Could a dimensional perspective on psychosis have advantages over a categorical perspective? One of the early proponents of a dimensional approach to the nosology of psychosis was John Strauss,1 who argued that the category of schizophrenia and its subtypes were unwieldy for diagnosticians. He suggested that a dimensional scheme, which could capture important components of delusions and hallucinations such as conviction, preoccupation and plausibility, would improve our ability to evaluate meaningful similarities and distinctions between patients.
This observation lead to a number of symptom rating scales developed in the 1970s and 1980s2–4 that were designed not to improve diagnosis but to capture change over time and individual differences between psychiatric patients. The first wave of research using such a conceptualization suggested two dimensions of schizophrenia: a positive dimension encompassing hallucinations and delusions, and a negative dimension encompassing flat affect, poverty of speech, and other deficits.5–7 As the power and sophistication of these studies grew across the 1980s, a third dimension, disorganization, was added to include symptoms such as disorganized speech and inappropriate affect.8, 9
These factors form the basis of a dimensional approach that can enrich clinical case conceptualization. Consider the way diagnostic information is collected and conveyed when following a categorical model in a strict sense. Extensive interview data are reduced to decisions about whether the patient is above or below the symptom threshold within multiple mental disorder classes. For those classes for which the patient is above the threshold, the corresponding diagnoses are recorded.
As a result, a good deal of potentially important information is lost. How far above the threshold was the person for each diagnosis? Which symptoms were most salient, recent, or troubling? For which diagnoses did the patient come close to the threshold, and how close did the patient come? Why did the person fail to meet criteria? Was this due to specific psychopathological symptoms, the time frame of symptoms, or associated impairment criteria? Along these lines, a categorical diagnosis is insensitive to improvements or deterioration over time that may be insufficient to cross a diagnostic threshold (eg, from “diagnosable” to “undiagnosable”).
Discontentment with these limitations has lead some dimensional advocates to argue that this approach should replace the category of schizophrenia altogether.10 This is not a form of diagnostic nihilism. Indeed, in many respects, this approach is not radically different from a categorical approach. Both approaches rely on quantifying symptoms; the difference lies primarily in the way the approaches handle cutpoints.
In a categorical approach, rather than reporting the symptom count, information is reduced to a dichotomous decision about being above or below a threshold. This information is retained in more dimensional approach. A dimensional approach does not preclude identifying points on a scale that, if exceeded, may be of particular importance in terms of implications for the patient. For example, it could be the case that the number, duration, or intensity of psychotic symptoms is related to social and occupational dysfunction such that exceeding a certain level of symptomatology results in a marked decrease in the patient's ability to function. This does not mean that the number, duration, or intensity of symptoms are themselves discontinuous phenomena in nature. Rather, beyond a certain range, these phenomena may have clearly deleterious implications for a person's ability to function. In categorical approaches, these domains (symptomatology, dysfunction and duration) are confounded and reduced to a single dichotomous decision about presence or absence of diagnosis. In a dimensional approach, each of these domains is retained, and research can be pursued on the interrelation of the domains.
In addition, a dimensional approach can evaluate the effect of interventions on each domain separately. The dimensional approach also is well suited to address another set of nosological issues: those encompassed by the concept of “comorbidity.”
Comorbidity as a Challenge to Psychiatric Categories
Comorbidity refers to the co-occurrence of two different diseases in the same person. The categories of mental disorder described in current nosologies, such as the Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV),11 rarely are regarded as having the status of clearly delineated diseases. This is because we still have a great deal to learn about the etiology and pathophysiology of mental disorders. We lack compelling knowledge at the conjunction of etiology and pathophysiology of the kind that typically forms the basis for establishing a clearly delineated disease entity. For this reason, the term “comorbidity” might not be the best way to describe people meeting criteria for two putatively distinct mental disorders that, themselves, are provisional constructs (as opposed to clearly delineated diseases).12 Moreover, especially in the clinic, one frequently encounters patients who are not simply “comorbid” but who are more appropriately characterized as “multimorbid,” with numerous diagnoses that often transcend putatively distinct sections of DSM-IV, as discussed in the other articles in this issue.
