Clinical Informatics and Data Science in Psychiatry: Working with Industry for Patient Benefit
Summary
The prospective candidate will work in partnership between Dr Osimo (data science, risk prediction modelling, clinical psychiatry, genomics) and Akrivia Health Ltd, a company in the business of curating psychiatric electronic health records (EHRs) for research use. They will be co-supervised by Prof Murray, an expert in the field.
The project will be powered by Akrivia's rich dataset, including the anonymised EHRs of 6m+ patients, harmonised across 20 secondary care psychiatric healthcare organisations in England and Wales.
The Akrivia dataset is enriched with a bespoke natural language processing (NLP) pipeline to extract research-relevant information from free-text clinical notes, allowing for deep patient profiling according to transdiagnostic features such as symptoms, treatment response, and social determinants of health.
Akrivia is enriching the EHR data they curate with growing repository of multi-omics data, and a route to access linked national primary care data through a partnership with Optimum Patient Care.
The -omics data collection and linkage is supported by the Wellcome Trust and a pharma partner, and has linked 1,800 participants’ EHRs to blood samples for whole genome sequencing to date, with recruitment of a further 50,000 participants funded across major depressive disorder, bipolar disorder and schizophrenia.
The primary care linkage so far includes four HCOs in the Akrivia network and ~600,000 matched, linkage patients, with the remaining HCOs in the process of setting up the linkage process.
Project aims
The project will span several potential areas:
- Clinical risk prediction modelling: Dr Osimo developed a clinical risk prediction tool called MOZART (https://europepmc.org/article/MED/37034013), aimed at calculating the risk of treatment-resistant schizophrenia at the time of a first episode of psychosis – one project could for example look at adding genetic predictors.
- Epidemiology: work on using large population-based datasets to investigate the origins of mental illnesses such as depression and psychosis (see for example https://mentalhealth.bmj.com/content/28/1/e301506).
- Electronic health records: work on mining EHR for clues on the links between inflammation, cardiometabolic changes, and mental illnesses such as depression and psychosis – including pharmacological outcomes such as treatment response and condition subtyping (e.g., https://www.sciencedirect.com/science/article/pii/S088915912031521X).
Contact details
Contact details: Emanuele Osimo - efo22@cam.ac.uk
Opportunities
This project is open to applicants who want to do a:
- PhD