A Psychiatric Diagnosis Is Not a Disease
What DSM diagnoses can and can’t tell us.
In my first week as a psychiatry faculty member, an advanced psychiatry resident—I’ll call her Dr. G—staffed a case with me. That’s medical-speak for discussing a patient with a teacher. Dr. G gave me some demographic information and then began listing medications she was prescribing.
“Hold on,” I said. “What are we treating her for?”
“Anxiety.”
“How do you understand her anxiety?”
Dr. G cocked her head with a blank, non-comprehending look. I rephrased. “What do you think is making your patient anxious?”
She cocked her head to the other side.
“What is causing her anxiety?"
Dr. G pondered, then brightened. “She has generalized anxiety disorder.”
“Generalized anxiety disorder isn’t the cause of anxiety,” I explained. “That’s just the term we use to describe it.”
Another blank look. I tried a different tack. “What do you think is going on psychologically?”
“Psychologically?”
“Yes, psychologically.”
“I don’t think it’s psychological, I think it’s biological.”
“Okay, that’s a start,” I said. “Tell me why you think that.”
“Her mother was anxious.”
“This means your patient’s anxiety is biological?”
“Yes.”
It was my turn to cock my head.
“Let’s try a thought experiment. Suppose your patient was adopted at birth and is not biologically related to the mother who raised her. Do you think an anxious mother, who continually communicates that the world is unsafe, could make a child anxious?”
“I never thought about it that way.”
I suppressed a momentary urge to bang my head against the cinderblock wall. Then I signed Dr. G’s treatment plan and hoped I had planted at least a seed of curiosity.
Saying anxiety is caused by generalized anxiety disorder makes as much sense as saying anxiety is caused by anxiety.
Diagnoses listed in the DSM—the Diagnostic and Statistical Manual of Mental Disorders, the so-called bible of psychiatry—do not cause anything. They are not things. They are agreed-upon labels—a kind of shorthand—for describing clusters of symptoms.
Generalized anxiety disorder means a person has been anxious or worried for six months or longer and it’s bad enough to cause problems—little else. The diagnosis is description, not explanation. Saying anxiety is caused by generalized anxiety disorder makes as much sense as saying “anxiety is caused by anxiety.”
The same is true for other common DSM diagnoses. Major depressive disorder means a person has had continually depressed mood, or lack of interest or pleasure in activities, for two weeks or longer, along with several other symptoms that commonly co-occur. Major depressive disorder does not cause these symptoms; it is the term we use to describe them.
Here’s the circular logic: How do we know a patient has depression? Because they have certain symptoms. Why do they have these symptoms? Because they have depression.
Confusion arises because, in medicine, diagnoses often point to etiology—underlying causes. This is why “chest pain” is not a disease; it is a symptom. Atherosclerosis, myocarditis, and pneumonia are diseases. They are underlying conditions that can cause chest pain.
DSM diagnoses are categorically different because they are merely descriptive, never explanatory. It’s not that we don’t know their causes yet. It’s that DSM diagnoses cannot speak to causes, now or ever. The DSM was not designed to address causes, only to describe effects.
Most psychiatrists subscribe to a “biopsychosocial” model, which means they see the causes of mental health problems as biological, psychological, social, or an interaction between them. But the DSM is not organized by these domains. It does not address them, nor attempt to.
DSM diagnoses are merely descriptive, not explanatory.
The entries in the DSM sound like medical diseases, especially with the term disorder appended to every entry, but they are not. If we speak of generalized anxiety disorder or major depressive disorder as if they were equivalent to pneumonia or diabetes, we are committing a logical fallacy called a category error. A category error means ascribing a property to something that cannot possess it—like emotion to a rock.
The American Psychiatric Association, which publishes the DSM, made this very point. Until recently, its website included this crystal-clear caveat about the DSM:
“Diagnostic criteria provide a common language for clinical communication... Patients sharing the same diagnostic label do not necessarily have disturbances that share the same etiology, nor would they necessarily respond to the same treatment.”
