Abstract
Internet of Medical Things (IoMT) enable early detection and alerting of critical disease conditions through continuous monitoring of electro-physiological body parameters. However, large scale use, especially in rural and sparsely connected regions, poses challenges in terms of unavailability of sufficient data bandwidths and connectivity. Additionally, the physicians, who are already managing huge patient load, would be overwhelmed to take clinical decisions after viewing voluminous data pouring in from multiple IoMT devices. In our research, we developed a technique called Rapid Active Summarization for effective PROgnosis (RASPRO) that converts unwieldy multi-sensor time series data into summarized patient/disease specific trends in steps of progressive precision as demanded by the physician for patients personalized condition at hand and help in identifying and subsequently predictively alerting the onset of critical conditions. RASPRO generates clinically useful, yet extremely succinct, summary of a patient's medical condition represented as a series of symbols called motifs, which could be sent to remote physicians even over SMS or emerging narrow bandwidth Internet of Things (IoT) networks. We demonstrate that the diagnostic predictive power of RASPRO motifs is comparable to raw sensor data.