The “clinical high risk” (CHR) for psychosis syndrome is an antecedent period characterized by attenuated
psychotic symptoms that are marked by subtle deviations from normal development in thinking, motivation,
affect, behavior, and a decline in functioning. Early intervention in this CHR population is critical to prevent
psychosis onset as well as other adverse outcomes. However, the presentation of symptoms and subsequent
course is highly variable, and there is a paucity of biomarkers to guide treatment development. Thus, to improve
predictive models that are clinically relevant, several issues need to be addressed: 1) focusing on outcomes
beyond psychosis; 2) taking into account heterogeneity in samples and outcomes; and 3) integrating data sets
with a broad array of variables using innovative algorithms to overcome variability across studies. To address
these challenges, the proposed “Psychosis Risk Evaluation Data Integration and Computational Technologies:
Data Processing, Analysis, and Coordination Center” (PREDICT-DPACC) brings together a multidisciplinary
team of highly experienced researchers with proven capabilities in all aspects of large-scale studies, CHR
studies, as well as computational expertise. The ultimate goal is to identify new CHR biomarkers, and CHR
subtypes that will enhance future clinical trials. To do so, the PREDICT-DPACC will 1) aggregate extant CHR-
related data sets from legacy datasets; 2) provide collaborative management, direction, data processing and
coordination for new U01 multisite network(s); and 3) develop and apply advanced algorithms to identify
biomarkers that predict outcomes, and to stratify CHR into subtypes based on outcome trajectories, first from
the extant data and then refined and applied to the new data. The PREDICT-DPACC team has the broad,
comprehensive, and robust infrastructure that is sufficiently flexible to accommodate the inclusion of multiple
data types and to optimally address the needs of the CHR U01 network(s). Carefully selected extant data will be
rapidly obtained, processed, and uploaded to the NIMH Data Archive (NDA). Proposed analysis methods are
powerful and robust, leveraging the expertise and experience of computer scientist developers, and experienced
clinical researchers. The U01 network(s) will be coordinated by a team that is experienced in managing large
studies, familiar with the needs of such studies, flexible, and is knowledgeable in all aspects of CHR studies,
including measures, outcomes, biomarkers, and cohorts. Upon meeting the goals of this U24, and the supported
U01 network(s), the expected outcomes of the PREDICT-DPACC will be new predictive biomarkers for CHR
outcomes, new definitions of CHR subtypes that are clinically useful, and new curated and comprehensive CHR
datasets (extant and new) as well as processing tools and prediction algorithms that are shared with the research
community through the NIMH Data Archive.
