Navigating the Intersection of Health Data Access and Research Excellence to improve prediction in later life syndromes
Arlene Casey | Dunhill Medical Trust Research Fellow| Principal NLP Data Scientist DataLoch
Hello everyone! I am Arlene Casey I have been working with ACRC since 2021 as a NLP researcher and in the last year within DataLoch (NHS Lothian & South East Safe Haven) leading a collaboration of academic partners and Safe Havens in building a pathway for health free-text data access. I am thrilled to announce I have been awarded a Dunhill Medical Trust Fellowship; this will see me enrolled on to the Edinburgh Scientific Academic Track (ESAT) program in a 5-year fellowship program. I will continue to be partnered with ACRC in addition to my role within DataLoch as the Principal NLP Data Scientist
So, what is my fellowship about?
Unravelling the Power of Free-Text: Natural Language Processing
As we know, healthcare systems are predominantly organized to address individual diseases but older people often present with a myriad of conditions, compounded by factors like poor mobility. The conventional reliance on 'coded' data, standardized for specific conditions, falls short when dealing with the intricate web of problems documented in free-text notes by dedicated healthcare professionals.
Recognising the richness of information within free-text notes we know we can harness the potential of Natural Language Processing (NLP) to extract meaningful insights. However, a significant challenge stands in the way — personal information, embedded within these notes, poses a risk to patient confidentiality. The first element of my Fellowship embarks on a journey to develop methods, in collaboration with patients, carers, and NHS staff, to reliably remove personal information. This builds on my existing work leading a collaboration across the Scottish Safe haven network (Untapped potential - privacy risk in clinical free-text) and is a pivotal move forward to ensure that the invaluable data hidden within free-text notes can be shared without compromising patient privacy but also to address this at a national level.
My role as a research fellow and working within DataLoch allows me to combine my many years of experience in industry with academic scholarship. I will not only develop new methods to address privacy risk but also to work with Information Governance teams, NHS staff and infrastructure specialists to build an implementable pathway for release of large-scale unstructured data.
Predicting Outcomes in Older Populations
In the second element of my fellowship, the focus will shift to the development of new prediction tools tailored for older populations. Traditional prediction tools often fall short when applied to the unique health challenges faced by older individuals. The goal is to enhance the accuracy and applicability of these tools, through combining the new and rich sources of free-text and structured data. These tools play a crucial role in identifying individuals who may benefit from proactive care, potentially preventing or delaying future health crises.
A Collaborative Approach: Engaging the Community
At the heart of both parts of my fellowship is a commitment to collaboration. I look forward to further work engaging with clinicians, the public, patients, and carers, ensuring that privacy concerns are meticulously addressed. This collaborative approach will not only safeguard patient confidentiality but also ensures that the developed prediction tools are centered around the real, pressing concerns of the individuals they aim to assist. This will build on the existing work I have been doing (DOI 10.5281/zenodo.10055361)
Unlocking the Potential of Health Free-Text Notes
The future is exciting, we are building a bridge to close the gap between data and understanding, paving the way for new healthcare research. By facilitating access to large sets of health free-text notes, my research has the potential to open doors to a wealth of information that can fuel countless future projects and make a meaningful difference to healthcare for ageing populations.
Sounds interesting or do you want to know more, please get in touch. Arlene.Casey@ed.ac.uk