The ‘pragmatic’ secret of DSM revisions

The International Classification of Diseases, 11th Revision (ICD-11) is faced with the question of how it should coordinate with the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5). In the past, the opposite was also the case. As DSM-III was developed in 1980, its makers tried to ally it where possible with ICD-9 (originally established in 1975). In fact, in many psychiatric settings in the US, ICD-9 codes are still used for insurance reimbursement, rather than DSM codes, since all other medical specialties in the US use ICD codes.

This has been, in the US, a major defense of DSM revisions: they are needed for insurance reimbursement. But in fact, such American psychiatric exceptionalism seems unnecessary: we American psychiatrists should be able to use ICD codes just as all other American physicians do.

In reality, the radical third revision of DSM was developed in 1980 to put forward an agenda; initially there was some scientific impetus to that agenda – the laudable wish to base psychiatric diagnosis on scientific evidence that is valid (Decker, 2013). In other words, that we should practice the medical specialty of psychiatry based on scientific knowledge, not ideology. The ideology of the time was psychoanalysis, and thus the impetus was seen as being biological, which it was to some extent. This action met a reaction in the psychoanalytic establishment of the American Psychiatric Association: the result was a compromise we call DSM-III (Shorter, 2009). About a dozen scientifically valid diagnoses (as defined in the Research Diagnostic Criteria published in 1978) were expanded in two years to over 200 claimed diagnoses (as defined in DSM-III in 1980) including a whole category, personality disorders, which was created overnight. Most personality disorders had little to no scientific validation at the time, and still have very little, especially in comparison to the huge literature on dimensional nosology for personality, which continues to be rejected by DSM-5 nosology (Widiger and Lowe, 2008).

The defense was that even if the DSM-III diagnoses weren’t all scientifically valid, at least the new nosology was clear, and thus the diagnoses were reliable: we could all agree to define them a certain way. Then those reliable diagnoses could be further studied in research to see if they were valid or not; further revisions could then replace the invalid ones with more valid ones.

Over three decades later, we can say definitively that this wish has not come to pass. The fourth edition of DSM, DSM-IV (1994), was to the third edition like Stalin was to Lenin: an ideological hardening of the DSM categories, despite their poor scientific validity. The categories were expanded further to over 300, and the bar for making changes based on scientific changes was set so high that few changes were made.

Instead, a new philosophy for psychiatric diagnosis was put forward by the leadership of DSM-IV: ‘pragmatism’. Science was barely relevant anymore. Instead, what mattered were the ‘pragmatic’ consequences of defining psychiatric diagnoses one way versus another. Decisions were made by the DSM leadership based on what they deemed was good for patients or for the profession or for society (Frances, 2010). We can note for the moment that one might question who chose that leadership to have such power. A larger issue was that such pragmatic decisions need not have anything to do with what is true. Should we define hypomania as 4 days versus 2 days, or personality disorders as having or not having sexual abuse, or children as being diagnosed as having manic symptoms or not? On these and hundreds of other diagnostic issues, the central approach in DSM-IV was to ask the question: is it good for us and patients or not? The question was not: What is true?

This disregard for truth partly stems from disbelief. The DSM leadership reflected much of mainstream psychiatry, which reflects much of our culture: a postmodernist cultural zeitgeist has affected western society for about a century such that relativism about truth and skepticism about science have become as standard in a 55-year-old academic as in a 14-year-old teenager (Ghaemi, 2013). We devalue science and scientific truth; hence, it’s easy to be attracted to pragmatic sophistry in psychiatric diagnosis.

Yet, two decades after DSM-IV, we can say that such pragmatic nosology may well be a central reason why we have not had much progress in psychiatric knowledge and practice. We should ask ourselves this question: If we define our psychiatric diagnoses based on our professional and social preferences in 1980 or 1994 or 2013, why should nature go along with us? Why should the genetics or neurobiology or pharmacology of psychiatric diseases correspond with our cultural desires in a given society at a given time in human history?

This ‘pragmatic’ rejection of science is the deep conceptual problem with DSM-III and IV, and unfortunately, since it did not change in its core, DSM-5.

This is the key problem, in my view, with seeking to coordinate ICD-11 with DSM-5.

American psychiatry has put forward DSM-III through 5 based on certain cultural and social beliefs, and because of the influence of American culture in the world, DSM-III through 5 has overwhelmed psychiatric thinking about nosology throughout the entire world.

Instead of coordinating with a scientifically invalid nosology, world psychiatry would help the profession most by explicitly rejecting the American DSM ideology that pragmatism trumps science, and reversing the terms. Let’s seek scientific truth first and foremost. We don’t need to belabor whether we have ‘definitive’ scientific evidence (which is a way of ignoring science in favor of our ‘pragmatic’ preferences). We should commit to our best scientific knowledge at any moment in time, which is the way scientific medicine really operates. Knowing that science is a complex endeavor doesn’t mean we can feel free to replace science with our mere opinions under the guise of ‘pragmatism’.