Problems, Explanations, and Solutions

This was first posted on DXSUMMIT.ORG : A global platform for rethinking mental health on Wednesday 22nd May 2013

Politicians and the powerful often want you to think that “there is no alternative”. And one of the curious things about psychiatric diagnosis is that people seem to have great difficulty in contemplating an alternative way of doing things … despite the fact that alternatives are ubiquitous and straightforward.

As I set out to write this blog, I thought I would elaborate and detail the ‘problem-list, formulation’ approach. I was slightly deflated when I re-read my own entry from earlier – because it all seems quite clear. In the vision of services that I have in mind, we would eschew diagnosis and instead work as a multidisciplinary team to identify a person’s key problems, develop a case formulation – a set of working hypotheses about what might link the person’s problems, what might have caused them, and what might help – and plan care.

I must admit that it seems pretty straightforward to me, and I don’t see any need to add the unnecessary diagnostic element. What I mean by this is that there is no need to add – in the key summary statement above – the idea that we should also ‘take those key problems, apply arbitrary and unscientific rules for their supposed co-occurrence, make unscientific assumptions about underlying disease processes, add stigmatizing and misleading labels, and begin a process of obfuscation and biological reductionism’. I suggest we just identify the key problems, develop a case formulation and plan care.

This doesn’t seem in any way incompatible with a close working relationship with medical colleagues, and it doesn’t preclude medical explanations – or diagnoses – when these are appropriate. But it is a good basis to plan care, both for individuals and at a service level.

As I said in earlier pieces, there’s work to be done. We know that diagnoses are inadequately poor tools for identifying causal mechanisms, predicting prognosis, planning care – either for individuals or as policy-makers – or communicating, precisely because of their lack of reliability and validity. We also know that people have difficulty understanding this; indeed, a recent blogger suggested that diagnoses are necessary for communication “even if they’re wrong” - a really worrying indication that the malevolent effects of diagnostic thinking stretch as far as IQ!

People seem to worry that dropping the language of diagnosis means that we’ll be left in a vague soup of narrative accounts, making it difficult to give robust advice on “what works” for whom and when, making it impossible to communicate in clear, precise, language and impossible to plan. I don’t think that at all. We have a lot of work to do – psychiatry has been dominated by the ‘diagnosis-treat’ approach for so long that we will need to develop new systems and structures.

We will need a new lexicon of problems – it wouldn’t be realistic to say that we all agree on the most important social and psychological problems today, nor on the most efficient and effective ways to describe them. So we’ll need to put some work into that – remembering, of course, that diagnostic manuals require time and effort (and resources) to be developed. Our work will have to be democratic rather than secretive, and will have to be led by real people, service users, rather than the professionals, who so frequently have vested financial and ‘guild’ interests.

It’s important, once again, to stress that identifying problems is not the same as making diagnoses. One online commentator, trying energetically to defend DSM-5 (and, in my opinion posing more questions than she answered) suggested that the three new disorders of ‘Binge Eating Disorder’, ‘Excoriation Disorder’ and ‘Disruptive Mood Dysregulation Disorder’ were valid because she comes across the problems of binge eating, skin picking and temper tantrums in her clinical practice. I have no doubt she does. But while identifying and responding to these problems is absolutely a clinician’s job, it is unhelpful to think that adding the word ‘disorder’ after an identified problem solves it. As we’ve argued before, these ideas tend to lead away from genuine understanding towards the ludicrous idea that “your daughter is having all these temper tantrums because she suffers from disruptive mood dysregulation disorder” – a move away from explanation towards a combination of circular thinking and bio-reductionist thinking. And, of course, we fear that medication will naturally follow.

In these cases dropping the language of disorder merely makes sense. What it removes, here, is the assumption of underlying pathology, the danger of circular thinking, and the risk of medicalization. These are huge benefits (as we’ve previously argued), but the language of problem - binge eating, skin picking and temper tantrums – is clearly simpler (there is none of the additional complexity critics fear). The implications for clinical practice, for research, and for service planning, of a simpler and more straightforward language should be obvious.

