Response

I found much to agree with in Bilsland and Boyce’s article, ^{Reference Bilsland and Boyce1} but I do quibble with the claim in the subheading that psychiatrists should move to a transdiagnostic classification model. I agree that a single classification system is unlikely to work successfully for all researchers, psychiatrists and other clinicians, such as psychologists. As I have argued in general ^{Reference Huda2} and specifically for psychosis, ^{Reference Huda and Petch3} the needs of these different groups often vary and hence utility is not equivalent for each group. One underexplored aspect of the utility of classification systems is how it varies depending on how its user works. National Health Service (NHS) psychiatrists with large caseloads (often seeing 6 to 20 times the number of cases of colleagues such as nurses or psychologists ^{Reference Tyrer, Al Muderis and Gulbrandsen4} ) will have to see patients in follow-up appointments often of 15 to 30 min spaced apart by weeks or months ^{Reference Creed5} for their routine clinical work (a similar situation probably occurs for other public-sector psychiatrists internationally). Psychiatrists also assess people in emergency situations including in the middle of the night. The relatively simple categorical diagnostic system fits well with a human tendency to categorise concepts and utilises rapid cognitive processes of pattern recognition of prototypes. This system is fast with relatively low cognitive burden and so has a lot of use for most NHS (and other public-sector) psychiatrists given the constraints and pressures of their working environment, such as time, the need to make rapid decisions in emergencies and fatigue. Further, it is backed by an extensive research base using categorical diagnosis to guide clinical decision making.

Transdiagnostic or other dimensional systems are more complex and their use require a greater cognitive load given the constraints that many NHS (and other public-sector) staff work under. Other clinicians with smaller caseloads, working in office hours with more time for each patient, often being seen every week, and not working in emergency situations can take advantage of their strong points. These types of classification systems also tend to make covert diagnostic decisions. For example, positive psychotic symptoms experienced only during intoxication with lysergic acid (LSD) would be excluded from a transdiagnostic dimension of psychosis by clinicians and researchers. Depressive symptoms associated with underactive thyroid glands or after traumatic experiences or following a manic episode may share phenomenological (and may turn out to share some mechanistic) similarities, but the prognosis and treatment will differ, and this would be missed by a transdiagnostic depression construct.

As part of their clinical practice, once the initial diagnosis is established psychiatrists can expand their formulation of the patient’s difficulties using transdiagnostic and other dimensional classifications, but the initial diagnosis still serves a valuable role. As an example, when I see cases of psychosis I will make a diagnosis then assess the severity of dimensional psychopathology, as recommended by ICD-11: positive, negative, psychomotor, cognitive, depressive and manic symptoms.

Researchers can use these complex transdiagnostic or other dimensional systems to create better categorical diagnostic constructs with more validity and superior utility than the DSM-5 and ICD-11. Thus, it is not about public-sector psychiatrists moving to transdiagnostic classifications that have less utility for them but about creating better categorical diagnostic constructs (based on research using other classification systems) that they can use. We should also recognise that DSM-5 and ICD-11 are constructed primarily for the benefit of psychiatrists and that other professional groups will find alternative classification systems to have greater utility for the way they work.