Editor's Note: This is a reprint of Dr. Carey's remarks to participants at the NIH conference on ADHD, November, 1998.
William B. Carey, M.D.


Some colleagues have suggested that presenting a dissenting view on ADHD at this conference would be a bit like the situation of Daniel in the lion's den. The image does not quite fit; I am here of my own free will and I do not expect divine intervention. I prefer the example of David going out to slay the giant Philistine, Goliath, not the androgenous David of Donatello but the pensive, mighty David of Michelangelo. However, before the day is done I may look more like St. Sebastian. But remember that Sebastian survived his multiple perforations to confront the emperor once again with his heresies. And who remembers the emperor, Diocletian?

ADHD is defined as consisting of 6/9 inattention or 6/9 hyperactivity/impulsivity symptoms for 6 or more months, which have been present from before the age of 7 years, with impairment in 2 or more settings, and not due to other conditions. Additional common assumptions about ADHD include that it: is clearly distinguishable from normal; constitutes a neurodevelopmental disability; is relatively uninfluenced by the environment; and can be adequately diagnosed by brief questionnaires. All of these assumptions and some others must be challenged because of the weakness of empirical support and the strength of contrary evidence. (The 60 references supporting this presentation are listed in the 30 page full version.)

There does seem to be a general agreement on the existence of a small group of readily recognizable "hyperkinetic" children, about 1-2% of the population, with pervasive high activity and inattention. For them the activity and inattention are so extreme that they are the clinical disorder by themselves, not just the risk factors. But even for this group it is generally not clear whether the symptoms come from abnormal brains or adverse environments.

This discussion here today describes the problems in the diagnostic terminology of ADHD as it is currently applied to the other 5-15% of American children.

I) ADHD symptoms not clearly distinguishable from normal temperament variations.

The literature of ADHD defines the inattention and high activity behaviors as abnormal and easily differentiated from normal temperamental variations, using "cutpoints" in numbers of symptoms.

But- these same behaviors are also described as normal yet potentially aversive temperament risk factors, and many children who have them are free of dysfunction. Half of any population is more active and half less attentive than average. Temperament research informs us that dysfunction results neither from the numbers of risk factors present nor from extreme ratings of them but when any number of them induces a "poor fit" with the particular environment. Children with the "difficult" temperament cluster (low adaptability, negative mood, etc.) are more likely to develop social behavior problems, and those with the "low task orientation" cluster (high activity, low persistence-attention span, high distractibility) more likely to do poorly in academic achievement, but often they do not. No solid data support the current "cutpoints," where normal high activity and inattentiveness leave off and abnormal amounts begin.

II) Absence of clear evidence that ADHD symptoms related to brain malfunction.

The ADHD behaviors are assumed to be largely or entirely due to a neurodevelopmental disability. The DSM-IV does not say so, but textbooks and journals do. Some preliminary brain imaging studies have shown inconsistent differences in children with the ADHD diagnosis, but there is no proof that they are deviations. We do know that various brain insults like lead poisoning, fetal alcohol syndrome, traumatic brain injury, and low birth weight may lead to increased activity and decreased attention span.

But- several lines of evidence oppose this supposed link for ADHD: 1) No consistent pattern of high activity or inattention is seen in children with established brain injury; 2) No consistent pathological changes or structural, functional, or chemical neurological marker is found with the current ADHD diagnosis despite diligent searches with sophisticated techniques; 3) On the other hand, differences in brain function have been demonstrated in healthy children with normal temperamental variations (e.g. frontal electroencephalogram differences associated with temperament differences). Therefore, proof is needed that any test differences demonstrated with the ADHD diagnosis are signs of a disorder and not just of a temperamental predisposition. Evidence of a genetic basis for the current diagnosis of ADHD cannot be taken as proof of brain abnormality because normal temperamental variations and normal coping also reveal substantial genetic contributions.

III) Neglect of role of environment and interactions with it as factors in etiology.

The DSM-IV criteria for ADHD make no requirement to consider the environment in the etiology of the disorder. Since the problem is considered to be all in the child, the varying contributions of the setting are typically ignored.

