Is our concept of schizophrenia influenced by Berkson's bias?

Abstract

Background: If both positive and negative dimensions of schizophrenia independently influence need for care, a higher estimate of the comorbidity between these dimensions is expected in clinical samples than would be the case if non-clinical cases were investigated (i.e. Berkson's bias). The present paper investigates whether positive and negative dimensions independently contribute to mental health care use in a general population sample.

Method: A prospective cohort study was conducted, in which 7076 individuals were interviewed with the Composite International Diagnostic Interview Schedule at baseline and 1 and 3 years later. Lifetime positive and negative psychotic experiences (PPE, NPE) were assessed at baseline. Mental health care use (MHC) was assessed at baseline, and prospectively 1 and 3 years later.

Results: The rating of MHC was strongly associated with NPE and PPE, both retrospectively and prospectively and the effects of both variables remained strong and significant in the analyses with both variables included.

Conclusions: These results, therefore, suggest that the concept of schizophrenia, as a unitary entity with high comorbidity between positive and negative dimensions, is in part the result of Berkson's bias.