Prescribing high-dose antipsychotics is typically reserved for individuals with treatment-resistant severe mental illnesses, such as schizophrenia, bipolar disorder and psychotic depression. It carries an increased risk of adverse drug effects, necessitating regular monitoring. Non-mental health specialist clinicians may not always be aware when the maximum recommended dose of antipsychotics is exceeded, leading to unintentional high-dose prescribing without recognising the need for additional monitoring or understanding the associated risks. Therefore, providing clinical decision support (CDS) tools to support clinicians and improve the appropriate prescribing of antipsychotics is important. The aim of this study is to understand current prescribing practices and assess the impact of high-dose antipsychotic prescribing on clinical outcomes among hospitalised patients. The findings from this study will shape a future project focused on developing an integrated computerised CDS tool.

This retrospective cohort study will examine antipsychotic prescribing among hospitalised patients using Hospital Electronic Prescribing and Medicines Administration data in Scotland from 2019 to 2023, in linkage with hospital records, Scottish Morbidity Records and primary care prescribing (Prescribing Information System). Patients will be grouped into those prescribed high-dose (exposed), defined as exceeding the 100% maximum recommended British National Formulary dose and normal-dose (unexposed) antipsychotics, followed from their first ever antipsychotic prescription date (index date) until the end of the study, study outcomes or death, whichever happens first. We will quantify high-dose antipsychotic prescribing, profile patient characteristics and use machine learning techniques to assess associations of high-dose antipsychotic prescribing with clinical outcomes, including harms and benefits, but will not attempt to establish causality.

The Health and Social Care Public Benefit and Privacy Policy Panel (HSC-PBPP) has granted ethical approval (ref. 2024-0239) following a Data Protection Impact Assessment, with data securely held and accessed in the National Safe Haven. The results will be published in international peer-reviewed journals and will be shared with clinicians.