How stable is an early ASD diagnosis?


Turns out, pretty stable. While available diagnostic tools may not be the best for very early diagnoses (<12 months), this study on over 1000 toddlers sampled from the general population found a diagnosis of ASD made at 18 months by a clinician with expertise in ASD was more stable than diagnoses of language delay or typical development. 84% of children who received an early diagnosis of ASD still qualified for that diagnosis at 3–4 years of age. Other relevant takeaways include:

  • The most common diagnostic transition: ASD at first visit to ASD features at last visit (i.e., children no longer met full criteria for ASD on DSM-V, but still had some characteristics of ASD)

  • Only 1.8% of toddlers went from an initial diagnosis of ASD to typical development

  • ASD was missed in almost 1/4 of first evaluations; many of these later-diagnosed children (~43%) were initially suspected to have a language delay instead

We still need more research to figure out the best treatments and courses of action, so the intervention piece is still business-as-usual; but for now we can trust that early diagnosis of ASD is not only possible, it’s pretty reliable.


Pierce, K., Gazestani, V. H., Bacon, E., Carter Barnes, C., Cha, D., Nalabolu, S., …, & Courchesne, E. (2019). Evaluation of the diagnostic stability of the early autism spectrum disorder phenotype in the general population starting at 12 months. JAMA Pediatrics. doi:10.1001/ jamapediatrics.2019.0624.

We can’t catch ‘em all; or can we?

Have you ever worked with one of those “quirky” children and thought, “I wonder how he’ll do later?” I mean, sometimes a quirk is just that, but how do we know when a quirk is something more? And why does it matter? Well, we know that early identification (for purposes of securing early intervention) is important for children with autism. But the false-negative rate from screening might be as high as 60% in some cases. That means that some children with autism are not being caught before school age, either because their symptoms aren’t that bad, or the skills they do have are compensating for the deficits. Without a crystal ball, what can we do for those kids who would otherwise fall through the cracks? The authors of this study looked to compare results of the Autism Diagnostic Interview-Revised (ADI-R) in two groups of children with ASD: early diagnosed (before age 5), and late diagnosed (after age 5).


They found no statistically significant difference between groups in Autism Diagnostic Observation Schedule (ADOS) scores or ADI-R scores, although the early diagnosed group had marginally worse communication scores and more often flagged on items such as “used another’s body to communicate”, repetitive use of objects/parts of objects, and pronominal reversal (when children use "he", "she", "you", or their own name instead of “I” or “me”). The later diagnosed group had fewer symptoms and better communication, but again these differences were marginal.

The authors also found some trends between groups in terms of demographics. Children who were Hispanic, of racial minority, had an additional diagnosis such as ADHD, and/or had higher verbal and full-scale IQs were significantly more likely to be diagnosed later.

Now, if you forget the rest of this review, remember this part: in comparing individual items from the ADI-R, there were five symptoms that ALL of the later-diagnosed children displayed. “Deficits in:

  • reciprocal conversation
  • offering to share
  • imaginative play with peers
  • social verbalization/chatting and
  • showing/directing attention.”

What is interesting about these symptoms is that they were all negative. In other words, kids diagnosed later seemed to “not do” stuff they should have, rather than do extra stuff that would raise a concern (like flapping or showing repetitive interests). The authors suggest that “attention to deficits in these behaviors may facilitate earlier identification of ASD without other obvious impairments.” Completing or referring for a more formal assessment when you notice a child has these symptoms could make a difference in outcomes for these children.

*Note: some skills, like being able to have a back-and-forth conversation, aren’t a major concern if toddlers haven’t quite mastered them, but if a child is still demonstrating these deficits by preschool, make that referral!

Goodwin, A., Matthews, N. L., & Smith, C. J. (2018). Parent-reported early symptoms of autism spectrum disorder in children without intellectual disability who were diagnosed at school age. Autism. Advance online publication. doi: 10.1177/1362361318777243

Screening for ASD? There’s an app for that!

Screening for autism in early intervention. From 16–30 months, we have the M-CHAT-R. What do you use beyond 30 months? The SRS-2 is an option now, but there is not a lot of research on how well it works yet. The SCQ starts at 4 years. And all of those are completed through parent report.

What if we had a screener that combined parent report with objective measurement, that could be administered on a parent’s smart phone? Enter Cognoa!


This two-part screener, intended for ages 18–72 months, is contained in a mobile app. Parents answer 15 questions (similar to the M-CHAT-R or SCQ), and then are prompted to record 1- to 2-minute videos of their children in everyday interactions. The questionnaire is scored immediately, and the videos are reviewed by experienced clinicians for a rating score. All of those scores go into algorithms (yay technology) to determine level of risk. Results are then sent to the family, physician, EI team, etc. Kanne, et al. evaluated Cognoa against other screening measures like the M-CHAT-R, SCQ, SRS, and CBCL and found it to be comparatively sensitive and more specific than those screenings. In other words, Cognoa correctly flagged for ASD evaluation as well as the screenings we already use, and it had fewer false positive screenings. Not bad! Unfortunately, there is a downside. Currently, Cognoa is only really available for parents whose employers buy access to the screenings as a health incentive to their employees. In the meantime, we might have to stick with the measures we already know, but keep your eyes and ears peeled in case this app picks up steam!


