Categorical or Dimensional Models: Which is Best for ADHD?
Categorical or Dimensional Models: Which is Best for ADHD?
Sam Goldstein, Ph.D.
This is a SamGoldstein.com Monthly Article - November, 2000
Copyright ? 2000 Dr. Sam Goldstein - All Rights Reserved
This month’s article will provide an overview of the issue of categorical versus dimensional models in the evaluation of ADHD. Clinicians using categorical models to make diagnoses must think dimensionally in the treatment process, as well as in planning for positive long-term outcome for individuals with ADHD.
The diagnostic systems in place today, both the Diagnostic and Statistical Manual - 4th Edition of the American Psychiatric Association and the International Classification of Diseases - 10th Edition, are categorical models. That is, they present an all or nothing phenomenon. They are polythetic, such that to receive a diagnosis a child must manifest a minimal threshold of presenting symptoms as well as a number of additional criteria related to age of onset, impairment and possible related comorbid or co-occurring conditions. If the child is one symptom short, despite the fact that he/she may meet the other criteria, a diagnosis of ADHD is not provided. However, on a dimensional basis, it is assumed that everyone exhibits these symptoms from a minimal to a maximal degree. On a dimensional basis, having all but one symptom necessary to receive a categorical diagnosis, when other criteria are in place makes one quite different from the general populus.
With the increasing theoretical understanding that ADHD represents a condition that leads individuals to fall to the bottom of a normal distribution in their capacity to demonstrate and develop self-control and self-regulatory skills, the use of a categorical model in defining who has and does not have the condition does not appear to represent a good fit. On the other hand, a dimensionally based diagnostic system is much more reasonable. In that system, the diagnosis is based upon whether an individual’s presentation is significantly different from that of the general population; not necessarily if that difference falls at an exact threshold on the bell curve. One can imagine certain situations in which a large group of individuals may not be able to exhibit a certain behavior or skill which might categorically lead a diagnostician to conclude they were impaired. Yet on a dimensional basis if many others struggle and impairment is the norm rather than the exception the diagnosis is not made. Dimensional models also better predict outcomes in a complete and unbroken way for a particular skill or symptom such as impulsivity. A dimensional model orders individuals, for example, from those who are minimally impulsive to those that may be mildly, moderately or significantly impulsive. The dimensional model does not suggest that there are heterogenic differences among mildly, moderately and more severely affected individuals but that the differences among these individuals represent differences in severity of a homogeneous concept.
Multiple researchers have reported a consistent and generally linear relationship between symptom severity and outcome risk. Dimensionally scored symptoms appear to be better predictors of outcome than measures based on a categorical classification. For example, in 1995 in New Zealand, nearly a thousand, fifteen-year-olds were evaluated on a continuous variable ranging from none to severe in regard to symptom presentation of ADHD. At sixteen years of age the group was re-assessed on a series of outcome measures, including substance abuse, juvenile offense and school drop out. This comparison demonstrated evidence of continuous and generally linear dose-response functions between symptom severity and outcome risk. Variables scored on a dimension were consistently better predictors of outcome than measures based on a categorical classification. Although the DSM categorical criteria had value in the short run, they are not particularly effective in the long-run in regard to predicting future outcome. Further, other researchers have demonstrated that the exclusive use of categorical criteria often results in predictive errors. Though categorical diagnoses allow the identification of a constellation of co-varying behaviors that can be used to predict the relative success of a possible intervention, their reliability and validity of many of the diagnostic criteria, including those for ADHD, are not well established. The use of the DSM also requires one to possess a working knowledge of child development and an understanding of how to apply and conceptualize symptoms at different ages. Characteristic symptoms of ADHD appear to diminish in presentation and severity as children grow older. For example, research with the revised third edition of the DSM found that although eight of fourteen symptoms were required in children, this threshold for diagnosis resulted in over-inclusion of affected younger children and under-inclusion of affected adolescents. Furthermore, recent research finds that into the adult years the number of symptoms necessary to set an individual apart from others categorically in regard to inattentive and hyperactive-impulsive problems continues to diminish into the fifth decade of life.
It is also important for parents and educators to understand that the majority of the diagnostic process completed by clinicians is dimensional. Questionnaires they use rate children in comparison to normative or clinic referred samples. The tests used, whether specifically designed for ADHD, such as continuous performance tests or general psychological testing, such as an intelligence measure, also provide standard scores. These scores place children on a dimensional continuum. I would further suggest that clinical judgment is also dimensional. Clinicians evaluate a child based upon their observation and intuition in comparison to children whom they have evaluated and worked with in the past. This too is a dimensional comparison. The very last step in the diagnostic process requires almost a leap of faith from dimension to category. It is not to suggest that clinicians do not carefully count symptoms and make certain that individuals receiving a diagnosis of ADHD meet the required criteria but that this is often a very last step.
In 1997, neurologist Marcel Kinsbourne, was the first to suggest that temperament (those qualities that children bring to the world likely as the result of the interaction of genetics and biology) falls on a continuum of characteristics such as impulsivity and hyperactivity. Thus, children receiving the diagnosis of ADHD represent the extreme of the temperamental continuum. During the 1970's there was increasing support for the notion of hyperactivity on a dimensional basis, reflecting delayed maturation. The hyperactive child was viewed as abnormal in rate of acquisition of certain skills. The notion of hyperactivity as a developmental delay was supported by soft neurological signs and related biological research. The new millennium will, I believe, witness a continued focus on symptoms and impairment caused by symptoms in ADHD as phenomena falling on a dimensional basis. I believe that this process will not only improve our diagnostic capabilities but also our ability to accurately evaluate effective treatments and assess the criteria necessary to choose the right treatments for each individual.
