NADD Bulletin Volume XI Number 1 Article 2

Complete listing

Precision And Unfairness Of The Pdd-Autism Diagnosis

Jacques Thivierge M.D.

In this article, I argue against the general tendency of clinicians and researchers to behave as if the diagnosis of PDD-autism were an informative and precise medical diagnosis, entailing some specific kind of treatment. In my opinion, this has unfortunate consequences, including; (a) Fostering and maintaining an unfair situation for many children with a developmental disability; (b) Causing unnecessary losses of time and energy for clinicians and unnecessary difficulties for many families; and (c)  Maintaining artificial standards for publication and research grants in the field.

The medical evaluation of children with Developmental Disabilities (DD) remains uninformative in the great majority of the cases. These children do need a thorough, individualized evaluation of the major aspects of their developmental abilities (motor, communicative, social, and cognitive). These evaluations are often being performed by professionals who are not doctors, followed by properly adapted educational interventions, equally performed by professionals who are not doctors. The aim of this paper is to discuss the practical importance of the medical diagnosis and to explore the consequences of the medical diagnosis.

The children with Developmental Disorders (DD) that we see as a child psychiatrists have Pervasive Developmental Disorders (PDD), Mixed Language Disorders (MLD), and Mental Retardation (MR). The “understanding” of what goes on in a child once a diagnosis of PDD, MLD, or MR has been retained has no known neuropathophysiological foundation in the great majority of the cases. Our “understanding” essentially remains at a descriptive level. The same is found with the treatment orientations based on these diagnostic categories: no specific treatment orientation whatsoever is attached to any of them. The general line of intervention or “treatment” according to the vocabulary of our times is the same for the three conditions, namely, an early educational approach of optimal intensity addressed to the particular developmental disability profile of every child.


From a clinical point of view, there is no such a thing as a precise diagnosis of PDD-Autism (PDD-A), as does exist with a precise diagnosis of diabetes mellitus or lung cancer. To say it is as important to make the precise diagnosis in the former as in the latter case is, in my view, incorrect, abusive and misleading. Let us remind ourselves that in medicine, the diagnostic act is a process through which one aims at understanding the medical nature of the problem and/or defining a specific treatment. The more information a diagnostic label gives us in this respect, the more it can be described as precise, informative, and rich in content.

 One will find no exaggeration in the idea that  we live in a world in which we are assaulted by images of precision and objectivity regarding the psychiatric diseases, namely PDD-A, from researchers and from the parental associations. We are led to believe that the children having received “a precise diagnosis of PDD-A” will respond in “a specific way” given therapeutic interventions like secretine or intensive behavioral intervention.  This situation has greatly contributed to establishing an image of science for this diagnostic category. As a result of this situation, it is understood that the diagnostic category of PDD-A affords more services, opens many more doors to public help, and favors a more rapid access to behavioral educational services than is the case for the other DD (i.e., MLD and MR). From this perspective, this diagnostic category is a definite asset. There is no problem in such situations for the doctor taking responsibility for the diagnosis if there are valid medical reasons to justify it. But the fact is that those valid medical reasons do not exist at this point for PDD-A.  I suggest that every child with MLD and MR could benefit from the intensive behavioral educational approach (the major treatment recommendation for PDD-A). This approach is as necessary to them as it is to the children having a diagnosis of PDD-A. Offering this approach to the PDD-A, while denying to children with MLD and MR, on the ground of diagnostic precision and treatment specificity, turns out to be an unfair situation with which clinicians are daily confronted.

My  argument, that the PDD-A diagnosis is not very precise and does not open the door to consequent specific treatment interventions, will appear controversial to many readers. I will present five arguments to sustain it.


1. The PDD-A diagnostic category is a clinically variable one


Everyone of us whose profession it is to care for children presenting with DD could cite numerous cases where, as time goes by, a former PDD-A diagnosis has converted itself into another more appropriate diagnostic category for a given child; or cases in which there is disagreement among professionals at given times about which diagnostic label should be applied. This is of relatively frequent occurrence now that the children are being referred at an ever earlier age to our clinics. The meaning of the term PDD-A has evolved considerably with time and keeps on changing (Grandin,1995). It is essentially a description based on behavioral references, and is therefore an unstable entity by itself subjected to a variety of sociopolitical pressure points. The former have been numerous and somewhat turbulent in the past (e.g., vitamin or Secretine) and still are, given the amount of money (for services and research) attached to this diagnostic condition.  This social background is most important for the clinician to be aware of because it does influence us all, whether I want to believe it or not. Our first point is then that from an ongoing clinical experience, the PDD-A diagnosis is a socially influenced and unstable condition.