Kanne, S. M., Carpenter, L. A., & Warren. Z. (2018). Screening in toddlers and preschoolers at risk for autism spectrum disorder: Evaluating a novel mobile-health screening tool. Autism Research. Advance online publication. doi: 10.1002/aur.1959


Hey, wait! This isn’t the only article on autism screening tools found this month! Janvier et al. found that a simple picture-based developmental checklist could successfully differentiate children with and without ASD among low-income, minority families. This screening tool, The Developmental Check In, may be a useful way to screen for autism among underserved children, particularly when parents are not native English speakers.


Janvier, Y.M., Coffield, C.N., Harris, J.F., Mandell, D.S., & Cidav, Z. (2018). The Developmental Check-In: Development and initial testing of an autism screening tool targeting young children from underserved communities. Autism. Advance online publication: doi: 10.1177/1362361318770430

Parents' role in the identification of early signs of autism

There have been many advances in the early identification of autism in recent years, but did you know that the average age of diagnosis is still around 4 years old???!!!. I’m sure you’re thinking the same thing that I am—We need to do better than that!

When it comes to early identification of autism, previous research has shown that one way we can make sure we’re getting all of the information needed to make a well-informed decisions is by including more than one person’s input in the process. And, if we’re talking about the birth to 3 population and we’re talking about autism, we naturally think of including the parent, right? Many of the well-known, most frequently-used screening tools for identifying early signs of autism and communication delays are designed to be completed by the parent, or at least include a parent questionnaire component (think: M-CHAT and CSBS DP).


But, exactly how should we be using parent input? Should we be putting more or less weight on our clinical observations? When we’re talking about children under two, how accurately is a parent able to identify the early signs of ASD compared to an experienced clinician? If you’ve wondered these same kinds of things, I have some good news for you: the results from this study help to clarify the role that parents play in the early identification of autism. 

This study included 188 siblings of children with ASD and their primary caregivers. Parents completed the Autism Parent Screen for Infants [APSI], and SLPs completed the corresponding Autism Observation Scale for Infants [AOSI] when the children were 12 months, and then again at 18 months. Parent response to the 19 items included in both the APSI and the AOSI (e.g., vocalizing back and forth, showing interest and pleasure, imitating, eye contact) were compared to the clinicians’ responses. And, the results were not only interesting, but also extremely informative: parents identified more behaviors on the APSI that differentiated children diagnosed with ASD from typically-developing children compared to clinicians’ identification of behaviors on the AOSI.

So, based on this information, when it comes to early identification of autism, we really should be viewing parents as a first line of defense. The setting as well as the context of our assessments limit our ability to get all of the information. We need to rely on parents to help us fill in the gaps, adding specificity and depth to our understanding of the developing communication skills of children under 18 months of age. Now more than ever, we need to see parents as our allies, and as key contributors to the assessment process.

Sacrey, et al., (2018). Parent and clinician agreement regarding early behavioral signs in 12- and 18-month-old infants at-risk of Autism Spectrum Disorder. Autism Research, 11(3), 539–547.

Throwback Pub (2008): Early gestures to predict vocabulary

Note: This study was part of a larger longitudinal study, information available here.

Wouldn’t it be fantastic if we could screen infants at 8 months and predict later development? This study looked at gesture and object use at 8 months, then followed up at ages 1;0 and 2;0 to determine whether later vocabulary can be predicted by early gesture and object use. The authors mailed out questionnaire packages to 1,477 families when their children were ages 0;8, 1;0, and 2;0. The Macarthur-Bates Communicative Development Inventories (CDI) were included in the packet, the CDI Words and Gestures (CDI:W&G) was completed for the first two data collections, and the CDI Words and Sentences (CDI:W&S) was completed at the 2-year mark.

In addition to measuring gestures to predict vocabulary use, the authors also examined the effects of SES and gender on vocabulary. Here’s what they found:

  • Gestures at 0;8 didn’t predict vocabulary (at 2;0) all that well. But gestures at 1;0 year did. Also, gestures predicted 2;0 vocabulary comprehension better than production.
  • Girls produced more words than boys at 1;0 and 2;0, but there were no differences between boys and girls in comprehension
  • Children in the two higher SES groups (there were 5 groups total) reportedly understood fewer words than the children in the lower SES groups. And children in the second lowest SES group were reported to produce more words than children in the higher SES groups

Remember that these fun facts were all based on the CDI, which relies on parent report. So how can we use this information? The authors suggest that parent estimation of their children’s skills may differ by SES; that is, families in lower SES backgrounds may overestimate their children’s abilities, or families from higher SES backgrounds may underestimate their children’s knowledge. This is something to keep in mind when many of our assessments in EI are based, in some capacity, on parent report.


Bavin, E. L., Prior, M., Reilly, S., Bretherton, L., Williams, J., Eadie, P., ... & Ukoumunne, O. C. (2008). The early language in Victoria study: Predicting vocabulary at age one and two years from gesture and object use. Journal of Child Language35(3), 687-701.