New Faculty Seminar Series - William Van Woensel
Two Sides of the Same Coin: Improving Practitioner Decision Making and Empowering Patients through Decision Support and Process Mining
In this main theme, I will shortly discuss several research topics:
Why the Heck (Not)?! Explaining recommendations for improving practitioner decision making and patient self-management.
Health practitioners can be issued evidence-based recommendations by Clinical Decision Support (CDS), backed by Artificial Intelligence (AI) models. We explain recommendations in terms of visual, intuitive workflow diagrams that are guided by a patient profile at runtime. As a use case, we created a CDS system on lipid management for Chronic Kidney Disease (CKD), a challenging problem for many Primary Care Providers (PCPs). At the same time, AI models can issue recommendations to help patients self-manage their illnesses. Similarly, instead of merely positing an “edict,” the AI model can explain why the recommendation was issued: why one should stay indoors (e.g., risk of flare-ups), why further calorie intake should be avoided (e.g., met the daily limit), or why the care provider should be contacted (e.g., prescription change). We aim to study the impact of explanations on patient education and long-term behavior change for COPD and Rheumatology.
Reality Check! Process and Decision Mining in healthcare to guide real-world care as well as improve normative process models.
A process is defined as a series of progressive and interdependent steps by which an objective is attained. Many organization, including healthcare institutions, use information technology systems to support organizational (business) processes. However, the actual, likely imperfect, but possibly more robust, processes carried out in the real world are often different from the normative, likely idealized, and often imaginary, business processes existing in manager's minds. Process mining analyzes series of events, as recorded by information systems, to discover actual processes being carried out, compare them with normative processes, and generate insights to better direct real-world processes and/or improve normative processes.
About the Speaker
Dr. William Van Woensel received his PhD in Applied Computer Science at the Vrije Universiteit Brussel in Belgium. He was a Research Associate at Dalhousie University, and a Visiting Researcher at the Universidad Politécnica de Valencia, Spain. It is his ambition to develop novel methods for Health Informatics, Information Systems, and Artificial Intelligence, and translate them into practice to transform healthcare; themes of particular interest are Knowledge Representation and Reasoning, Mobile Computing and Machine Learning. His research program has produced innovative methods for Intelligent Clinical Decision Support, Medical Knowledge Bases and Knowledge Discovery, Patient Engagement, Self-Management and Mobility, and Activity Recognition for Assisted Living in Smart Environments.
Dr. Van Woensel is the co-chair of the W3C Community Group on Notation3, a declarative semantic language to implement decision making in open web environments. He is further involved with Punya, an intuitive, visual app development platform, and targets its use for the quick prototyping of mobile health apps, with direct involvement from clinician and patient stakeholders. He has been involved in the organization of the AIME and DeclarativeAI conferences, acting as application demo and scientific chair, and co-organized a number of workshops and tutorials. He has received 3 best paper awards, including the John Fox Memorial Award for the best paper in the field of Computer Interpretable Guidelines and Explainability.