Flow chart of assessment.

BackgroundBlood pressure, grip strength and lung function are frequently assessed in longitudinal population studies, but the measurement devices used differ between studies and within studies over time. We aimed to compare measurements ascertained from different commonly used devices.MethodsWe used a randomised cross-over study. Participants were 118 men and women aged 45–74 years whose blood pressure, grip strength and lung function were assessed using two sphygmomanometers (Omron 705-CP and Omron HEM-907), four handheld dynamometers (Jamar Hydraulic, Jamar Plus+ Digital, Nottingham Electronic and Smedley) and two spirometers (Micro Medical Plus turbine and ndd Easy on-PC ultrasonic flow-sensor) with multiple measurements taken on each device. Mean differences between pairs of devices were estimated along with limits of agreement from Bland-Altman plots. Sensitivity analyses were carried out using alternative exclusion criteria and summary measures, and using multilevel models to estimate mean differences.ResultsThe mean difference between sphygmomanometers was 3.9mmHg for systolic blood pressure (95% Confidence Interval (CI):2.5,5.2) and 1.4mmHg for diastolic blood pressure (95% CI:0.3,2.4), with the Omron HEM-907 measuring higher. For maximum grip strength, the mean difference when either one of the electronic dynamometers was compared with either the hydraulic or spring-gauge device was 4-5kg, with the electronic devices measuring higher. The differences were small when comparing the two electronic devices (difference = 0.3kg, 95% CI:-0.9,1.4), and when comparing the hydraulic and spring-gauge devices (difference = 0.2kg, 95% CI:-0.8,1.3). In all cases limits of agreement were wide. The mean difference in FEV1 between spirometers was close to zero (95% CI:-0.03,0.03), limits of agreement were reasonably narrow, but a difference of 0.47l was observed for FVC (95% CI:0.53,0.42), with the ndd Easy on-PC measuring higher.ConclusionOur study highlights potentially important differences in measurement of key functions when different devices are used. These differences need to be considered when interpreting results from modelling intra-individual changes in function and when carrying out cross-study comparisons, and sensitivity analyses using correction factors may be helpful.【研究背景】在纵向人群研究中，血压、握力与肺功能常被作为核心评估指标，但不同研究间以及同一研究随时间推移所使用的测量设备存在差异。本研究旨在对比不同常用设备获取的测量结果。【研究方法】本研究采用随机交叉试验设计，纳入118名年龄介于45~74岁的男女受试者，分别使用2台血压计（sphygmomanometer）：欧姆龙705-CP（Omron 705-CP）与欧姆龙HEM-907（Omron HEM-907）、4台手持握力计（handheld dynamometer）：Jamar液压型（Jamar Hydraulic）、Jamar Plus+数字型（Jamar Plus+ Digital）、诺丁汉电子型（Nottingham Electronic）与斯梅德利型（Smedley），以及2台肺量计（spirometer）：Micro Medical Plus涡轮型（Micro Medical Plus turbine）与ndd Easy on-PC超声流量传感器型（ndd Easy on-PC ultrasonic flow-sensor）对受试者的血压、握力与肺功能进行评估，且每台设备均开展多次测量。本研究估算了设备配对间的平均差值，并结合布兰德-奥特曼图（Bland-Altman plot）计算一致性界限；同时通过替换排除标准、汇总指标以及多水平模型来估算平均差值，以此完成敏感性分析。【研究结果】两台血压计的收缩压平均差值为3.9mmHg（95%置信区间（CI）：2.5,5.2），舒张压平均差值为1.4mmHg（95% CI：0.3,2.4），欧姆龙HEM-907的测量值更高。在最大握力方面，任意电子握力计与液压型或弹簧式握力计相比，平均差值为4~5kg，电子设备的测量值更高；而两台电子握力计之间的差值较小（差值=0.3kg，95% CI：-0.9,1.4），液压型与弹簧式握力计之间的差值同样较小（差值=0.2kg，95% CI：-0.8,1.3）。所有设备配对的一致性界限均较宽。两台肺量计的第一秒用力呼气容积（FEV1）平均差值接近0（95% CI：-0.03,0.03），一致性界限相对较窄；但用力肺活量（FVC）的平均差值为0.47L（95% CI：0.53,0.42），ndd Easy on-PC的测量值更高。【研究结论】本研究表明，使用不同设备对关键生理功能进行测量时，可能存在具有实际临床意义的差异。在解读个体内功能变化建模的结果以及开展跨研究比较时，需充分考虑此类差异；采用校正因子进行敏感性分析或可为相关研究提供有效帮助。