New measures in morphosyntactic development of children with developmental language disorders using spontaneous language sample analysis: a study of TARSP-P

Abstract

Background Language sample analysis (LSA) is seen as the gold standard for diagnosis and documentation of oral language skills in children with developmental language disorders (DLD). In the Netherlands, 'Taal Analyse Remediëring en Screening Procedure' (TARSP) is one of the methods used for spontaneous LSA. A limitation of this method is that it lacks the ability to detect and evaluate small changes when a child makes progress in morphosyntactic structures. For that reason, in 2016 Van Oorschot and Bruisman developed a new measure: 'Taal Analyse Remediëring en Screening Procedure met Punten' (TARSP-P). Aim This study aims to determine whether TARSP-P is more sensitive to small changes in morphosyntactic development than traditional TARSP. Methods Repeated measures at three time points in a four month period were conducted in a group of twelve children with DLD, age 4 and 5 years old. Language samples were collected and analysed and a sentence repetition task was performed. Results Significant differences in TARSP-P scores for all time points were found. This indicates that TARSP-P is more sensitive to small morphosyntactic changes at both group level and age group level than traditional TARSP, which is limited to measuring seven phases in language development. Significant strong positive correlations of TARSP-P with one other language measure were observed. No significant correlations between TARSP-P, TARSP and a sentence repetition task were found. Inter Rater Reliability of TARSP-P outcomes showed good agreement. Conclusion TARSP-P is more sensitive to small morphosyntactic changes when compared with the traditional TARSP and is therefore a better indicator for assessing treatment effects in children with DLD. Implications of key findings For use in clinical practice, both the practical feasibility and time-commitment involved in using TARSP-P should be further evaluated. A computerised method of speech recognition and automated analysis are recommended.