Psychiatric Annals

CME Article 

Classification of Schizoaffective Disorder: The History of a Vexing Concept

Oliver Freudenreich, MD, FACLP; Nicholas Kontos, MD, FACLP; John Querques, MD


Schizoaffective disorder, which straddles schizophrenia and mood disorders, is a problem for our diagnostic classification system. Its validity as a unique disorder independent from schizophrenia and affective illnesses has not been established, and its reliability in routine clinical care is poor. However, it clearly fills a gap in our nomenclature, particularly for good-prognosis schizophrenia-like cases that remit like mood disorders do. In this article, we provide some historical and conceptual background to help clinicians understand the epistemological challenges of the diagnosis of schizoaffective disorder. [Psychiatr Ann. 2020;50(5):186–189.]


Schizoaffective disorder, which straddles schizophrenia and mood disorders, is a problem for our diagnostic classification system. Its validity as a unique disorder independent from schizophrenia and affective illnesses has not been established, and its reliability in routine clinical care is poor. However, it clearly fills a gap in our nomenclature, particularly for good-prognosis schizophrenia-like cases that remit like mood disorders do. In this article, we provide some historical and conceptual background to help clinicians understand the epistemological challenges of the diagnosis of schizoaffective disorder. [Psychiatr Ann. 2020;50(5):186–189.]

Schizoaffective disorder is a problematic disease category. Its nosological status stands on clay feet in terms of both validity and reliability. Consequently, at regular intervals, when the profession revises its diagnostic manuals, the question arises if schizoaffective disorder should be retained in its current form, substantially revised, or simply abandoned.1 Thus far, schizoaffective disorder has withstood proposals for its abolition; it continues to be included in two current classifications, the Diagnostic and Statistical Manual of Mental Disorders, fifth edition (DSM-5)2 and The International Statistical Classification of Diseases and Related Health Problems, 11th revision (ICD-11).3

Schizoaffective disorder's continued existence speaks to a clinical need for a disease category that acknowledges the messiness of patient presentations and of experiences that do not fit into the relatively neat bins of psychosis and affective episodes. In this article, we trace the history of schizoaffective disorder, situating it in the larger context of why psychiatry (and medicine in general) classifies it at all, and discuss how what does and does not count as a disease is determined. We do not propose a new approach to schizoaffective disorder but merely sketch its historical background so that clinicians can use the schizoaffective disorder category rigorously, even if also skeptically.


Modern psychiatric nosology owes much to Emil Kraepelin, whose seminal textbook of psychiatry introduced a fundamental distinction between episodic bipolar disorder and chronic psychotic illnesses.4 All modern classification systems have preserved this split. However, diagnostic problems arise in cases that do not fit neatly into one of these Kraepelinian categories. Cases with admixtures of mood and psychotic symptoms are problematic. Equally problematic are cases of schizophrenia-like psychotic disorders that seem to follow a more episodic course or even remit (ie, remitting, nonaffective psychoses). Other classification systems for psychotic disorders have been developed (eg, the Frankfurt Kleist-Leonhard school5) but have not had the same impact that Kraepelin's system has had. Schneider6 favored a differential typology to deal with cases straddling the border between psychosis and mood disorders.

The term for atypical (with regard to symptoms, course, and prognosis) cases of psychosis that eventually stuck was “schizoaffective psychoses,” first introduced by Kasanin7 in 1933 when he described nine cases with “blending of schizophrenic and affective symptoms.” Notably, his cases were characterized by a sudden onset of symptoms in the setting of stressors followed by eventual recovery. Reading his detailed case notes today, Kasanin's good-prognosis cases have more in common with the bouffée délirante of French psychiatry8 or the reactive psychosis described by Scandinavian psychiatrists.9 In today's nomenclature, we would likely classify them as acute and transient psychotic disorders according to the ICD system10 rather than as narrowly defined schizophrenia or schizoaffective disorder according to DSM-5. In the modern era, the DSM system has always retained the term “schizoaffective” to allow for diagnosis of cases straddling schizophrenia and mood disorders, particularly episodic bipolar disorder.1

Theoretically, schizoaffective disorder might be any one of the following: (1) its own separate illness, distinct from schizophrenia or bipolar disorder (“true” schizoaffective disorder); (2) on a unitary psychosis spectrum between schizophrenia and bipolar disorder; (3) “true” comorbidity between schizophrenia and mood disorders; (4) a variant of schizophrenia; or (5) a variant of bipolar disorder.

Research has failed to resolve this conceptual challenge completely.11 Schizoaffective disorder, according to DSM-5 criteria (ie, emphasizing both the longitudinal and chronic illness course and the blending of psychotic and mood symptoms), is most consistent with a good-prognosis form of schizophrenia; consequently, schizoaffective disorder is considered one of the schizophrenia spectrum disorders in DSM and the ICD systems.


