Explainable decision support through the learning and visualization of preferences from a formal ontology of antibiotic treatments.

Abstract

The aim of eXplainable Artificial Intelligence (XAI) is to design intelligent systems that can explain their predictions or recommendations to humans. Such systems are particularly desirable for therapeutic decision support, because physicians need to understand rcommendations to have confidence in their application and to adapt them if required, e.g. in case of patient contraindication. We propose here an explainable and visual approach for decision support in antibiotic treatment, based on an ontology. There were three steps to our method. We first generated a tabular dataset from the ontology, containing features defined on various domains and n-ary features. A preference model was then learned from patient profiles, antibiotic features and expert recommendations found in clinical practice guidelines. This model made the implicit rationale of the expert explicit, including the way in which missing data was treated. We then visualized the preference model and its application to all antibiotics available on the market for a given clinical situation, using rainbow boxes, a recently developed technique for set visualization. The resulting preference model had an error rate of 3.5% on the learning data, and 5.2% on test data (10-fold validation). These findings suggest that our system can help physicians to prescribe antibiotics correctly, even for clinical situations not present in the guidelines (e.g. due to allergies or contraindications for the recommended treatment).