Does racial bias influence psychiatric diagnoses assigned at first hospitalization?

Background: Research on racial bias in psychiatric diagnosis has largely been limited to studies of admission diagnoses assigned to chronically ill patients. This study tests whether racial bias influences diagnoses assigned to patients at discharge from their first psychiatric hospitalization. Methods: In a county-wide sample of patients with psychosis, hospital diagnoses were compared with research diagnoses formulated using structured interviews and strict adherence to DSM-III-R. Symptom patterns were also examined. Results: Racial differences were observed in the distribution of both hospital and research diagnoses. Using research diagnoses as the gold standard, the sensitivities and specificities of hospital diagnoses were similar by race (for blacks the sensitivity and specificity of schizophrenia was 0.33 and 0.91, and for whites, 0.43 and 0.89). The only suggestion of possible bias was that more blacks were discharged without a definitive diagnosis (38.7% of blacks vs. 26.3% of whites, χ2 = 5.80, df = 1, p = 0.02). Conclusions: We did not observe the expected racial bias in the assignment of diagnoses of schizophrenia and affective disorders. While there was evidence that hospital clinicians had more difficulty diagnosing black patients, the low concordance between hospital and research diagnoses for both black and white patients demonstrates the need to better understand the clinical diagnostic process for all patients with psychotic disorders at their first hospitalization.