Interview with CAP-AI Healthcare Data Scientist Sophie Williams
What is your role and what does it involve?
The COVID-19 pandemic has meant that the Barts Health has a wealth of patient data from a diverse local population, which can be used to analyse why some patients had worse symptoms and complications than others.
Since March this year, I’ve been managing and analysing this data, which is used to support COVID-19 clinical trials and research.
My main research up to this point was looking at the predictors of the vascular complications of diabetes in conjunction with the CAP-AI programme of artificial intelligence and machine learning projects led by Barts Life Sciences. I did this by modelling data collected by Barts Health through the Electronic Health Records (EHR) and using what is known as Natural Language Processing (NLP).
NLP effectively allows us make sense of things like free text from real time clinical information in patient notes and turn it into understandable data that can be used for research.
How long have you worked for the Trust?
Before joining the NHS, I worked as a senior data scientist for the civil service. I also have a PhD in experimental neuroscience from University College London.
What challenges and opportunities are there in your role?
I really enjoy the challenges of the role and have been heavily involved with working alongside colleagues from across the world on projects using tools such as SNOMED.
It’s important that any data is secure and is assessed in a broader societal and clinical content. The data needs to be reliable, consistent and of sufficient quality to use effectively.
The creation of the East London Patient record has helped us develop a better understanding of health outcomes in this part of London. This data is invaluable in helping us find solutions and in matching the experiences of our patients with those of people across the world.
I’m particularly excited by the development of a Barts Life Sciences campus in Whitechapel.
What is the SNOMED project?
SNOMED is a clinical vocabulary that aims to describe any clinical term. It does this through the use of a robust system of coding. This helps search and categorise information, which supports the analysis of large amounts of data.
The SNOMED tool is useful in helping us search and categorise information using different descriptions, which in turn helps us analyse data. This can be particularly useful in helping us predict patterns and features in relation to particular conditions such as diabetic foot disease. For example, a researcher might be interested in using a certain type of symptom and its outcomes.
This is supporting clinical trials, research and is helping us understand the population of COVID-19 patients at Barts Health and how outcomes are related to demographics, comorbidities, virus symptoms and treatments.
I was delighted to be invited to present my work on a recent clinical webinar for SNOMED International. I was pleased to be joined by Dr Charles Gutteridge, Chief Clinical Information Officer who gave a separate and fascinating presentation about the use of data in the context of the Covid-19 pandemic.
How will you use the Covid-19 patient data and is it secure?
We have received a large number of requests for our data, but it’s not possible for us to allow access to all of our data. So we’ve created a standard dataset. This can support clinical trials, research and help us understand the population of COVID-19 patients at Barts Health and how outcomes are related to demographics, comorbidities, virus symptoms and treatments.
The full datasets are only available for internal purposes and certain clinical trials in which the patient has consented. An anonymised version is available for researchers and external requests. We are also linking these with open datasets covering primary care and demographics.
For further information about the CAP-AI programme and how you can get involved email firstname.lastname@example.org, follow us on Twitter @ai_barts; visit www.capitalenterprise.org/capai/ www.bartslifesciences.org
CAP-AI is funded by the European Regional Development Fund and Barts Charity.
SME of the month: Clinithink
Clinithink is a technology and services company built around CLiX, the world’s first Healthcare AI capable of truly understanding unstructured medical notes. CLiX can recognise idioms, metaphors, turns of phrase, and all 72 different ways of saying ‘fractured left neck of femur’. CLiX can read, understand and extrapolate from the meandering mass of deductions, hunches, opinions and observations that doctors use to describe conditions in real life. Through CLiX, Clinithink can generate deeply human insights at superhuman speeds, to save time, money and even lives.
Clinithink’s life science customers are using the innovative CLiX AI technology alongside their expert services to facilitate deeper, faster and more accurate insights into a worldwide network of electronic medical data. These insights are saving time in the product development lifecycle, saving money in the clinical trial, research and commercial space, and saving lives by expanding access to life changing treatment.
Clinithink also specialises in using innovative AI solutions to tackle the complexities of challenges in healthcare sector. The Clinithink team can enable the CLiX insight platform to answer any number of healthcare business queries, providing businesses with the opportunity to manage patient pathways, reduce costs and improve financial & resource management.