A Novel Visualization System for ICU Clinical Activity Tracking

Abstract

Patient and nurse interaction in the Intensive Care Unit (ICU) is important as it influences the patient outcomes. Optimizing the nurse-to-patient ratio can reduce the mortality of patient, prevent nurse burnout, and reduce costs in the ICU. However, there is a lack of methods to quantify and evaluate these bedside interactions. This paper presents a Clinical Activity Tracking System (CATS), which was designed to track and evaluate nursing motion at the patient bedside, aiming to quantify how nurses spent their working time. CATS utilizes the Microsoft Kinect, a motion sensing device containing an embedded camera and infrared sensor. For CATS, the Kinect is fixed on the ceiling, facing downwards to track clinical activity at the patient bedside. The system was set up in an experimental environment to simulate the ICU bedside activity, and different motion paths and test candidates were tested over 5 iterations to evaluate the performance of the system. The total tracking area for the CATS can reach 2.3 m × 1.6 m, which mimics to the ICU bedside area. The system can track candidates with different heights from 1.52 m to 1.90 m. The system can also track different motion patterns consistently, with median percentage tracking error 2.30% (Inter-quartile range (IQR): [0.72%, 4.25%]). The system can also track multiple candidates with median percentage error 1.75% (Inter-quartile range (IQR): 0.97%, 4.57%). The results show that the system can be used in real-time applications to track bedside clinical activity. This system is capable of evaluating the ICU nursing activity, with the ultimate aim to generate appropriate nurse-to-patient ratio to prevent nurse burnout and increase patient care. Also, it is able to track different candidate heights, adapt to different motion paths, different dwell time, and identify multiple people simultaneously. The results revealed that the system can be used to quantify and evaluate bedside clinical activity.