2. The so-called modern and objective methods of making the diagnosis do not make a difference in the previous state of affairs.


Relative to the first point, one could argue that the modern methods of making the diagnosis (for instance, the Autism Diagnostic Interview –ADIS-, or the Autism Diagnostic Observation Schedule-ADOS-) have put some order in the original instability of this situation, so that we now have at our disposal reliable and valid methods of making the PDD-A diagnosis. But this is not true. The point is that these methods have in no way modified the basic imprecision underlying the clinical approach of the problem (i.e., the behavioral reference itself). However, it was fair to believe that structuring the approach could have made it clinically reliable. We all kept hope that structure and training about these very behavioral references would lead us to unravel stable relationships between our diagnostic categories and some significant biological factor, or some specific treatment approach. But as of now, nothing of this sort has really happened, and certainly not for the PDD-A category. Curiously enough though, we keep on behaving as if these methods were more usefully precise than the clinical approach itself. The clinical approach shows a great advantage over the often too narrowly channeled modern diagnostic approach to the clinical problems, namely in actively considering the signs and symptoms within their different contexts of evolution. The modern instruments are quickly out of their own depth in this respect, and also the judgments based exclusively upon them. Modern psychiatry has unfortunately almost forgotten the word “context” along the way, especially now that the children are being referred at an early age. The clinical approach constitutes a definite asset over an essentially categorical approach to diagnosis.

No specific relationship has ever been described between any given operational definition of PDD-A and whatever biological factor (be it anatomical, genetic, or biochemical) or specific form of treatment. Despite the modern research effort to standardize the diagnostic procedure, Wing’s (1997) judgment is still valid about the continuing lack of any independent biological or psychological marker besetting our attempts to produce a reliable and valid sub grouping.


3. The ambiguous status of the PDD-A diagnostic category in research


Nobody will ever object to a researcher trying to obtain as homogeneous as possible a group of children with DD in order to further our understanding of these children. The problem as we see it for PDD-A currently resides in the ongoing confusion between “precision per se” and “useful precision,” that is a precision with specific problem solving capabilities. Establishing these problem solving capabilities could be looked upon as the ultimate aim of any research endeavor in medicine. I personally think that we are actually being hypnotized into believing that “precision per se” should be understood as “useful precision,” which, of course, is totally untrue. Even though the modern PDD-A diagnostic criteria have no useful established medical precision as such, everybody tends to behave as if they had. For instance, training sessions are being organized around them for clinicians; they are also being used to control publications in scientific journals, and are used for allocating research funds as well. This general situation stemming from the welcome efforts at organizing research in this area strangely appears out of tune with what one sees going on in some research areas.

      Take for example the recent epidemiological study concerning 289,456 children aged 3 to 10 years of age, in the Atlanta area (Yeargin-Allsopp et al., 2003). On the basis of charts obtained from multiple educational and clinical sources, the cases have been retained following a review of their diagnosis by a panel of four persons described as experienced in making the PDD-A diagnosis. The important idea here is that one accepts a great disparity of methods and persons in the observation of the children, and above all, one accepts evaluations made by unknown other persons without any experienced diagnostician seeing any of the children. This is very unsatisfactory from a clinical point of view and, in my view, equally questionable for research purposes.

However, Yeargin-Alsopp’s paper must be put in its proper context as a study of the Centers for Disease Control in the USA; that is, an administrative context. Within this context the study is acceptable as such. However, I argue against the use of these administrative data for answering fundamental medical research questions about a disease that, contrary to many other diseases (e.g., cancer, infections, or heart failures), lacks any precise biological definition.  For instance, according to our clinical experience as well as to what I see occurring in other communities (CDDS, 1999), it would be reasonable to think that many a child termed PDD-A in Yeargin-Allsopp’s study could be better described with the terms MR or MLD. I believe one cannot seriously draw any meaningful conclusion from such data about ways to conduct epidemiological studies in this population except for administrative purposes; if one is to use epidemiological data in relation to a medical question one would obviously ask for more coherence in the allocation of the cases.

