Continuous non-invasive blood pressure estimation.

血压是五个生命体征之一，也是患者健康的主要指标。高血压（高血压）是心血管疾病（CVD）的最强风险因素之一，心血管疾病是全球死亡的主要原因。目前用于测量血压的方法使用血压计，即放置在患者手臂周围并充气直到血流被阻塞的充气袖带。这提供了患者的收缩压（心脏收缩时的最高血压）和舒张压（心脏放松时的最低血压）的单一阅读。然而，血压是一个全天变化的信号，因此单一阅读不能提供足够的信息来准确诊断CVD或其他血压相关疾病。此外，由于临床医生进行测量，患者的血压升高，存在“白色外套”效应，这表明许多患者在检查血压时会被误诊。目前在连续无创血压估计领域的研究工作主要集中在脉搏传导时间（PTT）上，其被定义为血液从一个动脉部位行进到另一个动脉部位所需的时间。PTT的变化与适当校准后的血压变化相关。为了记录两个不同位置处的血流开始，大多数使用PTT来获得血压估计值的设备通过使用放置在患者身上的两个或更多个设备来这样做。这些传感器的示例包括链接到同一系统的心电图（ECG）和光电体积描记（PPG）传感器，或者两个PPG传感器。由于需要两个传感器，因此该设备不太可能一直使用，因为患者会感到不舒服，并会妨碍他们的日常生活;因此，很难将设备称为真正的连续设备。目前正在为该项目考虑两种方法（然而，由于该项目还很年轻，很可能会有更多的方法有待探索）。第一种是使用放置在患者手腕上的单个PPG传感器来导出PTT样信号或使用机器学习技术来从PPG估计血压。第二种方法将研究基于相机的血压估计方法。已经表明，患者的视频摄像机流能够导出患者的病理信号，这项工作可以扩展到考虑患者身上许多位置的血液流动，因此PTT分析可以用于从这些位置之间的时间差导出血压估计。最近的研究表明，50岁以上住院患者的收缩压夜间升高，而不是在年轻人中发现的通常下降模式。研究还发现了性别差异，以及一周中不同日子的差异。这些结果表明，跟踪血压随时间的变化不仅可以帮助评估高血压等临床疾病的患者，还可以提供有关心血管健康的新信息。在夜间使用血压计重复测量血压变化是困难的，因为存在唤醒患者的风险，这将导致血压升高。因此，连续的非侵入性血压估计（或遵循相同趋势的血压的导出替代物）将允许容易地识别血压不下降的患者以进行进一步治疗。

Blood pressure is one of the five vital signs and a major indicator of a patient's health. High blood pressure (hypertension) is one of the strongest risk factors for Cardio Vascular Disease (CVD), which is the leading cause of death worldwide. Current methods for measuring blood pressure use a sphygmomanometer, an inflatable cuff placed around the patients arm and inflated until the blood flow is occluded. This provides a single reading of a patient's systolic (highest blood pressure when the heart contracts) and diastolic (lowest blood pressure when the heart relaxes) blood pressure.However, blood pressure is a signal that changes throughout the day and so a single reading cannot provide sufficient information for accurate diagnosis of CVD or other blood pressure related diseases. In addition, the presence of a "white coat" effect, in which a patient's blood pressure increases due to the clinician taking the measurement, has shown that many patients will have been misdiagnosed when having their blood pressure examined. This has proven the need for a continuous non-invasive blood pressure estimation.Current research efforts in the field of continuous non-invasive blood pressure estimation are mostly focussed on Pulse Transit Time (PTT) which is defined as the time it takes for blood to travel from one arterial site to another. Changes in PTT are shown to correlate with changes in blood pressure after appropriate calibration. In order to record the onset of the blood flow at the two different locations most devices that use PTT to derive an estimate of blood pressure do so by using two or more devices placed on a patient. Examples of these include an electrogardiogram (ECG) and a photoplethysmography (PPG) sensor linked to the same system, or two PPG sensors. Due to the need for two sensors it is very unlikely that the device will be used at all times as patients will find them uncomfortable and will get in the way of their daily life; and therefore, it is difficult to call the devices truly continuous.This DPhil project aims to approach a novel implementation of continuous blood pressure estimation. There are two methods that are currently being considered for this project (however, as the project is still young it is very likely that there will be more methods to be explored). The first is to use a single PPG sensor placed on the wrist of a patient to derive either a PTT-like signal or use machine learning techniques to estimate blood pressure from the PPG. The second method will study camera-based methods of deriving a blood pressure estimation. It has already been shown that a video camera stream of a patient is able to derive pathological signals of the patient, this work can be extended to consider the flow of blood in many locations on the patient and thus PTT analysis can be used to derive a blood pressure estimation from the time differences between these locations.An ideal application for this project is in continuous monitoring of blood pressure during the night time. Recent research has shown a nocturnal rise in systolic BP for in-hospital patients above the age of 50 as opposed to the usual dipping patterns found in younger individuals. The research has also identified significant gender-specific differences, as well as variations according to the day of the week. These results suggest that tracking BP changes over time could not only help the assessment of patients with clinical conditions such as hypertension, but also provide new information about cardiovascular health. It is difficult to use a sphygmomanometer to repeatedly measure blood pressure changes during the night for risk of waking the patient which would result in a rise in blood pressure. Therefore, continuous non-invasive blood pressure estimation (or a derived surrogate of blood pressure that follows the same trends) will allow for a patient whose blood pressure does not dip to be easily identified for further treatment.