Summary
Objective: Development of a clinical sensor network system that automatically collects vital
sign and its supplemental data, and evaluation the effect of automatic vital sensor
value assignment to patients based on locations of sensors.
Methods: The sensor network estimates the data-source, a target patient, from the position
of a vital sign sensor obtained from a newly developed proximity sensing system. The
proximity sensing system estimates the positions of the devices using a Bluetooth
inquiry process. Using Bluetooth access points and the positioning system newly developed
in this project, the sensor network collects vital sign and its 4W (who, where, what,
and when) supplemental data from any Blue-tooth ready vital sign sensors such as Continua-ready
devices. The prototype was evaluated in a pseudo clinical setting at Kyoto University
Hospital using a cyclic paired comparison and statistical analysis.
Results: The result of the cyclic paired analysis shows the subjects evaluated the proposed
system is more effective and safer than POCS as well as paper-based operation. It
halves the times for vital signs input and eliminates input errors. On the other hand,
the prototype failed in its position estimation for 12.6% of all attempts, and the
nurses overlooked half of the errors. A detailed investigation clears that an advanced
interface to show the system’s “confidence”, i.e. the probability of estimation error,
must be effective to reduce the oversights.
Conclusions: This paper proposed a clinical sensor network system that relieves nurses from vital
signs input tasks. The result clearly shows that the proposed system increases the
efficiency and safety of the nursing process both subjectively and objectively. It
is a step toward new generation of point of nursing care systems where sensors take
over the tasks of data input from the nurses.
Keywords
Ubiquitous computing - context awareness - point of care system - sensor network -
Continua