These points notwithstanding, the phenomenon captured by the concept of comorbidity is of fundamental importance in psychopathology.13 The phenomenon clearly points toward some basic limitations of a categorical conception of mental disorder, especially in the absence of data indicating that specific mental disorders have attained the conceptual status of categorical disease entities. As described in DSM-IV, “A categorical approach to classification works best when all members of a diagnostic class are homogenous, when there are clear boundaries between classes, and when the different classes are mutually exclusive.”14
The prevalence of comorbidity raises questions about each of these three issues. First, frequent comorbidity means that members of a diagnostic class will not be homogenous with reference to the other disorders for which they meet criteria. (As a side note, this concern only adds to the issue of the heterogeneity within diagnoses created by the multiple ways in which a person can meet criteria for a specific diagnosis.) Second, frequent comorbidity suggests that clear boundaries are difficult to define, or at the very least, are not well-defined by current criteria. Third, comorbidity directly challenges mutual exclusivity, as persons easily can belong in more than one diagnostic class.
Many of these issues can be resolved by adopting a dimensional approach to collecting and conveying clinical data. This means simply describing on a continuous scale the extent of symptomatology, impairment, and duration within specific diagnostic groupings. For example, rather than recording a diagnosis of schizophrenia, the extent of symptomatology, impairment, and duration across components of the construct are recorded. Empirically supported symptom dimensions within the construct (eg, positive, negative and disorganization symptoms for schizophrenia) would constitute the first logical components. The other components would be the other traditional elements of a diagnosis, such as social and occupational dysfunction and symptom duration.
The Empirical Structure of Psychopathology
One potential objection to the kind of dimensional approach described above is that there may be bona fide categorical entities in psychopathology. If such entities actually exist, we would, of course, want to be in a position to identify these entities, to the extent that they correspond to the conjunctions of etiology and pathophysiology that represent the sine qua non of diseases (eg, one would need evidence of a specific lesion associated with a specific cause, such as a viral agent, gene, or environmental pathogen corresponding to a specific category of psychopathology). It might be said, for example, that counting psychotic symptoms is not unlike measuring body temperature and claiming that “extent of fever” is a legitimate disease construct.
The problem with this objection is that there are few examples of situations in which a particular categorical threshold actually succeeds in capturing discontinuity in nature, or in identifying a disease per se. Rather, the evidence tends to demonstrate that categorical cutpoints are relatively arbitrary and result in exactly the kind of information loss described above.15 With reference to psychosis, for example, van Os et al.16 demonstrated that dimensions of psychosis were better predictors of important clinical outcomes (eg, need for care and quality of life) than were categorical representations of psychosis. Hence, in the absence of data supporting the unique clinical value of specific thresholds, relying on existing thresholds to impose discontinuity where it may not exist is not a compelling, long-term approach.
Understanding the Mechanisms of Psychosis
If categorical distinctions are not inherent in nature, they may interfere in the process of uncovering the true nature of psychiatric entities. A dimensional approach, which focuses on core symptom dimensions rather than the complex co-occurrence of symptoms, may reduce problems associated with misclassification, facilitate research by isolating relevant elements, and demonstrate the continuity of clinical and normal psychological phenomena.17
One illustrative question is whether the tripartite factor structure of positive, negative, and disorganization symptoms is observed in the absence of illness, or whether it is unique to decompensated schizophrenia patients. If this factor structure is observed in the absence of illness, this further erodes the case for the importance of a categorical distinction between people with and without schizophrenia. Support for such continuity has been reported and replicated,18,19 which increases the plausibility that schizophrenia is an exacerbation or co-occurrence of a number of more common psychological processes.
Understanding the Etiology of Psychosis
A dimensional conceptualization of psychosis also may be important for understanding the causes of the illness. Across many studies, monozygotic twins of schizophrenia patients have an approximately 50% concordance rate. On the one hand, this highlights the importance of liability genes for schizophrenia. On the other hand, it suggests that many people with a genetic liability to schizophrenia never fully express the illness. This insight was captured in the concepts of schizotypy and schizotaxia, first popularized by Paul Meehl,20,21 which led to two complementary areas of research, one into the schizophrenia spectrum disorders22 and a second into intermediate indicators, or endophenotypes.23
The schizophrenia spectrum concept captures the notion that liability genes may be related to schizophrenia-like unusual behaviors, such as perceptual aberrations and social isolation, that do not interfere with functioning or rise to the threshold of a diagnosable illness. The Axis II diagnosis of schizotypal personality disorder in effect lists a number of such odd behaviors that were found in the relatives of schizophrenia patients.22 A second popular way to measure the schizophrenia spectrum has been through the use of personality scales, which score the level of deviance for a particular domain of schizophrenia-like behavior.24–27 Interestingly, some of these phenomena also may be caused by chance or environmental factors, rather than liability genes28,29
Endophenotypes are measurable biological or psychological indicators that lie between the liability genes for schizophrenia and the manifestation of the disorder.30 Useful endophenotypes are stable over time and relatively insensitive to stochastic and environmental influences, and thus may indicate the presence of a liability gene. Although not necessarily dimensional, endophenotypes are generally measured as continuous traits. In the search for liability genes, the endophenotype approach has become an important adjunct to diagnosis-based association and linkage studies.31 There has been some notable progress using this approach, including the use of measures of working memory, sensory gating, and glial abnormalities.30 There may be many more useful endophenotypes waiting to be uncovered.
