Physical Analytics - Healthcare
Description: The example of the application of Physical Analytics to healthcare is in the area of compliance measurements and controls. Specifically for this project, we partnered with a major operator of hospitals and primary care facilities in the central region of Ohio, who has been constantly seeking to improve the quality of patient care using new and innovative approaches while at the same time improving operational efficiency of the hospitals. Among the factors for better care outcome, infectious disease control and patient comfort are especially important. Progress in patient care outcome can be gauged by the reduction in the average length of hospital stay, re-admittance rate, and improved patient satisfaction.
The project was based on our Physical Analytics Technology/MMT platform. In this work the technology has been developed further, deployed in a hospital environment, and tested with the goal of providing a system that allows obtaining near real-time key performance indicators (KPIs) related to compliance with clinical hygiene as well as patient comfort. Additional KPIs may include energy and operational efficiency metrics, which are being developed during the project.
In this project recent improvements in wireless mesh network technology have been been leveraged, like IBM's Low-power Mote Technology (LMT). This technology was for the first time being applied to a “real-life” hospital environment. One key component of the approach is that IBM Research has added RFID (Radio Frequency Identification) capabilities, which is, to the best of our knowledge, the first instance of an energy-efficient RFID-enabled wireless mesh network. The ability to use RFID reading in a mesh network setting allows for a reduction in deployment costs—especially for existing (legacy) buildings compared to installing fixed, wired RFID readers. This cost reduction makes the deployment of more sensors and RFID readers possible, which improves tracking accuracy and data quality.
Prior art for tracking hand-washing compliance often uses “star networks” to communicate the information from the readers and sensors inside of the hospital to a central gateway. In a star network each reader (or mote) requires generally a full power source and a high bandwidth communication infrastructure, which adds significantly to the cost of the solution. By comparison, the work here uses very energy efficient mesh networks, where each reader (or mote) receives and sends the data independently of the existing hospital communication infrastructure. With the exception of a central gateway unit, the whole system is operated by batteries. The approach can be exploited to bring additional sensing capabilities such as detecting volatile organic compounds (VOC), which are the main compounds of any disinfectant. Finally, the solution leverages sensor-based, novel statistical metrics for tracking hand-washing compliance.