Positive Predictive Values Associated With Adapting the Redmond Sentence Recall Measure Into a Kindergarten Screener for Developmental Language Disorder

Abstract

Purpose: Researchers estimate the prevalence of developmental language disorder (DLD) in 5-year-olds to be between 7% and 12%. Current identification systems in public schools typically favor referral identification formats over targeted or universal screenings. Public schools face unique challenges when assessing the value of screening measures for DLD that include real-world considerations such as administration, time, and resource constraints. This study used the positive predictive value (PPV) of the Redmond Sentence Recall (RSR) to assess its fidelity when administered by special education paraprofessionals. Our obtained PPV was compared across three areas: (a) previous studies that have utilized the RSR, (b) rates extrapolated from the participating school district's preexisting referral system from a previous study, and (c) expectations based on DLD prevalence. Method: Language screenings were conducted in two elementary schools using the RSR administered by school-based paraprofessionals trained on the screener protocol. One hundred sixty-four kindergarten students (age range: 5–6 years) were screened. Confirmatory testing was completed on all students who failed the screener. Results: Of the 164 students screened, 19 failed the RSR (11.5%), and 14 met criteria (8.5%) for DLD (PPV = .74). Our PPV was similar to previously published studies that utilized the RSR using research assistants and was higher than the PPV associated with teacher-based referrals from the participating school district. Conclusion: The RSR represents a potentially useful screener for identifying children at risk for previously unidentified language disorders in public schools. Supplemental Material: https://doi.org/10.23641/asha.22044479