Overview
Professor Murray will introduce two broad areas of digital psychiatry - the use of pseudonymised electronic health records as tools for improving practice, and potential of mobile phones for gathering clinical information - and will describe the session objectives (see below).
Ms Seyedsalehi will introduce the audience to technical terms such as discrimination, calibration and net benefit. She will talk in general about risk prediction in psychiatry in comparison to other branches of medicine, and describe her own work developing a model in 2000 patients (with external validation on 1500) to predict, at the time of the first presentation, which patients are at highest risk of additional future psychotic episodes.
Dr Osimo will talk in general about electronic health record research and give examples from his own work: developing risk prediction tools MOZART and PsyMetRiC, for prediction, at psychosis onset, of later adverse psychiatric outcomes (treatment resistance) and adverse cardio-metabolic outcomes using routinely collected predictors, including demographics and biomarkers. He will describe work towards implementing these tools into clinical practice, including further model updating and external validation in multiple countries. He will conclude with an outlook on attempts to combine routine electronic health record data with ‘omics biomarkers for personalised treatment.
Objectives
- To gain an overview into two of the main tools of digital psychiatry - mining information in electronic health records for quality improvement and the use of mobile phone based patient report outcome measures (PROMS).
- To gain insight into how to assess a predictive model and know whether it is 'good enough' for clinical practice. To understand the relevance of terms such as discrimination, calibration and net benefit.
- To understand current strengths and weakness of electronic health records based predictive models that attempt, at the time of a first presentation of psychosis, to predict future relapse, treatment resistance and adverse cardiometabolic outcomes.
- To become familiar with the latest national large-scale initiatives in digital psychiatry in early psychosis - the EPICare digital registry and clinical decision making support tool and the Mental Health Mission early psychosis biomarker and digital phenotyping project.
Speakers
Chair: Dr Christelle Langley, University of Cambridge, Cambridge
Ms Aida Seyedsalehi, University of Oxford, Oxford
Dr Emanuele Osimo, University of Cambridge, Cambridgeshire and Peterborough NHS Trust, Cambridge
Professor Graham Murray, University of Cambridge, Cambridgeshire and Peterborough NHS Trust, Cambridge
