Sensitivity of the familial high-risk approach for the prediction of future psychosis: a total population study

Abstract

Children who have a parent with a psychotic disorder present an increased risk of developing psychosis. It is unclear to date, however, what proportion of all psychosis cases in the population are captured by a familial high-risk for psychosis (FHR-P) approach. This is essential information for prevention research and health service planning, as it tells us the total proportion of psychosis cases that this high-risk approach would prevent if an effective intervention were developed. Through a prospective cohort study including all individuals born in Finland between January 1, 1987 and December 31, 1992, we examined the absolute risk and total proportion of psychosis cases captured by FHR-P and by a transdiagnostic familial risk approach (TDFR-P) based on parental inpatient hospitalization for any mental disorder. Outcomes of non-affective psychosis (ICD-10: F20-F29) and schizophrenia (ICD-10: F20) were identified in the index children up to December 31, 2016. Of the index children (N=368,937), 1.5% (N=5,544) met FHR-P criteria and 10.3% (N=38,040) met TDFR-P criteria. By the study endpoint, 1.9% (N=6,966) of the index children had been diagnosed with non-affective psychosis and 0.5% (N=1,846) with schizophrenia. In terms of sensitivity, of all non-affective psychosis cases in the index children, 5.2% (N=355) were captured by FHR-P and 20.6% (N=1,413) by TDFR-P approaches. The absolute risk of non-affective psychosis was 6.4% in those with FHR-P, and 3.7% in those with TDFR-P. There was notable variation in the sensitivity and total proportion of FHR-P and TDFR-P cases captured based on the age at which FHR-P/TDFR-P were determined. The absolute risk for psychosis, however, was relatively time invariant. These metrics are essential to inform intervention strategies for psychosis risk requiring pragmatic decision-making.