A recently published study proposed and evaluated a new autism screening checklist for the 1-year well baby check-up. I had a hard time getting my hands on the study (crazy but true–even pediatricians sometimes are boxed out of research studies), but read lots about it first in the press. I was excited about the potential for early screening. The checklist, designed to be administered by a pediatrician, intends to improve early detection and diagnosis of autism. The goal: to enable early intervention and treatment for at-risk children. It’s clear that early intervention improves autistic children’s outcome with autism spectrum disorders but the media may have provided false information and false hope. So before you expect this checklist at your baby’s next 1-year check up, let me explain what the study found, my concerns about the results, and how the checklist may become useful.

I take care in using any screening tool–any intervention can cause unnecessary harm. My main concern resides around false positives (when a test suggests there is a problem when in reality there isn’t) and the numbers from this study…

Background on Autism Screening

Since 2007, the AAP has recommended that pediatricians screen for autism formally using a parent questionnaire (usually a tool called the M-CHAT) with a follow-up discussion at 18 months and 24 months of age. Many of us were using this before 2007. I screen both at 18 and 24 months in my office routinely, but we also screen babies at each well-baby visit for development or communication challenges prior to and after those visits. Every pediatrician does this, even when you don’t notice it. We watch your baby from the minute we crack the door. A pediatrician’s exam is not just when we’re standing at the exam table with your child. How and what your child says and does while we talk is an essential part of the exam. Unfortunately, there is no perfect tool to pick up autism in infants, although some studies have found utility in even single milestones like asking if a 9 month-old baby responds to their name. Yet, no single question or single milestone can screen or identify autism.

The study in the Journal of Pediatrics outlines the use of the new proposed checklist with 137 pediatricians in San Diego.

  • 137 pediatricians used the checklist to screen 10,479 1 year-olds. A total of 1,318 12 month-olds failed the screen (12.5%). But only 346 of those 1,318 children were referred to a developmental researcher to be included in the study. This left 972 children with a positive screen result from the checklist, not referred and not included in the study. Various reasons are given as to why these children were not referred–from improper scoring or pediatrician’s opinion that child didn’t need further evaluation. But the far majority of children with a positive screen weren’t referred and studied. This makes me concerned about the tool itself. Already, the group referred for follow-up may represent a different group, negating the power of the screening checklist.
  • Children who failed the checklist and who were referred to the developmental group were followed by a very expert research group every 6 months until they were 3 years of age (184 enrolled and completed the intervention). At that time a final outcome of meeting criteria for autism was made or they were categorized as meeting diagnosis for language delay, developmental delay, or “false positive” screen.
  • 20% of the children (of the 346) who failed the checklist at 1 year of age were ultimately diagnosed with autism spectrum disorder. 55% of the children who failed the screening test were ultimately diagnosed with language or developmental delay, and 25% were false positives–meaning their test showed concern for autism, but at age 3 they didn’t meet criteria for developmental delay or autism.

It’s the numbers that concern me. The intent of the study is fantastic: to evaluate a new checklist for autism diagnosis at age 1. But, I didn’t read a report that showcased my concern. The checklist “Found suspected autism, autism spectrum disorder, language delays, or other developmental problems about 75% of the time” reported the author.

The hype and worry we would cause in administering this screening test concerns me. False positives can cause true harm–suffering for parents and excessive worry. One in four times the checklist is administered, the checklist will be wrong. In a summary I read from WebMD, the lead researcher stated, “The price to pay for that is very tiny” when compared to intervention afforded. As a general pediatrician I disagree. Correct me here if you think I’m off.

As I’ve been thinking about this research the last few weeks, it’s the 25% false positives that continues to nag and concern me–those families that were told at that their 12 month old may have autism and who ultimately didn’t. How did those 2 years unfold? How uneasy were these parents and families made to feel? Twenty five percent false positive is extremely high for a screening exam.

How do those numbers make you feel?

However, another and more hopeful explanation exists, of course. Is it possible that some of those children who screened positive at 12 months of age avoided the challenges in communication that led to a diagnosis of autism because of the early intervention and close followup. It’s hard to know if they were a false alarm or really, a perfect result of intervention—avoidance of a diagnosis because of early intervention. This could be extraordinary, but we need to know more with follow up research and study.

As a parent and pediatrician, I not inclined to use this test without more instruction and explanation. Although the researchers report that pediatricians who used the survey evaluated the program positively and that all practices still use the screening tool today. So I may be missing something. My further hesitation resides in the numbers–that over 900 times the test was positive, the far majority of children weren’t referred. So something about pediatrician judgment and physical exam still provided additional and relevant screening. And it’s possible this is why the practicing docs like it so much. Maybe the checklist functions as a starting place.

Not An Autism Specific Screen

The lead author asserts this is not an autism specific screen although that’s what the media reported. It’s a screening for autism, language delay, developmental delay, and autism spectrum disorders. Therefore, the title of the article may be misleading. Although many media reports touted this as the new diagnostic manual for autism, this may be more widely applied as a screening checklist. Further, the researchers in this study didn’t test the specificity or sensitivity of the screening checklist. This will have to come with future research. And what to do with results? As the researchers assert, “The utility of a a screening program can only fully be realized in effective treatment options exist for test positive cases.” Many areas of the country don’t have organized specialized centers for helping support children and their family when facing a concern or diagnosis of autism.

Bottom Line: There’s No Universal Autism Screening Checklist for 1 Year-Olds. Yet.

There is no perfect and universal screening test for autism at 12 months of age, but this checklist may provide insight and hope for developing checklists and protocols for pediatricians earlier than 18 to 24 months. And more, this data may hint that intervention helped children avoid a diagnosis. Time and more study will tell.

What do you think? Would you complete the parent checklist knowing the data I shared? Am I wrong; should I be using this checklist?