Diagnostic stability of autism spectrum disorder in toddlers prospectively identified in a community-based setting: Behavioural characteristics and predictors of change over time.
Community ASD labels given at age two stayed put in 88% of kids; good eye contact, vocal turn-taking, and higher non-verbal scores at two years marked the few who later moved off the spectrum.
01Research in Context
What this study did
Barbaro et al. (2017) followed toddlers who got an autism diagnosis in a regular community clinic. They tracked the kids every six months from age two to four years.
The team wanted to know how many children kept the diagnosis and which early skills hinted at later change.
What they found
Eight out of every ten toddlers still met ASD criteria at age four. The kids who later 'moved off' the spectrum had better eye contact, more back-and-forth sounds, and higher non-verbal scores at two years.
How this fits with other research
Grigore et al. (2024) looked at every toddler screening study and say the evidence is still too shaky to prove early screening helps. That big-picture view includes Josephine's 88% stability figure, so the two papers fit together: stability is high, yet screening benefit remains unclear.
Glenn et al. (2003) tracked older, high-functioning kids for two years and saw almost no change. Their 'flat line' result pairs with Josephine's 88% stability, showing that little movement is the norm across age groups.
Wu et al. (2021) pushed reliable ASD detection even younger, down to 18-24 months with the T-STAT tool. Their work extends Josephine's window and suggests you can spot most cases before the two-year mark.
Why it matters
You can feel confident that a careful community diagnosis at two years usually holds. Watch for strong eye contact, shared sounds, and solid non-verbal problem-solving; these signs may flag kids who need less intensive support later. Pair this with newer tools like T-STAT if you screen before 24 months, and keep monitoring either way—because stability is high, but not 100%.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →During reassessment, pull video or session notes from the child's second-year eye contact and turn-taking; if those skills are strong, plan a shorter ABA re-evaluation cycle to check for possible diagnostic change.
02At a glance
03Original abstract
Autism spectrum disorder diagnoses in toddlers have been established as accurate and stable across time in high-risk siblings and clinic-referred samples. Few studies have investigated diagnostic stability in children prospective identified in community-based settings. Furthermore, there is a dearth of evidence on the individual behaviours that predict diagnostic change over time. The stability and change of autism spectrum disorder diagnoses were investigated from 24 to 48 months in 77 children drawn from the Social Attention and Communication Study. Diagnostic stability was high, with 88.3% overall stability and 85.5% autism spectrum disorder stability. The behavioural markers at 24 months that contributed to diagnostic shift off the autism spectrum by 48 months included better eye contact, more directed vocalisations, the integration of gaze and directed vocalisations/gestures and higher non-verbal developmental quotient. These four variables correctly predicted 88.7% of children into the autism spectrum disorder-stable and autism spectrum disorder-crossover groups overall, with excellent prediction for the stable group (96.2%) and modest prediction for the crossover group (44.4%). Furthermore, non-verbal developmental quotient at 24 months accounted for the significant improvement across time in 'Social Affect' scores on the Autism Diagnostic Observation Schedule for both groups and was the only unique predictor of diagnostic crossover. These findings contribute to the body of evidence on the feasibility of diagnoses at earlier ages to facilitate children's access to interventions to promote positive developmental outcomes.
Autism : the international journal of research and practice, 2017 · doi:10.1177/1362361316654084