Monday, 9 May 2011

SES and 1 in 38 autism prevalence

As I have mentioned, there is quite a lot of interest in the various figures on the prevalence (and incidence) of autism. Not least because, compared to little over 20 years ago, many more people are diagnosed with autism in all its various forms. In the coming years (now?) there will also be a significant increase in the need for services (health, social, education) and the provision of appropriate opportunities for people with autism.
By my last count (last week) we were looking at 1 in 100 adults presenting with an autism spectrum condition in the UK, and 1 in 77 children in Utah with autism.
Two further publications have entered the fray in recent days. I will discuss them one at a time.
The first by King and Bearman suggested that, at least some of the increase, specifically in California, might be due to socioeconomic factors such as the availability of community-level resources (author words not mine).
What this translates as is, for example, by using income status derived from whether or not a person is in receipt of Medi-Cal (a sort of healthcare service for those on low income to my non-US audience) there is a degree of predictability as to whether or not an autism diagnosis is received.
Medi-Cal = decreased 'risk' of autism diagnosis (by anything between 20 - 40%).
The flip side: more income = more access to services = greater likelihood to receive an autism diagnosis.
This trend follows the so-called negative socio-economic status (SES) health gradient: the less money/resources you have, the less likely you are able to access appropriate services, professionals, new technologies, etc.
The authors also make an interesting suggestion from their data. In spite of the relationship between SES status and autism diagnosis (low SES = less likely to be diagnosed), the prevalence of autism diagnosed in families with high SES (lots of dosh, nice house, etc) has stagnated around the 0.4% prevalence rate and the increase in prevalence in autism is now being driven by those children from a lower SES. I have my Mr Spock raised eyebrow upon reading this.
The paper is comprehensive in its detail and analysis, and I have to admit I found it a little hard going at times. I can't however argue too much with the findings and conclusions given that the authors seemed to cover quite a few issues with the figures including some of the more well-known relationships with autism such as gender and low birth weight.
The authors do also accept a few limitations to their analysis, not least that it is built on California data and hence is applicable to California.
I would perhaps add speaking from the European side of the Atlantic, from a country where universal healthcare via our wonderful NHS is available at the point of need, free of charge, such a negative health gradient is perhaps not as 'steep' here.
Certainly here in the UK the normal procedure for the assessment referral 'pathway' would initially be either through a Health Visitor at the various developmental checks, the GP (doctor) or through the school or nursery. All of these 'options' are generally-speaking free at the point of use; so whether you are as wealthy as Alan Sugar or not so as Yosser Hughes, you would still get to use the resources if required.
Might this suggest that different local/regional/national factors are driving some of the prevalence increase according to different countries/regions?
Speaking of national and ethnic factors, a second paper has also cropped up on my radar.
This paper based on data from a South Korean community reports an overall estimated prevalence of autism spectrum disorders of 2.6% in 7-12 year old children, which has been calculated as 1 in 38 children in this cohort.
I don't have the full study details yet but this is quite a figure - the 2006 UK estimate of 1 in 100 children seems to pale in comparison. The text of the abstract provides some methodological information which seems to follow a similar 'try-and-find' study design to that of the 1 in 100 adults study.
It appears that the authors looked at two groups: both the general population and also an 'at-risk' group. The prevalence rates between the groups differed quite considerably (general population = 1.89% vs. at-risk = 0.75%) as did the sex ratios (2.5:1 vs. 5.1:1 males:females). Interesting also that intellectual disability was present in nearly 60% of the at-risk group, compared with 16% in the general population group. This might suggest that there are quite a lot more people with Asperger syndrome or high-functioning autism than first thought.
I will need a little more time to digest these findings from this new prevalence paper and so perhaps expect follow-up posts to this at some point. Suffice to say that the epidemiology of autism continues to generate great interest.