Autism rate 1 in 100 in American Children?

5 Oct

A new study in the Journal Pediatrics claims a prevalence of Autism Spectrum Disorders of 110 per 10,000 (1.1%). This is notably higher than previous estimates based on the CDC’s Autism and Developmental Disabilities Monitoring (ADDM) Network.

There is little doubt in my mind that this will get a lot of media attention.

The paper is:

Prevalence of Parent-Reported Diagnosis of Autism Spectrum Disorder Among Children in the US, 2007


Michael D. Kogan, PhD,Stephen J. Blumberg, PhD, Laura A. Schieve, PhD, Coleen A. Boyle, PhD, James M. Perrin, MD, Reem M. Ghandour, DrPH, Gopal K. Singh, PhD, Bonnie B. Strickland, PhD, Edwin Trevathan, MD, MPH, and Peter C. van Dyck, MD, MPH

Here is the abstract:

OBJECTIVES: The reported increasing prevalence of autism spectrum disorder (ASD) and attendant health and family impact make monitoring of ASD prevalence a public health priority.

METHODS: The prevalence of parent-reported diagnosis of ASD among US children aged 3 to 17 years was estimated from the 2007 National Survey of Children’s Health (sample size: 78 037). A child was considered to have ASD if a parent/guardian reported that a doctor or other health care provider had ever said that the child had ASD and that the child currently had the condition. The point-prevalence for ASD was calculated for those children meeting both criteria. We examined sociodemographic factors associated with current ASD and with a past (but not current) ASD diagnosis. The health care experiences for children in both ASD groups were explored.

RESULTS: The weighted current ASD point-prevalence was 110 per 10,000. We estimate that 673,000 US children have ASD. Odds of having ASD were 4 times as large for boys than girls. Non-Hispanic (NH) black and multiracial children had lower odds of ASD than NH white children. Nearly 40% of those ever diagnosed with ASD did not currently have the condition; NH black children were more likely than NH white children to not have current ASD. Children in both ASD groups were less likely than children without ASD to receive care within a medical home.

CONCLUSIONS: The observed point-prevalence is higher than previous US estimates. More inclusive survey questions, increased population awareness, and improved screening and identification by providers may partly explain this finding.

The authors used the National Survey of Children’s Health. We have discussed the NSCH data previously on this blog.

The data are based on a phone survey, and use parent reports, not direct observations. For example, the researchers didn’t actually test any children, but they instead relied upon the phone survey answers to questions like:

For each condition, please tell me if a doctor or other health care provider ever told you that [your child] had the condition, even if he/she does not have thecondition now….Autism, Asperger disorder, pervasive developmental disorder, or other ASD?

The 1412 who answered yes to that question were asked:

Does [your child] currently have autism or ASD?

With 913 answering yes.

78,037 families were polled. This was a response rate of about 51% (assuming that all calls that were unanswered were residences and not busisnesses).

I have discussed raw data from the 2007 NCHS. By contrast, the researchers corrected the raw data for various factors that could cause systematic errors.

Estimates were weighted by using survey sampling weights available from the National Center for Health Statistics. These weights reflect the inverse of the probability of being selected for the survey and were adjusted to account for nonresponse and noncoverage of households without landline telephones. Weighted estimates are representative of the US noninstitutionalized population of children aged 3 to 17 years.

After correcting for these factors, the authors came up with a prevalence value of 110 per 10,000 children. (1.1%)

Parents were asked about the severity of their child’s ASD. Most answered mild. The breakdown is

Mild: 54 per 10,000
Moderate:37 per 10,000
Severe: 18 per 10,000

The age distribution wasn’t flat, but neither did it show the huge variation one would expect from special ed data, CDDS data or other sources often cited.

3-5 years old: 85
6-8 years old: 132
9-11 years old: 138
12-14 years old: 118
15-17 years old: 77

There is a higher prevalence amongst males than females, as expected. The prevalence amongst Hispanics (103/10,000) and non-Hispanic African Americans (61/10,000) are both lower than amongst whites (125/10,000).

Many of the results are shown in Table 1 from the study, shown below.

