www.medscape.com
March 09, 2015
I am Judi Miller, a psychologist at the Children’s Hospital of Philadelphia. Autism spectrum disorder is a developmental
disorder that for many individuals leads to significant impairments throughout life.
There is no blood test for autism, but the behaviors are sometimes noticeable at around 1824 months of age. The
American Academy of Pediatrics recommends formal autism screening at those ages.
A study published in Pediatrics
[1] grew out of two clinical observations. First, many families told us that the first time
they asked the doctor about autism, they were told that their child couldn’t have autism because the child just looked at
the doctor during the examination or just responded to his or her own name. For these families, then, it sometimes
took years and multiple evaluations before autism was seriously considered and the correct diagnosis was made. And,
of course, that is very frustrating for families. We know that early intervention can lead to better outcomes.
Second, there is a saying among autism specialists that “young children with autism show some of the features some
of the time,” so this suggests that the opposite is also true.
These two observations, together, led us to hypothesize that the ratio of typical to atypical behaviors may be important
and that if we understood that better, we might help clinicians make better referral decisions and also learn about the
early course of autism spectrum disorder as researchers.
We looked at three groups of children: some with early signs of autism, some with language or other developmental
delays, and a group who weren’t developing typically. We asked the question, “If you only had 10 to 20 minutes with
this child, what types of behaviors might you see?”
A strength of our sample is that these children were screened and identified through routine care, so we don’t have as
many selection biases as some studies. These children are fairly representative of what comes in front of the average
pediatrician. We found that children in all three groups showed a lot of typical behavior and that children in all three
groups showed at least some atypical behavior.
When we compared the groups statistically, we found that children with autism showed significantly more atypical
behavior, but clinically, in a short observation period, those behaviors might not really stand out. It is basically a signaltonoise
problem. In fact, our raters, who were autism specialists, made incorrect referral decisions about 40% of the
time, on the basis of short observations alone.
Let me describe one of our best examples. We found that children with autism responded to their own names about
50% of the time and that children with language delays or those who were developing typically responded to their
names only about 80% of the time. That suggested to us that children with autism sometimes do respond to their
names, and children without autism sometimes don’t respond to their names. It is maybe the rate of responding that is
more informative.
Can Autism Be Ruled Out in a Brief Observation?
Judith S. Miller, PhD
The other behaviors that we looked at in this study had even lower ratios. For every instance of an unusual behavior,
the children with autism might show many examples of behaviors that looked typical.
We think our results suggest two things. First, if you only have 10 or 20 minutes with a child, that may not be enough
time to detect the signal of autistic behaviors against the background noise of behavior that looks typical, so we must
be very cautious not to rule out autism too quickly, especially if we are ruling it out on the basis of a brief observation.
Second, we must be very careful about letting our clinical judgment override formal screening results. Screening
instruments are not perfect, but in the area of child development, our study and several others have found that formal
screening results are better at identifying potential problems than clinical judgment alone. We encourage clinicians to
take advantage of formal screening instruments to provide the best care for their patients.
If you would like more information about autism, visit the Center for Autism Research. We have developed a lot of very
helpful information for families and providers, and we think you will find it very helpful as you provide care for your
patients.
© 2015 Children’s Hospital of Philadelphia and Medscape
Cite this article: Judith S. Miller. Can Autism Be Ruled Out in a Brief Observation? Medscape. Mar 09, 2015.
References
1. Gabrielsen TP, Farley M, Speer L, Villalobos M, Baker CN, Miller M. Identifying autism in a brief observation.
Pediatrics. 2015 Jan 12. [Epub ahead of print]
Medscape: Can Autism Be Ruled Out in a Brief Observation?
