Pediatric Office Visits May Miss Autism Diagnosis

Terisa P. Gabrielsen, PhD, NCSP Assistant Professor, School Psychology Dept. of Counseling Psychology and Special Education Brigham Young University, Provo, UT Interview with:
Terisa P. Gabrielsen, PhD, NCSP

Assistant Professor, School Psychology
Dept. of Counseling Psychology and Special Education
Brigham Young University, Provo, UT 84602

Medical Research: What is the background for this study? What are the main findings?

Dr. Gabrielsen: One of the keys to improving outcomes for individuals with outcomes is to begin intervention as early as possible, which means we need to identify autism symptoms as early as possible, preferably during the early toddler years.  The current study grew out of a screening feasibility study to see what would happen if pediatricians followed the AAP guidelines for screening every child for autism at ages 18 and 24 months as part of their regular pediatric care appointments.  That study  was conducted in a large, independent community pediatrics practice.  We found that universal screening of 796 patients helped to identify 10 toddlers with autism who had not previously been referred for evaluations.  Physicians had previously identified 3 others with autism in the group, and toddlers with other delays, such as language delays, were also identified through the screening process.  We wondered what some possible causes were for the low rate of autism referrals and designed the current study to look for what information was available to a pediatrician during the timespan of a typical pediatric exam.  We found that even in toddlers with autism, a brief (10-minute) sample contains an overwhelming ratio of typical behaviors (averaging 89%) compared to infrequent atypical behaviors (11%)  that would indicate the presence of autism.  We had autism experts identifying the behaviors from videos of the evaluations of children in the previous study, so they had many luxuries that a clinician doesn’t have during an exam (i.e., ability to focus on one aspect of development, ability to rewind and re-view behaviors).  After watching the 10-minute video observations, we asked our experts, “Would you refer this child for an autism evaluation?”  We found that even the experts missed referring a child for an autism evaluation 39% of the time when the only data available were the brief observations.

Medical Research: What should clinicians and patients take away from your report?

Dr. Gabrielsen:

For clinicians, the findings indicate that brief observation alone is not sufficiently reliable for detecting autism symptoms in very young children.  We think this may be due to the very high ratio of typical behaviors that may overshadow the more subtle atypical behaviors.  In addition to in-person evaluations of behavior, we recommend implementation of the AAP guidelines for autism screening at 18 and 24 months as an critical source of data for decision making.

We also think it is very important that the clinician carefully consider parent-reported symptoms that might not occur during brief observations.  This means parents need to tell their pediatricians about atypical behaviors they have observed at home over time and what their concerns are about their child’s development.   If autism screening is not yet offered through your pediatric care provider, autism screeners (such as the M-CHAT-R) are available for free, online ( ), and parents can take the results to their pediatrician.  The CDC website “Learn the Signs. Act Early.” ( provides developmental milestones to look for between ages 2 months and 5 years, and also offers lists of “reg flags” at each age that indicate a possible problem with development.

Medical Research: What recommendations do you have for future research as a result of this study?

Dr. Gabrielsen: Our goal is to identify autism as early as possible.  Many researchers are looking at ways to pick up the subtle early signs of autism in all aspects of behavior, development, and physiology.  We are looking at how reliable other single sources of data are (commonly used screeners and parent-reported concerns) and how they compare to the decisions made after direct observation.  It may be that no single method is reliable, but that a combination of data sources would be most accurate.  Looking at what factors drive decision-making about whether to refer a child for further autism evaluations is another important direction for research in hopes of improving the outcomes of individuals with autism spectrum disorder.

The reference to the previous study is:

  1. Miller J, Gabrielsen T, Villalobos M, et al. The Each Child Study: Systematic screening for autism spectrum disorders in a pediatric setting. Pediatrics. 2011;127:866-871


Identifying Autism in a Brief Observation

Terisa P. Gabrielsen, PhD, NCSP Megan Farley, PhD Leslie Speer, PhD, NCSP Michele Villalobos, PhD Courtney N. Baker, PhD Judith Miller, PhD

Pediatrics Published online January 12, 2015

(doi: 10.1542/peds.2014-1428)

[wysija_form id=”1″]


Last Updated on January 13, 2015 by Marie Benz MD FAAD