DSM-5In the next few days the American Psychiatric Association is meeting in San Francisco and will announce the latest version of its diagnostic and statistical manual: DSM5.

Tom Insel, the Chief of the National Institute for Mental Health, created quite a stir last week by seeming to announce that the new diagnostic manual was an emperor with no clothes.

The questions that he raised are certainly important ones.  He pointed out that almost all of the diagnoses are based on descriptions of people’s mental processes or mental states, rather than a precise specification of an abnormal biological state (which you could get from an abnormal blood test, or a brain scan finding, for example).

However, as a number of commentators pointed out, this is not that unusual in medicine.

Even where we are able to go in and examine abnormalities in the body, diagnosis is most often not made by direct examination, but rather based on symptoms and tests which are suggestive, but not conclusive.  For instance, the diagnosis of a sore throat that requires treatment because it is likely caused by a bacteria, is most often based on symptoms and signs, rather than on a culture.

Tom was mostly saying that using this type of diagnosis is not good enough to rely on it in judging data about abnormal findings from research studies.

If depression is defined by the description of depression in DSM, and a researcher identifies a pattern of findings that are present in only some people with depression, then that finding may be viewed as unimportant.

We know that the current description of depression is inadequate, and almost certainly includes many different types of disorders. So, for research purposes we should be open to the possibility that biological findings may not perfectly relate to the current diagnostic criteria.

But for treatment purposes DSM 5, as imperfect as it is, is still the state of the art. And treatment guided by DSM 5 is going to be better than treatment guided by a clinician’s “hunch” about what is wrong.