Instituto Superior Técnico, University of Lisbon, 1049-001 Lisbon, Portugal.
Pulmonology Department, Santa Maria University Hospital (CHULN), Santa Maria Local Health Unit, Av. Prof. Egas Moniz, 1649-028 Lisbon, Portugal.
Sensors (Basel). 2024 Aug 10;24(16):5164. doi: 10.3390/s24165164.
The aim of this work was to validate the measurements of three physiological parameters, namely, body temperature, heart rate, and peripheral oxygen saturation, captured with an out-of-the-lab device using measurements taken with clinically proven devices. The out-of-the-lab specialized device was integrated into a customized mHealth application, e-CoVig, developed within the AIM Health project. To perform the analysis, single consecutive measurements of the three vital parameters obtained with e-CoVig and with the standard devices from patients in an intensive care unit were collected, preprocessed, and then analyzed through classical agreement analysis, where we used Lin's concordance coefficient to assess the agreement correlation and Bland-Altman plots with exact confidence intervals for the limits of agreement to analyze the paired data readings. The existence of possible systematic errors was also addressed, where we found the presence of additive errors, which were corrected, and weak proportional biases. We obtained the mean overall agreement between the measurements taken with the novel e-CoVig device and the reference devices for the measured quantities. Although some limitations in this study were encountered, we present more advanced methods for their further assessment.
本工作旨在验证使用临床验证设备获取的测量值，对实验室外设备获取的三个生理参数（即体温、心率和外周血氧饱和度）的测量值进行验证。实验室外专用设备集成到名为 e-CoVig 的定制移动健康应用程序中，该应用程序是在 AIM Health 项目中开发的。为了进行分析，从重症监护病房的患者中收集了 e-CoVig 和标准设备单次连续测量的三个重要参数，对这些数据进行预处理，然后通过经典一致性分析进行分析，其中我们使用林氏一致性系数来评估一致性相关性，并使用带有精确置信区间的 Bland-Altman 图来分析配对数据读数的一致性界限。还解决了可能存在系统误差的问题，发现存在附加误差，并进行了修正，同时存在微弱的比例偏差。我们得出了新型 e-CoVig 设备和参考设备测量值之间的总体平均一致性。尽管在这项研究中遇到了一些限制，但我们提出了更先进的方法来进一步评估。
