Computer Processing of Ambulatory Blood Pressure Monitoring as Multivariate Data

摘要

The blood pressure (systolic and diastolic) and heart rate of a patient were tested hourly during about 24-hours. Such a test is well-known in many clinics as ambulatory blood pressure monitoring. It is a standard for hypertension diagnostics. Usually, these three data sets are considering separately as different time series. Each series delivers its own information. We propose another approach, considering each measuring as a three-dimensional vector of the patient state at the given moment. Indeed, this vector is an image, that is a moment describing of patient's state. Thereby, the entire test might show us the dynamic of the state in the phase space during the observation period. The idea is borrowed from the pattern recognition and perhaps will sound well for the processing and visualization of monitoring results. According to the presented findings, the model of information system based on Micro-Electro-Mechanical System (MEMS) sensor for data collection and processing. The mathematical apparatus for the system has been adopted to the requirements for devices with low computational power. Thus, it can be implemented using modern microcontrollers.