Table 1 of prevalence report using NSCH data

Table 1 of prevalence report using NSCH data

Aside from the overall prevalence number, the second hot topic from the survey is the large number of children whose parent answered yes to “have you ever been told that your child has an ASD” but who answered no to “does you child now have an ASD”.

People are referring to these children as having “lost” their diagnoses. They account for about 0.4% of the total survey.

The authors spend a significant amount of time discussing the possible reasons why some children may have “lost” a diagnosis. We have already discussed some of the reasons on this blog.

The bottom line is, there isn’t enough information in the survey to tell. We do know that most of the children who “lost” an autism diagnosis had some other disability.

It needs to be stressed–there is no confirmation that these children ever had an official diagnosis of autism. That isn’t just for the group that “lost” autism, but for the entire group.

This is a significant limitation of this study. I have stated before, I think this sort of survey is good for generating questions for future studies. For that purpose, these surveys are valuable. They are faster and cheaper than a full study. But it is far from the sort of data that should be used to claim a definitive prevalence.

I expect to see a lot of press coverage of the new “prevalence”. While this study is a good, and it is consistent with many studies that show a prevalence in the 1-1.5% range, the data just aren’t strong enough to make such definitive statements.

Edited to add:

The USA Today story on this study has a quote that I think sums it up quite well:

Pediatrician Susan Levy, founder and director of the Regional Autism Center at Children’s Hospital of Philadelphia, notes that Kogan’s study is based only on what parents said about their children, not information from doctors involved in their care.

“For what it is, it’s very well done, and I don’t mean that critically,” she said in an interview. “It is one method of getting a handle of how many people in the country have this disorder. Is it the best method? No, but it is one way of approaching it.”

24 Responses to “Autism rate 1 in 100 in American Children?”

  1. Sullivan October 5, 2009 at 05:13 #

    I would like to point out an

    David Kirby on AoA and Huffington Post, blogs from Friday and mid-day Sunday:

    “Washington loves to dump its bad news on a Friday afternoon”

    Dan Olmsted on the Age of Autism Blog:

    In a conference call, the government just confirmed that 1 in 100 American children has an autistic spectrum diagnosis. They’re still saying they don’t know why, or whether the increase is real. And HHS Secretary Sebelius showed where the government is still going wrong by talking about “more research in the genes underlying autism.” The rest was bureaucratic blah-blah-blah

    from the Edmonton Sun:

    The other government estimate has not been formally released yet. But because of the new published findings, officials from the Centers for Disease Control and Prevention decided to announce Friday during an embargoed press briefing that their preliminary findings also show about 1 in 100 children have the disorders.

    Yes. The conference call information was embargoed. That means that the reporters on the call weren’t supposed to disclose the information until 12:01 am Monday, the same time that the Pediatrics paper was published.

    I guess they missed that part of the phone call 😉

    • Sullivan October 5, 2009 at 05:30 #

      I meant to point out–

      Washington didn’t dump the news on Friday. They were prepping the news sources to be ready once the embargo lifted Monday Morning. The news broke today…for everyone who respected the embargo.

      I guess there is no embargo on misrepresenting the truth.

  2. Laurentius Rex October 5, 2009 at 07:39 #

    Worst kind of survey possible though, just asking for parents self reports, wheres the academic credibility in that, there needs to be some methodology, some checks.

    However I suspect the figures are probably still lower than will ultimately turn out to be the case when a proper adult survey is done as well.

    The USA is still seriously disabled by the condition of looking at Autism as a problem in children only, not helped at all by the focus of groups like Autism Speaks.

  3. Visitor October 5, 2009 at 10:27 #

    So Kirby, being the unethical scumbag that he is, not only breaks an embargo, plainly put in place so that reporters would have a chance to read the study, call their contacts for comment etc, and then write their stories, but he then goes on to suggest that this is some evidence of wrongdoing on the CDC’s part.

    The mindfuck of these people is beyond belief.

  4. K James October 5, 2009 at 10:52 #

    There are many medical health care centers in America and they are in operation in order to provide medication, therapy, counseling and relief to people suffering from various diseases

  5. mike stanton October 5, 2009 at 11:32 #

    I am surprised at AoA’s response. I would have thought this represented the gold standard in autism research for them – a telephonic survey of parents. And if you take it at face value it is quite good news: a third of children have lost their diagnosis. Most of the remaining children are mildly autistic and the smallest group are those with severe autism.

