5-minute Fellowships: Dr Arlene Casey – Unlocking hidden information in clinical free text

The complex conditions older adults experience are often poorly coded in healthcare records. This is because healthcare systems focus on distinct diseases and not syndromes such as mobility. Syndromes tend to be recorded in clinical notes which cannot be easily made available for research because they contain identifying information. As a DMT Senior Research Fellow, Arlene is developing tools to address the privacy risks of using free text data in research and using natural language processing (NLP) to identify later-life conditions such as fall risks from clinical notes. 

So far, Arlene has used workshops and surveys to understand what the public and other stakeholders think about the risks of making free text available for use in research. She then worked closely with the four Scottish NHS data providers to develop a privacy labelling schema and label a sample of 16,000 datasets to begin to see where exactly the risks lie. Now, Arlene is laying the foundation for guidelines for health free text provision in research as well as the national language processing coding needed to tackle privacy risks in the data.