Imperfect: there are significant limitations to current classification

All can agree that psychiatric diagnosis is not perfect. ^{Reference Kendell7-Reference Insel, Cuthbert, Garvey, Heinssen, Pine and Quinn11} A fundamental issue in psychiatry is that current classification schemes are of clinical syndromes. Diagnosis is based on descriptive data elicited from clinical observation rather than measurements that relate directly to brain function and pathology. This arises because our understanding of the brain, particularly its higher functions, is less well understood than for many other areas of medicine where, of course, far less complex organs are the focus of interest.

It is important to note that there are differences between the DSM classification system (used in the USA) and the ICD system used in the rest of the world. The DSM is underpinned by a research culture that has sought to have diagnoses as homogeneous as possible for the investigation of treatment and prognosis. This inevitably excludes many patients who do not meet strict diagnostic criteria and creates the need for multiple ‘comorbid’ diagnoses when, for example, patients with a depressive disorder are also diagnosed as having a range of anxiety disorders. The problem has been stated amusingly and clearly by Dr Steve Hyman, a former Director of the National Institute of Mental Health: ‘DSM… has the miraculous bad property of being too broad and too narrow at the same time’ (cited in Aldhous & Coghlan). ^{Reference Aldhous and Coghlan12} The ICD has always been more pragmatic by being less rigid about diagnostic criteria. It facilitates sensible clinical diagnoses but at the expense of including within the categories more heterogeneous patient groups.

Psychiatrists should not be defensive that the diagnostic categories alter over time - this is no different to alterations in the definition of, for example, hypertension and diabetes. Also, as with hypertension and diabetes, many of our disorders represent an extreme of variation within the population. And, as for blood glucose level or diastolic blood pressure, this does not mean that the extremes do not cause significant problems for individuals or that they do not require careful assessment and treatment.

Impersonal: diagnosis alone is not sufficient

The higher functions of the brain are complex and reflect the interplay of diverse processes from the molecular to the social. It is, therefore, natural to expect that brain dysfunctions will reflect this rich complexity and defy simple, reductionist models (be they molecular, psychological, social or any simplistic admixture). To most psychiatrists, this is one of the great appeals of the specialty - the challenge of helping a patient by embracing complexity and using pragmatism to guide the help that we provide. A good psychiatrist (as any good doctor) must be aware of the limitations of the diagnostic categories used and decisions made. It is important to think beyond diagnosis and make assessments of causation, the range and pattern of symptoms, severity and impairment in functioning (sometimes called ‘domains of psychopathology’ ^{Reference Craddock and Owen13} ). The key is to produce a formulation that goes beyond a simple list of facts and that complements the diagnosis by including information about important clinical variables that have relevance for the management plan. So for bipolar disorder, this might include such variables as the lifetime preponderance of mania v. depression, tendency to develop suicidal ideation, lifetime experience of psychotic features, inter-episode functioning (including cognitive), co-occurrence of panic/anxiety, co-occurrence of alcohol and substance misuse, history of triggering of mood instability by antidepressants, etc. Together with assessments of the patient’s life circumstances, relationships, experiences and personality, these detailed clinical assessments feed into the formulation and management plan.

Important: diagnosis is a key component of high-quality, safe care

Diagnosis is a key part of how we communicate with our patients and each other. ^{Reference Craddock, Kerr and Thapar14} Indeed, in any situation in which more than one intervention is available some form of classification/diagnosis is needed to guide logical decisions about which intervention is better (or whether no intervention is the optimal option). If an adult presents with lethargy, weight loss, reduced activity and reduced interested in life, we need to distinguish between such diverse possibilities as (a) temporary adjustment to a changing life situation (i.e. understandable normal response to life’s difficulties), (b) cancer, (c) heart failure, (d) severe depressive episode with immediate risk of suicide. Diagnosis is a key to determining which interventions will be of value and underpins the entire opus of evidence-based practice.

