition, only showed significant improvements in Articulation (p < .001) with a moderate effect size.

Conclusions: This randomized controlled pilot was one of the first to evaluate an application to treat prosodic deficits in students with ASD. SLPs implemented the software with their students; however, usage data highlights that the software was implemented less than activities used to target social-communication and other language goals. Given the complexity of ASD, it is unsurprising that prosody might not be on the top of a clinician’s intervention priority list. Although no significant differences were observed in post-treatment ratings of prosody between treatment groups, trends in improvement of intensity and rhythm were emerging for the treatment condition. While SpeechPrompts may not be more effective than standard prosody interventions administered at greater intensity, its use appears to be associated with positive changes in prosody. These results suggest that SpeechPrompts might be an appropriate intervention option for those with prosodic deficits secondary to neurodevelopmental disorders.