The Observation Scale for Autism (OSA): A New Screening Method to Detect Autism Spectrum Disorder before Age Three Years

Nils Haglund, SvenOlof Dahlgren, Karin Källén, Peik Gustafsson, Maria Råstam


Background: There is an increasing body of evidence that early interventions for children with Autism Spectrum Disorder (ASD) promote a positive development of social interaction. Thus, tools for early detection of ASD are warranted.

Aim: Development of, and deciding cut-off-levels for, a new screening tool for ASD, the Observation Scale for Autism (OSA). The OSA was designed to be used at the free health check-up at 30 months, offered to all children in Sweden.

Method: The OSA consists of 12 observations and takes less than 10 minutes to use. The performance of the test was investigated by assessing 37 children previously diagnosed ASD, 23 with Down Syndrome (DS) and 26 typically developing children (TD).

Results: Children diagnosed with ASD showed statistically significant higher scores in all 12 items compared to TD children, and significantly higher in 10 items compared to the children with DS. Most of the observations in OSA seemed to cover specific symptoms of ASD, but two of the observations were more related to developmental level. The nine most discriminative items for ASD were identified, and among those, a cut-off limit was chosen (≥3 items). Among children with ASD, 34/37 reached the proposed cut off, compared to 0/26 and 4/23 among children in the TD and DS groups, respectively.

Conclusion: The results suggest that the OSA discriminates children with ASD from TD children and children with DS. Using the suggested cut off, OSA provides high sensitivity for ASD (92%) with a very low false positive rate.


Screening, Autism, sensitivity, Intellectual disability, typically developing children

Full Text:



  • There are currently no refbacks.

ISSN: 2292-2598