Chances and risks of predicting psychosis

Abstract

Prevention is currently regarded a promising strategy for fighting the unfavorable consequences of psychosis. Yet, for the error probability inherent in any predictive approach, benefits and costs must be carefully weighed against each other. False attribution of risk may unnecessarily provoke stress and anxiety, and lead to unwarranted intervention exposure. However, clinical risk samples already exhibit psychopathological symptoms, cognitive and functional impairments, and help-seeking for mental problems. Thus, the risk of futile interventions is low as long as preventive measures also provide treatment for current complaints. Differentiation between still normal and clinically relevant mental states is another challenge as psychotic-like phenomena occur frequently in the general population, especially in younger adolescents. Reported prevalence rates vary with age, and if severe in terms of frequency and persistence, these phenomena considerably increase risk of psychosis in clinical as well as general population samples. Stigmatization is another concern, though insufficiently studied. Yet, at least more severe states of risk, which are accompanied by changes in thinking, feeling, and behavior, might lead to unfavorable, (self-) stigmatizing effects already by themselves, independent of any diagnostic "label," and to stress and confusion for the lack of understanding of what is going on. To further improve validity of risk criteria, advanced risk algorithms combining multi-step detection and risk stratification procedures should be developed. However, all prediction models possess a certain error probability. Thus, whether a risk model justifies preventive measures can only be decided by weighing the costs of unnecessary intervention and the benefits of avoiding a potentially devastating outcome.