Psychiatry Does Not Need a Softer DSM. It Needs a Smarter One.

COMMENTARY

Psychiatry has always been caught between 2 impulses: the desire to understand human suffering in all its complexity, and the need to define mental illness with the kind of precision expected in every other branch of medicine. DSM-6 now forces the field to decide which impulse wins. The stakes are not small. The central problem confronting psychiatry is not stigma, access, or even treatment innovation; it is credibility. And the source of the credibility problem is real: too many diagnoses still rest more on expert consensus than on clear scientific validity.

Psychiatric diagnosis in the United States did not emerge from biologically grounded disease classification. Early American psychiatry was shaped by psychoanalytic thinking, which treated symptoms as expressions of psychogenic conflict and symbolic meaning. DSM-I and DSM-II reflected that worldview. DSM-III was a major corrective; it introduced clearer criteria, a shared diagnostic language, and a more medical model. But it also dramatically expanded the number of psychiatric disorders, many built on uncertain foundations. The result was progress in reliability, clinicians could use the same language and arrive at similar labels, without enough progress in validity: whether those labels correspond to something real in nature, with a consistent course, meaningful biological correlates, predictable treatment response, and ideally genetic or biomarker support.

That is why DSM-6 matters. The American Psychiatric Association has signaled that DSM-6 may become a “living document,” updated as new evidence emerges. In principle, that is a good idea. But the direction of evolution matters far more than the format. If DSM-6 becomes more flexible, more inclusive, and more layered without becoming more scientifically grounded, it will make the same old problems more elaborate.

Some proposed shifts do make sense. Moving beyond a purely categorical model and recognizing spectrum phenomena better reflects clinical reality, particularly in bipolar disorder. Anyone who practices psychiatry knows that many patients do not present in neat boxes. But acknowledging dimensionality is not the same thing as improving validity. A looser framework is not automatically a better one.

The proposed 4-part diagnostic assessment, including contextual factors, severity ratings, and transdiagnostic features, illustrates both the promise and the danger of the current direction. Context matters. Social determinants of health matter. No serious clinician denies that. But these factors do not, by themselves, establish the validity of a psychiatric disease. If a disorder is real in the medical sense, it should show recognizable patterns across cultures and socioeconomic conditions. It should reveal coherence in symptoms, onset, family history, biology, or treatment response. The more DSM-6 shifts its center of gravity toward contextual description without improving the core science of diagnosis, the more it risks confusing distress with disease.

That is a dangerous mistake, and one psychiatry is already making. In daily practice, clinicians assign diagnoses to patients who do not cleanly meet criteria, or whose presentations are heavily confounded by substance use, medical illness, personality structure, or environmental chaos. Many patients clearly need help. But not all of them have a discrete psychiatric disease in the way we speak of bipolar disorder, schizophrenia, or obsessive-compulsive disorder. DSM-6 appears poised to allow broader and less specific labels, recording “psychosis” or “trauma” in place of schizophrenia or posttraumatic stress disorder in some settings. The intent may be to reduce stigma. But psychiatry increasingly acknowledges uncertainty without clearly defining where illness ends and nonspecific suffering begins. Broadening labels may make documentation easier. It does not solve the scientific problem. It obscures it.

DSM-6 should also confront diagnostic hierarchy more directly. Psychiatry remains plagued by overlapping symptom clusters—depression, anxiety, impulsivity, psychosis, irritability, sleep disruption, concentration problems—that surface across radically different conditions. In many cases, the most scientifically valid diagnosis is the one driving the rest. But our diagnostic systems often fail to establish that hierarchy, leading to bloated problem lists and clinical confusion. Transdiagnostic language may help describe overlap, but description is not enough. The field needs a cleaner way to determine what is primary, what is secondary, and what may not represent disease at all.

This concern is especially relevant when diagnostic criteria expand in ways that increase inclusivity without strengthening validity. Adult attention-deficit/hyperactivity disorder is a case in point. The solution to decades of underdiagnosis cannot simply be to make criteria more accommodating in an era already marked by rising diagnosis rates and widespread uncertainty about the boundaries of normal variation, stress-related dysfunction, and true neuropsychiatric illness. When a field is still debating the validity of many of its categories, expanding them should not be treated as progress by default.

The same discipline should apply to biomarkers. Psychiatry has spent years promising biological precision, yet very few biomarkers have meaningfully improved routine clinical care. For a biomarker to matter in psychiatry, it should be cheap, easy to obtain, reproducible, and supported by evidence that it improves diagnosis or treatment decisions. Otherwise, the field risks repeating the mistakes made with pharmacogenetic testing: impressive marketing, intuitive appeal, enthusiastic early adoption, and far less clinical value than advertised when the evidence is scrutinized carefully.

DSM-6 should be judged by a simple question: will it help psychiatry identify real mental illness more clearly, more reliably, and with greater scientific confidence? If the answer is yes, it will strengthen the field. If the answer is no, if it mainly offers broader frameworks, softer boundaries, and a larger vocabulary for distress, psychiatry will continue to drift from the standards of diagnostic rigor expected elsewhere in medicine.

DSM-6 should not be remembered for making diagnosis broader, softer, or more fashionable. It should be remembered for making psychiatry harder to dismiss as anything less than real medicine.

Dr Rossi is a board-certified psychiatrist specializing in inpatient and consultation-liaison psychiatry. His work focuses on evidence-based treatment, complex mood and psychotic disorders, and practical clinical decision-making. He is passionate about education, thoughtful skepticism, and advancing psychiatry through honest, nuanced discussion.