Review of language impairment definitions in 75 speaking autistic adolescents and young adults

Purpose: Co-occurring language impairment (LI) in autism is common and predicts long-term academic, occupational, and social outcomes. Yet little is known about LI in autism beyond childhood. One challenge to closing this gap is the lack of consensus regarding how LI should be operationally defined in adolescents and adults. This study examines how different epidemiological definitions of LI influence clinical classification and observed language profiles in speaking autistic adolescents and young adults. Method: Participants (N = 75; ages 13-30) varying in levels of autism traits completed norm-referenced measures of overall expressive language, overall receptive language, receptive vocabulary, expressive vocabulary, nonword repetition, and NVIQ. Scores were compared to epidemiological definitions for LI varying in criteria and cutoffs from -1 SD to -1.5 SD. Data were analyzed using descriptives and clustering. Results: More stringent definitions yielded a greater proportion of participants meeting LI criteria, and more stringent cutoffs for LI yielded greater overall consistency in clinical classification across individual language measures, but there was no one-to-one ratio between cutoff and clinical classification. Clustering indicated three profiles differentiated on the basis of language and nonverbal cognitive skills, but each cluster was heterogeneous. Individual performance also varied across language measures. Discussion: Findings support multi-domain approaches to characterizing language skills in autistic adolescents and adults, including those with LI. Future work is needed to understand language skills in autism beyond childhood and how to develop effective assessment practices.