The frustrating lack of progress in psychiatric research and treatment may be partly due to the limitations of the dominant classification systems generally used in psychiatry: the Diagnostic and Statistical Manual of Mental Disorders (published by the American Psychiatric Association) and the International Statistical Classification of Diseases and Related Health Problems (ICD published by the World Health Organization). The ICD is an initiative of the World Health Organization (WHO), first published in 1948 and it is not restricted to mental disorders. Mental disorders are listed under chapter 5. A categorical system classifies a disorder as being present or not, depending on whether the patient fulfills a certain set of criteria (Chmura Kraemer et al., 2004) and is designed to make communication between researchers and clinicians worldwide simple and efficient (Sheehan et al., 1998).
The very first version of the DSM was published in 1952 (Grob, 1991). It was divided into two major categories: disorders that were clearly caused by an impairment in brain function (trauma, infection, intoxication, etcetera) and disorders characterized by the inability of an individual to adjust. The latter was divided into psychotic and neurotic disorders (Grob, 1991). The fourth edition is divided into 5 axes: clinical disorders, personality disorders and intellectual disabilities, acute medical conditions and physical disorders, psychosocial and environmental factors contributing to the disorder (problems at home, at work, etcetera), and finally a global assessment of functioning, which is scored from 0 to 100 and encompasses both overall functioning and severity of symptoms. Each disorder is characterized by a set of criteria that a patient must fulfill before receiving that diagnosis. The DSM is commercialized and encompasses an ever increasing number of diagnostic categories (more than 400 in the DSM-IV). A fifth edition has been published in June 2013 with new diagnoseses such as ‘Voyeuristic disorder’ and ‘Tobacco use disorder’, i.e. smoking. Over the years, more and more objections to the DSM have been raised.
In psychiatry, nosology has been defined by the DSM (and ICD) since the Second World War. It is important for any researcher to realize that the DSM does not take into account any (hypothesized) cause or process supposed to underlie the diagnostic categories. The DSM system defines ‘diagnostic categories’ by operational criteria based on the presence of symptoms, without reference to supposed psychological or biological processes associated with the diagnostic categories. As such, the DSM categories do clearly not resemble mature medical diagnostic concepts validated by an underlying biological disturbance. This, of course, does not imply that DSM concepts are completely arbitrary. They have been chosen because of their apparent clinical validity, reinforced over a century of psychiatric thinking and practice (van Beveren and Hoogendijk, 2011).
However, the key concepts, notably schizophrenia, were developed in late stage samples from tertiary settings and before modern treatments were developed. They were essentially descriptive, with an attempt to capture prognostic utility which proved challenging. Consequently, they have not only a weak relationship to treatment selection (one of the major purposes of diagnosis), but also poor utility when retro-fitted to the early stages of mental ill health. Critically, they set the bar too high for commencement of treatment and provide little guidance as to safe and appropriate care. Although the DSM has improved the reliability (under research conditions only) of categorical diagnoses, these diagnoses represent expert consensus with all its failings, and not discrete valid medical diseases in the traditional sense (Nieman and McGorry, 2015). In the case of medical diseases, e.g. ischemic heart disease, diabetes or glioblastoma multiforme, it is possible to have the disease without symptoms. In contrast, is not possible to have a psychiatric disorder without symptoms (Borsboom and Cramer, 2013).
The DSM works under the assumption that psychiatric disorders are discrete entities that have natural boundaries delimiting them (Dalal and Sivakumar, 2009). However, over the years, comorbidity has proven otherwise. The DSM allows multiple diagnoses per patient, which reflects the fact that disorders such as anxiety and depression often go together. Surveys have shown that comorbidity is more frequently present than absent. In the United States 51 per cent of patients diagnosed with major depressive disorder suffered from at least one comorbid disorder (Maj, 2005; Kessler et al., 2011). In Spain, 30.3 per cent of all patients surveyed had more than one current mental disorder (Roca et al., 2009).
The DSM also assumes the concept of multiple parallel pathways each leading to distinct diagnostic categories. This assumption is not supported by longitudinal clinical nor neurobiological studies (Hickie et al., 2013; Lichtenstein et al., 2009; Sullivan et al., 2012). How do psychiatric symptoms develop? Data from studies that assess patients longitudinally from childhood or adolescence show that symptoms are shared and sequenced across disorders (Merikangas et al., 2012; Kelleher et al., 2012). Mental disorders are not fixed and independent entities, but each diagnosis is robustly related to other diagnoses in a correlational structure that is manifested both concurrently and in patterns across time (Lahey et al., 2014). Prototypically, anxiety disorders that are evident in children before the age of 12 years predict later depressive, bipolar and psychotic disorders (Kim-Cohen et al., 2003). This sequential pattern opens up the opportunity for prevention of secondary disorders. Stress, and its accompanying physiological changes and anxiety, may be at the core of many psychiatric disorders, as corroborated by aberrant oxidative stress and inflammation biomarkers in evolving severity (Kuloglu et al., 2002; Bouayed et al., 2009; Dean et al., 2011; Nieman and McGorry, 2015).
Another hotly debated topic when discussing categorical diagnosis is the fact that DSM imposes very specific thresholds, above which a patient qualifies to be diagnosed with the disorder. It can be a time frame, such as presence of a certain severity or number of symptoms in an uninterrupted fashion for an arbitrary period such as two weeks. Someone who has had the symptoms for 10 days will not be considered ill and thus will not receive treatment, whereas someone who has had them for 15 days might receive the same treatment as a patient who has suffered for a year. Although the DSM was not primarily created for making treatment decisions but for facilitating communication and research, in many countries it plays an important role in treatment allocation in clinical practice. This can lead to some patients with a need for care not getting treated at all, as well as to possible over-treatment of others (Helzer et al., 2006). Yet this is obviously a set of somewhat arbitrary decisions, and represents a spuriously objective definition of the boundary of ‘need for care’. We do not have such thresholds for conditions such as allergy or rheumatoid arthritis. The threshold here is lower, though intensity of treatment will vary.
Furthermore, recent studies have raised the stakes when it comes to sub threshold symptoms reported by patients: they have shown that particularly in psychosis, depression and anxiety, sub threshold conditions can impair psychosocial functioning and may lead to the full blown disorder as arbitrarily defined. This shows that the criteria listed in the DSM may be insufficient when it comes to detecting patients in need of care (Shankman et al., 2009; McGorry, 2010; Karsten et al., 2010). On the other hand, there has been concern about broadening the boundaries of mental ill-health (Frances, 2013). A definition of need for care or an approach to prevent any harm arising from ‘soft entry’ is needed and in fact achievable (Nieman and McGorry, 2015).
Finally, DSM diagnoses are polythetic. A polythetic class is defined in terms of a broad set of criteria that are neither necessary nor sufficient for class membership. Each member of the category must possess a certain minimal number of defining characteristics, but none of the features has to be found in each member of the category. For example in the DSM, 9 criteria are listed for a major depressive disorder but only 5 are required for the diagnosis, and thus two patients can receive the same diagnosis while sharing only one symptom (Hyman, 2010). In addition, sleep difficulties are listed under major depressive disorder, post-traumatic stress disorder, generalized anxiety disorder and a number of other diagnostic categories. This way of defining classes (polythetic) is associated with Wittgenstein’s concept of ‘family resemblances’. Polythetic classification has been seriously criticized by, e.g. Sutcliffe (1993, 1994, 1996) as logically incoherent.
A monothetic class is defined in terms of characteristics that are both necessary and sufficient in order to identify members of that class. This way of defining a class is termed the Aristotelian definition of a class.
Thus, the DSM categories were conceived by consensus among psychiatrists, and not by nature. It is similar to creating a category by consensus of ‘pain on the chest’. All the patients in this category have pain on the chest, but one subject has angina pectoris and the next a simple muscle ache. The optimal treatment and prognosis is diverse for the individual patients in this category and it would not be possible to find one biological substrate in research. Therefore, a polythetic category of ‘pain on the chest’ would lead to worldwide consistency and reproducibility of diagnosis, but not to validity. It is more useful to use imaging, plasma enzymes, etcetera to narrow down the diagnosis to a point where it does provide information about optimal treatment and prognosis. In many other fields of medicine, categorization is based on biological abnormalities, i.e. determined by nature and not by expert consensus.
In clinical practice, it is not uncommon to do an intake with a patient who has previously received four or more different DSM diagnoses. Unfortunately, the clinician’s interpretation of the patient’s symptoms with respect to the DSM criteria is the basis for the diagnosis (Stein et al., 201...