Towards Adaptive Decision Support for Individual Persons Suffering from Chronic Diseases

This paper presents the concept of an adaptive decision support system for individual persons (ADSIP). ADSIP will serve as a diagnosis assistant for a smart wearable computing system and aims to provide elderly people suffering from chronic diseases with reliable health status diagnosis in real time. The smart wearable computing system addresses unobtrusive and long-term home and outdoor health monitoring, for the purpose of supporting elderly people in leading an independent life with more mobility, flexibility, and safety. Chronic Obstructive Pulmonary Disease (COPD) is chosen as the initial application domain. To achieve reliable health status diagnosis, the proposed approach for ADSIP is to individualize the decision support system by means of learning an individual person´s health characteristics repeatedly at certain time intervals from measured sensor data and user input, so that a dynamic decision-theoretic model can be adapted gradually to this person´s individual health condition. It is expected that with ADSIP, measured deviations from the norm can be categorized either as a critical situation or as a normal uncritical variation, depending on an elderly person´s individual health characteristics. As a consequence, false positive health status diagnoses will be reduced and false negative diagnoses will be avoided.