Development of an intelligent blood pressure measurement device to reduce measurement variability

我们建议研究血压测量变异性的原因，从而开发和评估改进的测量技术，从而开发一种新型原型智能血压测量设备，我们将在临床验证和评估。高血压是冠状动脉疾病、充血性心力衰竭、肾脏疾病和中风的主要心血管危险因素之一，是英国30%的死亡人数和每年400万NHS病床日的一个因素。尽管血压测量的重要性及其广泛使用，但它是临床实践中表现最差的诊断测量之一。单次血压测量通常决定接受治疗（或不接受治疗），尽管测量结果之间存在很大差异。然而，无论是手动还是自动测量，同一个人连续的血压测量结果差异很大。我们自己的研究表明，在连续记录之间，手动测量的血压通常相差超过10毫米汞柱。测量误差会严重影响诊断。《美国医学协会杂志》（JAMA）的一篇主要评论估计，5毫米汞柱的错误将导致2100万美国人被拒绝治疗，或27人暴露于不必要的治疗，这取决于错误的方向。这项研究计划的核心是我们所做的观察，这些观察将能够探测到潜在的测量变异性。我们已经看到在血压测量时，血压变异性和臂袖中出现的压力脉冲变异性之间有很强的联系。这些观察结果导致了本应用的建议，即研究和开发一种智能血压测量设备的技术，该设备将使用从袖带获得的变异性信息。所开发的技术将拥有知识产权（目前正在寻求），并有可能将其纳入电子手动和自动血压设备中。测量稳定的临床血压将是重要成果，给予临床测量信心。可影响临床血压测量变异性的常见干扰包括心率改变、频繁异位搏动、心律失常、患者运动、呼吸障碍、咳嗽、说话和肌肉紧张。我们自己的研究表明，这些干扰与袖带压力的振荡脉冲的变化有关。有了稳定的数据，我们看到袖带压力平稳下降，同时在袖带压力曲线的主要下降上叠加的小振荡脉冲的振幅平稳变化。（这些脉冲是由自动血压设备分析的，在袖带放气时可以看到水银柱的脉动。）当血压变化时，我们看到这些平滑的脉搏特征的偏差。我们在研究干扰对血压测量变异性的影响方面处于独特的地位。我们有一个广泛的数据库，包括1300多个预先记录的袖带压力和振荡脉冲压力波形，在各种不同的血压组中进行临床记录，以及由我们研究小组的两名训练有素的成员同时独立测量的听诊压力。目前还没有可公开访问的振荡波形数据库。我们的数据库是作为欧盟资助的多中心国际研究联盟的一部分获得的，该联盟开发了一个模拟器，通过使真实的、先前记录的振荡波形能够再生，来评估非侵入性血压（NIBP）设备的准确性。数据和模拟器在该项目中起着重要的作用。

We propose to research the causes of blood pressure measurement variability, and hence develop and evaluate improved measurement techniques, leading to the development of a prototype novel intelligent blood pressure measurement device that we will validate and assess clinically. High blood pressure is one of the leading cardiovascular risk factors for coronary artery disease, congestive heart failure, renal disease and stroke, and is a contributory factor in 30% of all deaths in the UK, and 4 million NHS bed days annually. Despite the importance of blood pressure measurement and its very widespread use, it is one of the most poorly performed diagnostic measurements in clinical practice. A single blood pressure measurement often determines the treatment (or non-treatment) received, in spite of high variability between measurements. However, consecutive blood pressure measurements in the same individual vary significantly, whether the measurements are taken manually or automatically. Our own research has shown that manual blood pressure measurements often vary by more than 10 mmHg between consecutive recordings. Measurement errors can seriously compromise diagnosis. A major review in the Journal of the American Medical Association (JAMA) estimated that a 5 mmHg error would result in 21 million Americans being denied treatment or 27 being exposed to unnecessary treatment, depending on the direction of the error. At the heart of this research proposal are observations we have made which will be able to detect potential measurement variability. We have seen a strong association between blood pressure variability and variability of the pressure pulses present in the arm cuff during blood pressure measurement. These observations have led to the proposal in this application, to research and develop techniques for an intelligent blood pressure measurement device that will use information about variability obtained from the cuff. The techniques developed will have intellectual property (which is currently being pursued) and potential for incorporation in electronic manual and automatic blood pressure devices. Measurement of stable clinical blood pressure will be the important achievement, giving clinical confidence in the measurement. Common disturbances that can influence clinical blood pressure measurement variability include heart rate changes, frequent ectopic beats, arrhythmias, patient movement, respiratory disturbances, coughing, talking and muscle tension. Our own research has shown that these disturbances are associated with changes in the oscillometric pulses in the cuff pressure. With stable data we see a smooth decrease in cuff pressure, along with a smooth variation in the amplitude of the small oscillometric pulses superimposed on the main descent of the cuff pressure curve. (These are the pulses analysed by automated blood pressure devices, and visible as pulsations of the mercury column during cuff deflation for manual measurements.) When blood pressure is varying we see deviations from these smooth pulse characteristics. We are in a unique position to investigate the influence of disturbances on blood pressure measurement variability. We have an extensive database comprising more than 1300 pre-recorded cuff pressure and oscillometric pulse pressure waveforms, recorded clinically in a variety of different subject groups with a wide range of blood pressures, together with auscultatory pressures measured simultaneously and independently by two trained members of our research group. Currently there are no publicly accessible databases of oscillometric waveforms. Our database was obtained as part of a European Union funded multi-centre international research consortium to develop a simulator for evaluating the accuracy of non-invasive blood pressure (NIBP) devices by enabling real, previously-recorded oscillometric waveforms to be regenerated. The data and simulator have important roles in the proposed project.