Since few studies have focused on the diagnostic process foregoing a case record diagnosis of schizophrenia, the present study was undertaken with the aims to examine the time interval between the onset of the illness, the first admission as inpatient, and the timepoint of a case record diagnosis of schizophrenia. Further, the study aimed at analyzing the influence of demographic and clinical variables on an 'early' or 'late' diagnosis of schizophrenia. The records of 84 inpatients treated during the calendar year 1986, were retrospectively analyzed. Only 21.4% of the patients had received a case record diagnosis of schizophrenia at their first admission into hospital, and additionally 16.7% after two admissions. Those patients with 'early' diagnosis differ significantly from patients with 'late' diagnosis, being more frequently men, having longer admission latency, higher frequency of Schneiderian first-rank and negative symptoms, a longer stay at hospital at their first admission, and lower levels of working capacity at the onset of the illness. The data suggest a propensity to include course and social functioning aspects of poor outcome in the diagnostic process. Confining the diagnosis of schizophrenia to the severe cases indicates a conservative, perhaps tautological, approach to this diagnosis.