Treating Dimensions of Psychosis
Following a long precedent, we have reviewed arguments for understanding psychotic phenomena along a number of dimensions, rather than as several categories. Psychiatrists often treat psychosis in this manner. For example, antipsychotics frequently are used to treat delusional symptoms, whether the diagnosis is schizophrenia, borderline personality disorder, or Alzheimer's disease. Nevertheless, it is worthwhile to highlight clinical implications that have been drawn from several considerations.
If psychotic dimensions depend on distinct cognitive and neural processes, then it would be overly optimistic to assume these processes rely on a single neural or neurotransmitter abnormality. It might be more fruitful to judge the success of a given medication by how well it treats a single dimension of psychosis.32 This observation has further implications for the development of pharmaceuticals because novel agents might be best evaluated based on their ability to treat an a priori dimension of psychosis, rather than their ability to alleviate all symptoms. While treating an isolated dimension of psychosis, it is important that such medicines do not adversely affect the other symptoms.
Cognitive impairments are an important aspect of disorganization in schizophrenia and may be related to hypoactivity of dopamine D1 receptors in prefrontal cortex.33 A number of novel antipsychotic agents include dopamine D1 agonists to address these symptoms.34 It may be appropriate that the delivery of such agents should not be linked to the delivery of agents that act on different neurotransmitter systems, such as dopamine D2 antagonists. It also might be useful to titrate these agents independently to monitor their effectiveness for different dimensions of the illness.
- Strauss JS. Hallucinations and delusions as points on continua function. Arch Gen Psychiatry. 1969;21(5):581–586. doi:10.1001/archpsyc.1969.01740230069010 [CrossRef]5823480
- Andreasen NC. The scale for the assessment of positive symptoms (SAPS). Technical Report. University of Iowa; 1983.
- Andreasen NC. The scale for the assessment of negative symptoms (SANS). Technical Report. University of Iowa; 1983.
- Kay SR, Fiszbein A, Opler LA. The Positive and Negative Syndrome Scale (PANSS) for schizophrenia. Schizophr Bull. 1987;13(2): 261–276. doi:10.1093/schbul/13.2.261 [CrossRef]3616518
- Crow TJ. Molecular pathology of schizophrenia: More than one dimension of pathology?Br Med J. 1980;280(6207):66–68. doi:10.1136/bmj.280.6207.66 [CrossRef]6101544
- Harvey PD, Walker EE, eds. Positive and Negative Symptoms in Psychosis: Description, Research, and Future Directions. Hillsdale, NJ: L. Erlbaum Assoc.; 1987.
- Opler LA, Hwang MY. Schizophrenia: a multidimensional disorder. Psychiatric Ann. 1994;24(9):491–495. doi:10.3928/0048-5713-19940901-12 [CrossRef]
- Liddle PF. Syndromes of chronic schizophrenia: a re-examination of the positive–negative dichotomy. Br J Psychiatry. 1987Aug; 151:145–151. doi:10.1192/bjp.151.2.145 [CrossRef]3690102
- Arndt S, Alliger RJ, Andreasen NC. The distinction of positive and negative symptoms: the failure of the two dimensional model. Br J Psychiatry. 1991;158:317–322. doi:10.1192/bjp.158.3.317 [CrossRef]2036528
- Bentall RP, Jackson HF, Pilgrim D. Abandoning the concept of ‘schizophrenia’: some implications of validity arguments for psychological research into psychotic phenomenon. Br J Clin Psychol. 1988;27(Pt 4):303–324. doi:10.1111/j.2044-8260.1988.tb00795.x [CrossRef]3063319
- American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. Washington, DC: American Psychiatric Publishing; 1994.
- Lilienfeld SO, Waldman ID, Israel AC. A critical examination of the use of the term and concept of comorbidity in psychopathology research. Clinical Psychology: Science and Practice. 1994;1(1):71–83.
- Rutter M. Comorbidity: meanings and mechanisms. Clinical Psychology: Science and Practice. 1994;1(1):100–103.
