Doctors Agree On A Psychiatric Diagnosis Only 55% Of The Time, Study Finds

Why Two Psychiatrists Can Look At The Same Patient And See Two Different Illnesses

A patient walks into one psychiatrist’s office and leaves with a diagnosis of schizophrenia. A different patient, describing textbook symptoms of the same disorder to a different doctor, walks out labeled with OCD, a personality disorder, or bipolar disorder instead. That’s what a massive new international study found when it tested more than a thousand doctors using identical written patient cases.

Published in Frontiers in Psychiatry, the study gathered 1,038 medical doctors working in adult psychiatry across 19 countries in Europe and South America. Each one read written descriptions of fictional patients and assigned a diagnosis. Because everyone read the same words describing the same patients, disagreement couldn’t be blamed on one doctor asking different questions than another, or a patient having a rough day. Judgment alone was the variable, and it turns out to be wildly inconsistent when it comes to certain mental illnesses.

Two doctors looking at the same case agreed on a diagnosis only about 55% of the time. Doctors landed on the reference diagnosis established by an expert panel just 66% of the time overall. For cases involving schizophrenia, that figure sank much lower, with doctors frequently mistaking it for OCD, a personality disorder, or bipolar disorder, depending on which version of the case they saw.

Nine Written Cases, 1,038 Doctors, Two Diagnoses Each

Researchers built nine fictional patient cases reflecting disorders psychiatrists see regularly: three variations of schizophrenia, plus single cases representing schizotypal disorder, bipolar disorder, depression, generalized anxiety disorder, OCD, and a personality disorder. Each doctor was randomly given two of the nine cases to diagnose, choosing from a list of 30 common mental health diagnoses, producing 1,902 diagnostic decisions total. Most participants were psychiatrists, though psychiatric residents and other doctors working in adult psychiatry took part too, with experience ranging from brand-new doctors to those with 15 or more years in the field.

Experience made almost no difference to how often doctors agreed with each other. A doctor practicing for two years fared about the same as one with twenty, and psychiatrists matched residents and other medical doctors at similar rates. Job title and years on the job showed no meaningful link to agreement, pointing away from a simple lack of training. Something else drives the split.

That something else appears to be which disorder is on the table. Some cases produced near-unanimous agreement: bipolar disorder matched the reference diagnosis for 94% of doctors, OCD for 92%, depression for 88%. But schizophrenia cases told a different story. Across the study’s three schizophrenia vignettes, doctors often split between schizophrenia and alternatives, with one case dividing almost evenly between a personality disorder (45%) and schizophrenia (43%), and another seeing 41% call it OCD while only 37% matched schizophrenia.

Wrong Diagnoses Clustered Around the Same Alternatives

Wrong answers weren’t scattered randomly. When doctors missed a schizophrenia diagnosis, they tended to land on the same alternatives again and again, mainly OCD, a personality disorder, or bipolar disorder. Study authors argue doctors weren’t confused or guessing blindly. Instead, the pattern suggests they were weighing the same symptoms differently, often anchoring on obvious features like obsessive behaviors while giving less weight to the psychotic symptoms that should have pointed toward schizophrenia under standard classification rules.

Comparing these results to older research adds weight to the concern. The famous US-UK Diagnostic Project from the 1970s first exposed how differently American and British psychiatrists diagnosed the same patients, prompting a major overhaul of diagnostic manuals in the following decades. Those changes were supposed to fix the problem. This new study, along with agreement scores measured during the 2013 update to the American diagnostic manual, suggests it never fully took. Doctors here performed worse than doctors did during field trials for the third edition of the American diagnostic manual back in the 1980s, landing closer to the shakier scores seen in the decades before that manual existed.

A Wrong Label Can Follow a Patient Through Treatment

Diagnostic labels aren’t just paperwork. They shape treatment, what a doctor tells a family about the road ahead, and how a patient is understood by every future clinician who reads their chart. If similar disagreement plays out in real practice, it could affect treatment decisions, prognosis, and continuity of care.

It matters for research too. Scientific studies of mental illness assume a group of “schizophrenia patients” all share the same underlying condition. If doctors sort patients into that category inconsistently, research samples may be muddier than assumed, which could help explain why it’s been hard to find clean biological explanations or reliably effective treatments tied to specific psychiatric labels. Study authors note this may also explain why transdiagnostic approaches, treatments that don’t rely on a single precise label, have been gaining ground.

Real limits apply here, though. Written cases are a controlled, near best-case scenario compared to a live interview, where confusion, incomplete information, and rushed appointments can make things messier still. Researchers themselves suggest their agreement numbers likely mark a ceiling, meaning ordinary practice could produce even lower agreement than what showed up here.

Fifty years after the field’s reliability problem was first exposed, and decades after supposedly fixing it with stricter diagnostic rulebooks, psychiatry still hasn’t closed the gap between two doctors looking at the same case and reaching the same conclusion. That’s a crack running through how mental illness gets identified, treated, and studied.

---

Disclaimer: This article summarizes findings from a peer-reviewed study for general informational purposes and is not a substitute for professional medical or psychiatric evaluation. Anyone with concerns about a diagnosis should speak with a licensed mental health provider.

---