I personally think that the value of our PDD-A diagnosis in psychiatry is mostly on the sociopolitical side, an idea I always explain to the parents so as to give them the clearest picture of what I see as the actual meaning of that “diagnostic” category. Such a contradictory attitude between the alleged importance of a precise diagnosis of PDD-A in the clinical setting and the use of such varied and changing diagnostic methods for epidemiological research purposes sounds as if there is a law for the rich and a law for the poor. This speaks in favor of the precision of the PDD-A diagnosis as being somewhat of a myth.

 There is a definite pressure in the actual clinical settings to use the research-based criteria (ADIS, ADOS) to make a precise diagnosis of autism. This is inconsistent with the fact that researchers satisfy themselves with very loose diagnostic procedures in large-scale studies.(Yeargin & Allsopp , 2003). It also is inconsistent with the lack of any specific treatment recommendation attached to the “precise” diagnostic category.


4. The association of PDD-A to numerous other pathologies


I see no valid clinical reason to exclude epilepsy and sensory impairments when looking at the association of PDD-A to other medical conditions (Frombonne, 1999). We find a significant association to several such conditions (Gillberg, 2005). My fourth point: Talking PDD-A is talking about the imprecision of a syndrome rather than the precision of a disease.


5. And finally, an argument stemming from logic


It is established fact that in the course of recent years we have observed a significant increase in the number of persons identified as having PDD-A. This increase could be the result of a new way to identify a disease (which could be very interesting from a clinical and/or scientific viewpoint) or to a wider use of an old concept (which could be less interesting from a scientific viewpoint, but could be interesting from other perspectives). We necessarily have to accept that group P of patients (PDD-A for example) identified through the new method (for instance ADIS or ADOS) are being identified out of a larger group P’ of patients still sharing with patients of group P some common characteristics (developmental disabilities for instance). However, the P distinguish themselves from the P’ through special characteristics identified by applying the new diagnostic method.

A clear example of this would be the identification of a group of patients suffering from malaria (P) out of a group of patients (P’) presenting with fever whatever the cause of it. The identification of the Plasmodium in the blood of some members of group (P’) reveals itself as an important biological marker allowing their being properly separated to form a new group (P), so as to treat them specifically with quinine, a treatment of no use to the other members of the original (P’) or (P’-P). This gives a clear example of the use of a medical diagnosis being precise, informative, and rich in content, namely, helping us understand the nature of a medical problem so as to apply a specific treatment. This is the ultimate aim of any diagnostic endeavor in our profession.

In the cited case of malaria, one can describe the function of the diagnostic method (identification of the infectious agent) as being the one of excluding from the future group P to be every individual of group P’ whose fever can neither be understood nor treated on the basis of the presence of plasmodium in their blood. This process is perfectly in accordance with the concept of precision in the general area of human knowledge.  Let us recall that the Latin term “praecisus” means separated from, cut from, abrupt, or steep.

Going back to PDD-A, promoting a more precise diagnosis should mean having discovered a mean to form a PDD-A group (P) by excluding  from the DD group (P’) every child for whom the more precise diagnostic method cannot help us understand or treat better his/her developmental disability. I actually observe an unprecedented and phenomenal increase of the number of PDD-A cases. Group (P) is expanding itself to the profit of group (P’). Outside of an epidemic phenomenon, this is a rather disquieting occurrence as to the consistency of the terms in use. As a case in point, this very phenomenon has led some authors to talk about an autistic epidemic (CDDS 1999). It is as if the new diagnostic procedure brings understanding and/or specific treatment advantages to an ever-increasing number of children with DD. But the plain fact is that, from a medical viewpoint, this “new precise diagnostic procedure” brings no such significant understanding or specific treatment advantage whatsoever to the (P) children when compared to the (P’-P) children.

We are thus led to believe that the actual PDD-A epidemic is no consequence of a new discovered diagnostic procedure allowing us to identify more people from (P’) whose medical problem we understand better and treat more efficiently. Rather, it should be looked upon as the consequence of a conceptual widening. Such phenomenon would obviously not occur without some sort of underlying advantage.  Where could it be? We think that, capitalizing on the idea of medical diagnosis, it carries an idea of better understanding and one of specific treatment, ideas with great marketing power for buying a number of things, namely, appropriate services, those services not being medical but psychosocial. I think that this is one of the factors behind this modern conceptual widening of the term PDD-A. These educational approaches are not specific approaches for PDD-A children, but approaches that should be made available for every child presenting with a developmental disability.