The conundrum of schizoaffective disorder cannot be understood without some appreciation of the principles and purposes of classifications, including disease classifications (ie, nosologies). Classification systems serve multiple and, at times, divergent purposes. We cannot not classify: as physicians, we must make a diagnosis to treat optimally. However, not all systems of ordering and classifying are equal from a scientific viewpoint.12 A classification for animals, for example, that puts zebras and bees in the same category is a rather superficial system working only at the surface level of phenomenology; although sufficient for grammar school, animals in this category share little, if anything, at a deeper level. Having different classifications serve different purposes is, of course, not absurd. An avid gardener, for example, needs to classify plants according to flowering season and “looks,” whereas a botanist is interested in scientific aspects such as genetics. One recurrent criticism that has been leveled against the DSM system is that it is more a glorified nomenclature than a scientific classification, as our illness categories are based merely on superficial criteria.

Simply having difficulties determining boundaries does not negate the validity of a concept. Night and day is one such example. We may have difficulties determining exactly when day ends and night begins, but nobody would deny the practical aspects of the night-day distinction: our societies are arranged around it.


A diagnostic system in medicine (ie, a nosology) is merely a special case of a classification. Clinical diagnosis often begins with careful description of symptoms, both cross-sectionally and longitudinally. Eventually, however, biological markers representing a deeper understanding of etiology and pathophysiology are needed for further scientific progress.

Validity, reliability, and utility are all important characteristics of a diagnosis. In a seminal article, Robins and Guze13 proposed five diagnostic factors to establish the validity of a psychiatric illness (Table 1). At this point, their goal of well-validated psychiatric disorders remains aspirational as our current validators stem mostly from clinical description, family studies, and follow-up studies without help from laboratory testing.

Diagnostic Validators for Psychiatric Illness

Table 1:

Diagnostic Validators for Psychiatric Illness

A valid diagnosis is not necessarily reliable, and a reliable diagnosis may not be valid. Improving reliability, which has been a major effort (and a success) in psychiatric diagnosis, does not establish or guarantee validity.14

In addition to being valid and reliable, a diagnosis must have clinical utility (ie, be easy to use, provide actionable information). Without utility, clinicians, who must necessarily be practical, will simply not use unwieldy classifications that are inconsistent with their practice. Classifications that have utility, however, may not be particularly scientific. As noted earlier, a gardener's system of ordering plants has much utility for arranging beautiful gardens but will not facilitate botanists' discovery of different plants' genetic relatedness.

The Problems with Schizoaffective Disorder

Reliability and Clinical Utility

The continued use of schizoaffective disorder suggests some utility for this blended category. However, schizoaffective disorder is problematic for clinicians in that it is not a diagnosis that can be made easily, yet it can also be resorted to too easily. Its description remains complex (despite attempts to improve the wording); more importantly, it presumes knowledge about the illness course that is often unavailable. As a result, reliability suffers in routine clinical care compared to research settings, where more time can be dedicated to investigating the diagnostic criteria. In one study, clinicians selected the schizoaffective disorder category over schizophrenia for psychotic patients more often than researchers using structured interviews,15 which is consistent with a more impressionistic application of the criteria.

DSM-5 differs from ICD-11 in that it includes the course in its definition of schizoaffective disorder; by contrast, ICD-11 is meant to capture only the current episode of illness, without consideration of the longitudinal illness course.16 In a large field trial, clinicians had difficulties reliably distinguishing among schizoaffective disorder, schizophrenia, and psychotic mood disorders.17 Careful attention to specific criteria (eg, serious impairment for 6 months in the presence of significant psychosis and mood symptoms) may improve schizoaffective disorder's reliability and reduce diagnostic heterogeneity.18


At this point, most psychiatric syndromes, including schizoaffective disorder, are still mostly defined at the level of phenomenology (ie, symptoms), which has hampered further conceptual progress. Attempts to find essential or pathognomonic features (eg, Schneiderian first-rank symptoms) have been unsuccessful. We may simply have arrived at a point where further progress is impossible without a deeper understanding of how symptoms arise at the level of networks. We may have to conclude that we cannot expect categories based on traditional, descriptive phenomenology to map onto brain circuits.

