C. Keith Conners, Ph.D.
Presentation to the Ontario Psychiatric Society Toronto, Ontario, Canada April 16, 2011
Diagnosis in ADHD has had a rocky history. One only has to recall the events occurring at the NIMH Consensus Conference on ADHD in 1998. When asked his opinion about diagnosis, Mark Vonnegut, one of the experts on the panel, commented that “Diagnosis in ADHD is a mess.” This answer was reprinted in headlines the next day in newspapers around the world. When asked how you make the diagnosis, Dr. Vonnegut, threw up his hands, then took a line from Supreme Court justice Potter Stuart. The United States Supreme Court was in the midst of a controversy on how to define pornography. When asked for his definition , Justice Stuart replied, “I don’t know how to define it, but I recognize it when I see it.”
Dr. Vonnegut was probably correct in assuming that a fly-by- the-seat-of-your-pants recognition-of-ADHD-when-you-see-it, is the approach of most medical practitioners. Surveys of diagnosis in practice have repeatedly shown that many practitioners fail to follow the explicit guidelines for diagnosis,(1)such as those proposed by the American Academy of Pediatrics(2). This has led to repeated findings of over-diagnosis, based on the actual figures of ADHD from epidemiologic studies. (3) The late Atilla Turgay had developed similar guidelines for the Canadian Psychiatric Association, so we can assume that this audience will surely be familiar with those guidelines.
But there are other problems in making the diagnosis than simply ignoring or being unfamiliar with the guidelines. One major problem has to do with the criterion requiring impairment in at least two different settings. In early trials of DSM it was found that by using impairment in only one site, such as school or home, there was an excess of diagnosis, rising to as much as 40% of the normal population. Since most psychiatrists rely solely on the report of a parent about the child’s school behavior as well as their home behavior, the school setting is only indirectly being assessed. Mark Wolraich has reported that failure to get school information was one of the major limitations of diagnosis among pediatricians; and the evidence-based guidelines stress the importance of direct reports from classroom teachers.(4)
Another major challenge with diagnosis is the presence of comorbidities. Comorbidity with ADHD is not like seeing a diabetic who also has a broken leg, where the comorbid symptom is assumed to be caused independently from the diabetes. With ADHD two thirds of cases will have some associated problem with oppositional, conduct, anxiety, or depressive symptoms. These more “complicated cases” are the rule rather than the exception, so it places the burden on the diagnostician to have a good knowledge of the comorbid conditions. Factoring those comorbidities into the treatment plan is essential for knowing the type and sequence of appropriate treatments.
Many parents will show up at their first interview ready with an internet-based diagnosis that includes a wealth of behaviors that may or may not be relevant to ADHD. Like Mark Twain, after reading a medical text book and finding that he had “every malady except housemaid’s knee,” parents and young adults who have read “Driven to Distraction,” will allege symptoms covering all of the DSM disorders. Which ones rise to the level of a true comorbid disorder, which are significant but below the diagnostic threshold, and which are simply factitious provides a diagnostic challenge.
Finally, there is a special problem in making diagnoses for very young children. Many believe that such a practice is harmful to the child and family by providing an early stereotype that inflicts damage to self esteem, while others believe that such diagnoses are unreliable or unwarranted because of rapid developmental changes, which obscure meaningful diagnosis in preschoolers and younger children.
So, in summary, we see problems in ignoring evidence-based guidelines, problems in defining the pervasiveness of impairments, problems with the comorbid or associated conditions of ADHD, and problems in applying adult-based criteria to younger children. What resources are available to help with the problems we have identified here? Here is where my personal experience may be relevant. Permit me to recount a little of my own history in trying to deal with these problems.
When I began my first job at Johns Hopkins Medical Center in 1960, I was asked by my boss, Leon Eisenberg, to analyze the data from his just-completed randomized trial of Dexedrine and placebo in “delinquent” boys living in a cottage home run by social services.(5) The data I had to work with was a checklist of symptoms taken from Leo Kanner’s textbook of child psychiatry. The items in the checklist were simply being summed to give an overall symptom score. The data were striking in showing a dose-related improvement from the ratings done by the cottage “parents.” I became interested in the problem of monitoring drug effects with the checklist, and decided to convert it into a scale by having each of the 93 symptoms rated from 0-3, or “not at all, just a little, quite a bit, and very much.” Later I did a similar revision with a 39-item teacher report.
By factor analyzing (that is, clustering) the items according to their associations with each other, I was able to identify a small number of symptom clusters that were reliable and appeared to reflect some basic dimensions of behavior in our outpatient clientele. I began to collect normative data from children attending our clinic.
