Reconceptualizing developmental language disorder as a spectrum disorder: issues and evidence

Area of Science:

• Linguistics
• Developmental Psychology
• Clinical Research

Background:

• Developmental Language Disorder (DLD) presents with significant variability, leading to professional disagreement on its nature and clinical management.
• Characterizing this variability is crucial for advancing DLD research and establishing effective diagnostic and treatment frameworks.
• Three models—predictable subtypes, individual differences, and continuum/spectrum—were examined to explain DLD presentation variability.

Purpose of the Study:

• To statistically analyze three proposed models for characterizing variability in Developmental Language Disorder (DLD) presentation.
• To test these models using cluster analyses on a population-based sample of children with DLD.
• To determine if DLD presentation aligns with distinct subtypes, individual differences, or a continuum/spectrum.

Main Methods:

• Utilized cluster analyses (Ward's method and K-means) on a dataset of 505 children with DLD from the US Epidemiological Study of Language Impairment.
• Included all available language and cognitive measures for analysis.
• Employed Bayesian information criteria for optimal cluster size selection and bootstrapping/permutation methods to assess clustering randomness.

Main Results:

• Both cluster analyses generated over 10 clusters, lacking clear spatial distinction and clinical interpretability.
• Statistical tests confirmed that the identified cluster solutions were not the result of random aggregation.
• Findings indicated non-random clustering but a large number of clinically non-interpretable subtypes.

Conclusions:

• Empirical evidence supports the continuum/spectrum and individual differences models over distinct subtypes for DLD presentation.
• Substantial support was found for the continuum/spectrum model, with weaker support for individual differences.
• Clinicians should prioritize severity-based treatment plans for children with DLD, considering it a spectrum condition rather than focusing on discrete subtypes.