Dimensional Approaches in Diagnostic Classification: Refining the Research Agenda for DSM-V

The idea of “dimensional classification” may sound like an oxymoron at first. How can one “dimensionalize” something that is categorical? Dimensional Approaches in Diagnostic Classification reflects the efforts of the Workgroup/Conference on Dimensional Approaches to Diagnostic Classification toward answering this very question, i.e., How can a dimensional approach to psychiatric classification be applied without making fundamental changes in the current classification formats? It is truly a multidisciplinary, multinational effort.

The Workgroup/Conference on Dimensional Approaches to Diagnostic Classification, organized by the American Psychiatric Association, the World Health Organization, and the U.S. National Institutes of Health (NIH), was assembled with the intent of informing the development of classifications systems internationally. This is not a part of the DSM-V task force efforts but aims to stimulate the field to do the work necessary to inform the DSM-V and ICD-10 task forces. The goals of the group were to promote international collaboration, stimulate research to inform upcoming versions of psychiatric classification systems, and facilitate the development of “worldwide” criteria. To this end, the Workgroup on Dimensional Classification empanelled smaller workgroups structured around core categories reflected by DSM-IV and ICD-10 psychiatric diagnosis, including substance use disorders, depression, psychotic disorders, anxiety disorders, and personality disorders. The results of each subgroup’s efforts are presented in this book.

Psychiatric classification facilitates characterization, communication, and research about psychiatric illness. For a long time, there has been a debate regarding whether diagnostic categories that depend on symptom clusters are informative.

“Dimensionalizers” argue that there are dimensions of “functioning,” such as information processing, psychosis, affectivity, mood, and processing speed, that must be described and that different patterns among such measures reflect different psychiatric disorders. Each individual would have a dimensional profile. The “categorizers” support the process of determining specific groups or categories of symptoms that reflect psychiatric syndromes. Both approaches have value, and both approaches have limitations. In the first two chapters of the book, Dr. Lopez et al. (Chapter 1) and Dr. Kraemer (Chapter 2) discuss the pros and cons of each approach. With the goal of putting forth dimensional approaches, the book provides examples and models for each disorder. The authors are remarkably clear so that even those without statistical or psychometric knowledge can understand the issues and methods involved. Each chapter offers clinical examples and recommendations for consideration and discussion.

Given the constraint of maintaining the current diagnostic system structure and the fact that people cannot resist categorizing, several issues are important to take into account. The first consideration is whether one approaches psychiatric illness as a bundle of symptoms that can be tallied up or as syndromes that are qualitatively different from each other and from “normal” behavior. Throughout the book, some groups apply dimensions on a continuum from normal to pathological behavior, while others apply a severity continuum to behavior that is already defined as pathological. Each subgroup examines the application of dimensions differently. For example, in Chapter 3, Dr. Helzer et al. review the idea of summing symptoms of alcohol use so that abuse and dependence would be on a continuum based on a number of symptoms (although it would still be categorized), whereas Dr. Andrews et al., in Chapter 4, use more of a severity approach to major depression as well as a summing of symptoms. In Chapter 5, Dr. Allardyce et al. suggest using a continuum of psychosis from normal to pathological behavior. In Chapter 6, Dr. Shear et al. discuss a “hybrid” model that includes categorical and dimensional aspects to diagnosis. Dr. Krueger et al., in Chapter 7, recommend “facets” of personality, whereby those meeting criteria for a disorder receive a diagnosis on axis I, while those who do not have a personality disorder will have a description of “facets” on axis II. In Chapter 8, Dr. Hudziak et al. discuss the issues related to developmental psychopathology, suggesting a paradigm shift to thinking about the disorders of childhood as developmental rather than unique to infants and children. They use the example of conduct disorder to address developmental distinctions and how they can be applied clinically. They describe the changes in conduct disorder between an 11-year-old girl, a 17-year-old male, and an adult. They raise the role of contributory factors, such as age, gender, informant data, and comorbidity, and discuss how they can be included in dimensional ratings by providing the panoply of approaches. The reader is exposed to the diversity and implications of developing a dimensional system for psychiatric diagnosis.

Regardless of the numbering scheme, it is important to bear in mind that attaching a number to something doesn’t make it inherently numerical (i.e., dimensional). The underlying construct must be dimensional for such value to have meaning. Some constructs are qualitatively different, such as those that are either present or absent. In addition, just because something occurs in the general population does not mean that it is “normal,” as suggested by some of the authors. It just means that it is not necessarily rare. Additionally, a numerical or dimensional scale can have a normal distribution if it is assessed across the population or only in a pathological sample. This is important because if you are looking at a distribution in a general population, eventually there will be a determination of what is pathological (e.g., cut-off points are suggested in several chapters). This brings the field back to categorical designations. Finally, factors measured in a pathological sample may reflect severity or the specific nature of the pathology. The factor may be normally distributed within the sample of individuals with the psychiatric difficulty in question. Some pathological behaviors don’t exist in the “normal” population, yet they may be normally distributed in a pathological population. Empirical evaluation is critical.

Another consideration, perhaps one of the most overwhelming tasks, is the determination of which dimensions are relevant and how they can be most effectively measured. In Chapter 6, Dr. Shear et al. make reference to the plethora of psychiatric and psychological rating scales around the world that measure anxiety. Selecting the one psychometrically sound scale with construct validity for inclusion in a diagnostic system is a daunting task. Coupled with the discussion in Chapter 9, by Helzer et al., about “top-down” clinical approaches versus “bottom-up” empirical approaches, the choice of dimensions could take a lifetime. Whatever the choices for the new diagnostic systems, future task forces will be vetting them via clinical and empirical tests with the aim of refining their validity and utility.

The integration of clinical and statistical approaches is a true challenge as well. The process of applying dimensions to psychiatric classification requires input from both clinicians and statisticians, which means they have to learn each other’s languages as well as content to avoid making either anecdotal decisions that are not generalizable or statistical decisions that may not have clinical validity or utility. The effort by the Dimension Workgroup has already begun to address this issue by including clinicians and statisticians from around the world. The multinational composition of each workgroup represents a first step toward their goal of “worldwide” criteria for psychiatric classification.

The reaction from the clinical community that uses the instrument is critical. The system developed must have easy applicability, characterize people effectively, and provide ready communication among clinicians. A multinational system must also include considerations of applying the criteria in different cultural contexts. The authors of this book address complicated and sophisticated ideas, and their use of examples and their effort to make this accessible to all clinicians in the field bodes well for the likelihood that the new system(s) will be manageable for clinicians—albeit there will be a period of adjustment. This book opens the doors for clinicians to actively participate in the process of developing new diagnostic classification systems.

The book is applicable to a broad audience and holds the promise of expanding discussion of the use of dimensions in psychiatric classification. This was the aim of the Workgroup/Conference on Dimensional Approaches to Diagnostic Classification.