Bereavement, complicated grief, and the rationale for diagnosis in psychiatry.

Recent controversies over the DSM-5 raise a foundational question for all diagnostic classifications: what is their underlying purpose? The author raises this question in the context of the proposed elimination of the "bereavement exclusion" from the DSM-5; and the possible addition of the category called "Complicated Grief." The author argues that our psychiatric diagnostic scheme should not be aimed primarily at establishing boundaries among putative "natural types." Rather, it ought to be guided by the principle of "instrumental validity, " which focuses on reducing the suffering and incapacity of those who seek our care. In so far as the category of "Complicated Grief" helps achieve this goal, it will foster the humane and ethical values that underlie medical practice.

“Well, while I'm here I'll do the work—and what's the work? To ease the pain of living.”

Allen Ginsberg

The thoughtful articles in this issue highlight many of the scientific and diagnostic questions surrounding the concept of “complicated grief.” Along with the contentious issue of the “bereavement exclusion” in the still-developing DSM-5, complicated grief (CG) raises important questions regarding the boundaries between “normal” and “abnormal” grief, between grief and post-traumatic stress disorder (PTSD), and between CG and major depressive disorder (MDD).

And yet, important though these “boundary” issues are, I sometimes wonder if we lose sight of the underlying philosophical and ethical foundation of why we have a diagnostic classification in the first place. Perhaps the obvious answer may be derived from the Greek etymology of the term “diagnosis”; literally, the word means “knowing the difference between.” We create categories in psychiatry in order to help us tell the difference between conditions we presume exist not only in our patients but, in some sense, in “Nature.” Here we recognize the implicit Platonic underpinnings of medical diagnosis: we aim, as Plato put it in the Phaeadrus, to “carve Nature at its joints.” Against this notion of “natural kinds” is the school of thought known as “constructivism” or “conventionalism.” This holds that “natural kinds” do not exist independently of our own conceptions and mental categories. As philosopher Andrew Bird[1] colorfully summarizes the conventionalist argument: “The classifications of botanists do not carve nature at its joints any more than the classifications of cooks.”

Psychiatrists have been debating the ontological status of their diagnostic categories for decades—famously or infamously so, in the controversial work of Thomas Szasz.[2,3] But I would argue that our diagnostic categories—including those of “complicated grief,” major depression, etc—ought to be aimed only penultimately at demarcating boundaries among clinical syndromes. Ultimately, in my view, our diagnostic categories ought to serve a humane and ethical purpose: to reduce the amount of suffering, incapacity, and misery among those who seek our help.

Diagnostic categories should be our servants, not our masters. If our diagnostic criteria fail to improve the lives of those we treat, it matters little how many biomarkers we have linked to a particular set of signs and symptoms; or how high our rates of “inter-rater reliability” may be. We will have failed our patients, nonetheless. I have referred to this ethical-pragmatic approach to diagnosis as one of “instrumental validity”[4] On this view, a set of diagnostic criteria has high instrumental validity insofar as it helps us reduce the suffering and incapacity—however specified—of those to whom the diagnosis is applied.

The issue of instrumental validity has been brought vividly to the fore in the intense controversy over the so-called bereavement exclusion (BE), and its proposed elimination from DSM-5.5 In essence, one group[6] has argued that the BE confuses clinicians and interferes with the diagnosis and treatment of potentially serious depression (my own position); while the other[7] has insisted that eliminating the BE will “medicalize normal grief” and lead to widespread overprescription of antidepressants.

While some of this debate clearly touches on “boundary” issues—eg, “Where do you draw the line between normal grief and major depression?”—the crux of the controversy rests on divergent claims regarding the instrumental validity of the BE. In effect, the contesting camps view elimination of the BE as either increasing or decreasing the net suffering and incapacity of our bereaved patients. The “non-eliminationists” fear that by unnecessarily “medicalizing” normal grief and thereby exposing patients to potentially dangerous medication side effects, we will do more harm than good.[7] The “eliminationists” believe not only that the BE lacks a sound scientific foundation, but also that it discourages recognition and treatment of a potentially lethal condition—MDD.[6]

Unfortunately, this disagreement cannot be resolved definitively on the basis of existing studies of post-bereavement depression, which are based largely on uncontrolled data, or on epidemiological observations derived from nonclinical samples.[8] We are sorely in need of carefully controlled, long-term, prospective studies of recently bereaved (2 to 8 weeks post-loss) patients with major depressive symptoms, compared with comparably depressed, nonbereaved patients. These cohorts would be compared with respect to morbidity, mortality, vocational function, and response to psychosocial and somatic treatment. As far as I know, such studies have never been carried out. That said, several lines of clinical evidence suggest that post-bereavement depression meeting symptom and duration criteria for MDD does not differ substantially from MDD after other types of losses, or after no loss at all.[6,9]

Roughly analogous controversies may arise with respect to the construct of CG. To be sure, Prof Shear's paper in this issue[10] convincingly makes the case for considering CG as a discrete disorder, distinguishable from both MDD and PTSD, despite substantial areas of overlap. And yet, critics will undoubtedly complain that still another psychiatric category is being created in the service of “medicalizing” grief—what Thomas à Kempis aptly called, “the proper sorrows of the soul.” A subset of those critics will, predictably, see the reification of CG as another example of “disease-mongering”[11]—no doubt arguing that it represents yet another attempt to create a market for pharmacological “treatment.” For these critics, there may be no scientific argument that will persuade them of the contrary.

Nonetheless, several papers in this issue make a convincing case for viewing CG as a legitimate diagnostic category, worthy of effective and compassionate treatment. This is so, not because CG necessarily “carves Nature at its joints”; but because it usefully identifies a very real instantiation of human suffering and incapacity[12] To the extent the construct of CG permits us to reduce such misery in our grieving patients, it will gain “instrumental validity” in the sense I have described. In short, by recognizing and treating this condition, we may “ease the pain of living” for those whose grief has gone painfully awry.