Symptoms aren't enough to make psychiatric diagnoses. Symptom differences do not mean that diagnoses are different also. Just as pneumonia with chest pain is not a different diagnosis than no pneumonia without chest pain, so too psychiatric symptoms need to be validated by other non-symptom criteria, if they are to reflect different diagnoses.
In medicine, pathology and laboratory tests are those validators. In psychiatry, we have no useful pathology, so the classic diagnostic validators, besides symptoms, are genetics, course, and treatment effects. In a classic 1970 paper, Robins and Guze explained the reasons for this approach, reasons which are useful today for the practicing clinician with each and every diagnostic evaluation.
Access the first Psychiatry Letter newsletter to read more.