Quality Indices in Medical Alert Systems.

Numerous alert systems exist in healthcare domains but most of them produce too much false alerts, bad usage or disinterest. The need of better alert systems leads to develop context-aware alert systems. The alert system Tempas is a help-decision tool based on personalized alerts. It is adaptable to business environment, population targeted, expert user needs, and customized in real-time for immediate needs by end users. The adaptability is defined during the alert creation process. The customization is defined during the alert management process. It is based on the population targeted, activation conditions, and the alert behavior. It is supported by two quality indices: the applicability index expresses how much a patient (or a population) is concerned by the alert and the confidence index expresses how much the user can trust the alert. Both indices are used during the alert creation process (minimal thresholds for the population) and during the management process (minimal personalized threshold). The paper presents a summarized view of Tempas and focuses on the quality indices.