Skip to main content

Verified by Psychology Today


A Critique of DSM-5

Is it a license to diagnose anyone with anything?

The American Psychiatric Association released a revision of its diagnostic bible in May, the first major rewrite in two decades. "The Fifth Edition of the Diagnostic and Statistical Manual of Mental Disorders," or DSM-5, is the official guidebook for diagnosing every conceivable psychiatric ailment. Last week, the Wall Street Journal published my op-ed in which I called attention to the loosening of diagnostic criteria in DSM-5 (June 27 2013, p. A21). For my op-ed in the Wall Street Journal, my editor and I were limited to just 850 words. Here I am going to take a bit more space to defend my assertion, in the closing line of my piece for the Wall Street Journal, that the new DSM-5 provides “a license to diagnose anybody with anything.”

If I had heard someone make that assertion two months ago – before I had actually seen the new DSM-5 – I would have thought that person was silly and uninformed. That claim is simply implausible. Why would the expert psychiatrists who are responsible for the new DSM make it so vague that anybody could be diagnosed with anything?

And I’m not saying that the old DSM-IV was perfect, or even satisfactory. It wasn’t. I am both a prescribing physician (MD) and a PhD psychologist. Over the past 25 years I have signed off on more than 90,000 patient visits. As a practitioner, the main problem I saw with DSM-IV was that the criteria were too vague. The vagueness of the criteria gave rise to enormous variability in diagnosis across different regions of the United States. The same disruptive behavior which would get a boy diagnosed with ADHD in suburban Philadelphia, where I have lived for the past 5 years, or in suburban Washington DC, where I practiced for 19 years, would get the same boy diagnosed as having Conduct Disorder or Oppositional-Defiant Disorder in Minot North Dakota or in Dallas Texas.

Making the right diagnosis is important, for many reasons, the most important of which is: If you make the wrong diagnosis, then you are less likely to prescribe the right treatment.

Like many practitioners, I eagerly awaited the publication of DSM-5 last month. I hoped that the Fifth Edition of the DSM would tighten up the criteria and dispel the vagueness.

What a disappointment. Far from dispelling the vagueness, DSM-5 takes fuzzy to a whole new level.

In previous editions of the DSM, you the patient had to meet certain specified criteria in order to be diagnosed for most conditions. For example, if I were going to diagnose you as having Schizophrenia, then you had to have specific symptoms, such as delusions or hallucinations. If you didn't have those symptoms, then I couldn't make the diagnosis of Schizophrenia.

Not anymore. DSM-5 introduces a new diagnosis, "UNSPECIFIED Schizophrenia Spectrum Disorder." The only required criterion is that you have some distress from unspecified symptoms, but you "do not meet the full criteria for any of the disorders in the schizophrenia spectrum and other psychotic disorders diagnostic class." You don't have to have delusions. You don't have to have hallucinations. In fact if you do have delusions and hallucinations, then you probably don't qualify for UNSPECIFIED schizophrenia. Get it? (You will find the new diagnosis in one short paragraph at the bottom of page 122 of DSM-5.)

One might claim that the new DSM-5 diagnostic category of “Unspecified Schizophrenia Spectrum Disorder” is not really new, because the old DSM-IV contained the diagnostic category of "Psychotic disorder not otherwise specified." But let’s look more closely at the difference between DSM-IV and DSM-5 in the way they use the word “Schizophrenia” and also the diagnosis of “psychotic disorder not otherwise specified.”

