These brief summary notes revisit a number of issues including the question of raised incidence or changes in diagnostic practices, the source of autistic risk, identification and co-morbidity, the range of needs, and parental anxieties.

A continuing theme is that of seeking to avoid a “one size fits all” approach, but to recognise the wide individual differences among children all legitimately described as having autism or autistic spectrum disorder, and working towards criteria to determine how best to select interventions for given cases.


M.J.Connor                                                                                                   April 2005                   



Incidence (increase in rate or change in diagnostic practice ?)


There now seems to be a commonly-shared view that, while the rates of autism may have shown some upward trend over the last few years, the major basis for the apparently considerable rise in the number of children diagnosed with autism is the shift in diagnostic criteria.  

The concept of an autistic spectrum has led to more children being described as autistic or as having an autistic spectrum disorder when, previously, reference might have been made to pervasive developmental disorder, or learning difficulty, or language delay or disability.  There is evidence that as the numbers of autistic diagnoses has risen, so there has been a corresponding decrease in the number of these competing diagnoses. 


One further study has recently been completed with a reinforcement of this general conclusion. 


Barbaresi et al (2005) begin their paper with a reference to the anxiety brought about by the apparent rise in autistic disorders. 

Their study involved a calculation of the incidence of autism in a specific population served by a clinic in an area of Minnesota where there was access to comprehensive medical and educational records.  The cohort included all residents of this area younger than 22 years in each year from 1976 to 1997.


They noted that there was a clear increase in autistic disorder rates at each three-year interval, while the incidence of other neuro-psychiatric and developmental disorders also rose but by a much smaller degree.

However, it was also found that the increase in autism coincided with the 1987 publication of the revised Diagnostic and Statistical Manual of Mental Disorders which included much broader criteria for autistic disorder, and with the implementation of a federal law ensuring access to specialist support or provision for children with autism.


An assessment was made of alternative and environmentally-based explanations for the increase in (diagnosed) autism, with the conclusion that they did not offer sufficient basis for the changes observed; and, therefore, that increased “case-finding” (as a result of criteria changes) was most likely to be the primary cause.


Autistic Risk


While still on the subject of aetiology and incidence, one might take note of a recent study which has highlighted another area of apparently heightened risk for autism.


Croen et al (2005) have produced survey evidence to indicate that women with asthma and/or allergies during pregnancy appear to be at a somewhat enhanced risk for having a child with an autistic spectrum disorder. 


The study carried out in California examined associations between asthma, allergies, and maternal autoimmune disorders in the 4 years encompassing pregnancy and the eventual diagnosis of ASD.  The sample included 407 diagnosed children, and 2095 control children.


Asthma and allergies were significantly more common among the mothers of the children with ASD, and the risk appeared greater among those with asthma or allergy recorded during the latter half of the pregnancy.  However, they did not find evidence to support the hypothesised link between maternal autoimmune problems and autism risk.


The authors stress that their evidence is to be regarded as an initial finding, with the need for ongoing and larger scale studies.



Clinical Review


A comprehensive review of the field is set out by Baird et al (2003) who also begin by referring to the increase in public anxiety about enhanced risks of autism, including the postulated link between MMR and autism and inflammatory bowel disease. 

The context of their work is that of such anxiety which is further reflected in the quest for, and possible over-optimism about, treatment options, albeit with no controlled studies offering evidence for effectiveness.


In respect of the nature of autism, the authors describe the condition as behaviourally-defined  but as the possible end point of a number of aetiological pathways.

These include the prenatal impact of rubella, metabolic disorders, the effects of anticonvulsants taken during pregnancy, localise lesions, and postnatal infections. 


However, a specific medical cause is identified in only a small number of individuals with autism (between 6 and 10 % of cases), often those with severe learning difficulties.  (See, for example, Fombonne 1999)


Meanwhile, epilepsy is more common among an autistic population than among the population at large, with possible implications for some specific neurological basis for the condition. 

Genetic factors appear significant.