- Krueger RF, Piasecki TM. Toward a dimensional and psychometrically-informed approach to conceptualizing psychopathology. Behav Res Ther. 2002;40(5):485–499. doi:10.1016/S0005-7967(02)00016-5 [CrossRef]12038642
- Grove WM. When is diagnosis worth making? A statistical comparison of two prediction strategies. Psychol Rep. 1991;69(1):3–17.1961813
- van Os J, Gilvarry C, Bale R, et al. A comparison of the utility of dimensional and categorical representations of psychosis. Psychol Med. 1999;29(3):595–606. doi:10.1017/S0033291798008162 [CrossRef]10405080
- Persons JB. The advantages of studying psychological phenomena rather than psychiatric diagnoses. Am Psychol. 1986;41(11): 1252–1260. doi:10.1037/0003-066X.41.11.1252 [CrossRef]3813184
- Gruzelier J, Burgess A, Stygall J, Raine A. Patterns of cerebral assymetry and syndromes of schizotypal personality. Psychiatry Res. 1995 31;56(1):71–79. doi:10.1016/0165-1781(94)02564-Y [CrossRef]7792344
- Gruzelier JH. The factorial structure of schizotypy: part 1. Affinities with syndromes of schizophrenia. Schizophr Bull. 1996; 22(4):611–620. doi:10.1093/schbul/22.4.611 [CrossRef]8938915
- Meehl PE. Toward an integrated theory of schizotaxia, schizotypy and schizophrenia. J Personality Disorders. 1990;4(1):1–99. doi:10.1521/pedi.19126.96.36.199 [CrossRef]
- Meehl PE. Schizotaxia, schizotypy, schizophrenia. American Psychologist. 1962; 17(12):827–838. doi:10.1037/h0041029 [CrossRef]
- Kety SS, Rosenthal D, Wender PH, Schulsinger F. Mental Illness in the biological and adoptive families of adopted schizophrenics. Am J Psychiatry. 1971;128(3): 302–306. doi:10.1176/ajp.128.3.302 [CrossRef]5570994
- Gottesman I, Shields J. Schizophrenia: the Epigenetic Puzzle. New York: Cambridge University Press; 1982.
- Vollema MG, van den Bosch RJ. The multidimensionality of schizotypy. Schizophr Bull. 1995;21(1):19–31. doi:10.1093/schbul/21.1.19 [CrossRef]7770738
- Raine A, Benishay D. The SPQ-B: A brief screening instrument for schizotypal personality disorder. J Personal Disord. 1995;9(4): 346–355. doi:10.1521/pedi.19188.8.131.526 [CrossRef]
- Chapman LJ, Chapman JP, Kwapil TR, Eckblad M, Zinser MC. Putatively psychosis-prone subjects 10 years later. J Abnorm Psychol. 1994;103(2):171–183. doi:10.1037/0021-843X.103.2.171 [CrossRef]8040487
- Kwapil TR, Hegley DC, Chapman LJ, Chapman JP. Facilitation of word recognition by semantic priming in schizophrenia. J Abnorm Psychol. 1990;99(3):215–221. doi:10.1037/0021-843X.99.3.215 [CrossRef]2212270
- MacDonald AW III, Pogue-Geile MF, Debski TT, Manuck S. Genetic and environmental influences on schizotypy: A community-based twin study. Schizophr Bull. 2001;27(1):47–58. doi:10.1093/oxfordjournals.schbul.a006859 [CrossRef]11215549
- Berenbaum H. Peculiarity. In: Costello CG, ed. Personality Characteristics of the Personality Disordered. New York, NY: Wiley; 1995:206–241.
- Gottesman II, Gould TD. The endophenotype concept in psychiatry: etymology and strategic intentions. Am J Psychiatry. 2003; 160(4):636–645. doi:10.1176/appi.ajp.160.4.636 [CrossRef]12668349
- Moldin SO. Indicators of liability to schizophrenia: perspectives from genetic epidemiology. Schizophr Bull. 1994;20(1):169–184. doi:10.1093/schbul/20.1.169 [CrossRef]8197413
- Opler LA, Albert D, Ramirez PM. Psychopharmacologic treatment of negative schizophrenic symptoms. Compr Psychiatry. 1994;35(1):16–28. doi:10.1016/0010-440X(94)90165-1 [CrossRef]7908618
- Friedman JI, Temporini H, Davis KL. Pharmacological strategies for augmenting cognitive performance in schizophrenia. Biol Psychiatry. 1999;45(1):1–16. doi:10.1016/S0006-3223(98)00287-X [CrossRef]9894570
- Harvey PD, Siu CO, Romano S. Randomized, controlled, double-blind, multicenter comparison of the cognitive effects of ziprasidone versus olanzapine in acutely ill inpatients with schizophrenia or schizaffective disorder. Psychopharmacology. 2004;172(3):324–332. doi:10.1007/s00213-003-1652-2 [CrossRef]