The status quo as a good outcome: How the DSM-5 diagnostic criteria for generalized anxiety disorder remained unchanged from the DSM-IV criteria

Introduction

This year finally saw the publication of the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5). It was feared that the DSM-5 would take a ‘reckless’ approach by introducing arbitrary diagnostic criteria, lowering the diagnostic thresholds and unnecessarily increasing the number of diagnoses. Looking at the final product, one can see that, to some extent, these concerns were well founded (Malhi, 2013). However, there are also signs in the DSM-5 of a measured approach. A good example is generalized anxiety disorder (GAD), whose diagnostic criteria in the DSM-5 are the same as those in the DSM-IV.

The purpose of this article is to review the trajectory of GAD from the DSM-IV to the DSM-5 and, using GAD as an example, shed more light on the process of revising psychiatric diagnoses and classifications.

Statement of purpose: What should be changed in the DSM-IV conceptualization of generalized anxiety disorder?

Problems with the concept of GAD in the DSM-IV have been aptly summarized in the literature (Horwitz and Wakefield, 2012; Portman, 2009; Starcevic et al., 2012). Briefly, GAD does not have a single specific characteristic, it co-occurs with other disorders to the extent that their presence may be needed for it to be noticed or recognized and it appears to be neglected by clinicians who diagnose it relatively rarely, despite it being one of the most common mental disorders in the community. It is no wonder that questions about the validity and clinical utility of the diagnosis of GAD have often been raised.

Due to its tendency to split and polarize, the construct of GAD has also been seen as a great dichotomiser (Starcevic, 2012), as follows:

• In DSM-IV, GAD is conceptualized mainly as a worry disorder, whereas in the 10th revision of the International Classification of Diseases (ICD-10), it is largely considered a pathological emotional state (anxiety) and an autonomic arousal disorder.

• Non-psychiatric mental health professionals usually emphasize pathological worry and cognitive aspects of anxiety as the crucial features of GAD, whereas psychiatrists (and physicians in general) tend to emphasize anxiety, especially somatic anxiety, in GAD.

• In the course of the psychological assessment and treatment of GAD, especially cognitive behavioural therapy, levels of pathological worry are assessed by means of an instrument such as the Penn State Worry Questionnaire, and used as a treatment outcome criterion; in contrast, in the course of the psychiatric assessment and pharmacological treatment of GAD, levels of anxiety are assessed by means of an instrument such as the Hamilton Anxiety Rating Scale and used as a treatment outcome criterion.

When the work on the DSM-5 began, some of these problems were apparently recognized, and the DSM-IV concept of GAD was subjected to closer scrutiny.

Stage 1: Is generalized anxiety disorder a form of depression?

Studies of the structure of psychopathology (e.g. Krueger, 1999; Vollebergh et al., 2001) reported that GAD belonged to the anxious-misery group of internalizing disorders (also called distress or dysphoric disorders), along with major depressive disorder (MDD) and dysthymia. This research suggested that GAD was more closely related to depressive disorders than to the fear group of internalizing disorders that comprised other anxiety disorders – panic disorder, agoraphobia, social phobia and specific phobia. Once believed to be a basic anxiety disorder, GAD came to be viewed as the main challenger of the dichotomy between anxiety and depression (Starcevic, 2008).

Subsequent research examined the relationship between GAD and MDD and a proposition that GAD should more appropriately be classified as a depressive disorder. The findings of these studies were inconsistent, with some confirming a very close link between GAD and MDD at the genetic level and others reporting significant differences in other domains. A partial and unenthusiastic consensus emerged that, despite a close relationship, GAD and MDD did not appear to be different forms of the same disorder and that they should remain conceptualized and classified as separate conditions – GAD among other anxiety disorders and MDD along with other mood or depressive disorders (Goldberg et al., 2010).

Stage 2: Attempts to ‘fix’ the DSM-IV criteria

Once it was established with some certainty that GAD is an anxiety disorder after all, studies started to revolve around the DSM-IV diagnostic criteria for GAD. However, the developers of DSM-5 did not address the issues of the lack of specificity, problematic co-occurrence with other mental disorders and divisive nature of the DSM-IV GAD concept. For example, the status of pathological worry as a prerequisite for the DSM-IV diagnosis of GAD was not challenged, despite evidence that it is not specific to GAD, with levels of pathological worry failing to reliably discriminate between GAD and other anxiety disorders (Gladstone et al., 2005; Kertz et al., 2012; Mohlman et al., 2004; Starcevic et al., 2007), and between GAD and depression (Kertz et al., 2012; Starcevic, 1995). On the contrary, a proposal was made to rename GAD as generalized worry disorder (Andrews et al., 2010), thereby erroneously implying that worry is synonymous with anxiety and that pathological worry and GAD are interchangeable concepts.

In accordance with the principles of including in the DSM-5 diagnostically subthreshold and mild disorders as early signs of a more serious psychopathology (Kessler et al., 2003) and avoiding ‘false negatives’ at any cost (Batstra and Frances, 2012), the designers of the DSM-5 were concerned with the recognition of mental disorders. As the main problems with GAD were presumed to be its poor recognition and a high number of ‘false negative’ cases (Wittchen et al., 2002), the key goals were to improve detection of GAD and decrease the rate of ‘false negatives’ in the DSM-5.

