The validity and utility of psychiatric diagnoses have long been a bone of contention between and within different professional and patient groups. This was clearly shown by the nearly 70 rapid responses to a 2001 BMJ article that proposed that post-traumatic stress disorder was a social construct with little clinical utility. ^{Reference Summerfield1} The reponses were emotive and polarised, with an equal proportion of patients and professionals in each camp: those who felt diagnoses were important and life-changing, and those who felt outraged and negated by the medicalisation of social suffering. In their recent editorial Craddock and Mynors-Wallis ^{Reference Craddock and Mynors-Wallis2} frame this diagnostic debate in terms of ‘benefits and limitations’; possible ‘disadvantages’ are acknowledged but mention of potential harms is conspicuously absent.

They advocate ‘embracing complexity’, but for the rest of their article this does not ring true. They reel off the standard list of apparent advantages to diagnosis - providing reassurance and reducing blame, shame and stigma - but without reference to research findings. (Nowhere in their paper is any patient-led or collaborative research cited.) Also conspicuously absent in their list is the necessity of a diagnosis to guide treatment. Is this a tacit acknowledgement that there is little evidence to support such a claim and that, in mental healthcare at least, ‘common factors’ linked to the therapeutic alliance, alongside extra-therapeutic factors, explain the majority of treatment variance? ^{Reference Bracken, Thomas, Timimi, Asen, Behr and Beuster3} In spite of this, they then go on to assert ‘there are no issues about diagnosis (or indeed treatments) that are unique to psychiatry’ (for the counter-argument see Bracken et al ^{Reference Bracken, Thomas, Timimi, Asen, Behr and Beuster3} and related correspondence).

Their erroneous linkage between diagnosis and stigma reduction stands out as particularly misleading. There is now an abundance of evidence, including a comprehensive review published last year in this journal, ^{Reference Angermeyer, Holzinger, Carta and Schomerus4} that biomedical framing of mental illness tends to increase personal and social stigma and public desire for distance.

The authors may counter that a diagnosis does not imply biological causality, and they seem to endorse the standard biopsychosocial frame of reference. The problem is, as Roland Littlewood ^{Reference Littlewood5} points out, it is more or less impossible to hold a ‘personalistic’ view of the self as agentic and intentional while at the same time subscribing to a ‘naturalistic’ view of being a product of biology, or even of the environment. One position always elides into the other. If this is true for professionals, it is certainly true for patients. And the dominant cultural understanding of diagnosis is that of biology, as it is with de facto psychiatric practice. ^{Reference Barrett6}

Craddock and Mynors-Wallis seem to want to be reasonable; identifying themselves, with other psychiatrists, as ‘reflective and tolerant of strongly opposing views and ideologies’. First, however, they resort to an unsubstantiated moral and emotive appeal to their position: ‘This can be to our patients’ disadvantage if we allow these views [i.e. critical of standard diagnostic practices] to be unopposed by suggesting that our patients are somehow less deserving of a psychiatric diagnosis than a physical diagnosis’. Then, just in case we are still equivocating, using the College’s Good Psychiatric Practice to bring us into line (as if this too was some ahistorical and acultural document), they pronounce: ‘This [use of standardised diagnosis] is not an issue of personal choice for a practitioner. It is a professional responsibility to the patient’. Their penultimate reference (entitled ‘Time to end the distinction between mental and neurological illnesses’) betrays their own ideological foray.

Of course, if diagnosis is understood in the broader sense of a thoroughgoing, descriptive and summative attempt at understanding a patient’s struggles, respectful of personal meaning and unblinded to issues of power and social context (the latter often being harder to change than biology, in which it may then of course be reflected ^{Reference Krieger7} ), then we too might endorse Craddock and Mynors-Wallis’s position. But in terms of a reverence to standardised manuals (whether DSM or ICD) that lack true nosological validity, even by their own standards, and whose utility is at best questionable, ^{Reference Timimi8} and which in effect serve to obscure key psychosocial antecedents, ^{Reference Krieger7} we would also argue that our patients deserve better.

There is little space for wider critique (for this, see Timimi ^{Reference Timimi8} ) and discussion of alternatives here, but if mature science is comfortable with dissent and debate (and indeed sees this as necessary for progression) this editorial seems a misplaced attempt to close down discussion - first through unsubstantiated emotive appeal, then through the threat of professional censure - in order to maintain a façade of professional consensus. While we might wonder what lies behind such a move, we would advocate a more far-reaching attempt at embracing complexity. In particular, as we have argued elsewhere, ^{Reference Bracken, Thomas, Timimi, Asen, Behr and Beuster3} in attending to issues of power, meaning, social context and the therapeutic alliance, alongside but not reduced to biology, we have much to offer the rest of medicine, which is also beginning to grapple with related issues. ^{Reference Das9,Reference Sharpe10}