In this opinion piece, psychiatrist Samei Huda responds to a recent journal article published in Psychiatry Research that labeled psychiatric diagnoses as "scientifically meaningless".
When a paper claims to find that psychiatric diagnosis is “scientifically meaningless”, to judge if this conclusion is correct we need to examine what the study found and ask whether those findings prove the claim.
What did the study find?
The study examined the criteria used to define the following psychiatric diagnoses in the American psychiatric diagnostic manual DSM-5: schizophrenia spectrum and other psychotic disorders; bipolar and related disorders; depressive disorders; anxiety disorders; and trauma- and stressor-related disorders. Diagnostic criteria are the rules laid out in the DSM-5 on how to identify a particular diagnosis in a patient. The authors looked for evidence of what is called heterogeneity both within individual diagnoses and between different diagnoses (the meaning of the term “heterogeneity” means will be expanded on below).
The main differences found can be summarized as follows:
- Heterogeneity is present in the definition of individual diagnostic criteria compared to each other such as different standards to which symptoms are compared in different diagnoses or different minimum time periods used for different diagnoses. There is a partial acknowledgment of the role of clinical judgement; for example, in the absence of the person feeling distress if there is an impairment of functioning then a diagnosis can still be allocated to the patient (the classic example is someone in a manic state who may not regard themselves as distressed but is acting in a reckless or dangerous manner out of keeping with their usual selves).
- Heterogeneity is present across several different diagnostic criteria so that similar clinical problems are present across different diagnostic criteria such as hallucinations being part of the criteria for schizophrenia and also major depressive disorder with psychosis
- Trauma is listed as causative for PTSD but not listed as a cause for other conditions even when there is research evidence of association e.g. in depression.
- Concerns about how the criteria are used and constructed e.g. which perspective is used (patient or clinician)
- How to separate symptoms from ‘normal’ or expected responses.
- Diagnoses are organised as separate categories which often does not reflect how mental health problems are structured in real life.
We need to put these findings in context – it would seem the most suitable comparison is the use of diagnoses in general/ internal medicine. Are these problems isolated to psychiatry?
Heterogeneity within diagnoses
Heterogeneity within general/ internal medical conditions can also be marked – heart attacks can present differently in women compared to men and in ethnic minorities, autoimmune conditions and connective tissue disorders can vary greatly between patients. Therefore, expecting diagnoses in medicine including psychiatry to be largely homogenous in nature is unrealistic.
Heterogeneity between diagnostic constructs
Many different general/ internal medical conditions can have similar features: identifying the correct diagnosis can be challenging in fever, fatigue or even chest pain. Several medical conditions can present in different ways with non-specific symptoms making them hard to identify; tuberculosis and syphilis are regarded as “great imitators” which can be mistaken for other illnesses. In medicine there are often shared clinical features between very different diagnoses.
Lack of inclusion about specific causes in the diagnostic label
Many psychiatric diagnoses do not intend to explain the cause of symptoms. In the same way, the label hypertension does not explain why every individual with that diagnosis has high blood pressure; these are conditions with complex causes that can’t be summed up in a simple label. Many psychiatric conditions are associated with complicated causes with lots of contributing factors. These can be biological, psychological and social in nature and it would be impossible to include a label for each one, as they often vary between each individual with the same diagnosis. The diagnostic formulation is a summary of the relevant factors for an individual patient in terms of causing their problem, such as trauma, as well as the allocated diagnosis.
The absence of trauma from many diagnostic labels.
Although traumatic events are an important cause of mental health problems, they are not the only cause of many mental health problems. Whilst traumatic events are sufficient in themselves to cause mental health problems, there is usually an interaction with additional biological, psychological and social factors that determines the nature of the resulting mental health condition. Childhood trauma is associated with causing a third of psychosis, up to half of depression and anxiety and up to 90% of cases given the diagnosis of borderline/ emotionally unstable personality disorder. For the latter it has been suggested the term “complex trauma” may be better but for the different psychotic illnesses and anxiety/depression it makes no sense to put trauma in a label when it may not be involved in a high proportion of cases and would leave out other important contributory factors. Many medical conditions such as hypertension or cancer do not have the causes listed in the title, but the contributory factors are learnt by the doctor when they read about these conditions.
How diagnostic constructs are created, and which perspective is used
Many general/internal medical diagnoses are not based on conditions that are clearly separated from each other but are created by consensus based on research and historical traditions often involving committees of experts (many of whom have received grants from pharmaceutical companies for research and/or giving talks). The perspective historically has been driven by professionals who would say that their goal is to improve outcomes for patients but no doubt some other factors exist such as prestige. For all of medicine it is important to move to a more patient-centred focus.
Lack of clear-cut separations from normal or “expected” responses.
Many conditions in general/internal medicine do not exist as neat categories separated from healthy states. Medical conditions such as hypertension and type 2 diabetes are on a spectrum with normal blood pressure and blood glucose values in the population and the criteria used to define the conditions can appear arbitrary. One can disagree where the boundaries are drawn for defining hypertension and depression but that does not mean that a boundary does not have to be set, for example to identify people who may benefit from treatment.
In psychiatry judgements are often not based on objective measures but use clinical judgements involving factors such as level of distress and associated risks that take into account the cultural norms relevant to the patient. In the same way, in general medicine, several conditions require clinical judgement in the absence of more objective markers to decide whether the presenting problem is suitable for medical treatment rather than a “normal” phenomenon. As an example, the decision whether to treat acne vulgaris depends on associated distress and risk of scarring, not counting the number of lesions on a person’s face and body.
Using categories to classify mental health problems in the form of diagnostic constructs
That mental health problems are classified using categories in the form of diagnoses was described as “disingenuous” but this accusation is itself disingenuous. The DSM-5 (and DSM -IV) states that it recognises that mental health problems do not necessarily exist in discrete categories but that categories are used for practical reasons. In fact, in general/ internal medicine many conditions are not clearly separate from each other, such as myeloproliferative diseases, autoimmune diseases, metabolic syndrome conditions and connective tissue disorders.
In medicine as a whole, the problems that people present with are divided into categorical diagnostes because people tend to think in categories, which they find easier to learn and recall in clinical practise. Doctors often see patients under stressful situations such as limited time, in emergencies or in the middle of the night, which means they prefer classification systems that are simpler to use over complex but more accurate systems. In addition, much of the research psychiatrists use will be carried out using diagnostic classifications. Professionals who work in different ways, such as psychologists, may on the other hand prefer complex systems.
The description of psychiatric diagnoses as “scientifically meaningless” seems meaningless – the criticisms in this paper can be applied to many general medical diagnoses and I doubt people regard them as “scientifically meaningless”. Psychiatric diagnoses are often heterogenous and descriptive rather than referring to a distinct type of disease process, but they can still be used in scientific research to give us information on treatment and prognosis. They are also interim, and we are waiting for superior diagnoses. They may not be suitable for other professionals in the way they work and other classification systems may be better for research on causes and mechanisms in certain scenarios. Nevertheless, they are still useful and similar in nature to many general medical diagnoses.
Conflicts of Interest
I have written a book that defends the use of diagnosis in mental health (see Further Reading) to inform the reader of this article of my viewpoint and this article is based on this book.
Original article reference: Allsopp, K., Read, J., Corcoran, R., & Kinderman, P. (2019). Heterogeneity in psychiatric diagnostic classification. Psychiatry Research, 279, 15–22. https://doi.org/10.1016/j.psychres.2019.07.005
Further Reading: The Medical Model in Mental Health by Ahmed Samei Huda