But- there are indications that the environment can produce or at least worsen the ADHD symptoms, as it does for other problems in adjustment. Something else is needed beside the behavioral predisposition to cause a disorder; e.g. family problems with difficult temperament to produce behavior problems; family problems or inappropriate teaching (or other factors in child like learning differences) with high activity and low attention span to result in academic underachievement.

IV) Diagnostic questionnaires now in use highly subjective and impressionistic.

Current practice involves the widespread use of brief, vaguely worded parent and teacher questionnaires to diagnose the presumed complex neurodevelopmental disability of ADHD.

But- the scales generally consist of only small numbers of items, are nebulously worded ("often," "excessively," etc.), and place much of the responsibility not only for reporting but also for making clinical judgments as to deviation in the eyes of the beholder. Variations in experience, tolerance, or criteria used among observers are not allowed for. They probably measure parent or teacher discomfort as much as they do the actual behavior in the child. They have not met adequate psychometric criteria. Yet this vagueness leads to an all-or-nothing diagnosis of a neurodevelopmental dysfunction. The consequences have included poor inter-rater reliability, overdiagnosis, misdiagnosis, inclusion of other problems (the comorbidity issue), and various unvalidated techniques (e.g. electroencephalograms) have been proposed by some in an effort to improve the precision of the diagnosis. Certainly the traditional techniques of interviewing, observations, and examinations are still preferable.

V) Most important predisposing factors may be low adaptability and cognitive problems.

The DSM-IV definition says that high activity and low attention span are the disorder itself.

But- accumulating evidence is demonstrating that other factors may be more important in production of the behavioral or scholastic dysfunction: 1) A different behavioral predisposition, variously described as low adaptability, limited ability to modify behavior, a problem in regulation of responses, and a deficiency in response inhibition; and 2) A developmental predisposition- There is a high frequency of cognitive disabilities in children diagnosed with ADHD today.

VI) Lack of evolutionary perspective

Embodied in the current ADHD diagnosis is the assumption that a child not fitting into the modern classroom has a defective brain.

But-an evolutionary perspective informs us that the ADHD traits may have been highly adaptive in primitive times in a world full of predators, although less so now. They may also be adaptive outside the modern classroom.

VII) Small practical usefulness and possible harm from label.

Some observers maintain that the ADHD label represents real progress in mental health diagnosis in that it takes the blame off the parents and schools, helps children get services, and justifies the use of medication.

But- there are several negative aspects of the labeling: 1) It is of limited practical value to teachers, psychologists, and physicians in that it offers no articulation of the individual's problems and strengths and no suggestions for specific management other than medication. 2) It may be misleading as to the true nature of the problem. 3) The complex phenomenon of attention is analyzed in too simple a way. 4) The label may be harmful and stigmatizing by stating or implying brain malfunction when it is unproven. Labels stick.

Conclusions- Is ADHD a valid disorder? The DSM-IV's own definition of a mental disorder says that it should be a clinically significant behavioral syndrome resulting in present distress or disability and arising from a dysfunction in the individual. That may be true for that 1-2% of the population. However, what is now most often described as ADHD for the other 5-15% of children in the United States appears (at least to temperament researchers) to be the problems resulting from a dysfunction in the interaction between normal but sometimes aversive behavioral variations in neurologically intact individuals and their incompatible environments. This

discrepancy leaves the validity of the ADHD construct at least partially in doubt.

Research for a better diagnostic system should include: 1) DSM-V should finally acknowledge the existence and clinical importance of normal temperament differences. Both environmental and biological determinism should be avoided.

2) A revised ADHD diagnosis would be tenable if it is limited to the 1-2% who are truly "hyperkinetic."

3) The other 5-15% would best be defined in terms of areas of dysfunction rather than as to their risk factors. The diagnosis of Adjustment Disorder already exists and would probably fit most of them. Another option for explanation is normal temperament variations without dysfunction, which are a very common source of concern to parents or teachers and frequently subject to pathologization.

4) Any diagnosis of brain malfunction should be based only on some objective evidence of it.

5) In the meanwhile, comprehensive functional assessments, which encompass both child and setting and strengths and problems, are much preferable to overly simple categorical labels.

6) Much research is needed to clarify these points. One would hope that the consensus report of this conference will avoid a false certainty and correctly reflect these concerns about how little we really know.

Go to "Things to do at this website"