1 Study protocol for the development and internal validation of SPIRIT (Schizophrenia Prediction of Resistance to Treatment): A clinical tool for predicting risk of treatment resistance to anti-psychotics in First Episode Schizophrenia.

Treatment Resistant Schizophrenia (TRS) is associated with significant impairment of functioning and high treatment costs. Identification of patients at high risk of TRS at their initial diagnosis may significantly improve clinical outcomes and minimize social and functional disability. We aim to develop a prognostic model for predicting the risk of TRS in patients with First Episode Schizophrenia, and to examine its potential utility and acceptability as a clinical decision tool. We will use two well-characterised UK-based first episode psychosis cohorts: AESOP-10 and GAP for which data has been collected on sociodemographic and clinical characteristics. We will identify candidate predictors for the model based on current literature and stakeholder consultation. Model development will use all data, with the number of candidate predictors restricted according to available sample size and event rate. A model for predicting risk of TRS will be developed based on penalised regression, with missing data handled using multiple imputation. Internal validation will be undertaken via bootstrapping, obtaining optimism-adjusted estimates of the model’s performance. The clinical utility of the model in terms of clinically relevant risk thresholds will be evaluated using net benefit and decision curves (comparative to competing strategies). Consultation with patients and clinical stakeholders will determine potential thresholds of risk for treatment decision making. The acceptability of embedding the model as a clinical tool will be explored using focus groups with clinicians in early intervention services.