Background: The diagnosis of schizophrenia by clinicians is not always accurate in terms of operational diagnostic criteria despite the fact that these diagnoses form the basis of case registers and routine statistics. This poses a challenge to psychiatric research. We studied the reasons for diagnostic discordance between clinicians and researchers.
Methods: The Northern Finland 1966 Birth Cohort (n = 11,017) was followed from mid-gestation to the end of the 31st year. Psychiatric outcome was ascertained through linkage to the national hospital discharge register containing clinical diagnoses made by the attending physician. The hospital notes of all subjects admitted to hospital during the period 1982-1997 due to psychiatric disorder were reviewed and 475 research, operational DSM-III-R diagnoses were formulated.
Results: Ninety-six cases met operational criteria for schizophrenia. Fifty-five (57 %) had concordant diagnoses: both the clinical and research diagnoses were schizophrenia. Forty-one (43 %) had discordant diagnoses: the clinical diagnosis was other than schizophrenia (mainly schizophreniform or other psychosis). Discordant cases were more likely to be older at onset, experience a shorter treatment duration, fewer treatment episodes, and to have a comorbid diagnosis mental retardation.
Conclusions: Clinicians do not make the diagnosis of schizophrenia as often as the application of operational criteria would suggest they should. The discordance between clinical diagnosis and the research, operational diagnosis is especially likely in cases having late onset and few contacts to psychiatric hospital.