Abstract
BACKGROUND AND HYPOTHESIS: Subclinical psychotic symptoms (also known as psychotic experiences comprising positive features only, and schizotypy comprising positive, negative, and disorganized features) are important markers of schizophrenia liability. Different assessment methods detect different sources of meaningful variance and are vulnerable to different biases and sources of measurement error. Whereas interview-rated psychotic symptoms in childhood are known to predict adult schizophrenia diagnosis, the predictive value of parent-rated psychotic symptoms remains unknown. We tested whether clinician-rated psychotic symptoms and parent-rated positive, negative, and disorganized schizotypy in early adolescence are nonredundant predictors of schizophrenia diagnosis by age 38 years. STUDY DESIGN: In a representative birth cohort (n = 1037) from Dunedin, New Zealand, psychotic symptoms were assessed by clinical interview at age 11 years, schizotypy was assessed by parent or caregiver ratings at ages 13- and 15 years, and lifetime schizophrenia diagnosis was assessed throughout adulthood until age 38 years. We tested for redundancy using bootstrapped multivariable logistic regression. STUDY RESULTS: Clinician-rated psychotic symptoms at age 11 predicted adult schizophrenia diagnosis (OR = 2.68, 95% CI = 1.42, 5.06), as did parent-rated total schizotypy (OR = 1.83, 95% CI = 1.42, 2.36). In univariable models, clinician-rated psychotic experiences and parent-rated positive, negative, and disorganized schizotypy were significant predictors of schizophrenia diagnosis. In multivariable models where clinician- and parent-rated scores were entered, only parent-rated negative and disorganized schizotypy did not predict adult schizophrenia diagnosis. CONCLUSIONS: Parent-rated schizotypy and clinician-rated subclinical psychotic symptoms are valid, nonredundant indicators of lifetime risk for schizophrenia.