Gin S. Malhi: A critical analysis of concepts in psychiatry.

Gin S. Malhi and Erica Bell: Fake Views: Schizoaffective Disordes is not "SAD," just bad


Gin S. Malhi and Erica Bell’s reply to Janusz Rybakowski, Edward Shorter and Hector Warnes’ comments


Three… Two… One… Lift off!

Responding to the Comments on Schizoaffective Disorder


       We thank Tom Ban for inviting commentaries on our article in which we critiqued the diagnosis of schizoaffective disorder (SAD) and offered a practical, albeit interim, solution – involving keeping the same acronym, but using the letter “D” to represent dysfunction instead of disorder. Our basic concern was that the term is poorly defined in standard taxonomies such as DSM and has therefore been used in clinical practice without specificity; perhaps reflecting, in part, our lack of understanding as to its aetiology and pathogenesis. We also pointed out that in reality the moniker is often used to defer having to make a diagnosis of schizophrenia, even when it is evident that this is the illness from which the individual is suffering. Furthermore, we also warned that the designation is used to bypass recognized indications to gain access to medications, such as antipsychotics, that are limited to the treatment of psychotic disorders.

       The reason these points are important, and why we wrote the article in the first place, is that despite its poor definition and a lack of any guidelines for its management, SAD is a surprisingly popular diagnosis, with up to a third of patients in inpatient facilities attracting this descriptor. The problem now is that the field is stuck, because until the illness is properly defined, research into its management will necessarily be stymied (Kantrowitz and Citrome 2011). On the other hand, in the absence of any clear objective indicators concerning the nature of this illness, its definition remains wholly reliant on clinical symptomatology and although this is the case for nearly all psychiatric diagnoses, it is especially problematic in the case of SAD, because none of the symptoms used to “define” the disorder typifies the illness. In fact, all of them are key features of the very disorders from which SAD is being differentiated, namely schizophrenia and mood disorders. Thus, phenomenologically, SAD sits across both sets of disorders and their symptoms, and the only way in which it is characterised is by the pattern of these symptoms over time and their chronological relationship to one another. And as detailed in our paper, identifying these patterns with sufficient confidence in real-world clinical practice is extremely difficult and sometimes simply not possible.

       The three commentators on our paper each address different points, ranging for example, from concerns that the SAD acronym could be confused with seasonal affective disorder, to our core quandary, namely, whether schizoaffective disorder is a separate entity. Determining the unique substrates of SAD is difficult because very little is known about its biological basis and our neuroscientific knowledge of schizophrenia and mood disorders is also somewhat rudimentary. Interestingly, between them, the three commentators suggest that there are either three (Warnes), two (Rybakowski) or possibly even only one (Shorter) type of disorder. They contend that genetic evidence suggests separation between schizophrenia and mood disorders, and that while SAD shares elements with both, it also has validity as an independent concept. This bridging role that creates a spectrum is observed clinically and reflected by the overlap of treatments used to treat schizophrenia, mood disorders and SAD.

       We do not disagree with any of these hypotheses or indeed other possibilities, and if genetic or other biological evidence were to show discrete entities, then the matter would be readily resolved. But the problem we have at present is that all of these models are based on extrapolations of clinical impressions and the latter are determined as much by Shorter’s “deeper currents,” as by the presuppositions and expectations of those observing the phenomena in the first place. In other words, if we are expecting to find more than one disorder, then we are likely to do so. And indeed, when formulating clinical scenarios, it is often the case that we generate a hypothesis and then try to see if the evidence fits. In clinical practice, it is often possible to “make” things fit and, given the heterogeneity of both schizophrenia and mood disorders, it should not surprise us that the phenomenology of these illnesses can be adequately modelled as a unitary, dichotomous or trichotomous concept.

       The most intriguing possibility is that alluded to by Shorter, namely, that perhaps there is only one underlying illness that manifests at different times as mood, madness or agitation. This is certainly worthy of examination, although he too leaves open the possibility, that maintains the complexity, that there are several deeper currents and not just the one.

       From a therapeutic perspective, the fact that the broad-spectrum actions of antipsychotics that quell behaviour and counter psychotic symptoms also improve and perhaps even modulate mood, would fit with the hypothesis that there are underlying (deep) mechanisms that lead to different clinical manifestations. It would explain how, as an illness progresses over time, it evolves and different facets come to light. At the same time, underlying processes may well separate at some point, but to settle this, we agree that some kind of objective biomarkers are needed, at the very least they would be of great help provided they are specific.

       However, returning to the key issue that was not addressed head-on, is that such advances are only possible if, in our clinical practice, we have means of defining meaningful groups of patients. That is to say, groups that have at least some distinguishing clinical characteristics. Thus, we return to our fundamental concern that we spell out in our article, namely, that schizoaffective disorder, as defined by DSM-5, is almost impossible to characterise in clinical practice. To be able to identify periods of illness where the features of a mood disorder and those of psychosis occur both separately and together, is extremely challenging, and as a consequence most diagnoses of SAD in clinical practice rarely fulfill DSM criteria. Added to this diagnostic challenge, the contextual difficulties remain. Clinicians are (understandably) reluctant to make a diagnosis of schizophrenia too soon and so SAD “buys” some time. It also opens the door to broader prescribing while only having achieved a minimal threshold. And whilst this is not ideal, even these problems would be manageable were the diagnosis only temporarily applied. However, this is seldom the case and instead, once assigned, the diagnosis of SAD is rarely revised.

       These are powerful reasons as to why this “made-up” disorder remains and is widely used. Unfortunately, it is hindering research, perhaps because, as Shorter reports, amongst “insiders,” schizoaffective disorder has always been something of a joke. Thus, our proposal was intended to raise awareness and perhaps provide a steppingstone that allows a slight shift away from the futile path that has been pursued thus far. Our own view, that is purely speculative, is that it is possible there is only one illness that has many manifestations, but that it is more likely that there are two groups of disorders that that have different characteristics and trajectories, and that these groups require somewhat different approaches to management.

       However, on the basis of the clinical presentations that we see, we do not feel there is sufficient reason to believe that there are three types of illness and it is because of this, we have argued that schizoaffective disorder needs to be completely rethought. Otherwise, research and understanding will never “lift off” and we will be none the wiser as to whether there are three, two or one type of underlying disease.



Kantrowitz JT, Citrome L. Schizoaffective Disorder. CNS Drugs, 2011;25:317-31.


April 15, 2021