Psychosis proneness scales in schizophrenia spectrum personality disorders: familial vs. nonfamilial samples.Psychiatry Res. 1993 Jan; 46(1):47-57.PR
The authors evaluated the extent of overlap between DSM-III-R schizophrenia spectrum personality diagnoses (SSPD) and the Psychosis Proneness Scales of Chapman and his associates. The subjects were recruited from the family members of schizophrenic patients ("familial" subjects; n = 45) and members of the community with negative family histories for schizophrenia ("nonfamilial" subjects; n = 60). Clinical interviews were performed to obtain DSM-III-R Axis I and II diagnoses. In 105 individuals with no Axis I diagnosis, the five Chapman Scales were administered. The results suggest that the nonfamilial subjects with diagnoses of SSPD (n = 24) scored significantly higher on the Chapman Scales of Magical Ideation, Perceptual Aberration, and Impulsive Nonconformity compared with the familial SSPD subjects and the other non-SSPD groups. The familial SSPD subjects (n = 17) scored significantly higher than the nonfamilial, non-SSPD groups on the Physical Anhedonia Scale. Scores on the Social Anhedonia Scale were highest in the SSPD subjects, but only scores for the nonfamilial SSPD subjects were statistically different from those for the other non-SSPD groups. The data were reanalyzed by first dividing the scores from the Chapman Scales into high and low scores based on different cutoff points. Sensitivities, specificities, and predictive powers of the "high" Chapman scores for SSPD diagnoses were then calculated. These were done because the Chapman Scales are often used to identify individuals with schizophrenia-related personality disorders on the basis of scores that exceed arbitrary cutoff points. The results suggest that the Chapman Scales (other than the Physical Anhedonia Scale) and DSM-III-R criteria identified mostly the same subjects, when only nonfamilial subjects were considered (with sensitivities and specificities of about 0.70). However, the overlap between these two constructs was poor when only schizophrenia spectrum subjects recruited from the family members of schizophrenic patients were considered.