—American Psychiatric Association
When the National Institute of Mental Health concluded that DSM diagnostic categories do not and cannot map underlying causes—and therefore cannot be a foundation for mental health research—the American Psychiatric Association agreed: “DSM, at its core… is a guidebook to help clinicians describe,” the chair of the DSM-5 Task Force wrote in response. “It provides clinicians with a common language.”
How could Dr. G misunderstand something so basic? How did she come to think of generalized anxiety disorder as a disease that causes anxiety?
Our struggling students, in their moments of concreteness, hold up a mirror to the hypocrisies in our field.
The American Psychiatric Association says patients with the same diagnosis do not necessarily have the same disturbances or respond to the same treatments. Then researchers develop treatment manuals for DSM diagnoses. Professional organizations publish clinical practice guidelines for DSM diagnoses. Health insurers ask providers to follow treatment algorithms for DSM diagnoses. And pharmaceutical companies run television ads that say, “Depression is a distinct medical condition… It causes intense mood and physical symptoms.” Of course they do.
We say DSM diagnoses are constructs, not things—then blithely proceed to speak of them as if they were things. Doublethink, anyone?
Recently, I saw a self-exam in the American Psychiatric Association’s study guide for DSM-5. The format is case vignettes followed by multiple-choice diagnoses. One vignette described a patient with a fear of flying, followed by the question: “Which of the following disorders is the most likely cause of his anxiety?”
My write-in answer would have been: “None of the above, because DSM diagnoses are descriptive labels, not causes.”
The study guide answer was “c) Specific-phobia—situational type.” I would have failed that exam. My student, Dr. G, would have aced it.
“Mental disorders are constructs, not diseases. Descriptive, not explanatory.
Helpful in communication/treatment planning.
But no claims re: causality/homogeneity/clear boundaries.
We wrote all this in DSM-IV Intro—but no one read it.”
—Allen Frances, DSM-IV Editor-in-Chief, Twitter/X, July 29, 2019
More essays, interviews, clips, and reflections: linktr.ee/jonathanshedler
DSM labels describe patterns; they don’t explain why those patterns exist or persist. When diagnoses start standing in for causes, curiosity shuts down and formulation collapses into medication lists and algorithms. The reminder here is simple but consequential: understanding requires asking what functions symptoms serve, how they were learned, and what contexts sustain them. Without that, treatment becomes technically busy and conceptually empty.
We need more supervisors like this in psychiatry residencies. There is such a poverty of psychology didactics and supervision at most programs. They took almost all the psychoanalytic and psychodynamic didactics out of our curriculum because of student feedback from prior years. I dont think many residents understand the scope of the field enough to recognize what they need.
So so often during my residency training the health care system was pushing residents to see more and more and more patients such that the quality of care and supervision became seriously compromised. Depth and nuance was often hollowed away until all that was really prioritized for a vast majority of cases was agreeing on the DSM diagnosis and medications. Then even that fell away and it was mostly about putting out fires; or worse, making notes compliant with insurance reimbursement coding.
I went to a strong general psychiatry program (University of Michigan) and a decent child and adolescent program ( Loma Linda University) - graduated last year. My experience has led me to believe that most of our current health systems aren't really built to facilitate treating, nonetheless teaching, even basic biopsychosocial formulations and treatment. Which means most psychiatrists don't understand these fundamentals - unless they've done a substantial amount of self-study and found/created an environment that supports that way of practicing after training (usually private practice).
That dynamic then furthers the division between the psychiatry practiced at academic institutions vs more nuanced psychologically minded approaches. Im not saying it is impossible to teach psychiatry in a psychologicaly minded way at an academic institution - some people find ways to do it amidst the minimal time for supervision, unrealistic efficiency expectations, and smoldering moral distress inherent in working within those systems - but it is no small feat.