Nenad Macesic1,2,3, Sonika Tyagi1,4, Anton Peleg1,3,5
1Department of Infectious Diseases, The Alfred Hospital and School of Translational Medicine, Monash University, Melbourne, Australia, 2Infection Prevention & Healthcare Epidemiology, Alfred Health, Melbourne, Australia, 3Centre to Impact AMR, Monash University, Clayton, Australia
4School of Computing Technologies, RMIT University, Melbourne, Australia, 5Infection Program, Monash Biomedicine Discovery Institute, Department of Microbiology, Monash University, Clayton, Australia
Abstract:
Routine clinical care generates vast datasets stored in electronic health records (EHRs). These datasets are now being enriched with extensive outputs from ‘omics technologies, including genomics, proteomics, and transcriptomics. Recent breakthroughs in artificial intelligence (AI), particularly in machine learning and natural language processing, have allowed us to rapidly analyse these complex datasets, revealing critical insights that may significantly enhance infection prevention surveillance.
Through the MRFF-funded SuperbugAI project, we have integrated these technologies to improve patient safety and reduce healthcare costs. Our efforts have led to the development of AI models capable of using clinical, genomic, and proteomic data to predict antimicrobial resistance and sepsis outcomes. Despite these advancements, several challenges persist, such as managing diverse data formats, ensuring data quality and privacy, and achieving standardisation across institutions. These factors are crucial as they directly impact the performance of AI models, which depend on high-quality training data for accurate predictions. Addressing these challenges requires close collaboration between clinicians, data scientists, and bioinformaticians to fully realise the potential of these tools in everyday infection prevention practice.