the case against psychiatric diagnosis

when i was studying psychology, i kept waiting to feel convinced. convinced that the system worked, that the categories made sense, that the map matched the territory. that certainty never came. the more i learned, the more the framework of modern mental health diagnosis seemed like an elaborate structure built on a shaky foundation. and the more i understood why i couldn’t see myself practicing within it.

it turns out i’m not alone. a growing body of researchers and clinicians appear to share this concern. the way we categorize mental illness, they argue, is fundamentally misconceived. and the consequences of getting this wrong extend far beyond academic debate.

541 categories and counting

the diagnostic and statistical manual of mental disorders (dsm) is the definitive reference guide for mental health diagnosis in much of the world. its origins trace back to the late 19th century, when psychiatrists, inspired by classification systems emerging in biology and medicine, began grouping symptom clusters into named disease entities. the ambition was scientific credibility: if you could name it, describe its onset and course, you could study and treat it. the most recent edition, the dsm-5, contains 541 diagnostic categories.

the neo-kraepelinian model, adopted after psychiatry faced a credibility crisis in the 1960s and 70s, assumes that a cluster of symptoms reflects a single underlying disease entity. in the same way a persistent cough might point to a lung tumour, the symptoms are read as evidence of something deeper.

however, kraepelin himself, the founding father of this categorical model, acknowledged its limitations. while his descriptions captured what he observed in patients, the boundaries between syndromes were not clearly defined. over a century later, the boundaries remain just as blurred, and the categories have only multiplied.

consider major depressive disorder: research shows there are up to 10,377 unique symptom combinations that qualify for the same diagnosis. two people can receive identical labels while sharing almost no overlapping experiences. studies on common factors across diagnostic boundaries also reveal that very few biological or psychological processes are actually unique to individual diagnoses. the categories are simultaneously too broad and too rigid, capturing too much while explaining too little.

the most effective disorder-specific psychological interventions achieve recovery rates around 50-80%. when the treatment is built on a flawed conceptual model, its reach is inherently limited.

the cost of the wrong map

the limitations of the dsm play out in real clinical encounters, with real consequences for the people seeking help.

take the cluster of conditions labelled personality disorders. as psychologist aidan wright and colleagues have argued, the label is both stigmatizing and potentially misleading. imagine being told that your difficulties are rooted in an enduring feature of your personality. the implicit message is that the problem is you, rather than your experiences or circumstances. this kind of labelling can create shame, worsen outcomes, and be used by clinicians, consciously or not, to place the burden of dysfunction on the individual.

the defining criteria of personality disorders, that they are enduring and pervasive, also fails empirical scrutiny. for instance, many cases of depression and schizophrenia are equally long-standing. evidence shows that personality disorder symptoms frequently improve over time, contradicting its categorical definition. yet the label persists, carrying all its clinical and human weight, because the system has no adequate alternative.

when diagnosis-specific treatment is the standard, patients with multiple diagnoses face a fragmented, inefficient experience: multiple clinicians, treatment courses, and frameworks that may not speak to each other. and that is assuming they are among the fortunate ones. in practice, even accessing a single clinician and course of treatment is challenging enough, and for those with complex presentations, evidently insufficient.

symptoms talk to each other

a radical challenge to the dsm model comes from dutch psychometrician denny borsboom, whose network perspective reframes the premise of psychiatric diagnosis.

the dsm assumes that symptoms are caused by an underlying disorder. fatigue, hopelessness, and low mood are expressions of so-called “depression,” the way a fever is an expression of infection. borsboom inverts this. in his model, symptoms cause each other. fatigue worsens low mood; low mood deepens hopelessness; hopelessness disrupts sleep, which worsens fatigue. depression, in this account, is not the entity producing these experiences, but rather a system that emerges from their causal interactions.

this reframing has significant implications for how we think about the frequent co-occurrence of supposedly distinct conditions. when anxiety and depression appear together, the dsm treats this as two diseases happening simultaneously. the network perspective proposes that they share symptoms, and those symptoms reinforce each other.

