MONDAY, Jan. 12, 2015 (HealthDay News) -- The 10 to 20 minutes of a typical well-child visit isn't enough time to reliably detect a young child's risk of autism, a new study suggests.
"When decisions about autism referral are made based on brief observations alone, there is a substantial risk that even experts may miss a large percentage of children who need a referral for further evaluation," said lead study author Terisa Gabrielsen. She conducted the study while at the University of Utah but is now an assistant professor in the department of counseling, psychology and special education at Brigham Young University in Provo, Utah.
"In this study, the children with autism spectrum disorder were missed because they exhibited typical behavior much of the time during short video segments," explained one expert, Dr. Andrew Adesman, chief of developmental and behavioral pediatrics at Cohen Children's Medical Center of New York.
"Video clips without clinical context are not sufficient to make a diagnosis -- just like the presence of a fever and cough doesn't mean a child has pneumonia," said Adesman.
In the study, Gabrielsen's team videotaped two 10-minute segments of children, aged 15 months to 33 months, while they underwent three assessments for autism, including the "gold standard" test known as the Autism Diagnostic Observation Schedule.
The 42 children included 14 already diagnosed with early signs of an autism spectrum disorder, 14 without autism but with suspected language delays and 14 who were typically developing.
The researchers then showed the videos to two psychologists who specialized in autism spectrum disorders. These experts rated typical and atypical behaviors observed, and determined whether they would refer that child for an autism evaluation.
About 11 percent of the autistic children's video clips showed atypical behavior, compared to 2 percent of the typically developing children's video clips. But that meant 89 percent of the behavior seen among the children with autism was noted as typical, the study authors noted.
"With only a few atypical behaviors, and many more typical behaviors observed, we suspect that the predominance of typical behavior [in a short visit] may be influencing referral decisions, even when atypical behavior is present," Gabrielsen said.
When the autism experts picked out who they thought should be referred for an autism assessment, they missed 39 percent of the children with autism, the researchers found.
"We were surprised to find that even children with autism were showing predominantly typical behavior during brief observations," Gabrielsen said. "A brief observation doesn't allow for multiple occurrences of infrequent atypical behavior to become evident amidst all the typical behavior."
The findings, published online Jan. 12 in the journal Pediatrics, were less surprising to pediatric neuropsychologist Leandra Berry, associate director of clinical services for the Autism Center at Texas Children's Hospital.
"This is an interesting study that provides an important reminder of how difficult it can be to identify autism, particularly in very young children," Berry said. "While informative, these findings are not particularly surprising, particularly to autism specialists who have in-depth knowledge of autism symptoms and how symptoms may be present or absent, or more severe or milder, in different children and at different ages."
The observations in this study also differ from what a clinician might pick up during an in-person visit, Adesman suggested.
"It is important that information be gained from the child's parents and other caregivers," Adesman said. "Questions should ask about social engagement, verbal and non-verbal communication, interactions with the environment (especially toys) and behavior in general."
Gabrielsen agreed, adding that a thorough autism diagnosis requires a team.
"When we talk about an assessment for accurate diagnosis of autism, we are talking about a comprehensive assessment that often involves different disciplines -- psychologists, speech and language pathologists, occupational therapists, developmental pediatricians, other health care professionals and sometimes physical therapists," she said.
An evaluation also requires an extensive developmental and medical history and possibly hearing testing, Gabrielsen added, so false positives after a comprehensive diagnostic evaluation like this are uncommon.
Evaluating a child for autism can take as little as three to four hours or as much as several days, depending on the child's symptoms, she said. The cost can also vary from nothing (if a school district covers it), to a sliding scale at a state-sponsored clinic, to several thousand dollars at private clinics, Gabrielsen said.
The American Academy of Pediatrics now recommends universal screening for autism at 18 and 24 months.
"Some of the red flags include: no big smiles by 6 months, no back-and-forth communication by language or gestures by 12 months, no words by 16 months, no two-word phrases by 24 months and any loss of language or other social skills at any time," Gabrielsen said.
Parents can also use the Modified Checklist for Autism in Toddlers, Revised (M-CHAT-R), available at autismnow.org, to see whether their child shows signs of autism.
Spotting an autism spectrum disorder early is important because it can "lead to earlier intervention, which can often have a significant therapeutic benefit for a child with autism spectrum disorders," Adesman noted.
"That said," he added, "it is not uncommon for children to not be diagnosed until their preschool years, and mildly affected, higher-functioning kids with autism spectrum disorders may not be identified until their school-age years."
According to Berry, studies show about 30 to 50 percent of parents of children with autism do notice a problem in the child's first year of life, and 80 to 90 percent of parents notice problems by age 2.
Gabrielsen agreed that children whose autism is identified earlier and who receive earlier intervention tend to have better outcomes. Those not identified until they are older may develop behavioral problems in the meantime.
"We think our findings are an important clue as to why many children with autism are not identified until they are 4 to 5 years old or older," she said.
"If we can identify autism in the toddler years, we can begin intervention while the brain is still rapidly developing," she said. "Teaching skills such as language and social skills during the developmental period in which those skills are typically being acquired make intervention more effective and efficient."
There's more on the signs of autism spectrum disorders at the U.S. Centers for Disease Control and Prevention.