Remote health monitoring is an emerging discipline with strong potential. It is also a critical application of modern-day networks of wireless sensor nodes--namely, wireless sensor networks (WSNs). For healthcare professionals, it is significant to determine the accurate status of the health of a remotely located patient or an aged person, so that when there is a need, appropriate treatment is vetted in a timely manner.
In this paper, Fine proposes an efficient clustering technique for making a decision on the health status of a remote subject. The proposed technique uses minimum spanning trees as part of a clustering algorithm to differentiate between normal and anomalous readings from the subject, such as: is person A in room B of his or her house? If yes, for how long? The resulting information, when compared locally with similar readings from the subject during the day, can help construe the health status of the person. The proposed technique is more accurate in the presence of a lesser amount of information about the subject, as opposed to other techniques such as statistical analysis.
The tone of the paper is general. The paper is accessible to anyone with a minimum knowledge of e-health. It will be of interest to readers with a particular interest in using technology for healthcare, as well as e-health researchers.