Accurate and appropriate cognitive screening has the potential to significantly enhance early psychosis care, yet no screening tools have been validated for the early psychosis population and little is known about current screening practices, experiences, or factors that may influence implementation. Using a hybrid type 1 design, the CogScreen study aims to validate two promising screening tools with young people with first episode psychosis (primary aim) and to understand the context for the implementation of cognitive screening in early psychosis settings (secondary aim). The present protocol outlines the implementation study, which aims to explore the experiences and practices, acceptability, feasibility, and determinants of cognitive screening in early psychosis settings from the perspective of key stakeholders. Young people with first episode psychosis (n=350), caregivers (minimum n=10), and service providers (minimum n=12) will be recruited from primary and specialist early psychosis services in Melbourne, Adelaide, and Sydney, Australia. Grounded in implementation science, two theoretical frameworks inform data collection and analysis: the Theoretical Framework of Acceptability and the Consolidated Framework for Implementation Research. A mixed methods design will be employed to collect and analyse data from questionnaires with young people with first episode psychosis, interviews with all stakeholder groups, and administrative processes. Quantitative data will be analysed using descriptive statistics. Qualitative data will be analysed through content analysis using deductive and inductive coding. Together with accuracy findings, results from the present implementation study will provide new insights about the practices, experiences, enablers and barriers to cognitive screening in early psychosis services.

Background. Mental illnesses mostly occur before 25 years of age when early identification and intervention could delay, ameliorate, or prevent lifelong disability. However, several challenges hinder optimal psychiatric care in youth due to the diversity of presentations, comorbidities, illness courses, and treatment responses. It is increasingly clear that one-size-fits-all approaches will not be effective and personalised approaches are needed to guide therapeutic pathways for individuals. Following precedents in other medical fields, machine learning approaches may address these needs by providing decision support tools that predict diagnoses, prognoses, and therapies. This paper outlines how such a future of augmented care may be facilitated through the establishment of the world’s first ‘Prediction of Early Mental Disorder and Preventative Treatment Centre of Research Excellence’ (PRE-EMPT). Methods. Five key components of the Centre will be addressed, including interdisciplinary collaboration, data harmonization, methods sharing, development of decision support tools, and capacity building. Conclusion. Implementation of research methodology guidelines within the Centre will assist in the production of reproducible, transparent, and open-source techniques and models. When combined, this paper will address key challenges facing the future of psychiatric care for youth and a roadmap of how to address them.