Meeting Language Milestones May Not be Associated with Better Functioning at School Age If ASD Is Not Detected Early

Background: Previous studies have reported that early language milestones are useful for predicting prognosis of ASD (e.g., Kover et al., 2016; Mayo et al., 2013). However, these studies included only children whose ASD was detected during toddlerhood. There is evidence that ASD tends to be diagnosed earlier in children with language delay (LD) than in children with no language delay (NLD; e.g., Mandell et al., 2005). Therefore, children with LD may be more likely to receive early intervention, which could improve subsequent functioning (Estes et al., 2015). Thus, if ASD remains undetected longer in children with NLD, they may not experience better outcomes than children with LD.

Objectives: This study compared age of diagnosis, adaptive functioning, and ASD symptomatology of a group of children with LD and a well-matched group of children with NLD, whose ASD was not diagnosed during early toddlerhood.

Methods: Using retrospective record review, scores on gold standard diagnostic and developmental assessments were examined in a sample of 110 school-age children (ages 5-18 years) with ASD. All children had a clinical diagnosis of ASD and met criteria on the Autism Diagnostic Observation Schedule (ADOS) and Autism Diagnostic Interview-Revised (ADI-R). Fifty-five children had typical phrase speech onset (i.e., < 33 months, as reported on the ADI-R). These children were individually matched on age and sex (45 males, 10 females) to children with delayed phrase speech onset. Measures included ADI-R diagnostic algorithm and current behavior scores, ADOS calibrated severity scores (CSS), Vineland Adaptive Behavior Scales standard scores, and age of diagnosis. Normally distributed variables were compared using 2 (group) x 2 (sex) ANOVAs. Non-normally distributed variables were compared using Mann-Whitney U tests.

Results: There was a significant group difference in age of diagnosis (U = 968, p = .001, r = .395), with the NLD group being diagnosed relatively late (M = 8.39 years, median = 6.89) compared to the LD group (M = 5.28 years, median = 4.16). There were significant main effects of group for all ADI-R diagnostic algorithm domains (Fs > 11, ps < .001), as well as ADI-R current behavior scores in the communication and restricted repetitive behavior domains (i.e., fewer/less severe symptoms in the NLD group). However, there were no main effects of group for ADI-R current behavior in the social interaction domain, ADOS CSS or Vineland scores in any domain (all Fs < 3.7, ps > .05).

Conclusions: On average, children with NLD were diagnosed 3 years later than children with LD. Despite having typical language onset and fewer early ASD symptoms (i.e., lower ADI-R diagnostic algorithm scores), the NLD group resembled the LD group in several aspects of functioning at school-age. Specifically, at school age, both groups had moderate-to-severe ASD symptoms (i.e., ADOS CSS) and low adaptive functioning scores (i.e., below age expectations in all Vineland domains). Therefore, typical language onset was not an indicator of better prognosis in late-diagnosed children with ASD. This research highlights the importance of earlier ASD detection and intervention.