The definition of "diagnosis" has been monopolised by medicine. In this sense it is the art or act of identifying a disease from its symptoms and signs. The concept of diagnosis also has a broader definition - the analysis of the cause or nature of a condition, situation or problem. For example, a car mechanic may diagnose engine trouble.

Application of the narrower medical sense of diagnosis to problems of the mind in psychiatry is commonly recognised as controversial (Clare, 1976). Medical diagnosis generally implies a statement about bodily dysfunction. A conceptual distinction has therefore been made between a "disease", which is caused by underlying physical pathology, and an "illness", which arises out of patients' complaints (Kleinman, 1988).

Some critics suggest that the process of psychiatric diagnosis should be abandoned and replaced by alternatives (Boyle, 1990). This article shares the concerns about the excesses of medical diagnosis in psychiatry. It does not attempt to provide a full critique of the practice of diagnosis in psychiatry. It is written with the intention of highlighting the basis for this critique. It also aims to clarify different emphases amongst critics. The principle distinction made is between those that deny that there is any validity to the practice of psychiatric diagnosis and those that want to replace a biomedical approach with a biopsychological approach to diagnosis.

This article takes the latter position. It does not want to mislead by suggesting that there is anything amiss with the broader process of trying to establish the nature of a patient's problems. It is important to try and make sense of presentations of mental disorder. This may be difficult, particularly with the so-called more serious presentations, when people can be very hard to understand. This is particularly the case with psychotic symptoms, when people may have hallucinations, bizarre ideas or their talk may be difficult to follow. However difficult it may be to comprehend these people an attempt needs to be made to understand and diagnose them in this sense.

Nonetheless, there are real problems with what may be called an overemphasis on biomedical diagnosis in current practice. Mistakes in diagnosis can be made with stigmatising and devastating consequences. As an illustration, Kay Sheldon was a patient who was misdiagnosed as schizophrenic and managed to obtain an out of court settlement from her Health Authority for £58,000. Kay knew that she did not have schizophrenia and that her treatment was wrong, but she could not get people to listen. The only course of action she found she could take was a medical negligence claim. Simon Foster, Mind’s Principal Solicitor when commenting on this case said:

This article examines why this situation prevails in psychiatry. The intention is not to threaten the authority of psychiatry. There are real problems in psychiatric practice and it is acknowledged that treatment does not always improve patients' situations and that it can make them worse. This is inevitably the case - if psychiatric treatment has the power to improve people's well-being, it can also make matters worse. However, both patients and professionals are caught up in this process and it is important not to attach unnecessary blame. Enough blame is already being projected onto "others" in our risk culture (Douglas, 1992). The argument of this article is merely for more self-awareness within psychiatry of the nature of its diagnostic practices.

The dominant model of mental illness in psychiatry is biomedical (Double, 2001). This statement is not made to encourage an ideological battle about the nature of mental illness. What needs to be pointed out is that there has been an alternative minority position in psychiatry. This has been particularly represented by Adolf Meyer, who was the foremost American psychiatrist of the first half of the last century (Meyer, 1951/2). His theoretical model was called Psychobiology and he took a biopsychological approach to diagnosis in psychiatry.

The influence of Adolf Meyer has been eclipsed in the current biomedical dominance of psychiatry (Double, 1990). His reputation at the Phipps clinic, where he was the first director from when it opened in 1913, had faded by the mid-1950s when visited by Shepherd (1986) and, in general, little is now known about his contribution to psychiatry. A typical evaluation of his approach has been that it was "almost entirely sterile" (Slater & Roth, 1969).

