DERI Seminar with Dr Sophie Williams, Lead Data Scientist at Barts Life Sciences
When: Thursday, April 4, 2024, 11:00 AM - 12:00 PM
Where: Hybrid
Speaker: Sophie Williams who is Lead Data Scientist at Barts Life Sciences
Sophie will be presenting in our basement presentation space, so do join her if you are in the building; register here on Eventbrite if you are coming in person.
Zoom link: https://qmul-ac-uk.zoom.us/j/81148100921
Title: Use of real-world data to identify patients in the hospital with diabetic foot in real-time, and flag high-risk patients.
Abstract: Detailed clinical and social information about patients attending hospitals is found in free-text clinical notes. We compared commercial and open-source Natural Language Processing (NLP) tools to identify patients with diabetic foot and diabetes from this real-world data. A query is now being trialled -through a dashboard - by the diabetic foot Multidisciplinary Team Meeting (MDT) using near-real time patient data.
Deep phenotyping of 56000 patients with diabetes was conducted to extract symptoms, complications, and social factors from free-text clinical notes. Methods were developed to analyse these large SNOMED-CT coded datasets, which leverage the structure of SNOMED-CT knowledge graphs for feature reduction and clustering of similar conditions in a data driven way. The aim of this study is to predict which patients with diabetes will develop diabetic foot disease, and to understand risk factors in this population. This analysis will provide preventative advice and guide treatment pathways to be more personalised to individuals.
Bio: Sophie Williams PhD is Lead Data Scientist at Barts Life Sciences and a DERI fellow. Her team develop generalisable data tools and models to support clinical and operational needs in Barts Health NHS Trust, and she has 5 year’s experience in deploying models within the NHS.