Early Autism and Developmental Disabilities Monitoring says autism prevalence is still rising...

"The overall ASD [autism spectrum disorder] prevalence was 13.4 per 1,000 children aged 4 years in 2010, 15.3 in 2012, and 17.0 in 2014 for Early ADDM [Autism and Developmental Disabilities Monitoring] sites with data for the specific years."

So said the surveillance summary published by Deborah Christensen and colleagues [1]. Those of you who follow the US ADDM initiative (see here) will already know about the aims of this "group of programs funded by the CDC" looking at the (estimated) autism numbers, changes to the numbers and the impact of the numbers on various communities. Through initiatives like the ADDM, we already know that the estimated prevalence of autism in 8-year olds living in the United States is round about 1 in 59 (see here) and that the estimate continues to grow for pretty much every surveillance year examined. We are also starting to find out about how the change to DSM-5 from DSM-IV is likely to/not to impact on future figures (see here). And hopefully, at some point, we might have some further data on what happens to autism past childhood (see here) from such an initiative.

The Christensen paper adds another tier to the knowledge being acquired as per their analysis of the (estimated) prevalence rate of autism in 4 year olds "whose parents or guardians lived within designated sites." Those sites were: Arizona, Colorado, Missouri, New Jersey, North Carolina, Utah, and Wisconsin. The Early ADDM initiative does not cover the same area as its big brother/sister ADDM but "is conducted in two phases using the same methods and project staff members as the ADDM Network." Those phases include first "reviewing and abstracting data from children’s records, including comprehensive evaluations performed by community professionals" and then a second phase involving "a review of the abstracted evaluations by trained clinicians using a standardized case definition and method." DSM-IV criteria covers most of the time points examined but: "For 2014 only, prevalence estimates based on surveillance case definitions according to DSM-IV-TR and the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) were compared."

Results: as per the opening sentence of this post, the overall (estimated) autism prevalence across all sites for autism in 4 year olds was on the increase between 2010 and 2014. That's not to say that there weren't differences between the various sites - New Jersey, a favourite autism prevalence site (see here) showed the highest prevalence: 19.7, 22.1, and 28.4 per 1,000 for 2010, 2012 and 2014 respectively - but overall the prevalence rate was increasing not decreasing.

There were a number of other important points raised in the Christensen findings. So: "Among four sites with ≥60% data on cognitive test scores (Arizona, New Jersey, North Carolina, and Utah), the frequency of co-occurring intellectual disabilities was significantly higher among children aged 4 years than among those aged 8 years for each site in each surveillance year except Arizona in 2010." I don't think I need to say much more about that. Also: "The overall prevalence estimate using a DSM-IV-TR case definition was approximately 20% higher than the prevalence estimate based on DSM-5 criteria." Again, I don't think too much more discussion is needed on this point aside from saying that for 4-year olds, the switch to DSM-5 might have made more of a difference than for 8-year olds. Indeed in comparison to the Wiggins data [2] based on 8-year olds where "46.0% children met both DSM-IV-TR and DSM-5 surveillance status, 44.0% met neither the DSM-IV-TR nor DSM-5 surveillance status, 4.0% met DSM-IV-TR status, but not DSM-5 status, and 6.0% met DSM-5 status, but not DSM-IV-TR status of ASD" the Christensen data showed something a little different: "Among 1,237 children who met the surveillance case definition for either DSM-IV-TR or DSM-5, 974 (78.7%) met both case definitions, 234 (18.9%) met the DSM-IV-TR but not the DSM-5 case definition, and 29 (2.3%) met the DSM-5 but not the DSM-IV-TR case definition." Perhaps more study is required on the diagnostic changes?

What's more to say? Well, one of the authors - Walter Zahorodny - kinda said it all in a media comment: "There’s no letup. I really don’t understand why the rate is going up in this way." So maybe the next question, a question that really should have been examined a long, long time ago, needs to be 'Why?' rather than just a continual chain of studies saying autism prevalence is increasing...

----------

[1] Christensen DL. et al. Prevalence and Characteristics of Autism Spectrum Disorder Among Children Aged 4 Years - Early Autism and Developmental Disabilities Monitoring Network, Seven Sites, United States, 2010, 2012, and 2014. MMWR Surveill Summ. 2019 Apr 12;68(2):1-19.

