Biomedical Signals and Systems, University of Twente, Enschede, The Netherlands.
Cardiovascular and Respiratory Physiology, University of Twente, Postbox 217, 7500 AE, Enschede, The Netherlands.
J Clin Monit Comput. 2023 Oct;37(5):1387-1400. doi: 10.1007/s10877-023-00975-w. Epub 2023 Feb 2.
Wireless vital signs sensors are increasingly used for remote patient monitoring, but data analysis is often challenged by missing data periods. This study explored the performance of various imputation techniques for continuous vital signs measurements. Wireless vital signs measurements (heart rate, respiratory rate, blood oxygen saturation, axillary temperature) from surgical ward patients were used for repeated random simulation of missing data periods (gaps) of 5-60 min in two-hour windows. Gaps were imputed using linear interpolation, spline interpolation, last observation- and mean carried forwards technique, and cluster-based prognosis. Imputation performance was evaluated using the mean absolute error (MAE) between original and imputed gap samples. Besides, effects on signal features (window's slope, mean) and early warning scores (EWS) were explored. Gaps were simulated in 1743 data windows, obtained from 52 patients. Although MAE ranges overlapped, median MAE was structurally lowest for linear interpolation (heart rate: 0.9-2.6 beats/min, respiratory rate: 0.8-1.8 breaths/min, temperature: 0.04-0.17 °C, oxygen saturation: 0.3-0.7% for 5-60 min gaps) but up to twice as high for other techniques. Three techniques resulted in larger ranges of signal feature bias compared to no imputation. Imputation led to EWS misclassification in 1-8% of all simulations. Imputation error ranges vary between imputation techniques and increase with gap length. Imputation may result in larger signal feature bias compared to performing no imputation, and can affect patient risk assessment as illustrated by the EWS. Accordingly, careful implementation and selection of imputation techniques is warranted.
无线生命体征传感器越来越多地用于远程患者监测，但数据分析经常受到缺失数据期的挑战。本研究探讨了各种插补技术对连续生命体征测量的性能。使用手术病房患者的无线生命体征测量（心率、呼吸率、血氧饱和度、腋温），在两小时窗口中重复随机模拟 5-60 分钟的缺失数据期（间隙）。使用线性插值、样条插值、最后观察值和均值前推技术以及基于聚类的预测对间隙进行插补。使用原始和插补间隙样本之间的平均绝对误差 (MAE) 评估插补性能。此外，还探讨了对信号特征（窗口斜率、均值）和早期预警评分（EWS）的影响。在 52 名患者获得的 1743 个数据窗口中模拟了间隙。尽管 MAE 范围重叠，但线性插值的中位数 MAE 结构最低（心率：5-60 分钟间隙时为 0.9-2.6 次/分钟，呼吸率：0.8-1.8 次/分钟，温度：0.04-0.17°C，氧饱和度：0.3-0.7%），但其他技术的 MAE 高达两倍。与未插补相比，有三种技术导致信号特征偏差范围更大。插补导致所有模拟中 EWS 错误分类的比例为 1-8%。插补误差范围在插补技术之间变化，并随间隙长度增加而增加。与不进行插补相比，插补可能导致更大的信号特征偏差，并可能影响患者风险评估，如 EWS 所示。因此，有必要仔细实施和选择插补技术。
