Model Based System Development/Research Themes/Medical Decision Support

Medical decision support

Medicine and health care are the primary application fields of much of the model-based reasoning research described in the previous subsection. Part of the research is focused on clinical guidelines, and checking the quality of clinical guidelines using theorem proving and model checking techniques. Clinical guidelines are structured documents that are aimed at assisting medical practitioners and patients in making decisions about appropriate health care for specific clinical conditions. In particular in the context of the Protocure I and II projects, various formal methods were applied to formalized versions of real-world clinical guidelines in an attempt develop both framework, tools and techniques that supports checking the quality of the guidelines.

In addition to this focus on clinical guidelines, another part of the research is concerned with medical decision support systems using probabilistic graphical models, in particular Bayesian networks, as main technology. In the TimeBayes project, various techniques in Bayesian networks were explored, in particular ways to deal with temporal evolution and the specification of probability tables. The main domain in which the developed methods were explored was the diagnosis and antibiotic treatment selection of ventilator-associated pneumonia in the ICU.

ProBayes

B-Screen.