The history of psychiatry is replete with the expectation of being on the cusp of scientific breakthroughs [1]. Miracles are always around the corner. Occasionally, they even happen: lithium, antipsychotics, antidepressants were all developed by pharmaceutical companies and/or solitary psychiatric researchers; the mechanisms of those agents were discovered by academic basic scientists; psychoanalysis was discovered by a Viennese private practitioner.

It may be doubted whether DSM-III and IV ever led to any major discovery in psychiatry. Despite obvious advances within psychopharmacology [2], genetics [3], and neuroimaging [4], these fields have not produced impactful (i.e., replicated, large effect size) achievements in relation to psychiatric diagnoses, using DSM definitions for their clinical phenotypes, over the last three decades [5-7].

This stagnation in psychiatry since the introduction of DSM-III may not be mere coincidence. We tend to blame the brain: the brain is complex, it is said, hence our decades of failed nosological genetics and imaging and pharmacology studies.

Perhaps we should blame ourselves: Why should Nature order its genetics and pathophysiology and pharmacology along the lines of what the interest groups of the American Psychiatric Association may wish? Could it be that the DSM-III and IV revisions (as well as ICD-10) hindered, rather than aided, scientific progress?

If so, one might rightly wonder, with DSM-5 and similarly ICD-11, whether our future can be any different than our past.

Readers will forbear some explanation:

There are always and only two parties in the universe, said Ralph Waldo Emerson: the party of the past and the party of the future; the Establishment and the Movement [8]. There is no doubt that DSM-5 (and IV and III) is the Establishment; there is also no doubt it represents the past of psychiatry, since 1980, when DSM-III was born. The question is whether DSM-5 has a future, and whether there is any Movement in psychiatry that will promote progress for the profession.

Doctor Gordon Parker makes an extensive critique of the scientific rationale for DSM-5's various judgments on mood illnesses [9]. Doctor Jan Fawcett, chairman of the DSM-5 Mood Disorders Task Force, graciously provides commentary on Doctor Parker's essay [10]. The most important aspect of Doctor Fawcett's commentary, in my view, is his frank and honest acknowledgement that the party of the future is not going to be found in DSM-5 or its predecessors:

‘So, OK, maybe the whole concept of DSM definitions of psychiatric disorders is obsolete – maybe it has served its purpose and now it's time to move on. This is my personal opinion, but in the meantime, we need something….’

We seem to agree – the three of us – that DSM-5 is our past, not our future. The question now is: How can the future be born, without the past suffocating it?

The problem is all about ‘in the meantime’.

The solution to the problem may grow from this observation: over time, the threshold for the amount of science needed to make important changes in DSM definitions has been getting higher and higher. For example, in 1980, a huge radical decision was made in mood illnesses: Kraepelin's very broad unitary mood diagnosis (manic-depressive illness), which had been a century-long consensus, was cut into two separate and different diagnoses (bipolar disorder and unipolar depression, renamed ‘major depressive disorder’) [11, 12]. The science that existed to make this change was based on the work of Karl Leonhard and his followers [13], as well as two major studies by independent European researchers Jules Angst and Carlo Perris [14, 15]. Three decades later, Jules Angst has analyzed two more generations of prospective data from the same cohort, with the opposite conclusion: the bipolar spectrum should be broadened, unipolar depression narrowed, and all mood conditions seen as spectra, somewhat analogous to Kraepelin's original view [16-18]. This conclusion, by the same researcher whose data were considered good enough for radical change in DSM-III, is not considered good enough for less radical change in DSM-5. Even Angst's minor suggestion to reduce hypomania duration from four days (which is based on zero data) to two days (based on decades of follow-up data) [19, 20], though apparently endorsed by Doctor Fawcett's task force, was rejected by higher-level DSM-5 and APA leadership, because of the fear of ‘overdiagnosis’, a scientifically debatable claim [21].

Based on these non-scientific ‘pragmatic’ concerns, the DSM process has become super-conservative: A rivulet of change today, if unpopular with powers that be, is rejected, despite having much better evidence than was sufficient for an ocean of innovation in 1980. The same researcher's study (Angst's Zurich cohort), with more and better data three decades later, is rebuffed in favor of older findings that the same study now disproves.

This is not a matter of not having enough science. It is a matter of rejecting the science we have.

