A Step in the Right Direction: An Interview with Darrel Regier
What distinguishes DSM-5 from the previous editions of the Diagnostic and Statistical Manual?
Darrel Regier: As the Task Force began to meet in preparation for working on DSM-5, the main question we considered was whether the old psychiatric diagnostic category system embodied in DSM-IV was actually starting to hinder research by failing to reflect the latest science about psychiatric disorders. Further, in our discussions, it became clear that the emphasis on maintaining rigid boundaries between diagnoses was simply not reflecting the most up-to-date findings in genetics and family history research. For example, the research was showing that people who grew up in families with a high incidence of schizophrenia were likely to have kids with bipolar disorder, major depression, and neurodevelopmental disorders. Findings like that underscore the fact that the old idea of a strict separation of disorders and of a separate genetic vulnerability hypothesis for each separate disorder was no longer tenable, given our current knowledge.
We now know that there are hundreds or thousands of genes that predispose someone to a condition like schizophrenia, and these genes interact with environmental exposures that will either turn on or turn off the gene in order to express a particular syndrome. This understanding requires us to take a different approach to the task of categorizing mental disorders. The idea that psychiatric conditions are polygenetic is something that we never imagined back in 1980, when DSM-III was first published. So we wanted to take a more dimensional approach to diagnosis, rather than assume that psychiatric conditions all belonged in neat categories.
In practice, what does this idea of “taking a more dimensional approach” to diagnosis mean?
Regier: Probably the best example would be to start with major depressive disorder, one of the most common disorders seen in clinical practice. In DSM-III and DSM-IV, the diagnosis of major depression involved nine criteria that include depressed mood, anhedonia or lack of interest, as well as a series of symptoms including sleep difficulties, lack of concentration, somatic concerns, and suicidal ideation. Nowhere in that set of criteria was there any mention of anxiety, even though research has established that more than 50 percent of individuals with major depression have substantial anxiety symptoms. In fact, DSM-IV even prohibited you from considering the diagnosis of generalized anxiety disorder in the case of major depression.
While the old assumption was that anxiety disorders were completely separate from mood disorders, research has shown that depressed individuals with anxiety have a much poorer response to antidepressants and cognitive behavior therapies for depression. They also are at a higher risk for suicide. So we debated for a long time whether we should add a separate diagnosis of “anxious depression” in DSM-5, but ultimately, we decided against it because we didn’t want to reinforce the idea that there are rigid boundaries between diagnoses. Instead, we opted for symptom specifiers within various diagnoses to emphasize the wide variety of ways in which different people can experience the same psychiatric disorder. So someone can be diagnosed with major depressive disorder with a specifier of “with anxious distress” in a way that calls specific attention to the presence of anxiety symptoms without having to invent new diagnostic categories.
How did the doing away with the old five-axis system reflect the new interest in the dimensionality of diagnosis?
Regier: Well, there are several ways in which it reflects the shift to a more dimensional approach. The five-axis approach separated conditions of personality disorders and mental retardation to their own axis (Axis II) from all other mental disorders. But the idea that somehow Axis II conditions were fundamentally different from other mental disorders simply hasn’t held up. Certainly, conditions like borderline personality disorder or, for that matter, antisocial personality disorder have been shown to be as clinically significant as any of the disorders in Axis I. We not only thought that was important to recognize, but we were also concerned that Axis II diagnoses had achieved a certain amount of stigma. Some insurance companies wouldn’t even provide coverage for Axis II conditions, considering them either unchangeable or not amenable to effective treatment. In DSM-5, clinicians list all diagnoses contiguously, which removes any implied hierarchy of importance between the disorders.
In DSM-IV, Axis III referred to any medical condition related to a mental disorder. But we decided placing those conditions in a separate category wasn’t clinically or conceptually helpful. For instance, we wanted clinicians to pay attention to the fact that if someone with a major depression also had comorbid diabetes, it shouldn’t somehow be considered an entirely separate concern, especially in light of the research showing the high rates of comorbidity between major depressive disorder and diabetes.
Axis IV looked at how psychosocial and contextual issues influenced a patient’s coming in for treatment. In practice, it was a catch-all axis for explaining the impact of the social environment, and it was never used very effectively in terms of diagnosis. Clinicians were inconsistent in their integration of such contextual information into clinical care, as there seemed to be no universal understanding of how Axis IV content should be used. There also was no way of coding such conditions in the patient’s medical record. So we decided to discontinue its use as a separate axis.
Axis V was, of course, the Global Assessment of Functioning (GAF) scale that was introduced in 1987. Over time, it had become a kind of shorthand for giving an impression of someone’s overall resilience and ability to cope. The problem with Axis V was that it tried to encode complex information in too simplistic a manner, incorporating the issues of possible psychotic symptom severity, suicide ideation, and homicide risk with the question of level of disability. Subsequently, there was no way of knowing what was actually being measured. In fact, the GAF became a kind of code that clinicians commonly used to get someone hospitalized or make them eligible for insurance. A score of 40 on a GAF might get one patient into the hospital, while a score of 60 might get the same patient qualified for discharge. The reliability of these scores was not high unless clinicians received rigorous training, and the meaning of the score was often ambiguous. So it was subject to real manipulation.