JMIR Publications

ABSTRACT

Background:

Language acquisition is indeed a crucial developmental milestone for children, and it is characterized by significant inter-individual variability in the first four years of life, depending on several factors. Furthermore, language development disorders exhibit symptom overlap with other neurodevelopmental disorders (NDD) or Simple Language Delay (SLD). Hence, early detection of developmental language disorders (DLD) is still challenging for primary caregivers.

Objective:

To validate the effectiveness of the Gades platform, an adaptive screening system for language disorders, designed to support preschool teachers in early detection. This system enables the dissemination of specialized information related to language disorders to educators without the intervention of a Nursery School Language Therapist (NSLT) during the screening process. The experiment aims to validate the tool's technical and functional capabilities in uncontrolled educational environments.

Methods:

Gades’s validation was conducted through a pilot observational study deploying the platform in a nursery school and an Early Childhood Educational and Psycho-pedagogical Centre in Madrid, Spain. The study involved 24 preschool teachers, 2 NSLTs, and 218 Spanish-speaking children aged 6-36 months. The methodology encompasses two phases. Initially, NSLTs conducted meetings informed teachers about developmental language disorders. Also, Gades platform was enhanced to gather child’s individual details. It delineates an adaptive screening model accommodating language development variability. Subsequently, teachers carried out screenings without the support of NSLTs. It evaluates Gades’s utility in guiding non-experts, such as preschool teachers, in language observations and its effectiveness in unsupervised educational settings.

Results:

The system identifies children with language difficulties (19.72%), noting more cases in boys (13.30%) compared to girls (6.42%); and those with a family history of NDD. There is a disparity in the prevalence of language issues between nursery school (13.43%) and the early childhood centre (29.76%). Children with language difficulties are prevalent between months 15 and 27, gradually decreasing until three years of age. In terms of language categories, children under 24 months exhibit difficulty in production questions. Gades’s performance is assessed based on agreement between the system's decisions and educator viewpoints. Gades exhibits an accuracy rate of 97.41%. In 2.59% of evaluations, evaluators disagreed with Gades’s proposal.

Conclusions:

Gades serves as a screening solution for language impairments into schools without support from NSLT. As a screening tool, it may yield higher prevalence rates, suggesting that Gades can detect children who have language difficulties, but refining the system's knowledge base to discern DLD or SLD cases is essential. Validated by non-experts like preschool teachers, Gades provides valuable information to conduct language observations, facilitating the identification of linguistic challenges within uncontrolled educational environments. Educators have embraced Gades due to its user-friendliness and ease of use. Finally, it allows for ubiquitous therapy assessing children in daily settings through multidisciplinary collaboration between teachers and language disorders experts.