The predictive value of a preschool diagnosis of developmental language impairment

Objective: To evaluate language and cognitive outcomes in elementary school children with a prior preschool diagnosis of developmental language impairment (DLI). Design/methods: A cohort of preschool children, consecutively diagnosed with isolated language impairment, was reassessed in elementary school. Measures used were the communication domains of the Battelle Developmental Inventory (BDI) and the Vineland Adaptive Behavior Scale, Peabody Picture Vocabulary and Expressive One Word Picture Vocabulary. Cognition was assessed using the BDI cognitive domain. Language impairment was defined as performance more than 1.25 SD below normative means on a language measure. Specific language impairment (SLI) was defined as language impairment concurrent with a cognitive score not more than one SD below the normative mean. Results: A total of 43/70 (61%) children were reassessed. Mean age in preschool was 3.6 +/- 0.7 years and 7.4 +/- 0.7 years at follow up. A total of 36/43 (84%) showed persistent language impairment. The mean BDI cognitive domain score was 80.0 +/- 14.2 (15/42 below -2 SDs). Only 11/42 (26%) children met current research criteria for SLI, 24/42 (57%) had language impairment but had cognitive scores more than one SD below normative means, and 4/42 (10%) had normal language and cognitive skills. No factors could be identified at intake that predicted language outcome using univariate or multivariate analysis. Conclusions: While a preschool diagnosis of developmental language impairment predicted persisting language impairment, the specificity of this impairment did not persist. This suggests either undiagnosed cognitive impairment in preschool children with apparently isolated language impairment or an evolving profile of more global developmental impairment.