Theoretically, schizotypal features should be prevalent in patients at ultra high risk of psychosis. In connection to the European Prediction of Psychosis Study (EPOS), we could study their prevalence in this group using three different ways of assessing schizotypality.

EPOS dataset comprises a large sample (n=246) of UHR patients, who were followed up for 18 months. Schizotypal features were assessed in connection to SIPS interview (SIPS-STY, researcher assessment), and with PDQ-R and SPQ scales (self-assessment). Descriptive data and intercorrelations between different measures are described. Concurrent validity of these three measures is assessed by externals validators (genetic risk/ neuropsychology).

The prevalence of schizotypal pdo was 13.4% with SIPS-STY and 34.6% with PDQ-R-STY. These categorial measures were poorly correlated (k=0,11). Of continuous measures PDQ-R-STY and SPQ scores were highly correlated (r=0,78, P<,000), but SIPS-STY score was only weakly correlated with these other measures (r=0.24).

As to external validation, FDRs of psychotics did not differ from other subjects on the level of schizotypal features. PDQ-R-STY but not SIPS-STY was associated with lower verbal IQ (P=0.004). In verbal fluency test, both SIPS-STY and PDQ-R-STY contributed to poor performance, but SPQ did not add to this. Schizotypal status did not associate with results of the Spatial working memory paradigm (SWMT).

Different measures of schizotypality produce somewhat inconsistent results when studied in a high psychosis risk sample. PDQ-R questionnaire seemed to give results most consistent with the current notion of schizotypicality.