Sensor-Based Decision Support for the Allocation of Patient Attendants in Hospitals.

In hospitals, patient attendants are often necessary in order to closely monitor patients with high risk of self-endangering actions and reactions. However, such additional monitoring of patients is associated with high costs. In this paper, we describe a technical infrastructure for monitoring the patient's activities, which helps to assess whether an attendant should be requested. It was central to for us to use non-invasive sensors and to exploit a variety of patient data such as heart rate, micro-activity and oxygen saturation.