The DSM-IV diagnosis of “psychotic disorder not otherwise specified” could denote merely a transient psychotic break, perhaps in reaction to a colossal and unanticipated life stressor such as the sudden death of one’s own child, perhaps in a motor vehicle accident. That parent – babbling incoherently for a few minutes in the morgue as she stared at the disfigured corpse of her child – could meet criteria for a brief psychotic break, but she could NOT meet criteria for any form of Schizophrenia. DSM-IV did include the diagnosis of “Schizophrenia, Undifferentiated Type” 295.90, page 289 in the DSM-IV. However, the DSM-IV diagnosis of “Schizophrenia, Undifferentiated Type” REQUIRED that the patient meet Criterion A for Schizophrenia. Criterion A requires the sustained presence (over at least a one-month period) of two or more of the following:

  1. Delusions
  2. Hallucinations
  3. Disorganized speech (e.g. frequent derailment or incoherence)
  4. Grossly disorganized or catatonic behavior
  5. Negative symptoms, i.e. affective flattening, alogia, or avolition

(The above is the definition of Criterion A in DSM-IV, p. 285.) In other words: in DSM-IV, no form of schizophrenia may be diagnosed unless the patient meets Criterion A, above.

Now let’s look at a new diagnosis in DSM-5, “Unspecified Schizophrenia Spectrum Disorder”. Here is the description of that new diagnosis, bottom paragraph of p. 122 in DSM-5:

“This category applies to presentations in which symptoms characteristic off a schizophrenia spectrum and other psychotic disorder that cause clinically significant distress or impairment in social, occupational, or other important areas of functioning predominates but does not meet the full criteria for any of the disorders in the schizophrenia spectrum and other psychotic disorders diagnostic class.”

Note that the DSM-5 criteria for this new diagnosis are actually not criteria at all, but a LACK of criteria: the patient “does not meet the full criteria for any of the disorders in the schizophrenia spectrum and other psychotic disorders diagnostic class.” If the patient actually HAS had delusions and hallucinations for the past month, then the patient would probably meet criteria for Schizophrenia (p. 99 in DSM-5) and therefore would NOT meet the criteria for UNSPECIFIED Schizophrenia Spectrum Disorder. The criteria for Unspecified Schizophrenia Spectrum Disorder in DSM-5 do not state what “symptoms characteristic” must be present, or how long they must be present – a month, a week, a day, an hour, a minute? If one minute of incoherence caused “significant distress” to one’s spouse, then that would qualify the patient for the diagnosis. A patient who is babbling and incoherent for one minute in the presence of a spouse could conceivably be diagnosed with Unspecified Schizophrenia Spectrum Disorder. No required criterion for duration is mentioned in the new DSM-5 diagnosis of "Unspecified Schizophrenia Spectrum Disorder"; on the contrary, only the ABSENCE of criteria is mentioned.

And let’s consider the phrase “Schizophrenia Spectrum Disorder.” In DSM-IV, Schizophrenia is a disorder which you either have, or you don’t. DSM-5 revives the old theory (dating back to at least 1968) that Schizophrenia is a “spectrum” and that we might all fall somewhere on the “continuum” or “spectrum”. Such a notion is a theoretical conjecture about the nature of schizophrenia. The creators of DSM-IV – and particularly Allen Frances MD, chair of the DSM-IV Task Force – were careful to keep theoretical conjectures out of the diagnostic criteria for psychiatric disorders as defined in DSM-IV. The authors of DSM-5 show no such caution. Dr. Frances has written that the “APA flunked” in producing DSM-5 and that “quality control was surreptitiously canceled and a premature DSM-5 was rushed quickly to the printers to get sales moving and the cash register ringing”, so writes Dr. Frances in his book Saving Normal: an insider’s revolt against out-of-control psychiatric diagnosis, DSM-5, big pharma, and the medicalization of ordinary life (New York: William Morrow, 2013, p. 175).

The creation of “Unspecified Schizophrenia Spectrum Disorder” in DSM-5 is novel 1) in applying the term “Spectrum” to schizophrenia, which DSM-IV did not do; and 2) in allowing the application of the diagnosis “Unspecified Schizophrenia Spectrum Disorder” to a very brief, transient psychotic episode which under DSM-IV criteria would have been classified as a “psychotic disorder not otherwise specified.”

Heavy baggage attaches to the word “Schizophrenia” in the mind of the layperson, even the educated layperson. The phrase “psychotic disorder not otherwise specified” carries few terrors compared to “Schizophrenia.” For the non-psychiatrist, “Schizophrenia” implies a lifelong psychiatric illness, and very likely a lifelong disability, and almost certainly an illness requiring that the patient take antipsychotic medications indefinitely. “Psychotic disorder not otherwise specified”, the DSM-IV term, carried none of these connotations.