The importance of diagnosis for patients is growing rather than decreasing. ^{Reference Craddock, Kerr and Thapar14} A diagnosis provides reassurance that their situation is not unique, mysterious or inexplicable and that there is a body of knowledge and experience that can be brought to bear in providing help. A diagnosis can help reduce inappropriate feelings of blame (perhaps parents believing that their parenting is at fault in a child diagnosed as having autism). It can reduce stigma by explicitly acknowledging the presence of illness (and, thus, that the feelings or behaviour cannot be dismissed as character weakness or bloody-mindedness). Diagnosis can help an individual make sense of being different or of not functioning like most others and challenges the feelings of shame, loneliness and low self-esteem that may otherwise occur if they are simply treated as being ‘odd’, ‘bad’ or having a ‘character flaw’. Diagnosis helps to communicate information between public and professionals about the support and service needs (including, where needed, speech and language therapy, occupational therapy, schooling support, etc). Diagnosis allows patients and carers to talk with, and get advice from, people with similar problems (for example, through support organisations such as UK Bipolar), advocate for improved services for particular groups of patients, (for example, the recent Schizophrenia Commission report ¹⁵ ) and be able to read the most relevant educational and self-help material. Individuals and families may actively seek to be diagnosed with certain diagnoses such as Asperger syndrome as a way of both helping them to understand themselves and also seek therapeutic interventions.

Psychiatric and non-psychiatric illness

It is important to be clear that there are no issues about diagnosis (or indeed treatments) that are unique to psychiatry. ^{Reference Kendell7} As psychiatrists we have a tendency to be more reflective and tolerant of strongly opposing views and ideologies than in many other medical specialties and spheres of human endeavour (and, indeed, more tolerant than among some non-medical professions within mental health). This can be to our patients’ disadvantage if we allow these views to be unopposed by suggesting that our patients are somehow less deserving of a psychiatric diagnosis than a physical diagnosis if for example they had broken their leg. Despite the caveats we have given, using operational diagnostic classifications psychiatrists can make diagnoses more reliably than cardiologists diagnosing myocardial infarction based on clinical assessment. ^{Reference Farmer, Williams and Jones16} Of course, what psychiatry awaits is specific tests, such as electrocardiograms and cardiac enzymes, that cardiologists use to confirm or refute diagnoses based on clinical assessment alone. For all the well-recognised limitations of operational diagnostic categories, many define highly heritable clinical entities that have allowed the identification of risk genes that provide a clue to understanding pathophysiology. ^{Reference Craddock and Owen13} Further, in psychiatry we have many treatments that have been developed and tested against diagnoses that compare very favourably in efficacy with those for non-psychiatric illnesses. ^{Reference Leucht, Hierl, Kissling, Dold and Davis17}

Improvements: into the future

We can certainly expect that over the coming generation psychiatry will move towards a classification that is better informed by understanding of the normal and abnormal workings of the brain. ^{Reference Insel, Cuthbert, Garvey, Heinssen, Pine and Quinn11,Reference Craddock and Owen13,Reference Casey, Craddock, Cuthbert, Hyman, Lee and Ressler18} The current Director of the National Institute for Mental Health (NIMH), Dr Tom Insel, has made explicit that future NIMH-funded research will need to use approaches that are based on domains of psychopathology rather than DSM categories. ^{Reference Insel19} He said that the long-term future for mental health is in the detection of biomarkers and that diagnosis should be informed by objective laboratory measures. The future is likely to require a willingness to use both categorical and dimensional approaches. It will also be necessary to ensure consistency between the diagnoses used in all aspects of medicine that relate to brain and behavioural disorders. ^{Reference White, Rickards and Zeman20} Further, like all medical classifications, it is likely to involve a pragmatic mix of approaches that reflect the differing levels of understanding of each diagnostic entity. ^{Reference Craddock and Owen9}

Will the approach in 50 years’ time look like DSM and ICD do today? Almost certainly not. Advances in understanding of neuroscience and other disciplines relevant to psychiatric disorders are likely to lead to diagnostic systems that map much more clearly onto the functions and dysfunctions of the brain.

Pragmatism: the professional, practical and patient-focused approach

So, what should we, as practising clinical psychiatrists, do now? We need to be pragmatic, thoughtful and honest. As we have discussed, there are a variety of limitations and shortcomings of current diagnostic schemes in psychiatry (as there are for classifications in other areas of medicine). We must recognise and acknowledge these and the need to be working towards better approaches in the future. However, we must use the tools for communication, evidence evaluation and decision-making that are currently available - namely the internationally recognised operational classifications. We must use these in a manner that maximises their benefits while minimising their disadvantages. We should continue to make diagnoses complemented by formulations in which a range of additional factors are brought together that are relevant to management and prognosis.