Effects of pulse transit time and physiological differences on wearable device based blood pressure estimation

Abstract

Hypertension related diseases have highly threatened people's health around the globe. What is even worse is that not only the number of hypertension patients has grown dramatically, their ages of onset have also become younger. Blood pressure monitoring is important for hypertension patients. However the traditional blood pressure measurement is inconvenient which results in less people with habits to measure blood pressure. To improve people's incentive of measuring blood pressure, in this study we have developed a wristband prototype for measuring blood pressure. With our wristband prototype, it is easy for users to establish habits of blood pressure measurement and tracking, and subsequently detect hypertension in the early stage. Our wristband prototypes of wearable sphygmomanometer have been utilized to measure 120 subjects' blood pressures. This wristband prototype can simultaneously measure signals of Photoplethysmogram (PPG) and Electrocardiography (ECG), and then compute the interval of the peak time of these two signals as pulse transit time (PTT). Since PTT has been verified to be highly correlated to blood pressure in numerous researches, PTT is suitable for blood pressure estimation. However, since blood vessels in human body are very complicated and they are highly correlated with blood pressure, we have taken some parameters such as the heart rate (HR) and the subject's physiological differences (e.g., height, weight, age and gender) into account and evaluated the parameters together using multivariate regression analysis. We can subsequently derive testing results from this non-invasive blood pressure measurement method. The results show that in the normal blood pressure range, the error in systolic blood pressure (SBP) is 6.9mmHg ± 8.6mmHg. Although our results have not yet reached the AAMI standard, it has been verified that it is effective to integrate PTT and physiological parameters to estimate blood pressure.