The medical model has failed to reduce stigma

Time for ‘clinical psychological problems’ to re-emerge

What I have learned...

I studied and trained in clinical psychology in the 1970s. This was a time when CBT, with its superior face validity, was growing out of behaviour therapy, which was nicely evidence-based and relatively effective, but was seen as overly simplistic and even manipulative.

In this context we were taught how to assess psychological-level processes and mechanisms (operant and classical conditioning, modelling, etc.) and how to develop behavioural functional analyses and, later, case formulations that included self-reported cognitions – the ‘C’ in CBT.

The research papers on treatment that we read mostly searched for basic principles and processes of behaviour change or cognitive shift. Hence, we learned about techniques like the exposure therapies, thought stopping, and psychoeducation, and what sort clinical psychological problems (CPPs) they were helpful for. We focused on CPPs like public speaking anxiety or reactive depression irrespective of whether psychiatrists regarded them as mental disorders, because “clinical psychological problems” are, by definition, (i) deserving of professional intervention (“clinical”), (ii) best assessed, analysed, and treated at a psychological level (“psychological”), and because (iii) clinical psychology is a ‘remedial’ discipline and so addresses ‘problems’.

Then in 1980 the American Psychiatric Association published the DSM-III, the first DSM to be based on a medical model of CPPs, thereby subsuming CPPs along with psychiatric illnesses under their own concept of ‘mental disorders’. This concept included those problems that are clearly dysfunctions at an organic level, like multi-infarct dementia, but also those that, to us, are clearly dysfunctions at a psychological level, such as panic disorder, and suddenly every CPP became a biologically based and medically conceived problem.

With this, everything changed. All clinical psychologists had to learn how to ‘diagnose’ ‘mental disorders’ based on their ‘symptoms’. Mental disorders were presumed to be manifestations of underlying, latent conditions, neural pathologies, or disease processes. Psychotherapy outcome studies from thenceforth had to follow the drug trial format if they were to attract funding. Blocks of manualised treatment – so as to standardise interventions, just like a certain dose of a certain drug – were administered to treatment groups that had been uniformly diagnosed with a mental disorder, and rates of ‘cure’ were compared to a control group.

These comparisons told us only about the average efficacy of a packaged intervention on a group of mixed CPPs with some overlapping ‘symptoms’, that is, with one diagnosis. The results were then assembled in treatment guideline manuals that told us that, for example, CBT and some of the SSRIs are the most effective treatments for the (invented and psychologically heterogeneous) mental disorder called major depressive disorder.

But this new medical conception of CPPs entailed a whole gamut of major problems for psychologists, who operate at a different level of analysis and intervention – the psychological level of cognitive, behavioural, and emotional processes, not the biological level of neurological and chemical states. It implied that a functional analysis or case formulation to find the psychological-level processes or mechanisms underlying a CPP is no longer required. Instead, we could run through a checklist of symptoms, make a diagnosis, and then consult a relevant treatment guidelines booklet to choose a manualised treatment program to apply.

However, this is not what the overwhelming majority of experienced clinicians actually do. Other remedial psychologists, such as health psychologists, forensic psychologists, or educational and developmental psychologists may not be constrained by another profession’s diagnosis. Even clinical psychologists will invariably do their own assessment, develop their own case formulation, and then tailor a psychological treatment program to the individual before them, based on research-derived change principles.

But because almost all research in clinical psychology has had to apply manualised treatment packages to diagnosed groups, a ‘scientist-practitioner gap’ has developed, in some countries producing a profound schism between psychological scientists and practitioners.

Clinicians have been perfectly justified in going their own way because over the past 42 years the medical model of CPPs has failed both research psychologists and our clinicians’ clients in almost every way. These failures and incompatibilities have been thoroughly elaborated elsewhere (e.g. Bakker, 2019; DCP, 2013), and include the failure to identify any reliable biomarkers underlying any mental disorders despite millions of research dollars spent (Ivanets et al., 2021), the clear heterogeneity of CPPs that occur within any DSM-diagnosed treatment group (Zimmerman et al., 2015), the lack of theoretical or logical connection between diagnosis and treatment selection (First et al., 2018) unlike our case formulations, and the fact that between a third and a half of all CPPs presenting to psychologists do not formally qualify for a DSM diagnosis (Westen et al., 2004). For example, relationship problems are a large slice of our presenting problems, and even DSM admits that they cannot be conceived as ‘mental disorders’. But we assess, formulate, and address them in exactly the same way as other CPPs.

One of the major justifications proposed for the imposition of the medical model of CPPs onto clinical psychology’s adjacent psychological level of analysis and intervention has been the claim that it can reduce stigma and punitive responses, encourage the admission of problems, and increase medication compliance, at least compared to historical conceptions of CPPs such as the weak character, sinfulness, moral failing, or demon-possessed conceptions (Deacon, 2013; NIAAA, 2024). Hence the promotion of mottos such as “mental illness is like any other medical illness” or “brain not blame”.