To improve recognition of GAD, its diagnostic criteria were to become less restrictive and its diagnostic threshold was to be lowered. This was to be accomplished by making it easier to meet the criterion pertaining to pathological worry (so that it would not need to be uncontrollable as stipulated in the DSM-IV), decreasing the minimum number of required symptoms associated with the DSM-IV (restlessness or feeling keyed up or on edge, muscle tension, irritability, being easily fatigued, difficulty concentrating or mind going blank and sleep disturbance) from three to one and reducing the minimum duration of GAD from 6 months to 3 months. These criteria appeared on the DSM-5 website, along with a proposal to introduce the behavioural criteria (avoidance of situations in which a negative outcome could occur, marked time and effort preparing for situations in which a negative outcome could occur, procrastination due to worries and excessive reassurance seeking due to worries), of which the presence of at least one was postulated as the requirement for the diagnosis.

Support for these proposals was equivocal but overall not strong (for a review, see Starcevic et al., 2012). The behavioural criteria were criticized on the grounds that they were too general and non-specific for GAD and that no tool had been developed to adequately assess them (Starcevic et al., 2012). Also, some of these behaviours, such as excessive reassurance seeking, were reported to be transdiagnostic and therefore characterizing other disorders besides GAD.

In contrast to the unconvincing evidence for the proposed changes to the DSM-IV criteria for GAD, there was more evidence to support concerns that broadening the concept of GAD by lowering its diagnostic threshold would have a number of negative consequences. These included blurring the boundary between GAD and normal anxiety and worry and thereby ‘pathologizing’ ordinary worry, encompassing many ‘false positives’ (diagnostically subthreshold individuals who do not go on to develop full-blown GAD) and spuriously increasing the prevalence of GAD (Beesdo-Baum et al., 2011; Horwitz and Wakefield, 2012; Ruscio et al., 2007; Starcevic et al., 2012). Furthermore, the proposed DSM-5 criteria would not improve clinicians’ ability to recognize GAD, because the key issue for clinical practice – the non-specific nature of the diagnostic criteria for GAD – was not addressed.

Stage 3: Arriving at the point where the journey began

When the DSM-5 was unveiled, the criteria for GAD emerged as identical to those in the DSM-IV and a full circle was thereby made. This was a somewhat surprising outcome, regardless of whether one agrees or not with the proposed changes for DSM-5. In contrast to the special section in the DSM-5 on Highlights of changes from DSM-IV-TR to DSM-5 (American Psychiatric Association, 2013), the DSM-5 contains no information on the bumpy road for GAD from DSM-IV to DSM-5 and the reasons for choosing the status quo. This raises a broader issue of the need to explain all outcomes of the process of revising psychiatric diagnoses and classifications, not only the changes that are made, and underscores the insufficient transparency of the DSM-5 process.

Considering that problems inherent in the DSM-IV conceptualization of GAD have not been resolved, GAD continues to be haunted by the lack of specificity of its diagnostic criteria. In addition, GAD remains a polarizing construct and as such, it hampers communication between various mental health professionals. The status of GAD as an ‘orphan child’ of psychopathology is unlikely to change and GAD will probably continue to be marginalized, perceived as having little relevance or clinical utility and regarded as ‘inferior’ to disorders with which it co-occurs. This also has negative implications for treatment, because targets of intervention (cognitive or somatic manifestations of GAD or both) can be formulated only within a sound conceptual framework. With this in mind, the following needs to be urgently addressed by research:

• Ascertaining what it is about pathological worry that may be specific to GAD (e.g. whether it is an anxiety-escalating thinking pattern that involves a series of uncontrollable ‘what if’ scenarios);

• Re-examining somatic symptoms, including symptoms of autonomic arousal, as diagnostic criteria for GAD, because this presentation seems to be common in primary care where many individuals with GAD seek help.

Concluding remarks

The GAD saga teaches us that when there is no progress in the understanding of psychopathology, things are better left unchanged. In the case of GAD, the creators of the DSM-5 indeed followed a premise that only changes supported by evidence should be made. Although the DSM-IV concept of GAD is largely flawed, at present there seems to be no alternative that is both evidence-based and acceptable to all key stakeholders. As uncomfortable as we may feel about keeping such a concept in the DSM-5, more damage would have been done by introducing arbitrary changes to the diagnostic criteria. In that sense, the status quo represents a good outcome, but not a triumph. Indeed, it may be a pyrrhic victory. On a positive note, maintaining the DSM-IV GAD criteria allows continuity between the DSM-IV and the DSM-5. This is important because of the changing diagnostic criteria for GAD in the DSM-III, DSM-III-R and DSM-IV, with the consequent uncertainties as to whether GAD diagnosed on the basis of different diagnostic criteria actually refers to the same condition. We know now that the DSM-IV GAD and DSM-5 GAD are the same ‘thing’, albeit not the one that the edifice of modern descriptive psychopathology can display with certitude.