In later years, with the help of MHS here in Toronto, we were able to obtain a national database from thousands of children in North America and Canada. At that point we were able to develop normative data for separate ages and gender, along with other demographic data. So now we had a scale for measuring the position of any putative ADHD patient compared with the average child of the same age and gender in the population. Very early in the use of these scales I noticed a lot of pressure from respondents to have shorter scales, so I eventually developed a 48-item parent scale, and by taking the best items from there I was able to create a 10-item scale. We called it the “hyperactivity index” because it seemed to reflect the essential characteristics of the hyperkinetic child; it was especially favored by teachers who resisted the longer forms.
In pursuit of even greater brevity, I once dreamed I had found a one-item scale, but when I woke up I forgot what it was. (This joke somewhat reminds me of the true story of William James when he was taking nitrous oxide, during which he felt he could discover the greatest secret of the universe. When he awoke from his drug-induced insight he finally remembered it well enough to write it down: “Higgedly hogamous, woman is monogamous; hoggedly higamous, man is polygamous.” He was disappointed, though some believe he may have hit upon an eternal truth.) But I digress.
We still use the original 10-item index because it seems to be a very sensitive indicator of response to drugs or other treatments, particularly when trying to adjust dosage for an individual, or to track behavior over time. But needless to say such a brief scale cannot cover the scope of information required for diagnosis or initial assessments. Because of the complexity involved in ADHD diagnosis, the trend towards brevity has been reversed, as greater precision and clearer guidelines are needed for a comprehensive evaluation that includes differential diagnosis, measures of impairment, and much other relevant information. That is why our latest diagnostic aide for ADHD (the C-3), covers ADHD as well as its major comorbid conditions of ODD, CD, Anxiety, and Depression. The companion scale to the C-3 is a much more detailed scale of all the major pathologies identified by DSM in childhood, called the “CBRS”, or Comprehensive Behavior Rating Scale. Let me give you a brief overview of our reasoning and the process behind the development of these two major rating scales.
First, we had at hand a large body of data from previous studies, such as the national survey of major rating scales carried out by Tom Achenbach, Herb Quay, and myself, funded by grants from the American Psychological Association.(6) This survey identified eleven major clusters of symptoms that cut across all socio-economic and ethnic levels. Second, we had the symptoms and impairments identified by DSM-IV for major childhood disorders.
We pooled all of this information and set about culling the data for redundancies, removing items of very low frequency, un-readability; and translations for the Spanish version carefully adjusted for the cultural meaning of the items. This pool of items was submitted to various experts in the field for suggestions, omissions, and relevance. On a test sample we performed several Exploratory Factor Analyses of the data, finally creating a “model” structure which was then verified by a Confirmatory factor analysis from a large North American national sample.
(Now I recognize that psychiatrists’ eyes tend to glaze over when psychologists start talking about factor analyses, but as in Herman Hesse’s novel, The Glass Bead Game, we have to have our own priestly mysteries, that we indulge in for our own special version of reality. This is a version of reality which sees that underlying the multiform appearances of behaviors there is a causal structure which we can identify by looking for the orthogonal relationships among groups of symptoms or behaviors).
Note that there are two parallel sets of “diagnoses” in these results: one of course is just the diagnoses representing the DSM-IV categories, albeit simplified in language and re-worked to be readable and comprehensible to parents, teachers and patients, and converted into normative scales. These are the “rational diagnostic categories,” only partly based on empirical data as well as clinical acumen acquired over several decades. The other set of content scales are those empirical factors which emerge from the factor analyses, and although they overlap with DSM, they empirically identify broad patterns.
In practice, the psychiatrist constructing the diagnoses of a patient will certainly want to use the “official” DSM categories, for medico-legal and professional communications, and the rating scales add a considerable benefit over the short paragraphs from the DSM handbook: Each diagnostic category is normed against a national sample adjusted for age and gender across a census-based representative population. So the prevalence of any given diagnosis in the population allows the clinician to know where the patient falls in the statistical distribution. While somewhat arbitrary, we set a threshold of 1.5 standard deviations above the mean as a clinically meaningful cutoff point for each score, including those from parents, teachers, and self-report of the child.
Alternatively, the scoring for the DSM categories can be done exactly as required by DSM involving the number of symptoms, rather than the statistical distribution in the general population. For the empirical or content scales, which overlap the DSM scales, we have clusters of symptoms that “appear in nature,” not just those constructed from closed-door committee meetings of the DSM working groups. We also calculated an overall statistical index of the probability that the patient has a diagnosis, based upon a comparison of a large well-identified clinical sample with matched normal controls. This allows the clinician to have an empirically derived probability estimate that the patient lies within the range of previously diagnosed patients who had a full clinical workup.
