Using data to identify complex patients

MacHSR Future Leaders Fellowship program (2023 cohort).

People presenting to hospital, especially general medicine, are increasingly older, frailer, and more complex. Unfortunately, there is no simple diagnostic test for complexity, so how do we identify those at risk of prolonged stay and hospital-acquired complications? We propose a unique big data solution: an algorithm combining natural language processing with demographic and clinical data to identify these patients on admission to hospital. Through transparent, consistent and reliable identification, we can prioritise these complex patients for specialised allied health management, aiming to maximise days at home, quality of life and enabling the person to live well in the community. These outcomes matter to patients and their families and have the potential to improve hospital flow.

Health services collect large amounts of data and I’m interested in finding ways to use this data to refine our service and improve patient outcomes.  I hope to refine the identification and selection of complex patients for future research within this population.  Identifying patients with the most complex needs and matching them with experienced allied health clinicians means we can expedite planning and successful discharge, benefiting patients and the health service.