Meyer (1906) argued with Kraepelin about the introduction of the concept of dementia praecox. What Kraepelin (1921) thought he was identifying was a single morbid process. When revising his textbook for a sixth edition, he collected together, almost for convenience, several presentations of mental disorder under the single heading of dementia praecox (Kraepelin, 1899). The emphasis in describing the historical conceptual development of psychiatric classification has been on the origins of the concept of dementia praecox, which came to be called schizophrenia. This is because schizophrenia tends to be seen as the prototypical psychiatric illness (Boyle, 1990). Kraepelin, however, also introduced the term manic-depressive illness, in its modern day understanding. Again, he regarded manic-depressive illness as a single morbid process. The distinction he made was that dementia praecox led to personality deterioration, or dementia, in his terms, and therefore had a poor prognosis. Manic-depressive illness was a episodic illness in which patients were relatively well between episodes.

Kraepelin (1921) later came to think that both the terms dementia and praecox, meaning onset at a young age, did not necessarily apply to the process he had identified. This allowed Bleuler (1951) to introduce the concept of schizophrenia. Following the introduction of the third edition of the Diagnostic and Statistical Manual of the American Psychiatric Association (1980) (DSM-III), diagnostic ideas very much returned to Kraepelin's original notion that schizophrenia and manic-depressive illness should be distinguished on the basis of prognosis (Klerman 1978). The concept of schizophrenia was narrowed to include poor prognosis cases only. The introduction of DSM-III has therefore been seen as part of the neo-Kraepelinian movement in psychiatry (Double, 1990). This trend will be examined in more detail in the next section of the article

In its origins, Meyer questioned the biological basis of Kraepelin's concept of dementia praecox. As far as Meyer was concerned it was important to have a psychogenic understanding of dementia praecox as much as for any other psychiatric presentation. It did not really help the understanding of the reasons for why people became psychotic to suggest that it was due to dementia pracox, as though proposing a biological condition behind the symptoms provided that understanding.

Meyer's emphasis was on understanding the patient as a person. His advice in psychiatric assessment was to concentrate on what he called "the facts of the case". This relates to personal and social details in the psychiatric presentation. It was important not to foreclose an understanding of a person's problems by moving on too soon to a single-word diagnosis. In a way, this can be seen as a phenomenological process of assessment, in the sense that the conclusions about diagnosis are bracketed out of the analysis until the appropriate time (Sims, 1988).

As far as Meyer was concerned, it may not be possible to move on the stage of diagnosis. In which case, the person's problems still needed to be managed, whatever the diagnosis. Thus, in a way, medical diagnosis may be unimportant. Certainly the danger is that it oversimplifies the nature of a patient's problems. For Meyer, too much emphasis was placed on diagnosis in everyday psychiatric practice. This meant that psychiatrists did not make enough effort to produce a commonsense understanding of why patients had reacted in the way that they had to the situation in which they found themselves.

Psychiatrists do not want to admit the uncertainty that there is round diagnosis. One only needs to attend a psychiatric case conference to realise that diagnosis is not an exact science. Many different opinions will be expressed. Obviously diagnosis has something to do with the observer diagnostician and not just the observed diagnosed patient. This is partly why the biomedical hypothesis seems attractive. It holds out the hope of a certainty in brain processes (Double, 2001). But the real world of being human is more complex. Making a judgement and assesment about another person inevitably involves values as well as facts. This issue will be taken up again after the discussion about the development of DSM-III.

Current thinking about diagnosis is very much focused on modern psychiatric classifications like DSM-IV (American Psychiatric Association, 1994) and ICD-10 (World Health Organisation, 1992). There was a significant shift that occured between DSM-II and DSM-III, which had its effect on the change between ICD-9 and ICD-10. DSM-III represented an attempt to emphasise the importance of diagnosis, deliberately countering trends which it thought had devalued the role of diagnosis, such as the Meyerian approach and psychoanalysis.

The change in diagnostic classificatory systems between DSM-II and DSM-III was dramatic (Blashfield, 1984). This can be seen if only from the size of the manual. The chapter related to psychiatric disorders in DSM-II is a thin pamphlet. In DSM-III it is a large textbook. And a large textbook with a difference - no references at all. Presumably DSM-III speaks with authority - it does not need any references.

Why such a dramatic change in these externals? What are the internal changes which these represent?