The old hypothesis linking autistic characteristics with some inappropriate or inadequate parenting has long been discredited as these physical bases for autism have become more recognised, including the identification of those genes (on chromosomes 2, 7, 15, 16, and 19) which appear implicated in the aetiology of the condition … ie  autism is a neurological-physiological condition rather than psychological disorder.


Nevertheless, results from brain scans have not led to the identification of clear and consistent diagnostic markers …. although functional brain imaging has indicated a consistent anomaly in brain activation patterns in tasks involving the processing of visual information relating to face recognition.


Baird et al (op.cit) conclude, therefore, that despite the organic basis of autism, the diagnostic criteria are a matter of experience and consensus rather than of some unarguable organic criterion.  There is no biological test for autism.

Further, diagnostic cut-off points are not readily definable because of marked individual differences the nature, number, and severity of the signs and symptoms from one case to another. 

The symptoms will also show variation within a given individual, and changes in the presentation of the condition will occur over time given the developmental nature of autism.


The authors go on to suggest that, despite these uncertainties, a confident and reliable diagnosis of autism is usually possible at between 2 and 3 years of age, although a review of studies suggests that a diagnosis of a broader autistic spectrum disorder is less reliable at this young age than among older children.


Asperger Syndrome or higher functioning autism may only become evident when the child in question is admitted to a school setting where (limited) social and interactional skills are first recognised as significant.  Prior to this, the children may even have been thought relatively advanced because of some precocious use of vocabulary and their special interests. 


Meanwhile, screening for autism is not seen as reliable given the relative insensitivity of existing assessment instruments; and it is noted, anyway, that so-called screening tests for autism are usually administered when the child is already the subject of  concern because of some suspected or observed developmental problem.


Reference is made to the frequency of regression.  Around 25-30% of affected children show a halt or a regression in their development between 15 and 21 months of age. 

This may involve word loss along with social withdrawal, loss of eye contact or interest in play activities, and perhaps a change in eating and sleeping patterns.

These signs may be accompanied by an onset of mannerisms such as flapping or an over-focused attention upon certain objects.


However, in many cases, it is argued that there were (subtle) signs of problems in advance of this regression such that, for example, the words apparently lost may never have been used in a truly communicative way. 


A more significant regression – disintegrative developmental disorder - can occur in a small number of children (fewer than 10%) over the age of two years of age, and after language has developed beyond the level of just a few words. 

The outcome is similar to autism of early onset and is usually associated with severe learning difficulties.


There is no evidence that regression rates have increased over recent years.


When it comes to diagnosis, the authors emphasise that what makes autism special is the great variation in presentation, the wide range of skills and deficits, and the high rates of associated behavioural, emotional, and learning problems.


Common differential diagnoses include :


Associated and co-morbid difficulties frequently include one or more of the following:


In respect of intervention, reference is made to the wide range of treatments which have become available, such as Lovaas Applied Behaviour Analysis, PECS, Higashi treatment (intensive exercise programmes), SonRise, etc.. 

A problem with evaluation of the interventions available is that the therapists working in health or education settings may be under pressure to demonstrate effectiveness, but outcomes or rates of progress may not easily be measured.  In any event, the needs of the “whole child” have increasingly been recognised and personal or emotional growth is seen as no less significant than specific cognitive or educational advances.


The kind of targets that could be adopted with any intervention could be drawn from the following and hierarchical list :


The authors reiterate the problems associated with the range of types and levels of needs, and variability of functioning with implications for a similar range of specialist services, and educational placements which might cover the spectrum between mainstream schooling to 52 week per year specialist residential schooling.


Some of these themes were taken up at a recent conference where Jones (2005) held that planning an intervention is still based upon a kind of “best guess” …. there is still no criterion by which reliably to determine the approach to which a child will make the most positive response.  Meanwhile, it is argued that it is dangerous to place great faith in the efficacy of any single given package, especially if it is necessary to modify the environment and the way in which individuals approach the child or set expectations alongside the precise interventions used with the child him or herself.