[2] Wiggins L. et al. Comparison of autism spectrum disorder surveillance status based on two different diagnostic schemes: Findings from the Metropolitan Atlanta Developmental Disabilities Surveillance Program, 2012. PLoS ONE. 2018; 13(11): e0208079.

----------

DSM-5 autism is "largely comparable to" to DSM-IV-TR autism

"Merry New Year". Welcome back to Questioning Answers in 2019. So let's continue.

The quote heading this post comes from the findings reported by Lisa Wiggins and colleagues [1] and their research aiming to "replicate agreement between surveillance status based on DSM-IV-TR criteria and DSM-5 criteria for ASD [autism spectrum disorder]." This authorship group have some interest in the description of autism across the DSMs (Diagnostic and Statistical Manual of Mental Disorders) as a function of their involvement with the United States CDC and their other, quite recent publication, talking about the national US autism prevalence estimates in 8-year olds for 2014 (see here).

As well as arriving at the conclusion that 1 in 59 children aged 8-year old in the United States *might* have autism [2] (still potentially an underestimate according to other recent figures), their latest prevalence paper also, for the first time, introduced the DSM-5 to the Autism and Developmental Disabilities Monitoring (ADDM) Network. Previously, the DSM-IV was top (diagnostic) dog, but with the publication of the DSM-5 (see here) there was a new sheriff in town. And as the DSM-5 started to be utilised more and more widely in diagnostic circles, so the CDC autism estimating teams had to start taking this into account in their musings...

Wiggins et al relied on data for "8-year-old children who had health and education records reviewed for ASD surveillance in metropolitan Atlanta, GA in the 2012 surveillance year as a part of the Metropolitan Atlanta Developmental Disabilities Surveillance Program (MADDSP)." As per the aims and objectives of the MADDSP - "estimates the number of children with selected developmental disabilities in metropolitan Atlanta" - autism is one of the diagnoses included in the surveillance program. DSM-IV-TR criteria for autism were applied to the health and education records of a cohort of children to see if diagnostic thresholds were met. Importantly too we are told that: "a previous ASD diagnosis was not sufficient evidence to confirm DSM-IV-TR surveillance case status." Alongside: "A DSM-5 coding scheme for ASD surveillance was developed by an independent body of ADDM-affiliated experts and then adapted and refined by a CDC-led clinical workgroup." Application of this DSM-5 coding scheme was similarly used on those education and health records data and results were collected and compared.

"Similar to ADDM data reported for the 2014 surveillance period, we found substantial agreement between DSM-IV-TR and DSM-5 surveillance status of ASD in a sample of records reviewed for the 2012 surveillance period." The figures in a little more detail: "46.0% children met both DSM-IV-TR and DSM-5 surveillance status, 44.0% met neither the DSM-IV-TR nor DSM-5 surveillance status, 4.0% met DSM-IV-TR status, but not DSM-5 status, and 6.0% met DSM-5 status, but not DSM-IV-TR status of ASD." The importance of this data is what the quote heading this post talked about: DSM-5 autism is "largely comparable to" to DSM-IV-TR autism. So when the ADDM eventually makes the transition over to DSM-5 derived autism data only, it means that all those years of DSM-IV diagnosed autism are still very much relevant (and can be used as a comparator) in a historical context.

But... the authors do mention about how some other studies have observed that: "fewer children meet DSM-5 criteria for ASD than DSM-IV-TR criteria for ASD in similar service settings" (see here and see here). This requires quite a bit more investigation; in particular how "record-review surveillance [such as that utilised by Wiggins et al] relies on information contained in surveillance records and does not include an in-person evaluation of the child." As I've indicated in my very critical musings on armchair diagnosis of historical persons (see here for an extreme example), there is absolutely no substitute for talking to and directly observing a person who has been referred for a diagnostic assessment. None. I've also previously suggested that another related label that crops up near to the DSM-5 description of autism - social (pragmatic) communication disorder (SCD) - might also need some research attention too. Indeed, I wonder if the ADDM might eventually start reporting on autism and SCD separately...

----------

[1] Wiggins L. et al. Comparison of autism spectrum disorder surveillance status based on two different diagnostic schemes: Findings from the Metropolitan Atlanta Developmental Disabilities Surveillance Program, 2012. PLoS ONE. 2018; 13(11): e0208079.

