Complex challenges in healthcare

Abstract

Artificial intelligence (AI) could change medical care as we know it. For instance, AI can help clinicians analyse difficult problems by processing large amounts of information. This thesis explores how statistical models and AI can be used to improve the management of infectious diseases and the care for people who are most at risk of severe illness. <br/><br/>Firstly, I applied statistics and AI to deepen understanding of the COVID-19 pandemic and support decision-making. This included creating a novel AI-based method to improve the detection of past COVID-19 infections and estimating the costs of COVID-19 hospital stays. Second, this thesis explores how AI can improve care for people who live with multiple chronic diseases, by strengthening collaboration between medical specialties and identifying individuals who would benefit most from integrated care. Finally, I developed models that notify intensive care unit (ICU) clinicians up to eight hours in advance when patients may need an infection-related consultation and I enabled the early estimation of the clinical and economic impact of ICU-focused AI models. <br/><br/>In conclusion, by using information from different sources and combining several scientific fields, I demonstrated how statistics and AI can help address complex challenges in healthcare. Despite this promise, few models are implemented in practice due to various barriers. Overcoming these barriers will be necessary for AI to transform medical care. In that future, AI should complement rather than replace clinicians, making human-AI collaboration a promising path towards better healthcare.

Ähnliche Arbeiten