2020 Bio: Sophie is a research data scientist at Barts Health NHS Trust. Her primary research project looks for predictors of vascular complications of diabetes by modelling Electronic Health Record (EHR) data processed with Natural Language Processing (NLP). Since March, she has been developing an EHR-derived dataset to support COVID-19 clinical trials and research at Barts Health.
Sophie’s research interests are in applying data science techniques for real world public benefit. Prior to joining the NHS, she worked as a senior data scientist within UK civil service. Sophie has a PhD in experimental neuroscience from University College London.
Read Sophie’s blog where she shares her experience of involving patients in data research through an online panel discussion and her key learnings about citizen engagement.
Watch a video about Sophie’s project here.
Digital Pioneer Fellowship project summary: Sophie and her team are showcasing how real-world data can be used to answer clinical questions, to increase the practise of clinical informatics across the Trust with our own data.
In response to the COVID-19 pandemic, she is leading the development of a master dataset from our EHR to support clinical trials, internal audit and research projects. They use text processing tools to extract clinical content from free text documents, which enhances the SNOMED-CT coded structured clinical data within thier EHR. They use descriptive logic relationships within the SNOMED ontology to manipulate complex datasets into a format suitable for modelling and answering clinical questions. The COVID dataset is used to identify eligible patients and return data to NIHR-prioritised COVID-19 trials, and by local research teams.
The larger vision is that by better predicting conditions a patient might develop or how a disease might evolve, they can treat it more effectively. Through earlier and improved medical intervention, they can prevent people from becoming unwell in the first place and by developing personalised, targeted medicines, it can dramatically improve patient outcomes. The longer term aim is for these curated datasets to engage and support clinical teams to undertake their own analytics and improve data entry.
Estimated number of patients / staff impacted by the project: As part of Barts Life Sciences (BLS), the project runs across five hospital sites in Barts Health which all use the enterprise EHR, covering a population of 2.5million patients per year.
Goal(s) for the programme: Sophie and her team want to increase the use of this real-world data from their diverse local population within the Trust both for patient care and to improve understanding of disease. They plan to do this by getting tools into the hands of care teams and by enhancing data quality within the EHR