How To Avoid Medicalizing Normal Grief In DSM5

March 16, 2010

The recently posted draft of DSM5 makes a seemingly small suggestion that would profoundly impact how grief is handled by psychiatry. It would allow the diagnosis of Major Depression even if the person is grieving immediately after the loss of a loved one. Many people now considered to be experiencing a variation of normal grief would instead get a mental disorder label.

The recently posted draft of DSM5 makes a seemingly small suggestion that would profoundly impact how grief is handled by psychiatry. It would allow the diagnosis of Major Depression even if the person is grieving immediately after the loss of a loved one. Many people now considered to be experiencing a variation of normal grief would instead get a mental disorder label.        

       Undoubtedly, this would be helpful for some people who would receive much needed treatment earlier than would otherwise be the case. But for many others, an inaccurate and unnecessary psychiatric diagnosis could have many harmful effects. Medicalizing normal grief stigmatizes and reduces the normalcy and dignity of the pain, shortcircuits the expected existential processing of the loss, reduces reliance on the many well established cultural rituals for consoling grief, and would subject many people to unnecessary and potentially harmful medication treatment. 
       Grief is an inescapable part of the mammalian experience and a necessary correlate of our ability to attach so strongly to other people. Although grief is universal, there is no one right to grieve. Different cultures prescribe a wide variety of different behavioral and emotional reactions and rituals. Psychiatry needs to tread lightly and have compelling reasons before encroaching with its own rituals on such time honored and usually effective practices. 
       Within a given culture, normal individuals also vary enormously in the content, symptoms, duration,and impairment of their grief and in their ability to draw consolation and sustenance from others. There is no bright line separating those who are experiencing loss in their own necessary and particular way from those who will stay stuck in a depression unless they receive specialized psychiatric help.
       The numbers on each side of the normal/mental disorder divide are probably very lopsided. Ever since the origin of our species, every human has had frequent occasions to grieve. Almost all of us come to terms with the loss and the altered conditions of a new life without the benefits of psychiatry-- and do just fine on our own. The change in DSM5 would attempt to identify the very small percentage of people who have a complicated grief that goes beyond the average expectable in severity, symptom pattern, and duration - - those who would not remit as part of the natural evolution of their grief. But when you use a big shovel to capture a small needle in the haystack, what you mostly get is hay. Any change in the way DSM5 defines grief may gather a very large proportion of false positives who would do better avoiding psychiatric help.
      The rationale given by DSM5 for its radical proposal is brief, cryptic, and fails to provide anything like a risk/benefit analysis of potential effects. DSM5 states that there is no evidence that the stress of losing a loved one is different than  other stressors and thus no justification to withhold the diagnosis of major depression after a loss. This  rationale places the burden of proof in the wrong direction. DSM5 should make so consequential a change only after a careful and considered evaluation proves with compelling evidence that it will do more good than harm.
      Such evidence is simply not available. The research in this area is interesting but in very early development and we don't know many essential things. We have no idea how any proposed criteria set would work in the general population. What percentage of grieving individuals would get the diagnosis (especially once drug companies raise awareness of it)? Among the people who would be diagnosed, we don't know what percentage truly need psychiatric help, what percentage would do better without it.
     In their recent commentary, Pies and Zisook have  gone far beyond the meager DSM5 rationale to present the strongest possible case for allowing the diagnosis of Major Depression in grief situations. They cite several lines of argument:
1. There is a clinical need--  some individuals have severe, complicated grief that looks just like severe Major Depression and that does not get better spontaneously. The longer diagnosis and treatment are delayed, the greater their suffering, impairment, and risks (eg, job loss, injured relationships, lowered treatment response, suicide).
2. The loss of a loved one is not essentially different from the many other serious stressors that abound in life. 
3. It is impossible to predict the future misuse of the DSM5 system, so we should make decisions based only on the best possible science.  
4. The criteria for complicated grief could be tightened to reduce false positives. (They suggest 2 useful ways described below and I add 2 others).
5. Education can solve the problem of false positive diagnosis and the risk of providing medicine in milder cases when time, support, and/or psychotherapy would be more indicated.    
       The  excellent proposal made by Pies/Zisook to reduce false positives could be strengthened even further if 2 additional exclusions were added (#'s 1 and 2 below) that they suggested. The entire package differentiating grief from depression would require:
1. An extended duration of 1 month.
2. A particularly severe presentation that includes some combination of unreasonable guilt, worthlessness, hopelessness, self-loathing, anhedonia, a focus on negative memories of the departed, alienation from others, and inability to be consoled.
3. To recognize the different cultural expressions, the diagnosis of depression would not be made if the person's grief is within  cultural norms.
4. An exclusion could be added that would take into account the person's own past experience of grief and its previous outcomes. If the individual previously had severe grief symptoms, but recovered spontaneously (without going on to a major depression), this would suggest they are now grieving their own way and do not require diagnosis or treatment.