But what if your symptoms are too vague even to nudge the compass, however sluggishly, in any direction, toward any particular diagnosis whatsoever? What if you're just feeling down because you're not "living your best life," to borrow Oprah's phrase, and you want your doctor to fix the problem? No worries. The DSM-5 offers the diagnostic category "Unspecified Mental Disorder" (see page 708). The only requirement is that you "do not meet the full criteria for any mental disorder.

Let’s look at this a bit more closely. DSM-IV did include a diagnosis “unspecified mental disorder, 300.9”; here is that paragraph, in full (p. 687 in DSM-IV):

There are several circumstances in which It may be appropriate to assign this code [300.9]: 1) for a specific mental disorder not included in the DSM-IV Classification; 2) when none of the available Not Otherwise Specified categories is appropriate, or 3) when it is judged that a nonpsychotic mental disorder is present but there is not enough information available to diagnose one of the categories provided in the Classification. In some cases, the diagnosis can be changed to a specific disorder after more information is obtained.

Contrast that description with the description in DSM-5:

Unspecified Mental Disorder: This category applies to presentations in which symptoms characteristic of a mental disorder that cause clinically significant distress or impairment in social, occupational, or other important areas of functioning predominate but do not meet the full criteria for any mental disorder. The unspecified mental disorder category is used in situations in which the clinician chooses not to specify the reason that the criteria are not met for a specific mental disorder (p. 708, emphasis added).

The first line of the old DSM-IV description indicates that it is generally NOT appropriate to assign the diagnosis of “Unspecified Mental Disorder” unless one of three particular circumstances apply – one of which has to do with “Not Otherwise Specified”, about which I will say more below. And the last line of the DSM-IV description above suggests that the diagnosis should be “changed to a specific disorder after more information is obtained”. In other words: in DSM-IV, the residual category of “Unspecified Mental Disorder” should not be used often, and when used, should generally be a temporary designation while awaiting more complete details.

No such caution is found in the new DSM-5 category of “Unspecified Mental Disorder”, which reads completely differently from the DSM-IV category and therefore deserves to be designated as a complete break with DSM-IV. (When you compare DSM-IV with DSM-5, you find that in most cases the wording has been tweaked, loosened, and broadened, but not completely changed; in this case, however, the writers of DSM-5 have written a description which is completely new.) The DSM-5 category of “Unspecified Mental Disorder” contains no hint that this diagnosis should be applied cautiously, or that it should be changed to a more specific diagnosis as soon as additional information is warranted. Instead, DSM-5 specifically gives warrant to the diagnostician NOT to provide additional information, nor even to provide the REASON for not providing additional information. This invitation to vagueness is without precedent in DSM-IV.

How do the authors of DSM-5 justify such fuzziness? They invoke the analogy of high blood pressure and high cholesterol. As Dr. David Kupfer, chairman of the American Psychiatric Association's DSM-5 Task Force, and two of his colleagues wrote in the April 24 Journal of the American Medical Association, there is no sharp dividing line between normal and high blood pressure or between normal and high cholesterol, there is merely a "continuum of normality." A similar continuum, they argue, exists for mental health. "Thus DSM-5 provides a model that should be recognizable to nonpsychiatrists."

In other words, everybody's a little bit crazy, it just depends where you are on the "continuum of normality."

Sadly, this waffling is likely to contribute to a continued rise in mental-illness diagnoses in the United States. Over the past two decades, American culture has made almost any deviation from a continual smiley face into a psychopathology, especially for children.

As recently as 1994, it was rare for any child or adolescent in the U.S. under 20 years of age to be diagnosed as bipolar. But by 2003, it was much more common. There was a 40-fold increase in the diagnosis of bipolar disorder among American children and teenagers just between 1994 and 2003, even though the official diagnostic criteria for bipolar disorder did not change one iota during that interval. Most of the new diagnoses were for children under 15 years of age.

