How many psychiatric diagnoses fit on the head of a pin?

To the Editor

Trying to understand the nature of reality is as old as Plato and as new as quantum mechanics. Scholastic debate in medieval universities crystalized two opposing views that inform current disagreements about the meaning of psychiatric diagnosis.

The ‘realists’ argued that we directly perceive reality; the ‘nominalists’ that we can only name and classify things but never really know their essence. Even though they seemed to have common sense on their side, the realists turned out to be wrong. All the evidence from physics and psychology confirms that we construct our realities rather than just sensing them. As the Talmud says: ‘We don’t see things as they are, we see things as we are’.

Freud was a nominalist. At the core of his theory of the mind was the Darwinian and Kantian notion that we cannot help but distort external reality under the influence of inner drives and intrapsychic structures.

In sharp contrast, the pioneering classifiers and biological psychiatrists of 50 years ago were determined and convinced realists. They assumed they were on the hunt for the causes of psychiatric diseases, confidently assumed to be real entities fully equivalent to diabetes or cancer. Indeed, there was a fevered debate before publication of the Diagnostic and Statistical Manual of Mental Disorders, Third Edition (DSM-III) in 1980 about whether to label its syndromes ‘mental disorders’ or ‘diseases’. ‘Disease’ lost out, not for want of fervor on the part of the realists supporting it, but only because the American Psychological Association cried foul and threatened to boycott DSM if it made what they regarded as an unsupported and prejudicial terminological and conceptual switch.

Fifty years of subsequent research confirms that DSM is no more than a catalogue of syndromes – social constructs, not diseases. What we call ‘schizophrenia’ is a heterogeneous grab bag of symptoms and behaviors that overlap bewilderingly with its near neighbors, both in clinical presentation and on biological markers. There will not be one or a group of schizophrenias, but rather a small army of hundreds of different forms of pathogenesis and presentation. ‘Schizophrenia’ is just a name, not a thing – a useful social and medical construct but also one that carries considerable baggage and risks. ‘Schizophrenia’ will gradually be eaten away or replaced altogether as we gradually find a better way of organizing the data and naming its more fundamental constituent parts.

The DSM and ICD (International Classification of Diseases) have their uses, but no one should worship them or fight for the priority of one system over the other. The two are no more than different languages, or, more precisely, different dialects or pronunciations of the same language. Their similarities are striking, their differences trivial. The mental disorders they define are our provisional way of looking through a glass darkly and will be replaced by closer approximations of reality as we develop better tools and more knowledge.

I don’t trust clinicians who slavishly follow DSM/ICD diagnostic checklists and lose sight of the psychosocial factors that always play a large part in the onset, maintenance, and treatment of psychiatric problems. As Hippocrates put it 2500 years ago: ‘it is more important to know the patient who has the disease than the disease the patient has’.

But I also don’t trust clinicians who are ignorant, or dismissive, of DSM/ICD diagnostic criteria. They are idiosyncratic outliers who have fallen out of touch with current diagnostic consensus in the field and make decisions based on personal bias, uncorrected by consensual validation. It is best to know and use DSM/ICD, but not see it as a substitute for a rounded view of the patient.

The US National Institute of Mental Health is basing its research agenda on a new way of classifying clinical phenomena. Rather than study the biological correlates of complex and heterogeneous syndromes, the targets will be research domains that are much simpler. This system has no clinical relevance now and probably will not influence clinical practice for decades to come, if ever at all.

So, for now, and for the foreseeable future, we are stuck with a diagnostic system that has clinical limitations and has been something of a research dud (Malhi, 2013). The good news, though, is that the system does work reasonably well to inform treatment recommendations and to predict prognosis (Sachdev, 2013). We should certainly not despair over the fact that neither DSM nor ICD is ‘excellent’ and in the process ignore the fact that, for most clinical purposes, they are good enough. We may not understand psychiatric disorder very well, but fortunately we have reasonably effective treatments for it. The rest of medicine is in pretty much the same boat.