Pilot Project, Year 6
Addressing Speech Health Equity Issues in the Southwestern US Using a Telehealth Speech Therapy Game
This project will address the needs of people who receive, give, or are affected by speech therapy. Anyone using speech to communicate on a daily basis can likely imagine the difficulty, uncertainty, and anxiety they might experience with a speech impairment. Research has shown that people with speech impairments often have lower income, delayed independence, symptoms of depression, and lower social skills. Speech is a skill that often improves with speech therapy, but up to 70% of speech-language pathologists in the US have waiting lists, indicating a workforce shortage. Lack of intervention services disproportionately affects Blacks, Hispanics, and Women. SpokeIt is the first and only speech therapy game capable of diagnosing speech errors in children born with an orofacial cleft, including articulation, compensatory, and resonance errors.
Study aims
- Test whether SpokeIt can accurately diagnose cleft speech errors without biases against Hispanic and Indigenous populations’ accents in the Southwest US and
- Gather preliminary data for an R01 application to examine the longitudinal efficacy of SpokeIt.
Our long-term goal is to provide clinically accurate and effective telehealth interventions for cleft speech therapy to underserved populations with limited resources. Our central hypothesis is that the underlying machine learning models and speech recognition systems in SpokeIt have a higher incidence of false positives when diagnosing common cleft speech errors for people with an accent. The rationale is that this work will increase confidence that biases are removed from SpokeIt’s diagnostics before proposing an R01 largescale longitudinal efficacy study.
Funding: The study is funded by NIMHD/NIH 5U54MD012388
About the investigators
Jared Duval, PhD