[2] Baio J. et al. Prevalence of Autism Spectrum Disorder Among Children Aged 8 Years — Autism and Developmental Disabilities Monitoring Network, 11 Sites, United States, 2014. Morbidity and Mortality Weekly Report (MMWR). 2018; 67(6): 1-23.

----------

Estimated one in 69 children aged 8 years old with autism (not one in 68) in 2012

"On June 5, 2018, the authors informed MMWR [Morbidity & Mortality Weekly Report] about a number of inadvertent errors throughout the report that resulted from reporting of autism spectrum disorder cases among persons who did not live in the geographic surveillance area."

So said a brief note [1] recently listed in the CDC Morbidity & Mortality Weekly Report (MMWR) concerning an important article published in 2016 [2] that described the estimated autism prevalence rate in the United States in 2012 for children aged 8 years old.

The original paper by Deborah Christensen and colleagues [2] was fodder for this blog at the time of publication (see here), with their estimated figures for 2012 (one in 68) showing a potential plateauing of the autism estimated prevalence rate in the US (see here for some discussion on the previous figures for 2010). The 'plateau' proved to be short-lived; as more recent figures published this year (2018) for the surveillance year 2014 once again showed the continuation of the upward trend (see here) in the childhood autism prevalence rate, now up to an estimated 1 in 59 children. Other figures have suggested even 1 in 59 is likely an understatement (see here).

Christensen and colleagues have republished their 2016 paper [3] showing 'where they went wrong'. It's not a wildly different article from their original publication and to a large extent, does not alter the underlying figures in any hugely significant way: 'one in 68' is replaced by 'one in 69'. Looking at the tables accompanying the Christensen republication, I was first drawn to Table 2 showing the: "Estimated prevalence* of autism spectrum disorder [ASD] among 1,000 children aged 8 years, by sex —Autism and Developmental Disabilities Monitoring Network, 11 Sites, United States, 2012." Under the column labelled "Total no. with ASD" it looks like a few participating States had corrections. Arkansas in particular stood out; their total going from 170 with ASD down to 125. Their estimated autism prevalence subsequently dropped from 12 per 1,000 children (aged 8) to 8.8 per 1,000. Other States showed a correction in the opposite direction. New Jersey (which has had an important role to play in the CDC estimates) showed a slight increase (of 3 children) in their "Total no. with ASD" similar also to Missouri.

There's little more to say about the Christensen correction aside from reiterating that: (a) the CDC statistics citing figures like 'one in 69' or more recently 'one in 59' are estimates, and (b) how and what data you include for counting is going to have an important bearing on what (estimated) prevalence rate you arrive at. By saying all that, I've not changed my view that we are witnessing something of at least a partial 'real' increase in cases of autism (see here and see here and see here) as older 'better awareness' and 'diagnostic substitution' arguments become less and less relevant as the numbers (estimated) climb ever higher. And aside from keeping on asking 'why?' the powers-that-be should be putting a lot more money and resources into the services that will inevitably be required, to ensure that children and adults on the autism spectrum aren't (societal) disadvantaged by their diagnosis.

-----------

[1] No authors listed. Correction and Republication: Prevalence and Characteristics of Autism Spectrum Disorder Among Children Aged 8 Years - Autism and Developmental Disabilities Monitoring Network, 11 Sites, United States, 2012. MMWR Morb Mortal Wkly Rep. 2018 Nov 16;67(45):1279.

[2] Christensen DL. et al. Prevalence and Characteristics of Autism Spectrum Disorder Among Children Aged 8 Years--Autism and Developmental Disabilities Monitoring Network, 11 Sites, United States, 2012. MMWR Surveill Summ. 2016 Apr 1;65(3):1-23.

[3] Christensen DL. et al. PPrevalence and characteristics of autism spectrum disorder among children aged 8 years — Autism and Developmental Disabilities Monitoring Network, 11 sites, United States, 2012. MMWR Surveill Summ. 2018 Nov 16;65(13):1-23.

----------

Autism prevalence in California: 1931 to 2014

Credit: Nevison et al, 2018

'Autism prevalence' is a topic of real discussion both inside and outside of autism research circles. It's not so much a debate about whether there has been an increase in the numbers of people being diagnosed with autism or autism spectrum disorder (ASD) - there has and continues to be (see here and see here) - but rather the extent of the increase and the possible reason(s) behind such an increase.