A reason for the failure to make progress may be another one of Kraepelin's powerful ideas, namely, that there are natural disease entities (Krankheitseinheiten) in psychiatry,19 characterized by specific etiologies and pathophysiologies.20 Those have failed to materialize and the assumption that they exist may have led psychiatry down the wrong path for a long time simply because our diseases are multifactorial with fuzzy borders.21 Moreover, it appears that our disorders have more in common with one another (eg, at the genetic level) than is the case for the disorders of other medical specialties (eg, neurology);22 they may not represent diseases as entities at all. Moving away from Kraepelin's disease entities, contemporary research is increasingly using models that are trying to understand underlying brain circuitry. The National Institute of Mental Health's Research Domain Criteria project and similar transdiagnostic (ie, across clinical disease categories) approaches are examples.23 At the level of biology, it appears increasingly unlikely that schizoaffective disorder is a natural, unique disease entity distinct from schizophrenia or bipolar disorder,11 or even from other disorders of brain development (eg, autism).


The current situation with schizoaffective disorder reflects the intrinsic difficulties in classifying brain disorders that are rooted in brain development. Alfred Korzybski famously pointed out that “the map is not the territory.”24 Any map is only an approximation of the surface terrain and even less of the underlying rocks and minerals. At the level of clinical phenomenology, sticking with Schneider's concept of differential typology seems to match clinical need best, while connecting to the experience of patients. However, no further scientific progress can be expected using only the clinical method of looking at (surface) symptoms and course; here we may have reached the end of the line beyond fiddling on the edges. It appears that nature cannot be carved into neat, separable disease entities as envisioned by Kraepelin but, instead, that psychiatric disorders have fuzzy boundaries and require transdiagnostic approaches.

We suspect that, to reflect these realities, psychiatry will eventually fundamentally revise its nosology for neurodevelopmental disorders, including schizophrenia. Until then, we are left with using the current system optimally in the clinic, including adding dimensional assessments to our categories and considering a careful differential diagnosis for those cases with polymorphous psychopathology. Schizoaffective disorder should not be used simply as a catch-all category in these situations and ought to be a rare diagnosis when DSM criteria are applied rigorously. Tension may always exist between a symptom-level clinical diagnostic system that is close to patient experience and a scientific system based on genetics and neurobiology.


