摘要
目的在健康人中测试基于光电体积描记法和深度学习算法的无袖带连续血压测量设备的准确性。方法使用经过认证的动态血压设备对健康成年受试者进行24h动态血压监测,同时使用基于光电体积描记法和深度学习算法的无袖带连续血压测量设备进行连续血压监测。使用均数检验、相关分析以及Bland-Altman图评估设备之间的一致性,包括相同时间点(点对点)以及夜间时间段(段对段)的均值比较。结果共获取了36名受试者的有效血压数据。点对点分析中,连续测压组与动态血压测量组收缩压和舒张压的均值分别相差(1.22±8.30)mmHg和(1.61±9.27)mmHg,配对t检验表明两组均值差异均无统计学意义(P>0.05)。Pearson相关分析显示,两组相同时间点上的收缩压(r=0.670,P<0.001)和舒张压(r=0.503,P<0.001)测量值均显著相关。段对段分析中,夜间时间段两种测量方法收缩压的均值(117.15±14.30)mmHg与(114.73±13.35)mmHg(P>0.05)和舒张压的均值(73.50±11.70)mmHg与(69.96±9.64)mmHg(P>0.05)均差异无统计学意义。结论基于光电体积描记法和深度学习算法的指环测量设备提供的连续血压测量结果与动态血压测量结果相当,有助于在未来用于日常血压监测以及血压变异性分析,为高血压的预防和管理提供新的工具。
【Objective】To test the accuracy of a cuffless continuous blood pressure measurement(CBPM)device based on photoplethysmography(PPG)and deep learning algorithms in healthy people.【Methods】24-hour ambulatory blood pressure monitoring(ABPM)was performed on healthy adult subjects with certified equipment,and CBPM based on PPG and deep learning algorithm was used at the same time.Paired t test,correlation analysis,and Bland-Altman plots were used to assess the consistency between devices,including mean comparisons at the same time point(point-to-point)and night time periods(period-to-period).【Results】A total of 36 subjects with valid blood pressure data were obtained.In the point-to-point analysis,the mean values of systolic blood pressure(SBP)and diastolic blood pressure(DBP)in the CBPM group and the ABPM group were 1.22±8.30 mmHg and 1.61±9.27 mmHg,respectively,and the paired t test showed that there was no significant statistical difference between the two groups(P>0.05).Pearson correlation analysis showed that the measured values of SBP(r=0.670,P<0.001)and DBP(r=0.503,P<0.001)were significantly correlated between the two groups at the same time point.In the period-to-period analysis,the mean values of SBP(117.15±14.30 mmHg and 114.73±13.35 mmHg,P>0.05)and DBP(73.50±11.70 mmHg and 69.96±9.64 mmHg,P>0.05)were not statistically different between the two groups during the night time period.【Conclusion】The CBPM device based on PPG and deep learning algorithm provides continuous blood pressure measurements comparable to ABPM,and is useful for daily blood pressure monitoring and blood pressure variability in the future analysis to provide new tools for the prevention and management of hypertension.
作者
张睿珊
宋珍
邱娴
韵育英
唐亿琳
谭诗健
谭浩
朱冰坡
ZHANG Ruishan;SONG Zhen;QIU Xian;YUN Yuying;TANG Yiin;TAN Shijian;TAN Hao;ZHU Bingpo(Southern University of Science and Technology Hospital,Shenzhen,Guangdong 518055,China;Household Medical Products Department,Shanghai Lepu CloudMed Co.,Ltd.,Shanghai 200000,China)
出处
《中国医学工程》
2024年第2期1-4,共4页
China Medical Engineering
关键词
光电体积描记法
深度学习
无袖带
连续血压测量
photoplethysmography
deep learning
cuffless
continuous blood pressure measurement
