The Toddler Autism Symptom Inventory: Use in diagnostic evaluations of toddlers.

The sample was referred after positive screening or provider concern, so the results may not apply to children identified in other ways. Children with a previous autism diagnosis or significant sensory or motor impairments were excluded, so the results do not describe how the interview performs in those children. The development and cross-validation samples were not randomly assigned or matched; they were collected at different times and had small demographic differences. The sample was not representative of the current US Census or of the specific sites, and the authors call for validation in more diverse racial, ethnic, and socioeconomic groups. Children and parents were not separated during the interview, which may have influenced scoring. Specificity was lower in the cross-validation sample (67.61), and positive predictive value for autism was 62.67 in Sample 1 and 42.50 in Sample 2. One item about intense interests had only moderate inter-rater reliability. The authors say further work should compare TASI results with other diagnostic interviews such as the ADI-R in a common sample.

No author conflicts of interest or funding declaration is included in the supplied text; the article is tagged as supported by NIH extramural research.

Do not read a score of 7 or more out of 40 as a diagnosis of autism, or a score below 7 as proof that autism is absent. The paper says the TASI should be used with other diagnostic tools, and six of 72 children with autism did not meet the cutoff.