  1. Heckers S. Is schizoaffective disorder a useful diagnosis?Curr Psychiatry Rep. 2009;11(4):332–337. doi:10.1007/s11920-009-0048-3 [CrossRef] PMID:19635243
  2. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 5th ed. Arlington, VA: American Psychiatric Publishing; 2013.
  3. The International Statistical Classification of Diseases and Related Health Problems. 11th revision. Geneva, Switzerland: World Health Organization; 2018.
  4. Kraepelin E. Psychiatrie. Ein Lehrbuch fuer Studierende und Aerzte. Achte, vollstaendig umgearbeitete Auflage. 3. Band. Klinische Psychiatrie. 2. Teil [Psychiatry. A Textbook for Students and Doctors. 8th completely revised edition. 3rd volume. Clinical Psychiatry. 2nd part]. Leipzig, Germany: Verlag Von Johann Ambrosius Barth; 1913.
  5. Salvatore P, Bhuvaneswar C, Ebert D, Maggini C, Baldessarini RJ. Cycloid psychoses revisited: case reports, literature review, and commentary. Harv Rev Psychiatry. 2008;16(3):167–180. doi:10.1080/10673220802167899 [CrossRef] PMID:18569038
  6. Schneider K, ed. Klinische Psychopathologie [Clinical Psychopathology]. 15th ed. Stuttgart: GeorgThieme Verlag; 2007. doi:10.1055/b-002-44928 [CrossRef]
  7. Kasanin J. The acute schizoaffective psychoses. 1933. Am J Psychiatry. 1994;151(6)(suppl):144–154. doi:10.1176/ajp.151.6.144 [CrossRef] PMID:8192190
  8. Crocq MA. French perspectives on psychiatric classification. Dialogues Clin Neurosci. 2015;17(1):51–57. PMID:25987863
  9. Opjordsmoen S. Reactive psychosis and other brief psychotic episodes. Curr Psychiatry Rep. 2001;3(4):338–341. doi:10.1007/s11920-001-0031-0 [CrossRef] PMID:11470042
  10. Legesse B, Freudenreich O, Murray E, Price B. A case report of confusional psychosis with abrupt onset and rapid resolution of symptoms. Psychosomatics. 2011;52(5):468–471. doi:10.1016/j.psym.2011.01.030 [CrossRef] PMID:21907068
  11. Goodwin FK, Ghaemi SN. The conundrum of schizoaffective disorder: a review of the literature. Psychiatr Ann. 2010;40(3):168–171. doi:10.3928/00485713-20100303-08 [CrossRef]
  12. Wolf P. Of cabbages and kings: some considerations on classifications, diagnostic schemes, semiology, and concepts. Epilepsia. 2003;44(1):1–4. doi:10.1046/j.1528-1157.2003.09202_2.x [CrossRef] PMID:12581219
  13. Robins E, Guze SB. Establishment of diagnostic validity in psychiatric illness: its application to schizophrenia. Am J Psychiatry. 1970;126(7):983–987. doi:10.1176/ajp.126.7.983 [CrossRef] PMID:5409569
  14. Klerman GL, Vaillant GE, Spitzer RL, Michels R. A debate on DSM-III. Am J Psychiatry. 1984;141(4):539–553. doi:10.1176/ajp.141.4.539 [CrossRef] PMID:6703133
  15. Wilson JE, Nian H, Heckers S. The schizoaffective disorder diagnosis: a conundrum in the clinical setting. Eur Arch Psychiatry Clin Neurosci. 2014;264(1):29–34. doi:10.1007/s00406-013-0410-7 [CrossRef] PMID:23625467
  16. Reed GM, First MB, Kogan CS, et al. Innovations and changes in the ICD-11 classification of mental, behavioural and neurodevelopmental disorders. World Psychiatry. 2019;18(1):3–19. doi:10.1002/wps.20611 [CrossRef] PMID:30600616
  17. Peterson DL, Webb CA, Keeley JW, et al. The reliability and clinical utility of ICD-11 schizoaffective disorder: A field trial. Schizophr Res. 2019;208:235–241. doi:10.1016/j.schres.2019.02.011 [CrossRef] PMID:30799218
  18. Seldin K, Armstrong K, Schiff ML, Heckers S. Reducing the diagnostic heterogeneity of schizoaffective disorder. Front Psychiatry. 2017;8:18. doi:10.3389/fpsyt.2017.00018 [CrossRef] PMID:28239362
  19. Jablensky A. The disease entity in psychiatry: fact or fiction?Epidemiol Psychiatr Sci2012;21(3):255–264. doi:10.1017/S2045796012000339 [CrossRef] PMID:22794347
  20. Kendler KS, Jablensky A. Kraepelin's concept of psychiatric illness. Psychol Med. 2011;41(6):1119–1126. doi:10.1017/S0033291710001509 [CrossRef] PMID:20809997
  21. Kendler KS, Zachar P, Craver C. What kinds of things are psychiatric disorders?Psychol Med. 2011;41(6):1143–1150. doi:10.1017/S0033291710001844 [CrossRef] PMID:20860872
  22. Anttila V, Bulik-Sullivan B, Finucane HK, et al. Brainstorm Consortium. Analysis of shared heritability in common disorders of the brain. Science. 2018;360(6395):eaap8757. doi:10.1126/science.aap8757 [CrossRef] PMID:29930110
  23. Clark LA, Cuthbert B, Lewis-Fernández R, Narrow WE, Reed GM. Three approaches to understanding and classifying mental disorder: ICD-11, DSM-5, and the National Institute of Mental Health's Research Domain Criteria (RDoC). Psychol Sci Public Interest. 2017;18(2):72–145. doi:10.1177/1529100617727266 [CrossRef] PMID:29211974
  24. Korzybski A. Science and Sanity: An Introduction to Non-Aristotelian Systems and General Semantics. International Non-Aristotelian Library Publishing Company; 1933.

Diagnostic Validators for Psychiatric Illness

Clinical description

Laboratory studies

Delineation from other disorders

Follow-up study

Family study


Oliver Freudenreich, MD, FACLP, is the Co-Director, Schizophrenia Clinical and Research Program, Massachusetts General Hospital; and an Associate Professor of Psychiatry, Harvard Medical School. Nicholas Kontos, MD, FACLP, is the Director, Fellowship in Consultation-Liaison Psychiatry, Massachusetts General Hospital; and an Assistant Professor of Psychiatry, Harvard Medical School. John Querques, MD, is the Vice Chairman for Hospital Services and the Chief of Inpatient Services, Department of Psychiatry, Tufts Medical Center; and an Associate Professor of Psychiatry, Tufts University School of Medicine.

Disclosure: Oliver Freudenreich has been a consultant for Alkermes, Neurocrine, Janssen, Novartis, and Roche; has performed contracted research for Otsuka, Janssen, Avanir, and Alkermes; receives royalties from UpToDate, Wolters Kluwer, and Springer; and has received payments from Medscape, Elsevier, and Global Medical Education for writing. Nicholas Kontos discloses royalties from UpToDate. The remaining author has no relevant financial relationships to disclose.

Address correspondence to Oliver Freudenreich, MD, FACLP, Erich Lindemann Mental Health Center, 25 Staniford Street, Boston, MA 02114; email:


Sign up to receive

Journal E-contents