Artificial intelligence (AI) techniques are employed when there is uncertainty, complexity, high amount of data, etc. which makes the decision-making task challenging. In these situations, AI techniques can provide results and insight, without direct human input; by augmenting techniques by, for example, using heuristics, we can add support for interpretability of predictions.
At the University of Skövde, they have been working in this area since the early 90s. In this presentation Ainhoa Goienetxea and Jonas Mellin from the University of Skövde will introduce a sample of the work they have done related to the use of AI in healthcare:
- Early diagnostics of sepsis: This work is based on three different research fronts: (1) the application of machine learning methods on existing Electronic Health Record data; (2) the development of multimarker panels – selection of biological markers and clinical data, aiding earlier diagnosis since these are monitored directly from the patient’s blood; and (3) a general development of a Clinical Decision Support System – to provide early warnings to the healthcare professionals.
- Early risk detection of age-related health problems: In the EU project HELICOPTER (http://www.helicopter-aal.eu), an electronic triage system based on adaptive and individualized anomaly detection and probabilistic expert systems was designed. This triage system can make early predictions of the risk of diagnostics such as heart failure, hyperglycemia, hypoglycemia and depression. The employed sensors were unobtrusive and low-cost. This system was tested with 70 participants in the Netherlands and Sweden.
- Decision support for the design and improvement of healthcare systems: To be able to effectively design and improve the flow of patients, facility layouts, resource allocation, etc. simulation-based multi-objective optimization and data mining can be employed to provide optimal system configurations to decision makers. The University of Skövde has worked with this novel combination of techniques to design an emergency department and operation theatres of different hospitals of the region, providing with optimal system configurations and knowledge for evidence-based decision-making.
The presentation will primarily target healthcare practitioners, system developers, staff working with continuous improvement (verksamhetsutvecklare), researchers, and decision makers.
Register at the latest on 31 January. A light lunch will be served to registered participants.