The paper published by Cynthia Nevison and colleagues [1] adds something further to the autism prevalence discussions. They focused on how much of an increase in cases of autism has been noted in one part of the United States (US) and how some notable birth years have been seemingly driving the increase.

OK, first things first, Nevison et al are no strangers to the [peer-reviewed] debates around the autism numbers (see here and see here). Their previous analyses have concluded that the increase in cases of autism in the US is NOT solely driven by changing diagnostic criteria or diagnostic switching/substitution, although these factors have probably played some role. Yes folks, there might have been a very real increase in the numbers of people being diagnosed with autism (see here).

On this most recent research occasion, Nevison and co-author Mark Blaxill are joined by a noteworthy figure in the field of autism prevalence tracking, Walter Zahorodny. Zahorodny is intricately involved in tracking autism prevalence in New Jersey (see here) and was also listed as an author on the most recent CDC stats regarding estimated autism prevalence that were quite quietly published a few months back [2]. In short, he's an expert in such matters.

Authors started with the California Department of Developmental Services (CDDS) data on autism which "provides services to eligible individuals living in California who meet the DSM diagnostic criteria for autism." In the age of the (very) wide autism spectrum, the CDDS has been pretty reserved in terms of who gets the services it offers. The term 'code 1 autism' has been used to refer to CDDS recipients eligibility, where autism or autistic disorder was typically the diagnosis listed and importantly: "individuals applying for CDDS services must demonstrate significant functional disability in 3 out of 7 life challenges, which include self-care, language, learning, mobility, self-direction, capacity for independent living and economic self-sufficiency." We are told that: "Milder subtypes such as Asperger’s syndrome and PDD-NOS [pervasive developmental disorder - not otherwise specified] have not been eligible for services unless they have another qualifying disability" (authors words not mine). With the onset of the DSM-5 description of autism, where individual diagnoses like PDD-NOS and Asperger syndrome have been 'rolled' into one definition (autism spectrum disorder), the CDDS has had to adapt and change. The years 2016 and 2017 have been particularly important for the DSM-5 change to the CDDS.

Added to their examination of the CDDS data, authors also looked at a couple of other initiatives with autism prevalence data: The Individuals with Disabilities Education Act (IDEA) and the Autism and Developmental Disabilities Monitoring (ADDM) Network. As I've already mentioned, Zahorodny has been quite 'active' in ADDM research circles for some time.

From all this [estimated] prevalence of autism data, authors set about to 'visualise' autism prevalence data extracted from the CDDS and other information sources and further examine whether "ASD is truly a constant prevalence condition" taking into account the various methods for collecting data on autism prevalence.

Credit: Nevison et al. 2018

Results: "The data are consistent across methods in showing a strong upward trend over time." The long-term trend of autism prevalence based on the CDDS data show "an apparent ~ 1000-fold increase in CDDS autism prevalence between birth year 1931, when prevalence was only ~ 0.001%, and birth year 2012, when prevalence had increased to 1.18% among 5 year-olds born in that year." Nevison et al talk about "age-resolved snapshot and constant-age tracking method[s]" as tools in their research, but the long-and-short of it is that the only way was up when it came to autism prevalence across the years. They further note: "The increase from ~ 0.001 to 1.18% in the 2012 birth cohort has occurred gradually, with a slow upward creep starting as far back as the 1940s, but with several change points along the way, around ~ 1980, ~ 1990, and ~ 2007, when the rate of growth accelerated."

I don't think there is too much more to say about the Nevison findings aside from reiterating that autism prevalence seems to have quite dramatically changed over the last 80 years or so. Insofar as those upticks in diagnoses around birth years 1980, 1990 and 2007, I'd be interested to know a little more about what factors might have been contributory to those particular years' growth or whether they are just statistical blips.

And before you say it, no, it's probably not just better awareness and the like...

----------

[1] Nevison C. et al. California Autism Prevalence Trends from 1931 to 2014 and Comparison to National ASD Data from IDEA and ADDM. J Autism Dev Disord. 2018. July 5.

[2] Baio J. et al. Prevalence of Autism Spectrum Disorder Among Children Aged 8 Years — Autism and Developmental Disabilities Monitoring Network, 11 Sites, United States, 2014. Morbidity and Mortality Weekly Report (MMWR). 2018; 67(6): 1-23.

----------