Built into the scales is also a check on the reliability of the informant, based upon overly negative or overly positive reporting biases, and consistency of response. The specifics of the scales and the steps needed to interpret them are provided in a Quick Reference handout. Speaking of one’s eyes glazing over, try to score these scales by hand, consisting of hundreds of items, validity checks, content scales, symptom scales, impairment items, and a variety of other clinical and immediate action scales, for parents, teachers and self-report by the patient. No wonder Mark Vonnegut elected to just wait until he saw it, instead of all that bother. Well, thanks to modern technology, the burden of all that work lands squarely on the parents, teachers, and patients who fill out the scales, not the psychiatrist who has to interpret them. The scales can be scored electronically on- line or a software disc, resulting in a beautiful comprehensive report, complete with graphs and interpretive results. Results from teachers and parents can be obtained directly from their computer-based input, with the scored results and interpretive reports available directly to the clinician.
Regarding early childhood diagnoses, recent epidemiological studies have shown that valid and reliable DSM diagnoses can be made in preschoolers(7) and an early childhood subcommittee of the APA has developed a modified diagnostic algorhythm for preschoolers to deal with constructs derived from older children or young adults. (4) I have provided an overview of those studies, which is available on my blog at http://adhd-world.blogspot.com/. We recently published an Early Childhood Diagnostic and Developmental Scale which seeks to identify the major symptom patterns in preschoolers and younger children, along with a parallel set of items to identify the major childhood developmental issues. The Conners Early Childhood™ (Conners EC™) aids in the early identification of behavior, social, and emotional problems in preschool-aged children 2 to 6. It also measures whether or not a child is appropriately meeting major developmental milestones. This scale, like our others, is described on the MHS website at MHS.com/Conners.
Finally, how should one use these reports in making a diagnosis? Let us stipulate that the most essential ingredient in making the diagnosis of ADHD is a wise and experienced clinician, trained broadly in the medical arts, particularly in taking a history; and one who is practiced at gathering as much information about the patient as is practical within the time limits of the particular setting, the availability of informants, and reimbursement issues. The key ingredient in all of this is the skillful history by the clinician, which then allows him or her to formulate hypotheses about the clinical status of the patient. The rating scale results are only hypotheses to assist clinicians in making the final decisions, based upon their own hypotheses gained from a careful and comprehensive history, including the family psychiatric history, medical examination (for example to rule out the rare thyroid condition), family functioning, educational history, and a host of possible environmental contributions. But we believe that the rating scales add a solid empirical basis to the entire process, addressing the issues we outlined in the beginning: following the best-practice guidelines, dealing with comorbidities, addressing areas of impairment, and assessing very young children.
In summary, we started with quite a few useful items to measure treatment outcome. Then we progressed to a smaller set of items by refining the long original list. And finally we ended up with a much more comprehensive set of rating scales in response to the obvious need to cover a very large field of pathologies and diagnoses. This approach, which started long ago, and evolved over time, represents my approach to diagnosis with ADHD. I think that the rating scale tools are extremely helpful in assisting the clinician in dealing with the diagnostic challenges we have discussed.
But of course, helpful as these tools are as aids to the practical problem of making a diagnosis according to today’s guidelines, they do not deal with the fundamental problem, that there is no patho-physiological diagnosis. Etiological diagnoses in psychiatry have largely been abandoned because of the absence of a patho-physiologic or anatomic causal basis, in favor of DSM’s strictly behavioral and clinical approach. So where do we go from here with diagnosis of ADHD? I have heard that psychiatrists like to talk about dreams, (at least they used to; but perhaps they mostly talk about pills now); but I will tell you one of mine. I call this dream, “A revised personal history of my work on diagnosis: or a Path Not Taken”.
This dream appears to be a wish-fulfillment about how I might better have spent my time. In the first part of this dream it is 1911, and a few friends of mine (Marie Curie, Henre Poincare, Albert Einstein, and Max Planck are at the Solvay Conference in Brussels, standing around endlessly discussing particle physics and such, and I am wondering why they aren’t paying attention to ADHD, since Sir George Still had only recently identified (in 1902) a group of children we now call ADHD. (Now I know you are wondering why I, who was born in 1933, could know this; but obviously in psychoanalysis we know that time travels very peculiarly in dreams.
Later in this dream it is 1960, and I have just started work at Johns Hopkins. Two years later one of my great heroes, Rachel Carson, published Silent Spring. She tells me that “From 1945 when the use of synthetic pesticides began in the United States, to the time Silent Spring was published, pesticide use increased about sixfold. In the ten years between the publication of Silent Spring and the banning of DDT in 1972, pesticide use increased tenfold, to about one billion pounds annually”.
Since then the total quantity of pesticides in terms of pounds has not increased; however, the actual toxicity of pesticides has increased ten to twenty times. (8) Figures from today estimate that less than 0.01 percent of the pesticides that are applied reach the target pests, which means that 99.99 percent of the pesticides that are applied pollutes the environment. About 35 percent of the food that is purchased has measurable levels of pesticide residues, with 1 to 3 percent having residues that are above accepted tolerance levels.