This history can only be understood by appreciating the deeper problem with keeping DSM-5 ‘in the meantime’: one senses a certain lack of conviction to acknowledgements that our nosology should be based on our best science. Repeatedly our best science has been overruled, explicitly as documented in the citations[22, 24, 25] by the DSM-III, IV and 5 leadership. This wasn't intended in 1980 [22, 23], but it has come to be, and it has been stated most explicitly by DSM-IV leaders [24] and endorsed by DSM-5 leaders [25].

Stated succinctly:

The DSM approach to diagnostic decision-making is based on ‘pragmatic’ considerations of what is best for the profession or the public, in the estimation of DSM and APA leaders. If scientific evidence, to the best of our knowledge at any point in time, is not deemed to be in the best interest of the profession or the public, then it is a matter of fact, not opinion, that the final judgments in the DSM process are made by the American Psychiatric Association (APA) Board of Trustees and other groups (such as the APA General Assembly), not by scientific experts.

To put it another way: DSM-III in 1980 claimed that reliability (a common dictionary) would lead to validity (scientific truth) [26]. DSM-IV in 1994 gave up on validity (‘pragmatism’ overruled contemporary science) [24]. DSM-5 in 2014 admits the end of validity, but claims we have nothing better. A common language is seen as enough; whether it is true or false has become less and less relevant.

This is not to say that DSM-5 did not incorporate any changes based on sufficient scientific evidence. The removal of the antidepressant-induced mania exclusion for the diagnosis of bipolar disorder is an example of a DSM-IV item with zero scientific evidence that had been clearly disproven [27]. To its credit, DSM-5 removed that exclusion. Still, Doctor Parker demonstrates many examples where the science of mood illnesses is ignored or overruled [9]. Another major example involves personality: over 50 years of thousands of studies have clearly proven the validity of dimensional definitions of personality traits, as opposed to a much smaller and poorly validated literature on the eight personality disorders of DSM-III [28]. The DSM-5 personality task force recommended that change toward dimensional diagnosis; the APA Board of Trustees rejected that recommendation on ‘pragmatic’ grounds, despite the clear scientific support for dimensions. DSM-5 personality disorders are completely unchanged from DSM-IV, which were largely unchanged from DSM-III. Since 1980, DSM has changed not a whit on personality, despite a very large and definitive scientific evidence base that indicates otherwise.

It is not enough to be scientific sometimes, and not other times. If we are to have progress in psychiatry, we cannot pick and choose when we wish to be scientific. We have to be scientific all the time. Maybe the DSM process is, as some say, like science policy rather than science, a political process informed by science but sensitive to current social demands. If so, then let it be said to be so. And then, it should be entirely legitimate, in fact preferred, to ignore DSM when conducting scientific research, as the NIMH leadership has proposed [29].

Our ends cannot be different than our means. The answers for the long and short run are the same. If our long-term goal is a scientific profession, then our short-term methods have to be scientific. There is always enough science to be scientific; it may not be ideal, it may not be as much as will convince everyone, or even most people. But some science is better than no science.

If we are born 30 years too early now, as Doctor Fawcett comments, to revise DSM-5 based on sufficient science, then we must have been born 30 years too early back in 1980 when DSM-III was written. And we will still be born 30 years too early three decades from now, because the future is predestined to be like the past if the present does not change.

We always have enough science to make medical and nosological decisions. The best science we have, at the present time, is the science we should use. Instead, the DSM approach is to make the excuse that our science is not good enough now. So let's wait a few decades. No medical discipline based on science operates this way. Psychiatry, by accepting this attitude, makes itself less scientific than the rest of clinical medicine.

Thirty years is a generation. Generations are terrible things to waste.

The scientific attitude differs from the messianic: the truth is always in front of you, and you see it not. You may view it distortedly, but mistakes are always forgiven, for scientific truth is corrected error, and, pursued sincerely, such scientific work has produced true medical miracles for humankind. Psychiatry, by rejecting this approach, guarantees a future no different than our recent past; it guarantees that we will continue to wait for miracles that will never come, and that more generations will waste their efforts, instead of working productively, with science as the only — not merely an optional — criterion. Miracles in medicine do not just happen; they are brought about by an honest science which toils for years —occasionally despised, sometimes rejected — refusing to ignore its results just because they are unpopular.

We may already have miracles in front of us, but, blurred by DSM, we do not see them.


  1. Top of page
  2. References