Each child will represent a unique set of abilities and weaknesses and behaviours and anxieties, but the suggestion is that children of school age will all benefit from planning which focuses upon the major and common issues of communication and language, social and emotional understanding, friendships, flexibility of thought and behaviour, sensory issues, self esteem, strengths and individual interests, and life/leisure skills.


Jones, too, recognises the challenges for planning interventions which arise from the heterogeneity of the population, the shifts in diagnostic criteria, co-morbidities, and changing fashions in what is perceived as the most appropriate style of treatment.  

Evaluation may be confused by a lack of clarity of targets or by variation in the way an intervention is actually implemented, changes in the professionals involved, and idiosyncratic combinations of interventions or therapies. 

There is no clear research tradition, sample sizes tend to be small, and there may be debate about whose opinions or perspectives are most salient in determining the extent of progress or even the areas to monitor … teacher ?  parent ?  researcher ?  service provider ?


Jones’ advice has it that one size does not fit all, but there is a need for a range of options to offer families and the children.  Further, it is likely that the children will need a number of different elements within an intervention package; and a willingness to ensure an eclectic approach is the hallmark of those schools likely to be most successful.



Parental Perspectives


Finally, one notes the summary produced by Bovell (2005) of common anxieties expressed by parents over the understanding of, or provisions for, autistic disorders.


It is held that the anxieties can be related to three basic sources.


Firstly, the range of types and levels of difficulty among children with autistic spectrum disorders.  These will range from high functioning children with areas of specialised knowledge albeit still showing marked social and interactional problems, to children who have no language skills and appear to be experiencing severe learning disadvantages.  The symptoms will range from marked withdrawal, through endless and repetitious talk about some given topic, to aggressive and acting out behaviours including attempted or actual self-harm.

A common parental view has it that, despite this, LEAs offer a narrow range of educational provisions seeking to steer most children into one standard setting.


Secondly, there are large numbers of approaches available, such as ABA, TEACCH, SonRise, etc., which may appear attractive and which may be marketed energetically, but which lack reliable evidence for effectiveness; and there is no means of determining how best to match a child with an intervention. 

The resulting anxiety about how best to support the child is exacerbated by what is seen as competition and conflict among the different proponents, alongside the pressure faced by LEAs to ensure value for expenditure and to avoid high cost but low evaluation approaches. 


The third issue is about the number of children who have been diagnosed with ASD, which does not appear to have been matched, in the eyes of parents, by an increase in places in units or specialist centres even allowing for a growing proportion of children with autism achieving satisfactory inclusion in mainstream schools … albeit noting the variation among mainstream schools in the extent of positive inclusive practices and outcomes. 

There is also seen to be a need for a sustained and ongoing programme of training for teachers, in mainstream schools and specialist centres, about autism and ASD.


The implication again is for a range of provision to match the range of needs, and for individual planning of intervention and educational or social goals … with the concern lest working towards social interaction – even forcing social interaction upon the children – may be counterproductive and may lead to avoidance behaviour in at least some cases. 


                                    *          *          *          *          *          *



M.J.Connor                                                                                                   April 2005





Baird G., Cass H., and  Slonims V.  2003   Clinical review – diagnosis of autism.  BMJ  327  488-493


Barbaresi W., Katusic S., Colligan R., Weaver A., and Jacobsen S.  2005  

The incidence of autism in Olmsted County, Minnesota, 1976-1997 : results from a population-based study.   Archives of Pediatrics and Adolescent Medicine  159(1)  37-44


Bovell V.  2005   Reckless Abandon.   The Guardian (Schools’ Section, 8-9)  1.3.05


Croen L. et al  2005   Maternal asthma and allergies and the risk of autism.  Summary available in Reuters Health Information – original article in Archives of Pediatrics and Adolescent Medicine  159  151-157   


Fombonne E.  1999   The epidemiology of autism.   Psychological Medicine  29  769-786


Jones G.  2005   Recent advances in educational interventions.  Presentation to a national conference (“Recent Changes for Children with ASD”).  London, University College, April 18th. 



This article is reproduced by kind permission of the author.

© Mike Connor 2005.

Back to NAS Surrey Branch Welcome Page