How accurate are common autism tests for preschool children?

In short

Common autism assessment tools for preschool children can be helpful, especially for identifying children who may need further evaluation, but researchers say no single test should be treated as a final answer on its own.

What the researchers studied

Researchers reviewed studies looking at how accurately different autism assessment tools work in children under six years old. In these studies, the final diagnosis made by an experienced clinical team was treated as the main point of comparison.

The review focused on several widely used tools, including ADOS, CARS, and ADI-R. Some of these involve direct observation of the child, while others are based on detailed interviews with parents.

The researchers combined results from 13 scientific papers covering 21 different sets of data. However, they also noted that many of the studies had weaknesses in how they were designed, which may have made the tests appear more accurate than they really are in everyday practice.

What they found

Among the tools studied, ADOS was the most sensitive overall. In simple terms, it identified most children who were later diagnosed with autism. But it also sometimes flagged children who ultimately did not receive an autism diagnosis.

CARS appeared somewhat better at avoiding false positives, meaning children without autism were less likely to be incorrectly identified. However, it also missed more autistic children compared with ADOS.

The ADI-R interview showed lower sensitivity than the other tools. This means that some autistic children could receive a negative result even when autism was actually present.

Researchers also found that combining multiple tools did not always improve accuracy as much as people might expect.

An important finding was that when weaker studies were removed from the analysis, the reported accuracy of all tools became lower.

What this means for families

For parents waiting for an autism assessment, this review is an important reminder that no questionnaire or test score alone can confirm or rule out autism.

Specialists usually make decisions by combining many sources of information. They look at developmental history, observe the child directly, talk with parents, and sometimes follow the child over time before reaching conclusions.

This also means that one surprising result — either positive or negative — should not automatically be treated as final. A child may score highly on one test and still need a broader evaluation, while another child may receive a lower score despite clearly needing support.

What this means for clinicians

The review supports using tools like ADOS, CARS, and ADI-R as part of a larger assessment process rather than as standalone diagnostic decisions.

Researchers also point out that newer versions of some tools have appeared since many of the included studies were published, so real-world accuracy today may differ somewhat from these estimates.

Limitations and what researchers still do not know

Many of the studies included small numbers of children and used inconsistent methods. Most of the research was also conducted in Western clinical settings, so the results may not apply equally everywhere.

The review also could not fully answer how well these tools work for girls, minimally speaking children, or children with intellectual disabilities.

Researchers say more independent and higher-quality studies are still needed.

---

This is a plain-language summary of Diagnostic tests for autism spectrum disorder (ASD) in preschool children by Randall M, Egberts K.J, Samtani A, et al., Cochrane Database of Systematic Reviews (2018).

Source license: CC-BY-NC-4.0.

This article is not medical advice. Please consult a qualified healthcare professional before making decisions about diagnosis or treatment.