A diagnosis is informative insofar as it excludes. In the field of PDD-A, we currently see the opposite, namely an ever-greater inclusion of individuals from the departing group outside an epidemic phenomenon. We personally interpret this as a sign of the ever-decreasing informative character of what we call PDD-A in our society; medically, PDD-A is an uninformative category with poor content.


So why do we give such importance to the diagnosis?


We think that part of the answer to the question lies in the market value of the idea of diagnostic precision. The snobbery of our time (and there is not much I can do about it) has culturally installed itself on the side of the scientific and objective look, which should be demarcated from an appropriate scientific and objective attitude. Each and every one of us is bathing in this water, that is, we are naturally being inclined to give a high mark of virtue and/or value to anything bearing a nice suit of precision and objectivity often without having the idea to look beneath. This image does marvels for one to obtain votes, subventions and services. This constitutes an obvious sociopolitical advantage. This has to be considered as a potential important factor in explaining the modern increased prevalence of the condition (CDDS, 1999) and the decreased prevalence of other DD conditions (Croen et al., 2002).

We think it essential for all of us at this point to realize that the quest for precision should never become an aim to pursue for itself; the scientific area is no different from the other aspects of human knowledge. The principle should be that the level of precision sought for should not exceed the level of precision necessary to the solution of the problem at hand. For the clinician, this means that the level of precision of our concepts should depend upon the clinical problems to be solved. The quest for precision for its own sake is as useless as the quest for certainty. Karl Popper’s commentaries (2002) appear most appropriate here. Too much importance being given to definitions inevitably leads us to substitute verbal problems to the ones of the real world. The consequences are regrettable losses of time, money, and energy.  Should I give an example? All this energy lost trying to figure out is autism is a disease with growing prevalence in the world; when one has in mind the great imprecision of the measures from the start, the drifting effects of current sociopolitical forces on the category, the applying of precise p-value measures to answer such a question appears disheartening to us. The question applies well to the PDD-A administrative phenomenon, but not to the PDD-A medical aspects.



 First, we consider as an unchallenged belief the one that PDD-A is a precise medical diagnosis entailing specific treatments. My opinion is that this belief should seriously be challenged at the present time. Second, uncritically behaving as if the PDD-A category was a precise and medically informative entity leads us (by way of unintended consequence) to unfairly make it more difficult for some children with DD (MR and MLD) to access an equivalent quality level of rehabilitation services. Third, in clinical practice this leads to significant loss of time fighting with the rehabilitation services providers so that the MR and MLD children get their fair share of educational treatment, and to useless difficulties to families. Fourth, in research this contributes to unfortunately maintain artificial standards for publication and research grants.

As a possible solution, why not decide that all the children with DD (PDD-A, MR, LMD) have a right to equal services of psychosocial stimulation? The appropriate assessment of every child will necessarily specify his/her particular difficulties namely mostly of the side of language (MLD), or socialization (PDD-A), or rather homogeneous difficulties (MR) without these particularities having an incidence on the optimal quality of educational services every child should receive. Not only will such a position put an end to the actual unfairness regarding educational services for the children with developmental disabilities, but this will spare us all, including the families, many unnecessary time-consuming activities and also help liberate the field of autism (clinically and research wise) of an unfortunate constraint.



California Department of Developmental Services. (1999). Autism Epidemic Reaches Six Per Day. Scaramento: California DDS.

Croen, L. J. (2002). The Changing Prevalence of Autism in Califoarnia. The Journal of Autism and Developmental Disorders, (32) 207-215.

Frommbonne, E. (2003) Epidemiology of Pervasive Developmental Disorders, Trends in Evidenced Based Neuropsychiatry, 5(1) 29-36

Gillberg, C. (2001) Autisme: Troubles Medicaux Associes. In A. Berthoz, A. Christain, Barthelemy, C., Massion, J. & R. Bernadette (Eds.) L’Autisme, De la Recherche a la Pratique,, Odile Jacob, Paris: 27-49

Grandin, T. (2005) Thinking in Pictures. Doubleday, NY

Popper, K.. (2002) Unended Quest. Routledge Press, London

Wing, L. (1997) Syndromes of Autism and Atypical Development. In D. Cohen & F. Volkmar (Eds.) Handbook of Autism and PDD. John Wiley & Sons, NY

Yeargen-Allsop, M., Rice, C., Karapurkar,T., Doerneberg, N., Boyle, C. & Murphy, C. (2003). Prevalence of Autism in a US Metropolitan Area, Journal of the American Medical Association, 289(1): 49-55.