Stability (and instability) has many different meanings depending on what area you are looking at. For the purposes of this post, I am talking about what happens to a diagnosis of autism/Asperger syndrome/ASD or PDD-NOS as a function of changes to presented symptoms; and in particular, the factors potentially affecting diagnostic stability such as age, co-morbidity, symptom severity and various interventions.
I use the words 'potentially affecting' because as with many things, it is very difficult (impossible?) to say definitively that one or other factor alone/combinatorially contributed to change something, bearing in mind the basis of science being probability not absolutes.
Stability with regards to autism diagnoses is a surprisingly unstable thing. Think about it: we diagnose on the basis of a prescribed pattern of symptoms being present and observable and occurring within a set chronological time frame. On the basis of such controlled but ultimately subjective judgements we specify 'autism', 'Asperger syndrome' or one of the other sub-diagnoses as being present as detailed in the DSM or ICD schedules.
OK, you're right, we do have various standardised schedules to aid diagnosis (ADI, ADOS, etc) but remember, these are only complementary to the diagnostic (or assessment) process and not confirmatory per se.
Final diagnosis is therefore dependent on factors such as who makes the diagnostic decision (their skill, experience and diligence), where and under what circumstances the diagnosis is made (home, school, clinic, all of the above), what kind of contributory assessment tools were used, and at what age the diagnosis is made (young or older age). Lots of different variables and lots of room for variability to occur.
As I posted in my previous entry on the proposed DSM revisions, there is nothing currently or in the planned revisions to the diagnostic manuals to say that an autism diagnosis is anything but a fixed feature; autism is after all a lifelong condition according to organisations such as the National Autistic Society.
Taking the issue of chronological age first, there are a few things worth noting. Autism research has, frankly become quite obsessed with early diagnosis. Lots of different studies have been conducted to look for the 'magical' set of behaviours and characteristics which would allow screening or assessment measures to universally identify the 'autistic child' at 3 months, 6 months or 12 months, etc of age.
Why you may ask? Well because there is a suggestion (and it is quite a strong suggestion) that the earlier that symptoms are identified and a diagnosis given, the earlier that intervention can be adopted in order to somehow influence the course of development. Given the proposed plasticity of organs such as the brain at these critical early developmental periods, the logic is that through play, speech and language and other therapies and interventions (mainstream and complementary), it may be possible to affect brain development, directly or peripherally, and hence potentially affect symptoms presentation.
There appears to be little wrong with this logic given the evidence available for its component parts: the brain for example during early infancy is a busy little bee adding (and pruning) various neural connections at a pretty spectacular rate over those early years. One side-effect from this whole order of developmental events is perhaps the conflict with the view of autism 'being a purely genetic condition'; whereby genes, and only genes, dictate the course of developmental events and because they are genes, they are somehow immovable.
The contrary logic of early diagnosis and early intervention presenting behaviour as plastic and malleable suggest pre-determination may not be 'the' key element of autism or at least 'some' autisms (perhaps a blog entry of its own on this topic is merited).
The reality is that we do not have a universally clear idea of early autism presentation (we need only look at studies on tools such as the CHAT and M-CHAT to know this) accepting also that age is a modifying variable on autism symptom presentation and diagnosis. Several studies have shown that there is quite a large degree of diagnostic instability as a function of age, particularly in the early months/years; most probably as a function of the rapid changes in physical and psychological development that occur during these formative years.
Moving on to co-morbidity. We know that autism can occur alongside several other linked / non-linked conditions, some of which may have a significant effect on how and what symptoms are presented. Learning disability (LD) is perhaps the most widely cited co-morbidity but things like epilepsy / seizure-type disorders have also been noted. Like autism, LD ebbs and flows in terms of presentation. Epilepsy also is not a static entity with a suggestion that it can impact on autistic symptom presentation (although the nature of the relationship still requires further study).
The severity of symptoms and the connection to which diagnosis is given has also been suggested to further contribute to diagnostic instability. Recent meta-analysis has suggested that greater variability in symptom presentation is evident amongst the ASD / PDD-NOS diagnosis when compared to classical autism. Does this mean that classical 'Kanner' autism is governed more by immovable genetics than environment when compared to other sub-diagnoses?
Finally we have the potential role of intervention on diagnostic stability. There are many different types of intervention or management strategies suggested for autism; educational, behavioural, pharmacotherapy, etc. all with varying degrees of evidence for efficacy (and importantly safety). Time and time again reports similarly emerge of children whose symptoms abate (disappear?) and what interventions might be contributory. I am not getting into the nitty-gritty of whether such reports are accurate or not (I am hardly in a position to question a parent's view of their own child). What such reports do suggest is that movement across the diagnostic autism spectrum is potentially possible (at least in some cases) and correlates with intervention (remembering of course that correlation does not imply causation).
So there we have it; lots of different factors pulling, yanking, tearing at a diagnosis. I will finish with a reiteration: stability with regards to autism diagnoses is a surprisingly unstable thing.