Thursday, 29 March 2012

CDC and the autism numbers game

Static @ Paul Whiteley
The waiting is over. The speculating is done. They're here...

Three years after the last update in 2009*, the US Centers for Disease Control and Prevention (CDC) have revealed their latest estimate for US childhood autism prevalence.

The 2009 estimate, based on surveillance in 2006, suggested that 1 in 110 US children (roughly equating to 1%) aged 8 years old - so born in 1998 - has an autism spectrum disorder. The caveat to that data was that the estimates fluctuated to some degree across the 11 States included in the analysis ranging from 12.1 per 1000 in the States of Arizona and Missouri to 4.2 per 1000 in the sunshine State of Florida. The overall figures nevertheless represented a 57% increase in the numbers of diagnosed cases compared with data from 2002** which previously estimated a rate of about 1 in 150 US children.

The paper trail of recent autism prevalence studies by the CDC using the same DSM-IV - DSM-IV TR criteria makes for some nice standardised methods for screening and case ascertainment which are not to be sniffed at.

So what is the latest estimate?

1 in 88.
[Update: the full report can be found here***]


Childhood autism in the United States has officially broken through the 1% prevalence threshold.

Are these latest figures a great surprise? Probably not to anyone who has followed the numbers over the years. Some people will remember the South Korea (SK) study from last year (2011) which suggested that 1 in 38 children (2.64% cumulative) may present with an autism spectrum condition. The data from Utah, USA, however, and their suggestion of 1 in 77 children with autism, seemed not to have received the same level of media interest despite being closer to the recent CDC mark. Utah also having the benefit of a good history on estimating autism prevalence down the years.

So, where next?

Based on the new estimate, I assume a familiar story will be put out to account for the change (better awareness, diagnostic substitution, more willingness to diagnose, widening of diagnostic criteria, etc) which whilst perfectly acceptable explanations still might not provide the full picture; not at least according to the Nature article from last year (2011). I have to say that I am always struck by how sure some people are that they know what is driving the increase in cases of autism as per one of the commentators on this recent newspaper article on autism going up here in the UK.

Autism on the rise as a consequence of better awareness? Well we do have World Autism Awareness Day on April 2nd and every year it seems to get bigger and bigger in terms of events and participation. Personally I'd like to think that the over the years of the DSM-IV and ICD-10, most professionals, whether teachers, health visitors, psychologists / psychiatrists, SALTs, know what to look for when it comes to autism in the early years. I think also that we should perhaps not underestimate parents and their observations in the early days of autism manifesting itself as important to this process. I could perhaps argue therefore that the awareness issue is probably to some extent covered in terms of the expertise and the various screening programmes in operation around the globe for autism. I'm not saying that awareness might not be contributory to the increasing prevalence but find it very difficult to say this is the sole 'cause' of the increase.

Diagnostic substitution? Widening of the criteria? Again, possible reasons, although probably more important reasons about 10-15 years ago following the replacement of DSM-III for example. Nowadays I'm not as sure that they still carry the same sort of weight.

So what are we left with? I've talked quite a bit on this blog about the rise and rise of epigenetics with autism in mind; that is genes and environment interacting together. Environment covers a lot of ground and I don't really want to start speculating on what facets of environment may or may not be linked. I do however think that we need to start shifting research attention away from the purely 'autism is genetic' model which has persisted for quite a few years. Autism may well have genetic underpinnings, which variably act on different people to different extents, but I don't seriously think that anyone would entertain all autism as being solely genetic. Indeed, epigenetics is beginning to turn quite a few areas on their research head and, in coming years, I am sure we will hear much more about it.

One final point to make: if the changes being suggested to DSM-5 (and I assume ICD-11 for that matter) for autism are actually implemented, we could be looking at a totally new way of coding for autism in subsequent CDC and other reports which will break those years of standardised measure based on DSM-IV. Will these changes affect the numbers of cases, and if so, in what direction? Mmm, a tough one.

* Prevalence of autism spectrum disorders - Autism and Developmental Disabilities Monitoring Network, United States, 2006. Autism and Developmental Disabilities Monitoring Network
MMWR Surveill Summ. 2009; 58: 1-20.

** Prevalence of autism spectrum disorders - Autism and Developmental Disabilities Monitoring Network, 14 sites, United States, 2002. Autism and Developmental Disabilities Monitoring Network
MMWR Surveill Summ. 2009; 58: 1-20.

*** Prevalence of autism spectrum disorders - Autism and Developmental Disabilities Monitoring Network, 14 sites, United States 2008. Autism and Developmental Disabilities Monitoring Network
MMWR Surveill Summ. 2012; 61: 1-19