1 in 68: An Interview with CHOP Experts on the CDC’s New Autism Rate

The Centers for Disease Control and Prevention (CDC) recently announced a new prevalence estimate for Autism Spectrum Disorder: 1 in 68 individuals. This is 29 percent higher than the estimate released in 2012 (1 in 88), 64 percent higher than the estimate from 2009 (1 in 110), and a 124 percent increase from the report from 2007 (1 in 150).

Philip R. Johnson, MD, director of The Children’s Hospital of Philadelphia Research Institute, recently interviewed Center for Autism Research experts to get their take on the new CDC numbers.  

Dr. Johnson: What could account for such a dramatic rise from the initial reports?

Judith Miller, PhD: There are a few reasons that we can hypothesize that can account for this rise. For one, with each surveillance study, the participating sites get better and better at gathering records from their clinical and educational sources. Secondly, autism spectrum disorder (ASD) awareness has increased dramatically over the last 15 years, so with each year increasing numbers of parents, schools, and clinicians are considering whether a child’s difficulties might be due to autism. Third, the true prevalence may be rising due to real changes in the factors that cause autism.

Dr. Johnson: How are these studies conducted? 

Susan E. Levy, MD, MPH: These studies rely on record reviews rather than in-person evaluations. Records of any child who meets the age criteria, who lives in the specified geographic region, and who has come to the attention of a healthcare or special education agency are included. An anonymous file is compiled of that child’s comprehensive health and education evaluations, which is then reviewed by clinicians using a standard rating procedure.

The studies looking at prevalence of developmental disorders such as ASD are carried out in multiple locations in the United States, funded by U.S. Department of Health and Human Services through the CDC. Sites are awarded the opportunity to participate based on their ability to access the necessary records and conduct the necessary work, and each site works closely with their local healthcare and education agencies to collect the data.  

Dr. Johnson: Do these numbers seem to fit with clinicians’ and researchers’ sense of the prevalence?

Dr. Miller: It’s always better to have data than rely on anyone’s impression. I think this is the record review method is a probably the best one possible. However, record review will always have limitations, including the lack of gold standard clinical evaluations. The best approach would be to do a real epidemiological study, basically door to door with expert clinical evaluations but that is so expensive as to probably be impractical in this funding environment.

The CDC, however, did do a small study with in-person assessments, and it showed good correlation to the record-based study. But that was many years ago. I am hopeful the CDC can receive funds needed to continue doing validation studies. With increased awareness in the community, and recent changes to diagnostic criteria it’s important to understand how the diagnosis is being made in the community.

Marsha Gerdes, PhD: It’s probably most important to take a step back from the numbers.  Even if the diagnosis is not 100 percent accurate in all cases, these are children who are coming to the attention of their schools, parents, or healthcare providers. They are struggling in some way with social communication and behaviors. We as a community need to come together around how best to meet their individual needs and to support their families and educators.

Robert Schultz, PhD:  Given the CDC’s methods, there will clearly be errors of diagnosis, including misdiagnosing some children who really don’t have ASD. However, that does not mean that all or nearly all of the children identified don’t have a condition that causes them to struggle much more than a typical child.

It is also important to recognize that diagnostic categories such as ASD mask considerable heterogeneity. It is often said, “If you have seen one child with autism, you have seen one child.”  More recent research is showing that ASD presents on a continuum with typical development, but for certain skills and traits, just on the lower end of some continuous dimension. The point here is that there might not be sharp boundaries and a real category of autism.

The World Health Organization (WHO) and other organizations will report the “true” prevalence of brain based mental disorders is about 20 percent, and these often start at birth or prenatally. I would say 1.5 percent of children who present with ASD on record review very likely have something. The “essence” of what they have however is proving a bit more elusive than we use to understand, and many times cases do not fit neatly in current diagnostic groups. But that does not mean that these kids are not struggling and in need of interventions.

Dr. Gerdes: Let’s put the 1.5 percent in the context of other developmental disorders. The CDC’s report, “Trends in the prevalence of Developmental Disabilities in U.S. children, 1997-2008,” reported learning disabilities at 7.7 percent and ADHD at 6.7 percent, an increase of 33 percent. Many of the same factors of increasing awareness and accessing services may have played a role in increases of other developmental disabilities.  

Dr. Johnson: OK, I get all that. But then, how does one define “normal?” If 20 percent of people have brain based mental disorders, then it’s no longer “atypical.”

Dr. Schultz: Yes, it is a value judgment as to where you draw the line and say this kind of problem necessitates special attention. The values most commonly invoked are economic and quality of life. The WHO discusses the impact of mental disorders in terms of both dollars and DALYS (disability-adjusted life years). Where you draw the line affects who gets treatment and ultimately can save both dollars and DALYS.  

Dr. Johnson: Sounds reasonable. But still. Let’s say 1 in 68 children have “something.” Did they always have something?

Dr. Miller: At the Center for Autism Research, we see more and more children being referred for mild forms of autism (the group that saw the largest increase in the CDC study). We find that 20-25 percent of children who have been diagnosed with mild forms of ASD in the community don’t meet our gold standard criteria for an autism spectrum disorder. These children are struggling, but their clinical presentation is complex, and doesn’t always fit neatly under ASD.

Dr. Gerdes: Services funded by the community (e.g., Early Intervention, Special Education, Mental Health, etc.) are often, but not always, tied to both a diagnosis and a level of impairment.  This can make it difficult to help those individuals with milder symptoms or a less common presentation of symptoms. Obtaining services for those children and adults can be challenging.

Dr. Schultz: The field has really begun to move away from studying categories of mental disorders, as they don’t map onto to the brain or to the genome in a clean fashion. Tom Insel, MD, Director of the National Institute of Mental Health, has lead the way here, and is pushing for grants to propose studies that cut across diagnostic boundaries, that focus on processes rather than diagnostic categories (e.g., processes such as sociability — which can be linked to reward circuits in the brain, or executive functioning — which can be linked to functions of the frontal lobes). Deficient executive functioning and sociability are features shared by many different disorders. Ultimately profiles of functioning across these and other dimensions will likely prove both more accurate and more helpful in guiding treatment.

Dr. Johnson: Given the efforts to encourage early screening, were you surprised that the average age of diagnosis remained the same?

Dr. Levy: Not at all. The AAP’s recommendations for early screening came out in 2007. The CDC’s data is on children born in 2002 (turned 8 years old in 2010). Pediatricians have been working hard on early identification, but still confront barriers of the “perfect” tool for a busy primary care office and access to appropriate resources for treatment in the community. But early identification (by age alone) is not the only hurdle. Access to care by underserved groups of children is still a major problem. White children were more likely to be identified with autism than black or Hispanic children. High-risk groups such as former premature infants, children with acquired brain injury, genetic disorders and others need to be identified and screened or evaluated more consistently. Environmental exposures and factors need to be investigated carefully, in the context of well-designed studies as potential causes of ASD.

That said, 1 in 68 children probably do have something that deserves our clinical and research attention. Thus, the CDC work always cuts both ways — it is good to have the public’s attention on the welfare of these kids. We agree that more attention can be focused on earlier identification, and also on appropriate evidence based services that can reduce symptoms and the impact of symptoms on functioning at home and school.

To learn more about autism spectrum disorders and ASD research at The Children’s Hospital of Philadelphia, see the Center for Autism Research website.