Friday, 2 September 2011

Long-term outcome and autism

This is one of my longer posts so please bear with me. I will try and approach this topic with a greater degree of caution and sensitivity than usual because it potentially ties into some quite strongly held views and beliefs, particularly on the nature of autism spectrum conditions.

It is one of the primary questions to be asked when applying an intervention or management strategy to any condition, autism or otherwise: what are the long-term outcomes? By long-term 'positive' outcome I mean whether or not strategy A, drug B or diet C, etc will actually provide either long-term remission from the symptoms of a condition or long-term improvement in the presentation of a condition or long-term improved quality of life for a person or decreased risk of developing other things. At the very least, will employing an intervention 'do no harm' in the longer run to that individual or group?

I was brought to this post by a few studies fairly recently published. This one from right here in Blighty suggested that early intervention in autism was perhaps a longer term thing than 'a quick fix'. The study was based on quite a small group of children with autism and controls (N=41) and looked at follow-up after a couple of combinations of early intensive behavioural interventions (parent-led vs. supervised) vs. business as usual (see paper here). The results suggested that the early benefits from intervention noted after two years were rather more mixed - diluted - at follow-up after intervention had halted.

The second study is this one based on what happens to scores on one of the 'gold-standard' assessment measures for autism, the ADI-R. The main conclusion being that stability of the ADI-R is OK in certain areas, although in others it is perhaps a little shakier than expected. I have talked about stability and instability in autism diagnosis in a previous post so am not going into the detail again.

The final study is this one looking at long-term differences in outcome for autism vs. autism spectrum disorder. The findings suggested that although there was little difference in things like mortality, convictions for crime and marital status, a greater association with impaired psychosocial functioning (CGAS) predisposed the autism group to an increased receipt of social benefit (disability pension) than the more able PDD-NOS group studied. The study included a familiar name among the authorship team, Eili Sponheim, who has previously published results in a research area close to my heart.

From initial reading, what these studies, at least the first two, suggest is that 'change' can happen in autism in the early years. The questions are:

  • Is that change because of intervention? 
  • Can that change be sustained in the longer term? 

I will state at this point that there are quite a few accounts around of people using one intervention or another (or sometimes more than one) and observing, in some cases, some quite marked improvements in various areas either core or peripheral to autistic symptoms. My focus however in this post is on the published data of (mostly) group effects in the longer term (bearing in mind the notion of the n=1).

The question of long-term outcome following intervention in autism has perhaps not yet been adequately addressed by all areas of research. Research has for example started to look at some of the longer term implications of treating co-morbidities such as epilepsy using anti-epileptic medication but the focus has tended to be on what happens to the epilepsy rather than what happens to autistic symptoms and other related variables, acknowledging this recent study.

There are some studies that have looked at the implications of interventions on presenting autistic features. This study reported a 3-year follow-up of a toilet training schedule delivered to 2 children with autism. The use of LEGO therapy over a similar time period is discussed here. In my own field of interest looking at dietary intervention and autism, I am taken back to the early work of Ann-Mari Knivsberg and colleagues, who when they originally looked at the use of a gluten- and casein-free diet, went back after 4 years to see what had happened to their original cohort in terms of the gains made (our most recent trial of diet had a 2-year experimental period). There is therefore at least some evidence for some interventions moderating autistic symptoms (core and/or peripheral) during the longer-term period measured between 2-4 years.

But what about beyond the short(!) long-term? What about from childhood into adulthood into old age? Well, this is where the research dries up a bit. I found a few studies looking over extended periods, such as this one from Wendy Stone and colleagues, which provides some support for the early intervention model affecting language skills in contrast to the earlier paper discussed. The late Ivar Lovaas also produced some initial evidence of longer-term effects of his brainchild ABA on some cases of autism. But this is not a unanimous view by any means, as witnessed by this paper which paints a far less rosy picture. The caveat to this final paper is that it contrasts with this paper from the same author.

Just writing this entry I realise that outcome is a very complicated area to discuss with potentially many confounding variables. Aside from notions of the n=1, the often quite important areas of co-morbidity and the important influence of societal factors, some of the main issues to consider (as I see it) are these:

  • How do you measure outcome? I recently posted an entry about the need for reliable instruments to measure 'change' in autism. Until there is some consensus opinion on one or perhaps two more 'universal' assessment tools available to all, we will be constantly plagued by problems of inconsistency in looking at the outcome of autism as a consequence of the different measures used.  
  • What do you measure for outcome? Supposing you have a person who couldn't read or write properly but then landed a job as the new centre-forward for Manchester United with a salary to match. Would you say that this is a good outcome? Likewise you have a person who would most probably be destined to be the next Einstein in terms of their mental skills and prowess but ends up long-term unemployed either through economic, social or personal circumstances. What kind of outcome? I defined positive outcome at the start of this post as a reduction in the more debilitating aspects of a condition but you can see that there are other definitions in much the same way that 'success' and 'failure' are often subjective terms which might describe outcome.

I know this has been a little bit more of a philosophical post than my usual musings and I am mindful that there are people on opposites sides of the fence on whether or not intervention related to outcome are useful or necessary to autism. Long-term outcome is nevertheless an important area of inquiry which deserves much greater examination, and follow-up study.

To end and completely unrelated to the blog post in question, Sammy Davis Jnr at his very best...