DSM-III was produced by a task force of which Robert Spitzer was the Chair. He had already made a name for himself as one of the authors of the Research Diagnostic Criteria (RDC) (Spitzer et al, 1978). These are descriptions of diagnoses defined in such a way that there are rules which can be followed to create some consistency in diagnosis for research purposes - in other words the RDC took advantage of operational criteria as definitions of psychiatric disorders. The RDC have probably been the most widely used criteria in psychiatric research, certainly before the introduction of DSM-IV.

Spitzer was particularly concerned about the reliability of diagnosis (Spitzer and Fleiss, 1974). What had most perturbed him was a study by Rosenhan (1973), called On Being Sane in Insane Places. Rosenhan was a sociologist who was interested in the labelling effect of psychiatric diagnoses. What he did, in a classic study, was arrange for normal confederates of his to get admitted to a psychiatric hospital. He arranged for them to present themselves saying they were hearing a voice, saying a single word. There were three variations in the trial - either the pseudopatient said they were hearing the voice say "thud", "hollow" or "empty". This was the only symptom they had. No delusions or thought disorder or other symptoms of mental illness. Just a simple hallucination, and even then just one word, which is not particularly characteristic of mental illness.

What happened to these pseudopatients? All of them were admitted to psychiatric hospital. After admission they stopped feigning their symptom of hearing a voice. Some of the real patients detected that they were pseudopatients, because they saw them writing notes about their experience.

What diagnoses did they receive? All of them apart from one received a diagnosis of schizophrenia - the one who did not was diagnosed as manic-depressive. There is some qualification to this because although they had acquired a psychiatric diagnosis they were noted to be in remission, improved or asymptomatic.

The response of the psychiatric establishment to this study was disbelief. Rosenhan therefore informed the staff of a research and teaching hospital that at some time during the following three months, one or more pseudopatients would attempt to be admitted. No such attempt was actually made. Yet approximately 10% of real patients were suspected by two or more staff members to be pseudopatients.

Rosenhan concluded from this that psychiatric diagnosis is subjective and does not reflect inherent patient characteristics. Spitzer (1976) was one of the main critics in the literature of this study and its conclusion. In other words, there is a direct link between Rosenhan's study and DSM-III.

The link is through operational criteria. Spitzer was so panicked that psychiatric diagnoses may be unreliable that he made every effort to ensure that they were clearly defined. Transparent rules were laid out for making each psychiatric diagnosis. They have been called "chinese menu" definitions. If you order a set meal in a Chinese restaurant you order so many dishes from one part of the menu, then so many from another. Similarly, operational definitions require so many symptoms from section A, then so many more from section B and so on until you have enough to conclude the diagnosis.

The problem is that Spitzer missed the point in a way. What was being challenged more was the validity of psychiatric diagnosis not necesarily its reliability. It is all very well to create a consistent, reliable diagnostic definition. But suppose this reliable definition bears no relation at all to what it is supposed to be representing. Moreover, many different operational criteria were created at about the same time as RDC and DSM-III. The problem was that each of them had a slightly different coverage from each other for each psychiatric disorder (Blashfield, 1984). So which was the correct valid definition?

DSM-III has encouraged unthinking practice and an impersonal approach to diagnosis. Little attempt is made to understand personal and social factors in aetiology. Instead, in the worse scenario, the trainee psychiatrist sits interviewing a patient with DSM-III on his or her lap, ticking off the presence or otherwise of the different symptoms in each of the diagnostic criteria.

The whole movement behind DSM-III, which it has already been noted has been called neo-Kraepelinian, took the opportunity to reassert a biomedical model of mental disorder. The Meyerian approach was regarded as too vague. It was seen as important not to devalue diagnosis in the way Meyer was perceived to have done and to replace his approach with the certainty and apparent authority of DSM-III. DSM-IV has essentially only reinforced this trend.

Does the biopsychological view of diagnosis need to accept defeat merely because biomedical psychiatry has a powerful hold over the concept of diagnosis? Klerman (1978) expected a neo-Meyerian renaissance but it has been slow to be formulated. Is the classificatory system necessarily so tied to a biomedical model that its concepts should be abandoned?

However much we may want to create certainty in diagnostic practice, there is bound to be considerable variability and diversity. The need to learn to live with such uncertainty needs to be acknowledged.

Concepts of diagnosis are used culturally. There are ideological elements to any ascription of diagnosis. A diagnosis is not just a fact - there are values involved in making a diagnosis about another.

Of course, none of this should be taken to mean that the reality of mental suffering is being denied. There are adverse consequences of denying the reality of mental illness. For example, Thomas Szasz (1972) is well known for his view that mental illness is a myth. His reason for this is logical in that he defines illness as physical pathology. Therefore, as far as he is concerned, mental illness cannot exist.

Illness may not be the best word to describe an abnormal mental state. In fact over recent years, partly for this reason, it is much more common to use the term mental health problems than mental illness. Nonetheless, there do need to be terms for abnormal mental functioning. Although the term mental illness may commonly be assumed to be a statement about biological functioning, in fact it is no more than a metaphorical description of abnormal mental functioning (Farrell, 1979). It need not be abandoned for this reason.

Szasz's view is too rigid. It has had harmful effects on the critique of biomedical psychiatry, because it has confused the argument (Sedgwick, 1982). It has allowed biomedical psychiatry to marginalise any criticism by suggesting that it is an attack on mental illness itself. No such threat to the existence of psychiatry should be implied. Mental health services always have been and will continue to be required. Every society has recognised the presence of mental illness or whatever it is called.

The same argument applies to specific diagnoses such as schizophrenia and manic-depressive illness. Although they may have been monopolised as biomedical concepts, it does not mean that they should be dismissed as meaningless. They are ultimately psychological not biological concepts. They need to be restated as psychological concepts.

The use of the terms schizophrenia and manic-depression may differ between professionals. This makes it incumbent on professionals to try and be as clear as possible about how terms are being used. It is also important not to worry too much about any difference and disagreement. Diagnoses are only words - words which are trying to represent psychological concepts nonetheless. If schizophrenia or manic-depression are not the right words to use of particular presentations, language should be rich enough to make an attempt at describing someone's mental state. The problem is that in an attempt to simplify, there has been a tendency to reduce and avoid the effort needed to put into words the necessary understanding.

Concepts like schizophrenia and manic-depressive illness have been developed and used inappropriately. To use them in the sense of biomedical disorders is unscientific (Boyle, 1990). The temptation to conclude that the concepts should be abandoned is therefore evident. Boyle (1990) argues that retaining the concept of schizophrenia will prove obstructive to attempts to clarify brain-behaviour links and environment-behaviour links. She proposes a rethinking of how the phenomena of interest might be grouped or conceptualised.

Boyle's proposals are reasonable, but in fact a more radical awareness of the nature of diagnosis is required. There will inevitably be limitations in the application of psychiatric diagnosis, whatever way symptoms and signs are grouped and conceptualised. Psychiatric practice needs to acknowledge this state of uncertainty. The concepts of mental illness do not need to be abandoned for this reason. They merely need to be recognised for what they are - attempts to describe psychological states. They are therefore related to unobservables and therefore not capable of description in this scientific sense.

What matters in assessment, or psychiatric diagnosis in its wider sense, is an understanding of the patient as a person. There are always going to be limits on this process. The problem with biomedical diagnosis is that it potentially produces the facade that understanding has been created. This myth does need to be challenged more forcibly.

In summary, a case has been made for a positive re-evaluation of the work of Adolf Meyer. Over recent years a neo-Kraepelinian position has marginalised this view. Nonetheless there is a strength in the current critique of biomedical diagnosis which should not be underestimated.

Critics of psychiatric diagnosis such as Mary Boyle (1990) have contributed significantly to that strength. However, her critique has proposed abandoning the process of psychiatric diagnosis, which is potentially misleading because the process of assessment of the patient as a person, or psychiatric diagnosis in this wider sense, still needs to be undertaken in practice. The aim in psychiatric assessment should be to produce an understanding of the basis of psychopathology, which is rarely helped by the hypothesis of biomedical diagnosis. The process of assessment is primarily biopsychological not biomedical.

American Psychiatric Association (1980) Diagnostic and Statistical Manual of Mental Disorders (3rd edition). Washington: American Psychiatric Association.

American Psychiatric Association (1994) Diagnostic and Statistical Manual of Mental Disorders (4th edition). Washington: American Psychiatric Association.

Blashfield, R.K. (1984) The classification of psychopathology. Neo-Kraepelinian and quantitative approaches. New York: Plenum

Bleuler, E. (1951) Dementia praecox or the group of schizophrenias (translated by J Zinkin). London: Allen & Unwin

Boyle, M. (1990) Schizophrenia: A scientific delusion? London: Routledge

Clare, A. (1976) Psychiatry in dissent: Controversial issues in thought and practice. London: Tavistock

Double, D.B. (1990) What would Adolf Meyer have thought of the neo-Kraepelinian approach? Psychiatric Bulletin 1990; 14: 472-4

Double, D.B. (2001) Can psychiatry be retrieved from a biological approach? Journal of Critical Psychology, Counselling and Psychotherapy, 1, 28-31

Douglas, M. (1992) Risk and blame: Essays in cultural theory. London: Routledge

Farrell, B.A. (1979) Mental illness: a conceptual analysis. Psychological Medicine, 9, 21-35.

Kleinman, A. (1988) Rethinking psychiatry: From cultural category to personal expereince. New York: Free Press

Klerman, G.L. (1978) The evolution of a scientific nosology. In Schizophrenia: Science and Practice (Ed. J. C. Shershow).Cambridge, Mass: Harvard University Press.

Kraepelin, E. (1899) Psychiatrie. 6th Edition. Leipsig: Barth

Kraepelin, E. (1921) Manic-depressive insanity and paranoia. Edinburgh: Livingstone

Meyer, A. (1906) Fundamental concepts of dementia praecox. British Medical Journal, 2, 757-759

Meyer, A. (1951/52) Collected Papers (Four Volumes) (Ed. E. Winters). Baltimore: John Hopkins Press

Mind (2001) Landmark settlement for wrongful diagnosis of schizophrenia offers hope for hundreds misdiagnosed. Press release: 8 February 2001 (available at http://www.mind.org.uk/press-room/press_page.asp?ID=96)

Rosenhan, D.L. (1973) On being sane in insane places. Science, 179, 250-258.

Sedgwick, P. (1982) Psycho politics. London: Pluto Press

Shepherd, M. (1986) A representative psychiatrist: the career, contributions and legacies of Sir Aubrey Lewis. Psychological Medicine Monograph Supplement 10.

Sims, A. (1988) Symptoms in the mind: An introduction to desciptive psychopathology. London: Baillière Tindall

Slater, E. and Roth, M. (1969) Clinical Psychiatry (Third edition of Mayer-Gross textbook). London: Baillière, Tindall & Cassell

Spitzer, R.L. (1976) More on pseudoscience in science and the case for psychiatric diagnosis. Archives of General Psychiatry, 33, 459- 470

Spitzer, R.L. and Fleiss, J.L. (1974) A reanalysis of the reliability of psychiatric diagnosis. British Journal of Psychiatry, 125, 341-347.

Spizer R.L., Endicott, J. and Robins, E. (1975) Research diagnostic criteria (RDC) for a selected group of functional disorders. New York: New York State Psychiatric Institute

Szasz, T.S. (1972) The myth of mental illness. London: Paladin

World Health Organisation (1992) The ICD-10 classification of mental and behavioural disorders. Geneva: WHO