      DSM5 has made many poorly thought through suggestions that can be fairly easily dismissed. Though I continue to disagree with the  Pies/Zisook proposal, it is reasonable and deserves serious consideration. Here are the opposing points:
o Re clinical need: In appropriate cases displaying clinically significant impairment, distress, or risk, the diagnosis Depression Not Otherwise Specified covers their false negative problem.
o I believe there is a difference between losing a loved one and most other life stressors. This  explains why grief is the universal target of communal healing rituals. It would be unfortunate for psychiatry to prematurely roam into problems usually better handled by family and other cultural institutions. Cultural biases would be very hard to surmount in making this diagnosis.
      My disagreement with Pies and Zisook is strongest on this point. All decisions for DSM5 should follow the injunction-"First Do No Harm." Although it is impossible to predict precisely how any DSM5 change will eventually play out once the manual is in general use, that doesn't reduce DSM's responsibility for the problems that occur-- even if they are unintended. All potential risks have to be thought through and factored into a thorough risk/benefit analysis.  The argument that we should just go to where the science takes us ignores that the science is (as they point out) not definitive, is subject to different interpretations, and is not readily generalizable from research to real world settings. Once the genie is out of the bottle and DSM5 makes it easy to diagnose depression in grief situations, this could easily become an industry propelled fad diagnosis.
o The tightened criteria would help reduce, but certainly not eliminate, the grave potential harm caused by the massive misidentification of false positives.False positives and excessive treatment are not be a problem for skilled and cautious clinicians (like Pies and Zisook), but in the real world most of the prescriptions will be written by primary care physicians who have 6 minutes with each patient, don't know the fine points of the criteria sets,and want the fastest solution. The false positive problem is too unknown and potentially far too large to ignore. At a minimum, there would need to be field trials to determine prevalence, reliability, false positive, and false negative rates. I doubt that DSM5 has the time, money, and skill to pull this off.
o It would be naive and unwise to rest our hopes that any educational program would reduce overdiagnosis and the overprescription of medication in grief situations. To the contrary, most of the education would go the other way. The drug companies devote enormous resources to "educating" physicians to be quick on the draw in prescribing medication 

   I respect the arguments made by Pies and Zisook and believe they  work well when applied by experts like them. My worry is the misuse of even reasonable ideas in the real world  situations where most diagnosis and treatment is done. Loose diagnostic and treating habits could lead to the widespread medicalization of grief well beyond what Pies and Zisook would themselves recommend.
 There are 2 ways to avoid this.

o The first is to keep things as they are and not to diagnose Major Depressive Disorder in the first 2 months after the loss of a loved one. 
o The second is for DSM5 to allow for the diagnosis of complicated grief but with a criteria threshold set high and including all 4 protections against false positives outlined above. As recommended by Pies and Ziskook there should also be a  physician and public education campaign normalizing normal grief and sharply delimiting the small group of grievers who need psychiatric
  Weighing the pluses and minuses, my call is to keep things as they are and not risk an "epidemic" of psychiatric grief.  


Disentangling Grief and Depression: Rejoinder to Dr Frances

We appreciate Dr Frances’s thoughtful rejoinder to our essay, and we recognize that his disagreements with us are motivated by goals we all share -- ensuring the well-being of those who seek our help, and reducing the suffering and incapacity of seriously depressed patients.

The last thing we would want to do is pathologize "normal grief"-- just as we know that Dr Frances would not want to "'normalize" bona fide Major Depression. We all agree that finding the specific line demarcating "normal" depressive feelings and thoughts (depression with a small "d") from clinically meaningful Depression (with a big "D") is often challenging, even for the most astute and experienced clinicians. We also agree that the ideal research studies defining the full spectrum of normal depression through pathological Depression, and demarcating the boundaries of these conditions, have not been done. Thus, the best we can do is to be guided the best available data.

However, we respectfully disagree with Dr Frances on what the "best available evidence" tells us. We stand by our previous review1 of the most pertinent recent data, which informs us that continuing the bereavement exclusion (BE) for the diagnosis of major depression does more harm than good. The BE ignores much of this recent evidence, 2-4 and we agree with the DSM committees' pithy but astute rationale for discontinuing the BE; ie, that there is no credible evidence that bereavement-related depression is  different (in severity, course, morbidity, co-morbidity, consequences or treatment response) from other, non-bereavement related instances of major depressive episodes.

Notably, the International Classification of Diseases (ICD-10) also omits the bereavement exclusion. While rates of clinically-serious depression may be somewhat higher when ICD-10 (versus DSM-IV) criteria are used, we are not aware of any evidence demonstrating an “epidemic” of “psychiatric grief” resulting from the use of ICD depression criteria.5

We agree with Dr Frances that the honored maxim, “First, do no harm” should be one important guiding principle behind the DSM5’s categories.  However, we may differ with Dr Frances on how “harm” ought to be defined and assessed in this context. We believe that the harm that may ensue if ordinary grief is “over-diagnosed” as major depression pales in comparison with the harm that may follow a missed diagnosis of major depression.

If a patient with ordinary grief or bereavement is (mistakenly) given a diagnosis of major depressive disorder (MDD), the attendant “risks” are mainly those of entering the mental health system and receiving treatment-- whether psychotherapy or medication. Both “talk therapy” and medication may entail risk, and we are well aware of the unpleasant and sometimes harmful side effects antidepressants may produce in a minority of patients. However, given that completed suicide is the outcome in roughly 4% of patients with MDD--averaged for outpatients (2%) and inpatients (6%)6--we believe that far greater harm will befall our patients if the BE is retained. This would encourage practitioners to “extrude” from the category of MDD many patients whose symptoms of major depression occur within 2 months of a recent loss-- for which, we believe, there is insufficient scientific basis. Some evidence even suggests that use of the BE may “backfire”-identifying more severely depressed individuals than controls with major depression but without bereavement.4

Rather than “back-loading” specificity in diagnosing MDD (by tacking on the bereavement exclusion), we believe it is wiser to “front-load” specificity by tightening the initial entry criteria for MDD; for example, by increasing the minimum duration from 2 weeks to 3-4 weeks. In our experience, many bereaved and non-bereaved individuals presenting with acute depressive symptoms will show substantial improvement after the initial two weeks of distress and dysfunction.

One issue that bears further discussion is Dr Frances's conflation of 3 related but distinct constructs: grief, complicated grief, and Major Depression. Grief is the constellation of feelings, thoughts and behaviors following loss.7 As Dr. Frances states, grief is normal and probably also adaptive. It is not a “disorder” that needs treatment. Most recently, Complicated Grief has come to designate a syndrome of intense and prolonged mourning that is not adaptive and indeed is associated with chronic and persistent suffering and disability.7 Complicated grief usually does require treatment. Major Depression is different from both Grief and Complicated Grief.7

Like grief, Major Depression may be precipitated by the death of a loved one. And, like grief, Major Depression often involves intense sadness, problems sleeping, concentrating, eating and interacting with others. But unlike ordinary grief, Major Depression involves these and additional symptoms occurring together, most of the time, for at least 2 weeks, accompanied by clinically significant distress or impairment.

We have also described aspects of the patient’s conscious experience (“phenomenology”) that often help distinguish major depression from uncomplicated grief. For example, in ordinary grief-- unlike in major depression --the bereaved person may experience pleasant memories of the deceased; believe that the grief will eventually “pass”; and have a sense that “better times are ahead.” Unlike the person with MDD, the person with ordinary grief does not typically have the sense of being a “worthless” or “unlovable” person. These and other distinguishing phenomenological features are discussed in several papers.7,8

Grief is a "normal", mammalian, adaptive, and perhaps even instinctual reaction to loss; major depression is a non-adaptive, clinically meaningful, disorder. Because both grief and depression may follow a major loss, they often travel together-- and each may make the other worse. Treating depression does not suppress or prevent grief, nor does it interfere with "grief work"; if anything, treating depression allows grief to take its course; facilitates mourning; and may even help prevent complicated grief from occurring.7,8

Finally, the issue of “over-diagnosis” of depression cannot be considered without noting the abundance of evidence showing that depression is frequently under-diagnosed-particularly in primary care settings, where only about a third of depressed patients meeting PHQ-2 screening criteria are correctly diagnosed.9

We applaud the DSM5 committee’s proposal to separate the 2 distinct constructs of uncomplicated grief and major depression by eliminating the bereavement exclusion. We believe this will facilitate appropriate diagnosis and treatment of Major Depressive Episodes occurring in the context of bereavement or other recent loss. In our view, this will mean not only “doing no harm” but also doing genuine good for our patients.

1. Pies R, Zisook S. DSM5 Criteria Won’t “Medicalize” Grief, if Clinicians Understand Grief.
2. Zisook S, Kendler KS. Is bereavement-related depression different than non-bereavement-related depression? Psychological Medicine: 1-31, 2007
3. Zisook S, Shear K, Kendler K. Validity of the bereavement exclusion criterion for the diagnosis of major depressive episode. World Psychiatry. 2007 June; 6(2): 102–107. Accessed at:
4. Corruble E, Chouinard VA, Letierce A, et al. Is DSM-IV bereavement exclusion for major depressive episode relevant to severity and pattern of symptoms? A case-control, cross-sectional study. J Clin Psychiatry. 2009 Aug;70(8):1091-7.
5. Guerra M, Ferri CP, Sosa AL, et al. Late-life depression in Peru, Mexico and Venezuela: the 10/66 population-based study. Br J Psychiatry. 2009 Dec;195(6):510-5.
7. Zisook S, Shear K. Grief and bereavement: what psychiatrists need to know. World Psychiatry. 2009 Jun;8(2):67-74
8. Pies R. The anatomy of sorrow: a spiritual, phenomenological, and neurological perspective.Philos Ethics Humanit Med. 2008 Jun 17;3:17
9. Ani C, Bazargan M, Hindman D et al: Depression symptomatology and diagnosis: discordance between patients and physicians in primary care settings. BMC Fam Pract. 2008 Jan 3;9:1.