This research initiative aims at strengthening the responsible development and application of AI in healthcare. Designated a UvA Research Priority Area, it brings together UvA researchers from the fields of Science, Medicine, Law and the Humanities.
Over the next 15 years, health expenditures are expected to outpace GDP growth in almost every OECD country. With AI, we can increase efficiency in healthcare and improve quality and safety. Although still in its infancy, the use of AI in healthcare has enormous potential: it can be used to automate routine and simple tasks and to support decision-making for complex problems.
This Research Priority Area promotes interdisciplinary research on AI-driven decision-making in healthcare, combining expertise from fields such as computer science, medicine, law and ethics. The aim is to develop ethical, high-quality AI solutions that help patients.
Research has begun on two applications for AI-driven decision-making: 1) developing coagulation and transfusion strategies and 2) identifying the hospitalisation risk of cardiac patients.
When hospital patients are undergoing major surgery and have a high risk of bleeding, doctors need a treatment strategy for blood coagulation and blood transfusion. Currently, physicians make these decisions by interpreting a large number of results from various laboratory tests and clinical observations. The task is complex, performed under high time pressure, and the decision outcome is physician-dependent.
For this application, we will design an AI-derived predictive model to determine the optimal amount and kind of coagulation agent to guide the anaesthesiologist and intensivist during and after major surgery. The model will be validated by comparing blood loss levels, blood product administration, mortality, and hospital costs for patients undergoing major surgery when using this model compared to standard care.
Cardiac patients who aren’t hospitalised can be monitored at an outpatient clinic and at home. If their condition deteriorates, they must be hospitalised. Unfortunately, the path towards deterioration is subtle. Detecting it requires interpreting information from several sources, including medical images, ECG signals, and diverse clinical parameters.
For this application, we will develop an AI-based risk stratification method. We will investigate whether this enables early intervention in the high-risk group, which could prevent hospitalisation, reduce costs and improve quality of life.
When health decision-making is supported by machines and data, it affects the autonomy of health professionals and their patients. The legal relationship between patients and medical professionals is built on this autonomy. We will investigate how this legal-ethical relationship is affected by the introduction of AI-supported decision-making in healthcare. Ethical and legal research will be integrated into this Research Priority Area at every stage.