    Unfortunately you cannot take it at face value. This is an interesting survey. But, for all the reasons you point out, while it is useful for suggesting questions for further research, it lacks the methodological rigour that would be required to provide reliable answers to those questions.

  6. Dawn October 5, 2009 at 12:35 #

    Well, at least the USA Today story has some sense in it. As you pointed out long ago with the initial data, Sullivan, the Autism/ASD question is very poorly worded. Since I am a health-care professional, I could say to a parent, “I think your child has autism/ASD” and that parent could legitimately give a yes answer to the survey even though I know nothing about diagnosing autism! The question does not ask if the child was diagnosed by a provider who specializes in the developmental delays.

    As for children who “lost” their diagnosis, we can’t even say if they were correctly diagnosed.

  7. Dawn October 5, 2009 at 12:36 #

    Ooops! End italics fail…sorry!

    Hooray for the edit function.

  8. Joseph October 5, 2009 at 13:50 #

    The prevalence of parent-reported diagnosis of ASD among US children aged 3 to 17 years was estimated from the 2007 National Survey of Children’s Health (sample size: 78 037). A child was considered to have ASD if a parent/guardian reported that a doctor or other health care provider had ever said that the child had ASD and that the child currently had the condition.

    Wait. Isn’t this pretty much how the CDC counts autistic children?

    It’s not reliable. For one, a professional might have once had the view that a child is autistic. That doesn’t mean this is the label the child ended up being diagnosed with. The other question is asking for the current personal opinion of the parent. We all know of children whose parents claim to be “recovered,” but they don’t appear recovered.

    Additionally, this survey probably has a non-trivial rate of refusal, and it probably selects for people interested in health related matters. It sounds like they didn’t control for this.

    There’s also no reason to think this method isn’t still under-counting autistic children, despite awareness of autism in the US.

    Imagine if the same exact methodology were used to count autistic children in, say, Sri Lanka.

    Larry was saying the NHS study was done “on the cheap.” If that’s cheap, what is this?

    • Sullivan October 5, 2009 at 17:56 #

      Wait. Isn’t this pretty much how the CDC counts autistic children?

      Do you mean in the MMWR’s? In that case, no. The CDC counts (1 in 150, 1 in 166, etc) are based on a review of records, medical and educational. They don’t screen and test, though.

  9. Laurentius Rex October 5, 2009 at 17:22 #

    Almost goes against the fifth ammendment doesn’t it, self incriminating 🙂

    If I were constructing a survey it would be a damn sight more subtle in detecting what I wanted to know, because I would avoid asking such a direct question, at least the UK survey used a screening instrument of sorts.

    You don’t get very far by asking the subject what they were doing on the night of the offence “ooh er, I was watching TV in my mums house and I’ve got witnesses to prove it”

    No you ask a number of other questions to eliminate the obvious leaving the only possible deduction that proves it was a fair cop.

    Another way of looking at this survey’s sloppiness is to consider that it is like asking a bypasser at random if they have ever broken the law, the anser is going to be subjective unless one has defined the threshold of lawbreaking at something you could go to jail or, or something minor like dropping litter. So it is with the subjectivity of this survey as there is no rigour whatever in the definition of ASD, no chance of comparing like with like at all.

    The problem is that whoevever carried out this survey, whether they are competent or not in there particular field of expertise it is quite clear they have little knowlege of the basics of field research. This again is the social model at work. Doesn’t matter if the survey is poor, it will slip past the peer review because of the prominence of the names, or even because the peer reviewers are actually judging something they themselves don’t have the expertise in.

    I am plotting a new survey myself (with a difference focussing on clinicians) – can’t say too much until I get approval for the questionaire, but there may be a paper in it, when it is done 🙂

    What I aim to discover in it will not necessarily be transparent to the recipient hence less opportunity to jig the ansers to fit the expectations.

    Pediatricians are neither epidemiologists or social researchers and that should not be forgotten.

  10. Joseph October 5, 2009 at 18:10 #

    I meant like in this press report. This study is apparently just a formalization of what is usually referred to as the “CDC survey,” which is simply the NSCH.