This phenomenon is peculiar to North America. During roughly the same time period in which the diagnosis of bipolar disorder was exploding for children in the U.S., the proportion of children under 15 diagnosed with bipolar disorder in Germany decreased.

DSM-IV often blurred diagnostic criteria by using the modifier “Not Otherwise Specified”. The DSM-IV modifier “Not Otherwise Specified” allowed diagnosticians too much wiggle-room, robbing diagnostic labels of meaning and specificity. I have already mentioned the article written by Dr. David Kupfer, Chair of the DSM-5 Task Force, and two of his colleagues, which appeared in the Journal of the American Medical Association, April 24 2013, pp. 1691-1692. In that article, Dr. Kupfer and his colleagues wrote that one of the big problems with DSM-IV was “the proliferation of residual diagnoses (ie, ‘not otherwise specified’ disorders) from DSM-IV.” Dr. Kupfer and his colleagues assure the reader that DSM-5 eliminates the vagueness in DSM-IV by introducing “definable thresholds that exist on a continuum of normality.”

Here Dr. Kupfer and his colleagues are both disingenuous and misleading. It is true that some diagnostic categories – not all, and certainly not Schizophrenia as noted above - but some diagnostic categories in DSM-IV were blurred by the use of the modifier “Not Otherwise Specified.” Dr. Kupfer et al. specifically invoke the “not otherwise specified” blurriness in DSM-IV and assert that DSM-5 is free of such vagueness because DSM-5 deploys “definable thresholds” for diagnosis.

That statement is simply false. Instead of making diagnosis more specific, DSM-5 enormously loosens diagnostic criteria, by introducing the new category of “Unspecified” for every diagnosis, including diagnoses such as Schizophrenia which previously could not be diagnosed unless very specific criteria were met.

Dr. Frances concludes his book Saving Normal with case histories of real people whose lives were seriously damaged because over-eager psychiatrists made the wrong diagnosis. As Dr. Frances shows: Once the term “Bipolar Disorder” or “Schizophrenia” appears anywhere in your medical history, even just once, it may be difficult or impossible for you ever to get a security clearance, or fly a commercial aircraft, or direct a day-care facility for young children, even if multiple subsequent psychiatrists write letters showing that the previous diagnosis was made in error.

DSM-IV was not strict enough; the “not otherwise specified” modifier was the worst example of the looseness of DSM-IV. Dr. Kupfer was aware of that concern and specifically mentioned it in his article for JAMA; but Kupfer et al. claimed that DSM-5 fixed the problem, when in fact DSM-5 made it much worse. Many disorders in DSM-IV with the “not otherwise specified” still had multiple required criteria; DSM-5 introduces the category of “Unspecified” with NO required criteria, for ALL diagnoses. That’s a profound and unprecedented change. As Dr. Frances observes, the loosening of diagnostic criteria in DSM-5 was “egregiously reckless” (p. 225) and the people in charge of DSM-5, such as Dr. Kupfer, were “incredibly indifferent to external criticisms by professional groups” (again p. 225). Dr. Frances concludes that “the DSM-5 fiasco has had one positive impact – alerting the press and public to the importance of getting psychiatric diagnosis right and the dangers of getting it wrong.” (p. 226)

Dr. Frances explores in detail the question of WHY the new DSM-5 is so reckless and vague. Why would organized psychiatry undermine its own credibility in so dramatic a fashion? That question is beyond the purview of this essay. Take a look at Dr. Frances' page on the Psychology Today web site, where he addresses this question from several different perspectives. My hope is simply that you will recognize the shortcomings of the new DSM, whether you are a patient, a practitioner, or a researcher.

Stay well. And if a practitioner makes a psychiatric diagnosis, don’t hesitate to insist on a second opinion.

Leonard Sax MD PhD FAAFP

PS: My op-ed for the Wall Street Journal was published with the title "'Unspecified Mental Disorder'? That's Crazy." You can read the full text of the op-ed at this link.

More from Leonard Sax M.D., Ph.D.
More from Psychology Today