Early Screening for Developmental Language Disorder: Predictive Validity of German Parent Reports

Area of Science:

• Child development research
• Clinical psychology
• Speech-language pathology

Background:

• Developmental Language Disorder (DLD) is a common condition with lasting impacts.
• Early identification is crucial, but predictive tools for DLD in German-speaking regions are limited.
• Existing parent-report tools like FRAKIS-K and SBE-2-KT lack systematic validation for DLD prediction.

Purpose of the Study:

• To assess the predictive validity of FRAKIS-K and SBE-2-KT at age two for identifying DLD risk at age four.
• To determine if adjusting screening cut-off values can enhance predictive accuracy.

Main Methods:

• A longitudinal study of 186 children assessed at age two and 145 at age four in Hanover, Germany.
• Children's expressive vocabulary and grammar were measured at age four using the P-ITPA.
• Predictive validity was evaluated using sensitivity, specificity, AUC, and RIOC; ROC analysis explored cut-off adjustments.

Main Results:

• The SBE-2-KT had moderate predictive power (AUC=0.85), and FRAKIS-K vocabulary alone performed similarly (AUC=0.83).
• Both tools showed adequate specificity but moderate sensitivity.
• Adjusting cut-offs improved sensitivity up to 71% with minimal loss in specificity.

Conclusions:

• FRAKIS-K (vocabulary) and SBE-2-KT are adequate first-tier screening tools for DLD risk at age two in the German U7 context.
• These tools effectively identify children not at risk, but positive screens require comprehensive assessment.
• Optimizing cut-offs may enhance early detection, though findings require validation in larger clinical samples.