Grief and Depression: When Science and Terminology Get Confused

In his ongoing critique of the DSM-5 process, Dr Allen Frances started a brushfire recently in challenging the DSM-5 Mood Disorders Work Group proposal to remove the bereavement exclusion from the diagnostic criteria for a major depressive episode. Here’s a summary of the debate.

Dr Frances led off with a New York Times op-ed piece arguing that the DSM-5 Work Group proposal to eliminate the bereavement exclusion for the diagnosis of major depressive disorder (MDD) would lead to the overdiagnosing and medicalizing of normal grief.¹ Letters followed, including one by Kenneth Kendler, David Kupfer, and Carol Bernstein, who stated that that “scientific evidence shows that there are no systematic differences between individuals who develop major depression in response to bereavement and those who develop depression following other severe stressors...”² In a Psychiatric Times blog post, Dr Frances retorted, in effect agreeing that normal grief may at be times indistinguishable from MDD, but that the two are nevertheless different.³

Dr Ronald Pies then responded to Dr Frances’ blog with a multitude of points, but with a central argument for the safety of false negative MDD diagnoses over the risk of false positives.⁴ Finally, in another post, Dr Frances suggested a compromise with Dr Pies, agreeing that extending the length of time before diagnosing MDD from 2 to 4 weeks would reduce the number of false positives, but continuing to express less concern about the danger of false negatives.⁵

One very relevant study left out of this discussion was the 2008 study by Kendler, Myers, and Zisook: “Does bereavement-related major depression differ from major depression associated with other stressful life events?”⁶ From a scientific perspective, this study might be seen as definitive. MDD looks like MDD wherever you see it: following loss of a loved one, following a divorce, following loss of a job, following whatever stressful life event, or following none at all. The jury of empirical evidence is rendering its verdict on the side of removing the exclusion. Science prevails.

But here’s in fact where things get more complicated and science seems to confuse us. The problems emerge in terminological distinctions that appear thoroughly clouded. The confusion begins with the DSMs. “Uncomplicated bereavement” in DSM-III, and “bereavement” in DSM-IV,may or may not reach the level of MDD. The instruction from DSM-IV is, in its contorted language, not to diagnose the bereaved person with major depression if the MDD symptoms are mild, even though meeting criteria of a major depressive episode, but to diagnose MDD if the symptoms are severe.

On top of this confusion generated by DSM-IV, researchers have introduced other terminology. In the Kendler and colleagues study just cited,⁶ the authors identify a “bereavement-related depression” group, of which 28% meet criteria for “uncomplicated bereavement” or “normal grief,” in order to show that the bereavement-related depression does not differ from a control group of individuals with depression related to other stressful life events, of which 25% meet criteria for “normal grief.” In this study “normal grief” means, following DSM-IV,not ordinary, non-pathological grief, but rather major depression of the milder sort. The individuals in both groups not meeting “normal grief criteria” (also similar in their symptom characteristics) fail to meet “normal grief criteria” because they are in fact worse, with more severe MDD symptoms. This study attempts to follow the contorted logic of DSM-IV but differs in that “uncomplicated bereavement” in DSM-III (and “bereavement” in DSM-IV) describe a population that may or may not meet criteria for major depression, while in the Kendler study the individuals meeting “normal grief criteria” presumably all meet criteria for MDD. At the end of the article, the authors describe bereaved people suffering from “normal sadness” who are presumably not part of their bereavement-related depression group.

More terminological variation occurs in Zisook and Shear’s, “Grief and Depression: What Psychiatrists Need to Know,”⁷ in which the authors strive to provide clear descriptions (and distinctions) of normal, uncomplicated grief, complicated grief, and grief-related major depression. Their study may be the best effort yet to map out the terrain of grief and depression, but it is difficult to match up their terminology with that described above. Since DSM-IV bereavement may meet criteria for major depression, some of DSM-IV bereavement would map onto “grief-related major depression” and some would map onto either of the other 2 categories. On the other hand, the bereavement-related depression group from the Kendler and colleagues study would all fall into Zisook and Shear’s grief-related major depression group, because all of the bereavement-related depression individuals (by my understanding) met criteria for MDD-the “normal grief” group with milder symptoms, the rest with more severe symptoms.

What everyone agrees on (except the authors of DSM-III and DSM-IV), is that bereaved people meeting criteria of MDD should be diagnosed with MDD and treated accordingly. Where the confusion occurs is over how to sort out sub-MDD grief from grief with MDD-and how to deal with the former. Dr Frances is concerned that removing the bereavement exclusion will result in over-diagnosing and over-treating non-pathological grief by (presumably) labeling it MDD. His opponents respond that we are able to distinguish non-pathological grief, sub-threshold-MDD grief with depression, and grief meeting criteria of MDD. This debate is somewhat artificial in that it stems so much from the confusing terminology and categories of DSM-IV. If the manual had categories reflecting those of Zisook and Shear, the debate would go away.

We need a category for ordinary grief; one for grief that reaches a level of MDD; and a third, mid-range category for depressive reactions to the loss of a loved one that are more than ordinary grief and less than MDD. The current category of “Bereavement as a V code” would do fine for the first category-if the DSM-5 authors would not muddle it by throwing in major depressive episodes as a possible manifestation. We already have major depression to cover grief that becomes-major depression. So all we have left to cover is the mid-range, what Zisook and Shear call complicated grief-minor depression not reaching the level of MDD.

The DSMs have had trouble defining such a state. Depression NOS, adjustment disorder with depressive features, and dysthymia (with its emphasis on chronicity) are the current candidates-and none fit very well. If we worked this out, I don’t think we would be having the debate over the bereavement exclusion, and what at first looked like a conflict over the science of grief and depression would turn out to be a muddle created by clumsy and confusing terminology.

References1. Frances A. Good Grief.New York Times, August 14, 2010.
2. Kendler KS, Kupfer D, Bernstein CA. Letters to the Editor. New York Times, August 19, 2010.
3. Frances A. Good grief versus major depressive disorder. Psychiatric Times. http://www.psychiatrictimes.com/blog/dsm-5/content/article/10168/1645888. Accessed September 15, 2010.
4. Pies R. Response to Frances. Psychiatric Times.http://www.psychiatrictimes.com/blog/dsm-5/content/article/10168/1645888. Accessed September 15, 2010.
5. Frances A. A possible compromise on grief vs depression. Psychiatric Times. http://www.psychiatrictimes.com/display/article/10168/1656471. Accessed September 15, 2010.
6. Kendler KS, Myers J, Zisook S. Does bereavement-related major depression differ from major depression associated with other stressful life events? Am J Psychiatry. 2008;165:1449-1455.
7. Zisook S, Shear K. Grief and bereavement: what psychiatrists need to know. World Psychiatry. 2009;8:67-74.