In other areas, such as ‘schizophrenia’, these same benefits accrue, but there are others, too. As well as adding additional and unnecessary complexity (and we might well conclude that the language used is designed specifically to create a professional mystique – ‘excoriation disorder’…), making assumptions about underlying pathology, subtly promoting biomedical thinking and medical responses and falling into circular logic, many diagnoses combine individual problems into clusters believed to reflect ‘syndromes’ (with presumed, and usually biological, underlying pathology). These combination rules are, as we know, hugely problematic, fail to reflect the ‘joints’ of nature and leave clinicians frequently unable to agree. They also mean that the very many possible combinations of different problems means that there is a confusing mess of huge numbers of diagnosis. The DSM project, of course, was designed to rationalize and simplify the system. But we’ve been left with more, not fewer, diagnoses with worse, not better, reliability statistics.

Finally… treatment… Well, the idea is, of course, that diagnosis should guide treatment. Unfortunately we also know that treatment (as well as being best when guided by a more integrative formulation) is best when based on problems rather than diagnoses.

A little detail might be helpful. I have a client (a real person, but someone I will try to avoid identifying). Mary is a 56 year old woman. She has some moderate learning difficulties (finding it challenging to achieve the qualifications commonly received at the end of high school, struggling to pass a driving test) and some problems understanding the emotions, intentions and behaviours of other people (what is commonly called the autistic spectrum). She has never worked (she does some voluntary work at her local supermarket, stacking shelves and general cleaning) but lives entirely supported by state benefits. Nevertheless, she has married (to a man who also has some mental health issues), and lives independently with their dog. During her late teens and early 20s, she was very troubled by intrusive auditory hallucinations, but these have largely ceased to trouble her. It’s noticeable that these voices are still present, but Mary has learned to live with them.

These problems – the moderate learning difficulties, the autistic spectrum problems, and the auditory hallucinations – became significant recently as the UK Government implemented a programme to reduce the cost of benefit payments to people with long-term disabilities, meaning that Mary and her husband were assessed with a view to reducing their benefits (the alternatives, of course, being “get a job” or “tell the panel quite how crazy and inadequate you are”). Mary was anxious, her sleep suffered, she talked more about her auditory hallucinations, and even discussed taking her own life.

Mary’s problems would meet diagnostic criteria for moderate learning difficulties, autistic spectrum disorder and (probably) schizophrenia. In practice, the first two diagnoses are relatively uncontroversial in themselves – they have relatively high reliability and validity, probably because both relate to known underlying causal mechanisms. The diagnosis of ‘schizophrneia’ is much more problematic. It’s unreliable (two clinicians disagree) and invalid (there isn’t a coherent ‘syndrome’ in the real world that maps onto this diagnosis). It’s also very difficult to know how to address the fact that Mary’s problems come and go and she copes with them to a greater or lesser extent from time to time. Most importantly, in the traditional diagnosis-driven world, we have seen two significant clinical problems. First, different parts of the healthcare system disagree as a result of these diagnostic codes – should she be cared for by the learning disability or adult mental health services? More worryingly, in the first service, the voices are largely ignored, in the second the voices are seen as inevitable given her learning difficulties. The diagnoses don’t help treatment; they get in the way. But finally, to respond with the intelligent and humane package of care that Mary needs to service this current stressful time, we have had to battle against a diagnostic model – voices are symptoms of ‘schizophrenia’, so the underlying illness should be treated… or ignored, since Mary has a learning difficulty. The idea that her experiences are understandable for what they are is a difficult one to punch through the collective ignorance of a diagnosis-driven, reductionist system.

I am not trying to say that traditional, diagnostic, psychiatry cannot either describe Mary’s problems adequately or respond to them humanely. But I am saying it would be easier and more appropriate to respond with a problem-formulation approach. Mary’s problems can be described in great detail or very simply. To say that she has long-standing problems with moderate learning difficulties and autistic spectrum problems, and (as a result of specific stressors) now has problems with auditory hallucinations strikes me as parsimonious and elegant. Invoking the concept of ‘schizophrenia’ adds nothing that the problem-based approach can offer, and instead leads to potential confusion and inappropriate and unnecessary assumptions.