The fact that such public education campaigns have been ubiquitous and incessant for over 30 years, and are still deemed necessary, suggests that they have not worked very well. Examples include the disease model of alcoholism, the ‘chemical imbalance’ theory of depression, and the ‘obesity is a disease’ message. Each of these has been promoted as true, as well as useful, but recent research addressing their validity has contradicted them all (Hoyt et al., 2014; Lubman, 2020; Moncrieff et al., 2022).

And with regard to their usefulness, even the project to reduce stigma has clearly failed (Schomerus et al., 2012). Although people experiencing CPPs are today less seen as weak or sinful, their medicalisation has resulted in sufferers being socially excluded, given a negative prognosis, and feared more, as they are seen as powerless to change, dangerous, and unpredictable, which attracts stigma (Seeman et al., 2016). A diagnosis of a mental disorder labels the person rather than the problem.

Among the reasons for this failure to reduce CPP stigma is that the medical model says that all of the following have the same problem – a mental disorder: Smokers, people with schizophrenia, people caught in a panic cycle, psychopaths, spider phobics, people with Alzheimer’s disease, and women who rarely orgasm. The stigma of any of these is therefore transferred to the others.

It is reasonable to assume that most problems that people will present with will involve a dysfunction at either a biological level (with psychological consequences) or a psychological level (though of course physically located in the brain), but rarely by coincidence at both levels. An exception may be a person on the autism spectrum presenting with an acquired dog phobia.

A professional relationship analogous to that which should occur between psychiatrists and clinical psychologists is that between computer hardware engineers and software engineers. A PTSD is in this sense a software problem. Drugs may help and a neurological correlate may be detectable, but Prolonged Exposure and Cognitive Therapy are the frontline evidence-based treatments for this psychological-level problem. Autism Spectrum Disorder, on the other hand, is more a hardware problem. Social skills training may help, but it will never ‘cure’ autism, a problem of neurological development. DSM conflates both under the heading of biologically-based mental disorders.

Gratefully, the past 20 years have seen a natural sidestepping by psychologists of the ‘hegemony’ of psychiatric mental disorders (Anand & Malhi, 2011). Overlapping developments that have occurred in response have included:

• Transdiagnostic research, theory, and treatment development (McEvoy et al., 2009). This has focused on psychological processes underlying CPPs irrespective of what their surface diagnoses are. For example, the massive ‘comorbidity’ rates between GAD and MDD are explainable by the underlying cognitive processes that they have in common, which can subsequently be addressed by a broadly applicable transdiagnostic intervention or ‘unified protocol’ (Barlow et al., 2004)
• Much progress in research on Network Theory (McNally, 2021). This model regards mental disorders as cluster regions in a spatial network of interconnected, causally-related symptoms. This conception is much more consistent with a functional analytic approach to the detection and alteration of problematic psychological mechanisms. However, it is still about symptoms of mental disorders, equating them with CPPs.
• A new psychological-level conception and nascent taxonomy of CPPs – PMC Theory (Bakker, 2019, 2021a, 2021b) – has taken these two developments a step further. It moves away entirely from the medical model of CPPs by defining their essence as lying in the operation of psychological-level causal ‘problem-maintaining circles’ (PMCs) or vicious circles. Then the difference between a grief reaction and a clinical depression problem, or an acute stress reaction and a PTSD, lies not in their symptom profiles (which can be identical), but in whether a fully perpetuating cyclic causal process has begun to operate. This problem maintenance process is what justifies external professional intervention and makes the problem ‘clinical’. For example, with PTSD research has shown that coping by avoidance is the most frequent factor turning an acute stress reaction into a PTSD-like CPP (Keane & Barlow, 2002).

PMC Theory aims to provide the only two benefits that psychiatry’s DSM-based medical model has conferred for psychology, and that transdiagnostic and network approaches have not yet achieved: Communication through a shared nonmedical language, and codification of CPPs by the standardisation of the basic elements of a case formulation – the PMCs. A much more comprehensive taxonomy of PMCs than the brief listing available in Practical CBT (2021a) is being compiled, soon to be published in a case formulation handbook titled Standardised Case Formulation in Clinical Psychology.

PMC Theory has no possibility of supplanting the medical model of CPPs, which is far too entrenched in the public’s consciousness, and psychiatrists will naturally always assume that every CPP must be seen as a diagnosable mental disorder. That is their level of analysis, and psychiatry, being a medical specialty, will always have more administrative power and profile than psychology.

But clinical psychology must be able to construct its own models, language, and taxonomy of CPPs alongside. They are, after all, clinical psychological problems. In some quarters this has continued. Research, and therefore practice, has largely remained focused on process in health psychology and educational and developmental psychology, and it happens naturally among experienced case formulating clinicians. But it needs to happen much more in psychotherapy process and outcome research (Bakker, 2021b), the financing of which is controlled outside of psychology. We need to resume (a la the 1970s) the development of psychological change principles, theory, and processes based on psychological-level standardised case formulations (not diagnoses) that enable generalised conclusions and guidelines to be drawn (Bakker, 2008; 2021a).