Borges Gabriel, Brusamarello Valner
Electrical Engineering Department, Universidade Federal do Rio Grande do Sul, Porto Alegre, 90035-190, Brazil.
J Clin Monit Comput. 2016 Dec;30(6):859-867. doi: 10.1007/s10877-015-9786-4. Epub 2015 Oct 6.
Automatic patient monitoring is an essential resource in hospitals for good health care management. While alarms caused by abnormal physiological conditions are important for the delivery of fast treatment, they can be also a source of unnecessary noise because of false alarms caused by electromagnetic interference or motion artifacts. One significant source of false alarms is related to heart rate, which is triggered when the heart rhythm of the patient is too fast or too slow. In this work, the fusion of different physiological sensors is explored in order to create a robust heart rate estimation. A set of algorithms using heart rate variability index, Bayesian inference, neural networks, fuzzy logic and majority voting is proposed to fuse the information from the electrocardiogram, arterial blood pressure and photoplethysmogram. Three kinds of information are extracted from each source, namely, heart rate variability, the heart rate difference between sensors and the spectral analysis of low and high noise of each sensor. This information is used as input to the algorithms. Twenty recordings selected from the MIMIC database were used to validate the system. The results showed that neural networks fusion had the best false alarm reduction of 92.5 %, while the Bayesian technique had a reduction of 84.3 %, fuzzy logic 80.6 %, majority voter 72.5 % and the heart rate variability index 67.5 %. Therefore, the proposed algorithms showed good performance and could be useful in bedside monitors.
自动患者监测是医院进行良好医疗管理的一项重要资源。虽然由异常生理状况引发的警报对于快速治疗的实施很重要，但由于电磁干扰或运动伪影导致的误报，它们也可能成为不必要噪音的来源。误报的一个重要来源与心率有关，当患者的心律过快或过慢时就会触发。在这项工作中，探索了不同生理传感器的融合，以创建可靠的心率估计。提出了一组使用心率变异性指数、贝叶斯推理、神经网络、模糊逻辑和多数投票的算法，以融合来自心电图、动脉血压和光电容积脉搏波描记图的信息。从每个数据源提取三种信息，即心率变异性、传感器之间的心率差异以及每个传感器的低噪声和高噪声频谱分析。这些信息用作算法的输入。从MIMIC数据库中选取的20份记录用于验证该系统。结果表明，神经网络融合的误报减少效果最佳，为92.5%，而贝叶斯技术的减少率为84.3%，模糊逻辑为80.6%，多数投票者为72.5%，心率变异性指数为67.5%。因此，所提出的算法表现出良好的性能，可用于床边监护仪。