borsboom’s framework also removes what he deems an “ineffectual search for biomarkers”: decades of research trying to locate the biological fingerprint of each condition. instead, the focus shifts to how environmental and internal stressors activate symptoms, and how interventions can disrupt the feedback loops that sustain them.

the transdiagnostic model

while borsboom’s network perspective challenges the conceptual model of diagnosis, a parallel movement in clinical research is building an alternative treatment framework: the transdiagnostic approach.

according to this model, depression, anxiety, ocd, and trauma-related conditions, presented by the dsm as distinct diseases, share the same underlying processes. in 2004, psychologist allison harvey and colleagues identified a set of cognitive-behavioural mechanisms common across these conditions: selective attention toward negative stimuli, avoidance of distressing experiences, safety behaviours that provide short-term relief while reinforcing long-term fear, and negative interpretations of ambiguous situations. these are the common architecture of psychological distress.

also in 2004, psychologist david barlow and colleagues made the case for unified treatments for emotional disorders, arguing that the proliferation of separate treatment protocols was creating unnecessary complexity. if the underlying processes are shared, the treatment logic should be too.

the transdiagnostic model carries a destigmatising potential. by defining mental health along a continuum, one that all of us exist on to varying degrees, it de-pathologizes the experience of distress. the anxious thoughts, low moods, and irrational beliefs that characterise diagnosed conditions are human experiences, turned up. the difference is one of degree, rather than kind.

so why hasn’t anything changed?

the evidence for a better system has existed for decades. experts have been publishing compelling alternatives since at least the early 2000s. the former director of the us national institute of mental health openly questioned whether dsm diagnostic categories were anything more than reified labels. and yet the dsm remains the standard.

part of the issue is likely institutional inertia. classification systems embed themselves into insurance codes, research funding structures, training programmes, and clinical guidelines. changing the map means changing the infrastructure built around it, which would be slow and costly.

there is also a financial incentive worth considering. the dsm system is the foundation on which disorder-specific pharmaceutical development is built. drug pipelines are organized around diagnostic categories: a medication for depression, a medication for anxiety, a medication for ocd. if the categories themselves are scientifically unstable, the rationale for disorder-specific pharmacological solutions becomes considerably more complicated. whether this dynamic contributes to the system’s persistence is worth questioning, even if it resists a simple answer.

a simpler factor may also be at play: categorical diagnosis offers a shortcut, and shortcuts are appealing in under-resourced, time-pressured clinical environments. the problem arises when the shortcut leads somewhere that doesn’t quite fit.

where the system is moving

none of this is to dismiss the utility of diagnosis. it can provide patients with meaningful validation, a language for experiences that have felt isolating and confusing, and a pathway to peer support and tailored care. the system also serves logistical functions: identifying who needs treatment, grouping participants in clinical research, and informing resource allocation.

there are signs, however slow, that the conceptual model driving the categories is beginning to shift.

the hierarchical taxonomy of psychopathology (HiTOP), developed by more than 70 classification researchers, data scientists, psychologists, and psychiatrists, offers a promising structural alternative. rather than assigning patients to discrete categories, HiTOP describes mental health according to specific symptoms, broader syndromes, and higher-order factors that capture shared elements across conditions. modular approaches, like MATCH (modular approach to therapy for children with anxiety, depression, trauma, or conduct problems), allow clinicians to select evidence-based components and personalize treatment around an individual’s difficulties. emerging interventions like rumination-focused cbt and memory flexibility training target shared underlying processes rather than disorder-specific symptoms.

the nimh’s research domain criteria (RDoC) initiative explicitly aims to move research beyond categorical systems, driving investigation of transdiagnostic processes instead.

towards a better map

these systemic changes have a ripple effect. better classification has the potential to improve how research is designed, which biopsychosocial processes get studied, and how treatment protocols are developed and evaluated. at the individual level, a framework that reflects the actual complexity of psychological distress, rather than reducing it into categories that may not fit, could change what it feels like to seek help. the map and the territory might finally begin to align.