DSM-5 and the elimination of the major depression bereavement exclusion

As the American Psychiatric Association approved the final version of DSM-5, the Chair of the DSM-IV Task Force Allen Frances declared it to be the ‘saddest moment’ in his career, saying that the final version included ‘changes that seem clearly unsafe and scientifically unsound’. He went on in a blog (Frances, 2012) to list the 10 most potentially harmful changes, amongst which was the decision to eliminate the grief exclusion criterion from the definition of major depression in DSM-IV, arguing that the loss of the criterion is likely to lead to ‘medicalizing and trivializing our expectable and necessary emotional reactions to the loss of a loved one and substituting pills and superficial medical rituals for the deep consolations of family, friends, religion, and the resiliency that comes with time and the acceptance of the limitations of life’. As a proposal, this was so controversial that it precipitated a statement justifying this decision from Kenneth Kendler, published on the DSM-5 website (Kendler, 2012). What was a proposal is now an approved change, the implications of which we will have to deal with until the next DSM edition.

To examine the impact of this change it is first useful to consider the exact criterion that has been removed. This is as follows: ‘The symptoms are not better accounted for by Bereavement, i.e., after the loss of a loved one, the symptoms persist for longer than 2 months or are characterized by marked functional impairment, morbid preoccupation with worthlessness, suicidal ideation, psychotic symptoms, or psychomotor retardation’. Kendler argues that there are many problems with the criterion and that the best solution was to remove it.

To consider the implications of abandoning the criterion, the following questions have to be asked. First, has evidence accumulated that the criterion was useful and valid? Second, did clinicians understand and use the concept as it was designed? Third, was it useful in research settings? Fourth, are there administrative consequences of abandoning the exclusion, such as implications for insurance or private health care schemes?

Research evidence

Five studies examining aspects of the DSM-IV bereavement exclusion criterion have been frequently cited. Kendler et al. (2008) examined the characteristics of bereavement-related depression compared with those of depression related to other life events and reported no difference. Kendler argues, on the basis of this and other evidence, that the specificity of the exclusion to ‘bereavement’ is illogical (www.dsm5.org). Corruble et al. (2009, 2011a, 2011b) examined a large number of patients with symptoms of depression and compared those who had been given a diagnosis of DSM-IV major depression and those with depressive symptoms who had been excluded on the basis of the bereavement criterion. Physicians in this study were specifically not trained in the exact use of the criterion in order to examine its naturalistic use in clinical practice. It could be argued that this gives a critical insight into how clinicians use the bereavement caveat. Compared with the included group, the excluded group were similar on most measures including symptoms, response to treatment and cognitive function. However, an important feature of the study was that in a predominantly general practice setting, the baseline Montgomery–Asberg Depression Rating Scale score was around 30, a level of symptomatology which is surprisingly high and may imply that patients were self-selected based on severe symptomatology. The same could be said of the Kessing et al. (2010) study in which more than 60% of the patients were inpatients. Only the investigation by Karam et al. (2009) studied a community sample and DSM-IV bereavement-excluded patients. Characteristics were similar between those with depression and bereavement-excluded patients, but with the caveat that there were relatively few of the latter. In sum, there is little evidence of a difference between bereavement and other life events and their effects on mood, and little to suggest that there are systematic differences between excluded and included patients. However, there is a dearth of evidence regarding the characteristics of more mildly symptomatic bereaved patients who may be included under the diagnostic category of major depression when the criterion is abandoned.

Current clinical practice

None of the above may matter unless clinicians currently use the exclusion criterion. On this issue there is little evidence but we will speculate from our own practice and the observed practice of colleagues. First, in our experience, many clinicians have (perhaps strictly speaking in error given the exact wording) usually interpreted the criterion as referring to a wide range of losses and always seen their job as being one of attempting, based on a comprehensive history, to judge whether the symptoms expressed are within the bounds of a ‘normal’ reaction to a particular stressor in a particular person – or a more pathological process likely to respond to specific treatments for depression. Similar judgements regarding the distinction between psychopathology and ‘normal response’ to stressful events have been made on a daily basis by clinicians in Christchurch (NZ) following the recent earthquakes.

Second, if milder symptoms persist for longer than 2 months, clinicians have usually interpreted this criterion loosely and if they thought that the reaction a person was having was relatively normal, both in terms of symptoms and duration, then they would not diagnose major depression regardless of the duration criterion. This is particularly relevant when considering variations in the duration of appropriate mourning behaviour expected across different cultures.

Third, Kendler argues that the criterion states ‘that someone who has experienced a recent bereavement (ie within 2 months) is not eligible for a diagnosis of major depression’ – unless, of course, they meet one of five ‘conditional symptoms’ (functional impairment, morbid preoccupation with worthlessness, suicidal ideation, psychotic symptoms, or psychomotor retardation). We believe that most clinicians have never been significantly influenced by the exact time frame and nature of symptoms in this way. If a clinician felt that someone had sufficiently severe symptoms, regardless of the exact nature of the symptoms, within 2 months of bereavement then they would certainly have made a diagnosis of major depression. Indeed, we would argue that in someone with bipolar disorder, for instance, it would not be uncommon for an episode of depression to be precipitated fairly rapidly by bereavement.

Research

It is necessary to use strict criteria in research and in this setting the removal of bereavement may result in an increase in the diagnosis of depression – although many people suffering from bereavement are unlikely to present to depression studies. However, there is increasing recognition of the variability of the phenotype of major depression and large trials or observational studies which include people with recent losses may be examined in order to attempt to answer some of the very questions which make this issue so problematic. Do people with symptoms of major depression, who have suffered a recent life event, behave differently in treatment or differ in other important ways from those who have not? What is recent? Analyses including ‘life event scores’ and temporal distance from a life event may add significantly to our understanding of the relationship between stresses such as bereavement and depression. This is, in our opinion, likely to be a more useful approach than the exclusion of cases on the basis of arbitrary cut-off times.

Administrative issues

As always, perhaps it is in the hands of administrators that changes to diagnostic systems can be most problematic. However, at least this one seems to go in the right direction. It allows more people to be potentially eligible for services because of a diagnosis of major depression, without the possibility that a computer or an insurance clerk will raise the issue of a recent bereavement and disqualify them on this basis. It puts the power in the hands of the physician, who can code distress as depression, or not, depending on a careful assessment.

Conclusion

Major depression is associated with adverse life events, including bereavement. Depression does not always require treatment and often does not require pharmacological treatment while bereavement typically does not. However, clear criteria to guide clinicians regarding when a normal reaction to loss becomes pathological and requires treatment are lacking. What evidence there is suggests that the DSM-IV criterion was relatively unhelpful in this regard. Some would argue that abandoning it risks over-diagnosis and treatment of depression but we believe that with appropriate training and common sense, clinicians will make a reasonable judgement regardless of the criterion, adopting, as Kendler suggests, ‘a conservative watch and wait approach’ in many cases of bereavement. The abandonment of the criterion, in our opinion, may aid research into the relationship between life stresses and into the correct management of depression complicated by bereavement. It may also avoid the exclusion, by insurance companies, of patients’ genuinely needing treatment because of the bereavement exclusion.

Clinicians must continue to assess patients both in terms of symptomatology and in the context of their life and cultural experiences, and use this to guide appropriate and sensitive treatment. DSM-5 will provide a framework to guide but not to dictate treatment and clinical diagnosis.