Principal Investigator: Professor Blanca Gallego Luxan

Abstract: Suboptimal secondary prevention of cardiovascular disease and increasing uptake of mobile devices has led to the development of mobile health (mHealth) technologies that provide health advice, remote monitoring, and home-based cardiac rehabilitation to address known barriers to patient engagement with traditional secondary prevention. However, some of these applications have costs associated with equipment for self-recording of vital signs and remote monitoring by cardiac specialists and therefore need to be targeted at high-risk patients who are most likely to benefit. Underpinning suboptimal care is inaccurate estimation and poor management of patient risk. Traditional risk scores do not capture a multitude of factors including frailty, findings from atherosclerosis imaging and use of preventative therapies, and lack precision and discrimination. We will use electronic medical record data and artificial intelligence technologies to develop and implement an algorithm to identify cardiac patients who are at high risk of a further cardiovascular event before they are discharged from hospital. This will enable the targeting of more intensive interventions such as higher potency medications or home monitoring programs to those more likely to benefit. The algorithm and a prototype visualisation dashboard will be developed ready for implementation at South Eastern Sydney Local Health District (SESLHD).

Principal Investigator: Professor Blanca Gallego Luxan

Abstract: Timely identification of patients in hospital wards who are at increased risk of clinical deterioration is crucial to patient safety. Current early warning systems in Australia were developed before the large-scale adoption of electronic medical records and the maturation of artificial intelligence technologies, and rates of preventable adverse events remain unacceptably high. This project will examine the use and outcomes of current systems in the cardiovascular population and will build and test new algorithms to support the systematic and proactive identification, prevention, and management of clinical deterioration in hospital wards.

Principal Investigator: Victoria Blake (PhD Candidate)

Abstract: Clinical text contains key clinical information that is currently only accessible through manual chart review. Automated tools that can extract this information and present it in a structured format would have wide-ranging applications across clinical, population health and research domains including population-wide epidemiological research, pragmatic clinical trials, automated clinical quality registries, clinical dashboarding tools and clinical decision support. Heart failure is an ideal target for these automated tools given its lack of structured clinical metrics for disease status and progression, its symptom-based management and reliance on clinical notes for decision-making. This project aims to develop accurate and reliable tools that automatically extract key heart failure information from clinical documents by leveraging recent advances in natural language processing (NLP) techniques.

Chief Investigator: Professor Blanca Gallego Luxan

Amount: $544,979

Project Title: CardiacAI: Deep Learning to predict and prevent secondary cardiovascular eve

Chief Investigator: Professor Blanca Gallego Luxan

Amount: $146,384

Project Title: Cardiovascular Analytics and Innovation (CVAI) data repository: A big data resource for machine-learning enabled research and health systems quality improvement

DOI: https://doi.org/10.1093/ije/dyae040 

Blake V, Jorm L, Yu J, et al. Data Resource Profile: The Cardiac Analytics and Innovation (CardiacAI) Data Repository. International Journal of Epidemiology 2024;53:dyae040. doi: 10.1093/ije/dyae040

DOI: https://doi.org/10.2196/46322

Liu L, Perez-Concha O, Nguyen A, Bennett V, Blake V, Gallego B, Jorm L. Web-Based Application Based on Human-in-the-Loop Deep Learning for Deidentifying Free-Text Data in Electronic Medical Records: Development and Usability Study. Interactive Journal of Medical Research 2023;12:e46322

DOI: https://doi.org/10.1016/j.compbiomed.2024.108321 

Huberts LCE, Li S, Blake V, et al. Predictive analytics for cardiovascular patient readmission and mortality: An explainable approach. Computers in Biology and Medicine 2024;174:108321