Knowing all these facts in 1962, I began my scientific career in search of a physical pathology for ADHD (hyperkinesis, MBD, whatever) that would provide a true patho-physiologic bases for diagnosis. So what I did (in the dream) was the following: I never liked animal studies but knew it was essential to this program, so I studied the toxic effects of pesticides on animal behavior, and then showed by histo-pathological brain studies that the pesticides not only destroyed three of the dopamine receptor sites, but did so by transmission through the breast milk and amniotic fluid of the mother rats.
I also discovered that these toxic organophosphates worked to destroy acetyl cholinesterase, leaving acetylcholine nerve fibers in a constant state of excitability. It was then an easy step to human studies, where the concentrations of pesticides in different areas of the country were easily matched to prevalence levels of ADHD in pregnant mothers and in their offspring. The follow-up studies of the offspring showed conclusively that ADHD was tightly linked to levels of pesticides in the mothers and the child at birth.
Then, by analyzing organophosphate exposure in the urine of more than 1,100 children 8 to 15 years old, I found that those with highest levels of dialkyl phosphates which are the breakdown products of organophosphate pesticides also had the highest incidence of ADHD. (You might notice the suspicious similarity of my results to those of Bouchard, et al, just last year). Overall there was a 35% increase in the odds of developing ADHD with every 10-fold increase in urinary concentration of the pesticide residues. The effect was seen even at the low end of exposure: children who had any detectable, above-average level of pesticide metabolite in their urine were twice as likely as those with undetectable levels, to show symptoms of ADHD. (8)
With the aid of Russel Schachar and his group we were then able to relate particular endophenotypes of ADHD to organophosphates, accounting for the fact that there is a high rate of genetic loading in ADHD, and the fact that a certain proportion of pesticide exposures do not result in illness. The fact that about 300,000 humans are poisoned with pesticides annually in the United States, and about 26 million poisoned worldwide, leaves us with no doubt that ADHD and learning disabilities are but one of the many diseases caused by over use of pesticides.
Finally, the part of the dream which earned me the Nobel Prize for medicine came after I developed a simple blood test for recognizing critical levels of pesticides, and a chemical anti-toxin that cured patients of ADHD. Now that’s what I call a good diagnosis!
Rachel Carson taught us that pesticides “reach through the natural ecosystem, affecting not just the ‘target species,’ but humans and the animals in which we rejoice and the habitats on which our lives depend. Their effects reach not only across the land, but through time into future generations; toxins flow into eggs, through amniotic fluid and breast milk, into the tissues of developing children and the young of other species. The interdependence of life links us inextricably to the death-dealing effects of toxins.”(9)
As my mentor Leon Eisenberg once said, “Perhaps it’s time to worry less about how to describe the bodies we are trying to pull from the river, and see who is pushing them in upstream.”
1. Le Fever G, Arcona A, Antonuccio D. ADHD among American schoolchildren: Evidence of overdiagnosis and overuse of medication. Scientific reviews of mental health practice. 2003;2(1).
2. American Academy of Pediatrics: Committee on quality improvement soa-dhdCPG. Diagnosis and Evaluation of the child with ADHD. Pediatrics. 2000;105(5).
3. Costello EJ, Mustillo S, Erkanli A, Keeler G, Angold A. Prevalence and development of psychiatric disorders in childhood and adolescence. Arch Gen Psychiatry. Research Support, U.S. Gov’t, P.H.S.]. 2003 Aug;60(8):837-44.
4. Carlat H, Perrin J, Stein M. The child with ADHD: Using the AAP Clinical Practice Guideline. American Family Physician. 2001 63(9):1803-11.
5. Eisenberg L, Lachman R, Molling P, Lachner A, Mizelle J, Conners CK. A psychopharmacologic experiment in a training school for delinquent boys. American Journal of Orthopsychiatry. 1963;33:431-47.
6. Achenbach TM, Howell CT, Quay HC, Conners CK. National survey of problems and competencies among four- to sixteen-year-olds: Parents’ reports for normative and clinical samples. Monographs of the Society for Research in Child Development. 1991;56(3),):1991.
7. Egger HL, Angold A. Common emotional and behavioral disorders in preschool children: presentation, nosology, and epidemiology. J Child Psychol Psychiatry. [Review]. 2006 Mar-Apr;47(3-4):313-37.
8. Bouchard MF, Bellinger DC, Wright RO, Weisskopf MG. Attention-Deficit/Hyperactivity Disorder and Urinary Metabolites of Organophosphate Pesticides. Pediatrics. 2010 May 17, 2010:peds.2009-3058.
9. Sideris L, Moore K, editors. Rachel Carson: Legacy and Challenge. New York: State University of New York Press; 2008.