Diagnostic and Statistical Manual 5

Tags: , ,

By: Roger S. McIntyre, MD, FRCPC

In May 2013, the Diagnostic and Statistical Manual 5 (DSM-5) was officially launched by the American Psychiatric Association at its annual meeting in San Francisco. The DSM-5 is the product of a decade of activity that has been dedicated to defining and operationalizing psychiatric disorders as well as establishing their reliability. The DSM-5 has generated a significant degree of interest, controversy, and disquiet within the broad ecosystem of psychiatry as well as in the general population as evidenced by the international conversation that has occurred in various media outlets.

The DSM-I, which was launched in 1952, contained 106 codable diagnostic conditions. The DSM-IV Text Revision, launched in 1994, had 365 diagnoses. At this juncture, I have not completed the enviable task of adding up all of the disorders in the DSM-5, but safe to say, there has not been a significant reduction in the number of diagnoses. I share the concern raised by many that the number of diagnoses contained in the DSM has far exceeded, in many cases, their clinical relevance. I am, however, curious at the frequently leveled complaint that the DSM lacks validity. The DSM was never intended to validate psychiatric disorders, and instead was primarily intended to provide a lingua franca for clinicians, scientists, and other stakeholders, as well as a mechanism for registering and compensating for psychiatric visits. In my view, one of the top five papers ever published in psychiatry was the classic paper by Robins and Guze in 1970, wherein they outlined how psychiatric disorders can be validated. The first step toward validation is establishing reliability. (1)

Notwithstanding the foregoing, it is rather concerning that the reliability of several of the common disorders in DSM-5 are of questionable clinical significance. For example, the kappa statistic (a measure of the level of agreement between observers, over and above chance) was 0.28 for major depressive disorder (MDD) and 0.20 for generalized anxiety disorder.  It was reassuring that the kappas for bipolar I disorder (0.56), bipolar II disorder (0.40), schizophrenia (0.46), and borderline personality disorder (0.54) were moderate.

The DSM-5 evolved significantly in the area of bipolar disorder (BD). Notably, individuals who declare hypo/mania while exposed to, or withdrawing from an antidepressant, are now diagnosed as bipolar. This is a long overdue change supported by the empirical evidence. Wisely, the DSM-5 has supplanted mixed states with the admittedly broader notion of mixed features specifier, which will apply to both BD and MDD. This latter change, which could be conceptualized as “Neo-Kraeplenian,” reminds us of the original description of manic-depression which had posited that cyclical disorders can be grouped together and phenotypically different based on the presence or absence of hypo/manic/mixed features.

The National Institutes of Health has proposed the research domain criteria (RDoC) as an alternative diagnostic manual for mental disorders wherein the overarching aim is to define and categorize mental disorders on the basis of their underlying biological substrate. The proposal is long overdue and it is anticipated that efforts to refine the RDoC will progress the field forward with the hope that the mechanistic basis for mental disorders will be better understood. One is tempted to hypothesize that future iterations of DSM (or RDoC) may be an integrated manual wherein both phenomenological- and biologically-based criteria will apply.

In the interim, diagnosing and categorizing mental disorders with a DSM-5 is the most practical tool available to us at this point in time. This reminds us of Churchill’s famous quote, “Democracy is the worst form of government, except for all those other forms that have been tried from time to time.” In contradistinction to forms of government, all possible alternative approaches to diagnosing and categorizing mental disorders have not yet been tried. The next decade will be extraordinary and exciting as advances in cognitive neuroscience, “-omics,” biomedical computational methods, and nanotechnology, converge to unravel the complex substrates that subserve mental disorders. Such efforts will likely change the way psychiatry diagnoses and categorizes illnesses, which will be the first step towards treating and preventing these highly disabling disorders.

Roger S. McIntyre, MD, FRCPC
Psychiatry and Pharmacology,
University of Toronto
Head, Mood Disorders Psychopharmacology Unit,
University Health Network


1. Robins E, Guze SB, et al. Establishment of diagnostic validity in psychiatric illness: its application to schizophrenia. Am J Psychiatry. 1970; 126:983:987.