    Are there any indications that the CDC understands the methodological problems with these surveys, which are subsequently used by the David Kirbys of this world to scare the public?

    I mean, they say “it is not entirely clear what (the) increase is due to,” but are they aware of just how unreliable the methodology is?

    • Sullivan October 5, 2009 at 18:38 #

      thanks Joseph.

      I knew I must be misunderstanding something there–your depth of knowledge in the autism epidemiology studies is impressive.

      They must know that there are various levels to the quality of the data/study. Do they really have the time to try to explain that in a short webpage or press release?

  11. guest October 5, 2009 at 21:34 #

    my son was diagnosed with PDD NOS and mild autism at the age 2 years 10 months. once by the federally funded program and once by pediatric neurologists. both times he was diagnosed based on the answers i gave, not by carefully testing and observing his behavior. today 2 years latter he almost lost all of his “autistic treats”. he was two times misdiagnosed, although he is extremely social and looks everyone straight in their eyes. any explanations?
    i also happen to know lot’s of other parents that have the same destiny as mine.

  12. Joseph October 5, 2009 at 22:09 #

    @guest: It’s not that unusual that age 2 diagnoses are “unstable.” That seems to be particularly true of PDD-NOS.

  13. Emily Wilkes October 6, 2009 at 01:35 #

    I agree that the method of collecting data for this study may not have been the most reliable. They used a survey and asked parents, but I think that checks should be done to make the study more academically sound. Having every one in a hundred children being diagnosed somewhere on the autistic spectrum does seem a little frightening, especially since no explanations have stuck yet. My theory is that the cause is so different for every individual that one explanation won’t work for all cases.

  14. guest October 6, 2009 at 15:29 #

    there is a trend right now in US that lot of children since 3 months of age have been diagnosed with autism or PDD NOS, and the label sometimes is carried on as the child ages. if more children are misdiagnosed there will be a higher recovery rate, as they were not autistic to begin with. i personally find this as unethical practice and unfair to the children and parents.

  15. Chris October 6, 2009 at 16:14 #

    Okay, I’ll bite: How do you diagnose a 3 month old infant with autism?

    I am hoping it is a typo and you really meant 3 years of age.

  16. Joseph October 6, 2009 at 20:29 #

    Having every one in a hundred children being diagnosed somewhere on the autistic spectrum does seem a little frightening

    It’s only frightening if you think it should’ve been a lot lower a figure. Of course, this is what people have been led to believe.

    If I were to say that 3% of the population score in the mental retardation range of IQ tests, is that frightening? I guess if you didn’t know that’s what’s statistically expected, and a stable rate, you might think it is. No one is panicking about that, either way.

  17. David N. Brown October 6, 2009 at 21:13 #

    Unfortunately, society seems quite willing to accept preposterous inflation of a “problem” where it reinforces existing prejudices. Witness the acceptance of the “yerkes-Army” study, which “showed” that the vast majority of eastern European immigrants were clinical morons.

  18. guest October 6, 2009 at 23:11 #

    there is a number increase since 1996 from 1-150 to 1-100. it is all related with the funding available. for example if a child is speech delayed at about 16 months to receive free service for speech therapy (weather or not is needed) it is usually diagnosed with PDD NOS as the funding is already available under this diagnosis. sad but 100% true.


  1. Tweets that mention Autism Blog - Autism rate 1 in 100 in American Children? « Left Brain/Right Brain -- - October 5, 2009

    […] This post was mentioned on Twitter by Jason Brown, autism_hub and ArieleImm. ArieleImm said: Autism Blog – Autism rate 1 in 100 in American Children? « Left … […]

  2. Autism prevalence > 1%? at EBD Blog - October 5, 2009

    […] the Huffington Post. Over on Left Brain/Right Brain Kevin Leitch (or one of his fellow bloggers) has a report, too; it includes good discussion of some of the methods of the […]

  3. Is autism just the new ADHD? - SmartPlanet - October 5, 2009

    […] A report in the journal Pediatrics says that over 1% of all children born in 1996 may have been diagnosed with autism. It is causing enormous controversy. (Photograph by the author.) […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: