Wednesday, March 30, 2016

Dr.Ghaemi on Dr. Spitzer

Nassir Ghaemi, MD has a commentary on Robert Spitzer, MD in this month's Clinical Psychiatry News.  After citing quotes by Shakespeare and John Adams to suggest that the dead are often idealized, he settles down to criticism based on whether or not the DSM-III helped or harmed the profession and Spitzer's role in that process.  Ghaemi comes down firmly on the side of harm because an unscientific approach to the diagnostic criteria for major depressive disorder has resulted in a lack of reliability and validity.  He uses the often quoted kappa score of 0.32 for diagnostic reliability of major depressive disorder in DSM-5 field trials as the main source of evidence, as well as the fact that the diagnostic criteria are unchanged since DSM-III.

Ghaemi suggests that his viewpoint is unique because unlike other eulogists, he had no personal connection with Spitzer and therefore can speak "in forthright recognition of fact from the impersonal perspective of another generation."  I am closer to Ghaemi's generation than Spitzer's and can make the same claim, but come to an entirely different set of conclusions.

I don't see Spitzer's efforts as being as corrosive as Ghaemi does, probably because I recognize the fact that there will never be a set of written diagnostic criteria that are perfect, based on science, and unambiguous.   But before I address the scientific, let me take on the rhetorical.  I would hardly blame Spitzer for the fact that the DSM criteria for depression have changed "hardly an iota" in the intervening 40 years since DSM-III.  Over that same time span there have been hundreds if not thousands of articles on the reliability of the major depressive episode diagnosis, as well as articles that analyze the symptoms according to that diagnosis.  There have been articles on standardizing various psychiatric and psychological instruments to detect major depression.  In fact, one of the rating scales basically copies the DSM criteria and asks the patient to rate on a 0 to 3 point scale - the percentage of days that they experience the symptoms. The PHQ-9 has become the standard for depression diagnosis in many primary care clinics.  There is also the fact that Spitzer's original DSM-III effort resulted in much higher reliability figures - a kappa of 0.72 to be exact (2).

There is also the issue that there have been two intervening Task Forces for DSM-IV and now DSM-5.   The Chair of the DSM-IV Task Force has since become a prominent critic of the DSM process and psychiatry in general.  I may have missed it, but at the time that Task Force was convened, I did not notice him or other members advocating for major changes to the major depression diagnostic criteria.  These are supposedly the top minds in the field.  Highly motivated academics with one axe or another to grind.  The idea that everyone would defer to Dr. Spitzer based on his original approximate efforts seems unrealistic to me.  More than a few people would have noticed his bungled and unscientific approach.

My major problem is using a single reliability figure as the grounds for this criticism.  Every year outpatient based psychiatrists can see up to a thousand new people a year.  They may find that up to 50% of those patients have had a life-long sleep disturbance.  Many can recall nightmares and sleep terrors as children.  Another 20-30% will have generalized anxiety or social anxiety since childhood.  In some there will be a performance based anxiety that is comorbid with the social anxiety.  Another 10-20% will have post-traumatic stress disorder to some degree.  About one-third will have a significant substance use problem.  These percentages will vary by clinic location and referral base.  The majority will be referred for a diagnosis and treatment recommendations for depression.  A substantial number of people with depression have comorbid anxiety and anxious temperaments.  I don't think it is a stretch to say that on any given day, many of the identified depressives will identify themselves as primarily anxious.  It is not unexpected to find that many patients don't really understand the difference between anxiety and depression or they will overtly say that they are the same problem - indistinguishable from one another.  Unless there is a clear differentiating factor like a manic episode, the postpartum state, or psychotic symptoms I would not expect that anxiety and depression are distinct disorders for most people.  At the minimum anxiety might morph into depression, but in most cases they are coexisting chronic conditions.  A low kappa in this situation should be expected and not a shock.

Does that mean that psychiatrists should be wringing their hands and blaming Spitzer for it?  Neither response is appropriate.  Psychiatrists are highly successful in diagnosing and treating mental illness, not because of a DSM manual, but because of clinical training.  When it comes to anxiety and depression there are no known ways to parse all of the symptomatic possibilities.  The human brain is designed to realize all of the possible combinations of human experience.  Why would we expect it to be different when it comes to experiencing anxiety and depression?  The only chance that a psychiatrist has to make sense of the world is a number of patterns of diagnoses based on their training and practice experience that they can match against the patient they are currently seeing.  These patterns guide the diagnosis and treatment plan.  A clinically astute psychiatrist is not plowing through the interview to see if the patient "meets criteria".  A clinically astute psychiatrist carefully attending to the patient's conscious state and trying to figure out how they can be helpful.  That includes figuring out the real problems and prioritizing them in a complex matrix psychiatric and medical problems.  None of that flows from the DSM and none of that resembles research based on lay people interviews using DSM criteria.

In closing, any commentary on Dr. Spitzer should include his role in eliminating homosexuality from the diagnostic manual.  This detail and how it occurred is never taught to residents.  I had to learn it from public radio many years after residency.  This detail is significant any way you cut it.  It invites criticism that monolithic psychiatry is currently moving too slow in other areas or that monolithic psychiatry was just responding to public pressure.  There is also criticism directed at Dr. Spitzer for a paper based on self report that was withdrawn years later on this same issue.  There are always advocacy groups seeking publicity by their own spin on the issue.   In my opinion, none of that diminishes that significant achievement that put psychiatry four decades ahead of most people in the United States.  Say what you will about the DSM, that accomplishment alone is enough.  I am thankful that Dr. Spitzer was open minded enough to listen to the advocates and eventually side with them.              

George Dawson, MD, DLFAPA

1:  Nassir Ghaemi.  Commentary:  Dr. Robert L. Spitzer - An impersonal appraisal.  Clinical Psychiatry News.  March 2016. p 12-13.

2:  Riskind JH, Beck AT, Berchick RJ, Brown G, Steer RA. Reliability of DSM-III diagnoses for major depression and generalized anxiety disorder using the structured clinical interview for DSM-III. Arch Gen Psychiatry. 1987 Sep;44(9):817-20. PubMed PMID: 3632255.


  1. Ghaemi is one of the folks who disagrees with the duration criteria for many mood disorders. Granted they are somewhat arbitrary, but they are put in there for a reason. The DSM Committees have rightly refused to monkey with them.

    BTW, Hamilton, the author of the HAM-D, had himself stated that his instrument is invalid for distinguishing anxiety and depression.

    Another big reason for lack of reliability in diagnosis is crappy psychiatric interviewing technique by some of the involved clinicians.

    1. "Ghaemi is one of the folks who disagrees with the duration criteria for many mood disorders."

      Unfortunate - the single most important criteria in the DSM is the duration necessary for a manic episode.

  2. I'm with Ghaemi on this one. He could have said, "Except for the major disorders described by Feighner, psychiatry for the most part treats symptoms, not diseases or any easily definable categorial disorders". That would have been intellectually honest and conform to the actual experience of being a psychiatrist.

    Instead he went for the money and prestige of trying to shoehorn it all into a disease model in DSM. And even people like Jeffrey Lieberman now believe that “the book precisely defines every known mental illness”. You're talking a sensible nuanced approach but the KOLs are doubling down on overstatement.

    It's great that Spitzer tossed out homosexuality as a mental disorder. However, in an uncloseted homosexual, the diagnosis has a perfect reliability kappa. He excluded on the basis of lack of construct validity, a concept he failed to apply to such things as adjustment disorder.

    1. A couple of opposing points:

      Depression was one of the major disorders proposed by Feighner. The basis for that proposal is outlined in this reference: - It includes the table: "TABLE 1. Historical Origins of the Symptomatic Criteria for Major Depression: Criteria Proposed 1950–1980" - with the research papers on which it is based. Feighner's 1972 paper is in column 3.

      That historical a article BTW, points out that the real tension at the time had nothing to do with money - it was clearly a battle between the Washington University group and their emphasis on diagnosis and psychoanalysts who wanted nothing to do with diagnosis. The origins or a depression diagnosis go back to 1950.

      I don't recall any discussion of construct vaildity and homsexuality. The MPR piece linked to my post on the subject suggests that he made the decision based on the number of functional and successful psychiatrists he encountered at the GayPA meeting at the time. I would appreciate any academic reference discussing the reasons because I clearly missed it.

  3. That was my point about reliability. He and the KOLs (not you) tended to use reliability to make their point but homosexuality when it was a disorder had a high kappa. The issue of course was construct validity.

    I certainly agree that major depression with melancholia is a valid psychological construct. I have loads of problems with mood disorder and adjustment disorder which not only have validity issues, but reliability issues.

    In other words, he could have stopped at 15 or so, but now we have 265 "precisely defined mental disorders" according to the Lawrence Kolb professor of psychiatry at Columbia and former APA president or at least his ghostwriter. Which every practicing psychiatrist in the community deep down knows is a load of bull.

    1. How is it possible that Lieberman could say something like that?

    2. I would imagine he has said a lot of things. Here for example in a joint statement he endorses it as the "best information available" and acknowledges that researchers currently need more than that.