Study Raises Questions About Universal Screening for ASD

Veronica Hackethal, MD

September 27, 2019

"Don't throw the baby out with the bathwater," Karen Pierce, PhD, codirector of Autism Center of Excellence, University of California, San Diego, cautioned about a new study that could potentially undermine screening for autism spectrum disorder (ASD).

The study, published online today in Pediatrics, was conducted in a real-world setting at Children's Hospital of Philadelphia primary care clinics, one of the nation's top pediatric care networks.

Results suggest that routine screening of every child for autism at 18- and 24-month well-child visits is feasible, but the results may not be very accurate.

As conducted, universal screening detected only 38.8% of children who were eventually diagnosed with ASD, and only 14.6% of children who initially screened positive truly had ASD. Accuracy was even lower for minority children, as well as those from homes with lower incomes.

"Although some may interpret these findings as evidence against universal screening, we caution against this interpretation given the earlier age of diagnosis for screen positives. Instead, results suggest that augmentative screening methods should be developed to detect more children through universal screening efforts and reduce disparities," write Whitney Guthrie, PhD, of the Children's Hospital of Philadelphia, and colleagues.

Pierce agreed. The study, although not perfect, provides important lessons, she said.

"If people reduce their screening because of this study, we're at huge risks of going backwards to when children weren't being diagnosed until 7, and the idea of recovery was unthinkable. Now there are many cases when that is possible," she said. "This study doesn't discourage me from screening whatsoever, but it does tell me that we need to improve."

Pierce is an advocate of universal early screening for ASD. She is also the coauthor of a widely publicized study that suggested some children may be reliably diagnosed with autism as early as age 14 months.

Symptoms of ASD can manifest themselves in the first years of life, although the average age at diagnosis is 4 years. Children of ethnic and racial minorities and those in rural settings or of lower income backgrounds may experience longer delays in diagnosis.

Many experts believe that early diagnosis is crucial because it enables earlier intervention, which, in turn, may improve outcomes. The first 18 months of life are critical for brain development. During this time, brain plasticity is enhanced, which may provide a window of opportunity to improve outcomes for children with ASD, according to Pierce.

"I have seen a huge amount of benefit in the sense that some children in San Diego are detected as early as 12, 13, 14, months and gotten into very early treatment," she said.

Children are usually screened only on the basis of parental report of symptoms, and that may not be enough to detect all cases of ASD, she continued.

"Screening is quick, usually 5 to 10 minutes, so it's not surprising that you're not detecting a lot of kids with such a quick process," she said.

In her own research, Pierce uses a broader screening tool that is not specific for ASD. That's because there is a lot of overlap among children with ASD and those with other types of developmental delay. Also, some children with ASD present with a single delay, such as not meeting their language milestones; using a broader screening tool may identify them before they meet full ASD diagnostic criteria.

Other ways to improve screening include combining diagnostic approaches, such as eye tracking, or risk factors, such as male sex, being born prematurely, or having a sibling who has autism.

Still, universal screening remains hotly debated.

The American Academy of Pediatrics recommends universal screening at age 18 and 24 months. But 2016 guidelines from the US Preventive Services Task Force neither recommend for nor against screening for ASD in children who do not have symptoms, owing to the fact that there is not enough evidence of potential benefits and harms of screening.

"Autism awareness and research is a constantly evolving field. In 2016, the Task Force called for more research specifically on the impact of screening in young children who do not have symptoms of autism spectrum disorder, in the hopes that it would help us to ultimately make a recommendation when reviewing this topic in the future," Task Force Vice Chair Alex Krist, MD, MPH, told Medscape Medical News. Krist is a professor at Virginia Commonwealth University in Richmond.

Ninety Percent of Children Screened

In the newly reported study, Guthrie and colleagues conducted primary care–based screening with the Modified Checklist for Autism in Toddlers With Follow-Up (M-CHAT/F), the most widely used and studied screening tool for ASD.

The M-CHAT/F was available in Spanish and English and was administered electronically to all children as part of routine 18- and 24-month well-child visits. The study took place through the Children's Hospital of Philadelphia's network of 31 pediatric community-based primary care clinics in Pennsylvania and New Jersey. Follow-up occurred at least to age 4 years and up to 8 years. The analysis included 25,999 children.

Results showed that near universal screening using the M-CHAT/F is possible in primary care: 90.9% of children were initially screened, and 2.2% were ultimately diagnosed with ASD.

But results also suggested low accuracy for the M-CHAT/F. Overall, the test had a sensitivity of 38.8%, missing the majority of children with ASD.

Likewise, the overall positive predictive value (PPV) was 14.6%, so only a small percentage of children who screened positive truly had ASD. But that statistic only included children who received a diagnosis of ASD. When all children with developmental delays were included, the PPV jumped to 72.4%.

Older age was associated with better sensitivity (48.8% at 21 to 26 months vs 35.1% at 16 to 20 months; P = .009), as was repeat screening (second, 39.8%, vs first, 31.8%; comparison of accuracy of screenings, first vs both P = < .001).

The test's PPV was lower among children of racial/ethnic minorities (white, 24.0%, vs black, 11.7%; P < .001; white, 24.0%, vs Asian 90.4%; P < .001; and Hispanic 92.%, vs non-Hispanic, 95.1%; P < .001), as well as among those from homes with lower vs higher incomes (92.3% vs 97.0%; P < .001).

However, the M-CHAT/F was effective at weeding out false negatives, with an overall specificity of 94.9% and an overall negative predictive value of 98.6%.

In a linked editorial, Lonnie Zwaigenbaum, MD, from the University of Alberta, Edmonton, Canada, and Jonathan Maguire, MD, from the University of Toronto, Ontario, Canada, note that results may not translate to smaller, less integrated healthcare systems that do not have a common electronic medical record system.

"It is difficult to know whether there might have been additional challenges if data collection had been attempted outside of a large urban centre," Zwaigenbaum told Medscape Medical News via email.

Also, children who initially screen positive for ASD on the M-CHAT/F are supposed to receive a second screening at 24 months, but in this study, only 41.2% of children who initially screened positive completed the second screening, Zwaigenbaum and Maguire note.

"One of the unique strengths of the M-CHAT/F is the 2-step assessment process, which is designed to maximize both sensitivity and specificity. Inconsistent administration of the follow-up interview undermined this strength," they write.

The authors suggest that providers may have gone straight to referral when they suspected ASD, rather than wait to conduct a second test. Alternatively, conducting a second screening may have been burdensome, according to So Hyun Kim, PhD, an assistant professor affiliated with the Center for Autism and the Developing Brain at Weill Cornell Medical College/New York–Presbyterian Hospital in New York City.

"Follow-up interview rates might be even lower in other settings with providers with limited resources and support," she said via email. "It will be important to identify potential burden for the follow-up interviews and develop strategies on when and how to proceed with the interviews to increase accuracy of the measure."

The study was funded by the Allerton Foundation, the Eagles Charitable Foundation, and the National Institute of Mental Health. The authors, the editorialists, and Pierce and Kim have disclosed no relevant financial relationships.

Pediatrics. Published online September 27, 2019. Abstract, Editorial

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: