Abstract
Delusion: a traditional psychopathological category
Delusions have long been recognized as central to the notion of psychosis [1]. As such what constitutes a delusion lies at the heart of psychiatric practice and thought. The traditional distinction between the presence and absence of delusions is still assumed in current systems of psychiatric classification [2, 3]. However it is increasingly asserted, sometimes as if it were beyond question, that the psychopathological category of delusions has been superceded, and that delusions are best considered as being continuous with more normal beliefs and experiences [4–7].
Which, if either, approach is to be preferred? The question has potentially profound implications for our view of mental disorders. A categorical approach allows for simplicity in case finding and measurement in psychosis. Categorical decisions about clinical interventions, and about criminal or civil responsibility, are served by the clear categorization of an individual as deluded or not. A strict category of delusions facilitates the recognition that some unpopular or otherwise deviant beliefs may best be considered as healthy; proper subjects for politicians, historians and other social scientists rather than for clinical psychiatry. It may also safeguard against incorrect diagnosis of severe mental disorder, with all the hazards that may entail. The categorical concept of delusions has underpinned the category of schizophrenia, and of psychosis generally, and this in turn has supported the rationale behind the use of various treatment modalities and theoretical approaches in psychiatry.
The potential advantages of a continuum approach to delusional phenomena are several. A continuum approach encourages recognition of the differences in delusions between individuals, and of change across the course of a period of mental disorder for an individual. It facilitates development of screening instruments and other means of early detection of psychoses. A continuum approach may assist a distressed psychotic person to understand and articulate their experience by extrapolating from the more normal or mundane. It may mitigate against stigma, by emphasizing the similarities between psychotic and normal experiences rather than differences.
In this paper I will outline the emergence of the continuity hypothesis and critique the two main arguments put forward in support of it: that delusional conviction is often not as absolute as has been thought and that delusions appear to be commonplace in healthy populations. The case will be made that the category of delusions continues to be valid. Finally the case is put that the continuity versus category debate may not require particular resolution, as both approaches have validity according to the immediate clinical or academic purpose.
Recent emphasis on a ‘continuity’ model
The traditional categorical distinction between delusions and other morbid or non-morbid beliefs and experiences has its origins in Jaspers' categorical descriptions of delusions [1, 8]. The view that delusions are qualitatively different from normal beliefs is implicit in Jaspers' emphasis on the non-understandability of true (primary) delusions (pp. 195–196) [1]. However Jaspers' description of delusion-like ideas clearly considered them as continuous with normal mental life in many respects, indicating that they ‘emerge comprehensibly from other psychic events’ (pp. 106–107). Jaspers considered that a ‘jealous man can develop into someone with delusion-like jealousy… a suspicious person into someone with delusion-like ideas of persecution’ (p. 640).
Influential texts, leaning heavily on Jaspers' description, have reinforced the consensus about the category of delusions [9, 10]. This mainstream view had endured despite the occasional non-categorical exploration of delusional experiences [11]. Definitions of delusions have often been phrased as to imply a clear category, as in the definition offered in the current DSM-IV-RT: ‘A false belief based on incorrect inference about external reality that is firmly sustained despite what constitutes incontrovertible and obvious proof or evidence to the contrary. The belief is not one ordinarily accepted by other members of the person's culture or subculture (e.g. it is not an article of religious faith). When a false belief involves a value judgement, it is regarded as a delusion only when the judgement is so extreme as to defy credibility’ [3].
Consideration of the nature of delusions is made more complex by inconstant conceptions of over-valued ideas. Considered as forms of delusion-like ideas by Jaspers (p. 107) [1], they have more recently been considered as quite distinct from delusions [12], and as a subset of delusions [13]. Despite the above definition the DSM-IV appears to consider over-valued ideas as continuous with delusions, differing only in the presence of lesser degrees of conviction [3].
Recent interest in viewing delusions from a continuity perspective began with the important study by Strauss [14]. In a study of 119 subjects in the International Pilot >Study of Schizophrenia, Strauss reported there were half as many ‘questionable’ delusions as ‘definite’ delusions. Reasons for rating delusions as questionable included the presence of depressive phenomena, religious content, bizarre experiences without associated belief, and most importantly for this discussion, less than full conviction for a belief that would otherwise be considered as delusional. Strauss was critical of the dichotomous view of psychotic symptoms as either present or absent and suggested four criteria for delusions, covering conviction, preoccupation, plausibility, and cultural context; each as a continuum.
The treatment of mental disorders characterized by delusions has changed from being dominated by the allocation of a categorical diagnosis, and prescription of an agent from the antipsychotic drug category, to encompass a range of psychological treatments. Such psychological approaches to delusions readily consider delusional conviction, distress and preoccupation as being on a continuum with normal worries and concerns [4, 15]. In contrast to the drug treatment of delusional illnesses, psychological treatments are adaptations of methods used in anxiety and affective disorders, again inviting consideration that these disorders may be less distinct from one another than has previously been supposed [4, 15]. Delusions construed cognitively may be considered as maladaptive cognitive schemata, with implications for the appropriate therapy [16].
Many aetiological theories of how delusions are formed and maintained suggest an unspecified and potentially variable degree of abnormality in one or more brain functions, readily allowing for delusions to be considered as continuous with normal beliefs. For some the view that delusions are continuous with normal beliefs forms part of the case that the diagnostic category of schizophrenia lacks validity [4].
Debate about the appropriate definition of a delusion often centres around whether or not delusions are to be regarded as categorically different from other beliefs or as part of a continuum. The problems inherent in the various definitions of delusions have been widely commented on [13, 17–19]. Definitions that recognize that different aspects of a delusion can be considered separately, and may not all be fully present, are increasingly prominent [20, 21].
However the main arguments advanced in favour of considering delusions as continuous with normal beliefs are two:
1 Delusional conviction is not as absolute as was once assumed.
2 Delusions, in full or partial form, are highly prevalent in normal community populations.
These arguments will be considered in turn.
Conviction is not absolute
Karl Jaspers' account of delusions are that they were held with ‘extraordinary conviction and subjective certainty’ [1]. Subsequent descriptions have also remarked on the incorrigible conviction that characterizes delusions. Delusions are ‘firmly sustained’ [3], ‘unshakable’ [22], or ‘beyond doubt’ [23]. It is increasingly observed that delusions are not all held with such complete conviction [24, 25]. Conviction varies, and can readily be assessed and represented on a continuum, and conviction in delusional phenomena may sometimes be of the same order as in non-delusional phenomena.
DSM-IV-RT contributes to the idea that excessive conviction is the sufficient condition for a delusion to be considered as present. In its description of obsessive compulsive disorder, DSM-IV allows for an additional diagnosis of a delusional disorder to be made when the obsession reaches ‘delusional proportions’. In body dysmorphic disorder an additional diagnosis of a delusional disorder is allowed by the DSM-IV if the preoccupation with the imagined defect is ‘held with delusional intensity’. Such delusional variants of these disorders are occasionally described, but such descriptions of delusional variants of non-psychotic illness generally depend on identifying delusions based only on the presence of extreme or absolute conviction. Reports of a nonpsychotic belief, such as may occur in a body dysmorphic disorder, developing into a delusion and the diagnosis changing to that of a delusional disorder, or other psychosis, also rely on changes in the reported conviction of the subject. Phillips and colleagues considered 100 cases of body dysmorphic disorder, 52 of which they classified as delusional on the basis of the subjects' stated convictions. The two groups did not show major differences on a variety of measures used and responded largely to the same treatments, primarily serotonergic drugs. Notably neither group responded well to antipsychotic drugs [26].
This emphasis on conviction has been considered naïve [13]. It ignores the other elements that comprise the delusion construct and which are referred to in the various available definitions of a delusion, including the DSM-IV definition above. In particular delusions are said to be ‘ununderstandable’ [27], occur in an ‘inappropriate context’ [28], or are ‘out of keeping with the patient's educational, cultural, and social background’ [22]. These are references to the striking incongruence between a delusional belief, and the range of what a person will plausibly believe given their personality, particular circumstances, and background. Certainly it would be clinically naïve to suppose that reported conviction was independent of personality, mood, and other circumstances of the assessment.
So-called partial delusions, where an otherwise delusional belief is expressed with less than complete conviction, have been recognized and are said to have similar diagnostic significance to full delusions [29, 30]. No useful purpose will be served if the term ‘delusional’ is allowed to become a simple synonym for ‘severe’.
Delusions, or delusion-like phenomena, are common in normal populations
Several large studies claim that delusions, in full or attenuated form, are highly prevalent in healthy populations. These studies suggest that such results indicate, or even confirm, that delusions are continuous with normal experience.
Peters and colleagues applied the Peters Delusional Inventory (PDI) to a healthy community sample of adults [5]. This self-report scale was developed from items concerned with psychotic phenomena from the Present State Examination [29]. The authors acknowledge that the items were ‘toned down and cast into a format that was thought to capture their normal equivalents’. Most of the items start ‘do you ever feel as if…’
The normal equivalent of a delusion, for example that one is being talked about may be a feeling that one is being talked about. This is a long way from indicating that delusions are on a continuum with such normal beliefs. Individual items in the PDI were endorsed, on average, by 25% of 272 healthy subjects. Peters and colleagues conclude that psychotic symptoms are the severe end of traits that are present in the general population.
Verdoux et al. used an earlier version of the PDI to assess delusions in 1053 general practice attendees [31, 32]. Individual items were endorsed by 5–70% of respondents. Those with a history of psychotic illness endorsed more than those without such a history. Their conclusion that such self-assessment may prove useful for screening for psychosis is entirely reasonable. Their interpretation that the findings confirm the hypothesis that delusions lie on a continuum with normality is premature. Verdoux et al. acknowledge the questionable validity of using such a measure to assess positive symptoms in a primary care population, and that the design of the questionnaire may encourage over-reporting.
Van Os et al. suggest that psychosis may exist as a continuous phenotype, implicating the potential usefulness of searching for multiple genes in the search for the biological causation of psychosis [33]. In this study, trained lay observers interviewed 7076 healthy adults. Where evidence of psychosis was found, subjects underwent telephone re-interview by a psychiatrist using questions from the Structured Clinical Interview for DSM-III-R (SCID). Van Os et al. reported that clinically not-relevant delusions, and psychiatrist-rated delusions, had prevalence rates of 8.7% and 3.3%, respectively, and concluded in support of the continuity hypothesis [33, 34]. Van Os et al. rely on the instruments reported for validity in diagnosing schizophrenia; however, those involved in development of the SCID expressed concern that low base-rates of the disorders of interest could compromise the use of the instrument [35].
Similar studies have obtained comparable results but the apparently high prevalence rates of psychotic symptoms in the general population (28.4% in the study by Kendler and colleagues [36]) have been interpreted cautiously by the authors who question the validity of such an approach to detection of psychotic phenomena [36, 37]. The high prevalence rates of apparent psychotic symptoms, particularly delusions, in such studies can be interpreted in several ways which do not support the continuity hypothesis; misunderstanding, culturally or subculturally sanctioned beliefs, and idiosyncratic but individually understandable beliefs, may all contribute to the findings. In addition it may be a misunderstanding of delusion, mistaking the part for the whole, to consider that prevalent beliefs and worries are similar to delusions on the basis of similar content. Kendler et al. indicate that assessment of psychosis in epidemiological samples requires more extensive contextual information. Instruments and even classifications of mental disorder derived from clinical populations are not known to be valid in community samples [38]. Such questionnaires may have a useful role as screening instruments for detection of cases, but such a function is separate from, and does not depend on, considerations of delusions as categorical or continuous.
Longitudinal studies of psychotic-like phenomena in non-clinical populations show they are weak predictors of clinical psychosis [39, 40]. Such studies may have an important role in characterizing the prodrome of psychotic illnesses and identifying strategies for early intervention. It is premature to conclude that the psychoticlike experiences of the majority who do not subsequently develop a psychosis represent a partial expression of the process that does result in a psychosis.
The case that delusions are on a continuum with more normal beliefs appears to be made partly on the grounds of similarities in the content of such beliefs. For example Johns and Van Os refer to the finding of Cox and Cowling, that 50% of British adults believe in thought transference, 25% in ghosts, and 25% in reincarnation [7, 41]. Such beliefs may be unscientific, and perhaps illogical, but to equate this with delusionality is to ignore much else of relevance in the delusion construct. In particular, delusions are characteristically idiosyncratic, preoccupying, and of particular importance to the deluded individual. Perhaps most importantly delusions have usually been considered not to be comprehensible as an expression of a person's personality and life situation. It follows that assessment of the presence of a delusion requires not only an assessment of the belief, but of the person who expresses the belief. If these other aspects of being deluded are ignored then every odd or incorrect beliefs can be conceived as a form of delusion.
A conceptual muddle
Since Strauss' publication, attempts to measure delusions on different dimensions have become commonplace [21, 42–45]. These studies make it clear that delusions can be considered as having multiple measurable facets, which vary within deluded individuals and across populations of deluded people. Some go further and claim that this dimensionality suggests delusions are on a continuum with normal beliefs [43]. Combining the concepts of dimensions and continua to imply that delusions and non-delusional beliefs are ideally considered as similar phenomena, may not be justified, and appears to depend on the use of the term ‘dimensions’ as synonymous with continuum.
There is an implicit presumption in much of the literature that properly determining delusions as categorical or dimensional phenomena will lead to new insights about their underlying nature. Any such presumption is flawed, as examples from elsewhere in medicine attest. Huntington's disease and tuberous sclerosis, autosomally dominant genetically determined illnesses (categories) have psychological manifestations that merge with the symptoms of other illnesses, and with normality. The pyrexia and malaise associated with many infectious illnesses is continuous with normal body temperature and sense of wellbeing, yet the categories of influenza and bacterial septicaemia still have validity. The characteristics of some breast lumps on palpation do not clearly distinguish them from the normal breast texture. Nonetheless the category of breast carcinoma has validity. Conversely, an underlying continuity, or interaction of continuous processes, may manifest externally as categorical events, for example febrile fits, or any complex behaviour such as an attempted suicide, or adulterous act. Intelligence is generally represented as a continuous distribution, but for purposes of planning services for the intellectually impaired, a simple category may suffice.
Across psychiatry there are conflicting views about the relative merits of categorical and other approaches to syndromes and their constituent phenomena [38, 46, 47]. Such debate may be of relevance only if we are determined to reify our constructs of mental disorder and symptomatology. However if we are content to recognize the provisional nature of our concepts then the question becomes a pragmatic matter [46].
Are delusions a valid category?
Delusions are for the most part private experiences, partially accessible to others by indirect means. A definition of a delusion should be no more than is required to allow clear discussion among clinicians and researchers. That such definitions are adequate, despite their myriad imperfections, is apparent from the high interater reliability found for delusions in clinical populations [29]. Asking more of a definition is sliding into essentialism, as is expecting a clear answer to questions regarding the categorical or continuous nature of delusions.
Validation of diagnostic categories falls into a number of stages, classically described by Robins and Guze: clinical description; laboratory and psychological assessment; exclusion of other disorders; follow up study; and family study [48]. It does not depend on any particular resolution of the category versus dimension debate. Validation of individual psychotic symptoms has never been a focus in the way that putative syndromes have been; nonetheless, a similar approach can be called for.
In clinical scenarios the most important validator is response to treatment. In this regard there is evidence of the value of the category of delusions. It has been repeatedly observed, in controlled and uncontrolled series, that depressed patients with delusions respond poorly to tricyclic antidepressants compared to nondeluded depressed patients [49–51].
Delusional depressed patients may respond better to electroconvulsive therapy [52, 53]. In addition delusional depressed patients may differ from non-deluded depressed patients on a variety of biochemical parameters [54–56]. Charney and Nelson report that in repeated episodes of depression, in both deluded and-non-deluded patients, the same phenomena tend to recur [51]. Delusions may not all be of a kind. Possibly some are more usefully considered as on a continuum with normal beliefs than others. The view that the abnormal beliefs of a person with a delusional disorder may be understood as a personality development, different in nature to the delusions in a psychosis-like schizophrenia has a long history [57]. In delusional disorders the abnormal beliefs lack the extraordinary bizarreness and obvious falsity of the delusions in schizophrenia. The content of delusions in a delusional disorder show ‘a remarkable agreement with those fears, wishes, and hopes, which even in normal individuals proceed from the feeling of uncertainty and the endeavour after happiness’ (p. 223) [58]. Commonly recognized examples of delusional disorder include cases of delusional jealousy, delusions of ill health, persecutory delusions and litigious paranoia, all cases in which equivalent non-delusional beliefs may readily occur in otherwise healthy individuals. In the case of morbid jealousy there is as good a case as for any disorder that delusional and non-delusional forms are similar in kind, putting in question the very plausibility of distinguishing the normal from the pathological [59].
In case of severe body dysmorphic disorder, evidence regarding the value of using antipsychotic drug treatment, or serotonergic drug treatment is inconsistent, and may depend on how delusions are conceived, as considered above [26, 60]. Specialists consider that delusional jealousy must be recognized as it requires different, or at least additional, treatment to non-delusional forms [59].
Validity in psychiatry, where understanding of aetiology and the basic relationships between brain events and conscious experience is so limited, must be a matter of practical utility [61]. Kendell has suggested that a point of rarity between two related syndromes validates a distinction [61]. Despite the appeal of sophisticated statistical techniques clinical judgement is still required to determine what potential discriminators are, and to determine the separation between data points which may be ordinal, but which for mental phenomena cannot be considered as interval data. Consideration of questions of validity in psychiatry risk being caught up in speculations or presumptions about aetiology, biology, and psychological mechanisms. A non-psychiatric example is given to illustrate how simple and flexible ideas about categories and continua may be, and how appropriate classification depends on our purpose. Suppose I am renovating a kitchen on a limited budget, and need a new toaster and a new fridge, along with some other items. Toasters are generally a lot cheaper than refrigerators, although both vary in price and some particularly good toasters are more expensive than some very cheap refrigerators: toasters and refrigerators can be considered to be on a continuum of cost. When I am subsequently deciding which items I can carry home on the bus and which will need to be delivered by van, refrigerators and toasters belong in different categories. When I am having the kitchen rewired they both belong in the same category, that of electrical items I must consider when instructing the electrician.
Conclusion
Recent arguments put forward in support of the continuity hypothesis do not take into account that the delusion construct entails more than unlikely content and excessive conviction. Conclusions that a categorical approach has been superceded are premature.
Further it is unhelpful to think of a single continuity ‘hypothesis’ for delusions. There are no useful general continuity or category hypotheses; only specific hypotheses about, for example, underlying biology, treatment response, outcome, and so on, which may be constructed as to yield results most usefully expressed in terms of continua or categories.
The shortage of clear aetiological understanding in most psychiatric conditions, and the readiness with which such issues become controversial, should make us cautious of pronouncements about what is fundamentally continuous or categorical. In the final analysis, understanding and description of human experience is always incomplete, and the ways that such understanding can be organized may tell us nothing meaningful about the experience itself.
Either categorical or dimensional approaches to delusional phenomena can be justified. The utility of the one approach in certain circumstances does not preclude, and should not prejudice, the use of the other approach. Hypotheses about the cause, treatment, or nature, of delusions can be generated within either framework, but the question of continuity or category may not itself be profound, and may not yield particularly useful hypotheses.
The yardstick against which we may measure a belief or experience in order to determine if it is delusional, is to a great extent the individual. As such the use of selfreport questionnaire methods to identify delusions is always going to be questionable, particularly in community samples where there is presumably a diverse range of normal beliefs and experience. Debate about continuity or dichotomy may help to identify causes, or treatments, by generating ideas and testable hypotheses. However conclusions about which approach is superior are premature, and may represent an inappropriate reification of concepts of delusion.
Footnotes
Acknowledgements
I am grateful to Gavin Cape, Grant Gillett, Richard Linscott and Lisa Turner for their help.