Unexpected sawtooth artifact in beat-to-beat pulse transit time measured from patient monitor data

RESEARCH ARTICLE Unexpected sawtooth artifact in beat-to-beat pulse transit time measured from patient monitor data Yu-Ting Lin1, Yu-Lun Lo2, Chen-Yun Lin3, Martin G. Frasch ID4*, Hau-Tieng Wu3,5,6* a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 1 Department of Anesthesiology, Taipei Veteran General Hospital, Taipei, Taiwan, 2 Department of Thoracic Medicine, Chang Gung Memorial Hospital, Chang Gung University, College of Medicine, Taipei, Taiwan, 3 Department of Mathematics, Duke University, Durham, NC, United States of America, 4 Department of Obstetrics and Gynecology and Center on Human Development and Disability (CHDD), University of Washington, Seattle, WA, United States of America, 5 Department of Statistical Science, Duke University, Durham, NC, United States of America, 6 Mathematics Division, National Center for Theoretical Sciences, Taipei, Taiwan * (H-TW); (MGF) Abstract OPEN ACCESS Citation: Lin Y-T, Lo Y-L, Lin C-Y, Frasch MG, Wu H-T (2019) Unexpected sawtooth artifact in beatto-beat pulse transit time measured from patient monitor data. PLoS ONE 14(9): e0221319. https:// doi.org/10.1371/journal.pone.0221319 Editor: Yan Li, Cleveland Clinic, UNITED STATES Received: June 28, 2019 Accepted: August 5, 2019 Object It is increasingly popular to collect as much data as possible in the hospital setting from clinical monitors for research purposes. However, in this setup the data calibration issue is often not discussed and, rather, implicitly assumed, while the clinical monitors might not be designed for the data analysis purpose. We hypothesize that this calibration issue for a secondary analysis may become an important source of artifacts in patient monitor data. We test an off-the-shelf integrated photoplethysmography (PPG) and electrocardiogram (ECG) monitoring device for its ability to yield a reliable pulse transit time (PTT) signal. Published: September 9, 2019 Copyright: © 2019 Lin et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: Data underlying the study are shared in the Harvard Dataverse (https:// dataverse.harvard.edu/dataset.xhtml?persistentId= doi:10.7910/DVN/OJBZ67). Funding: The work of YTL was supported by the National Science and Technology Development Fund (MOST 106-2115-M-075-001) of Ministry of Science and Technology, Taipei, Taiwan. MGF gratefully acknowledges funding support from the Canadian Institutes of Health Research (CIHR #1). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Approach This is a retrospective clinical study using two databases: one containing 35 subjects who underwent laparoscopic cholecystectomy, another containing 22 subjects who underwent spontaneous breathing test in the intensive care unit. All data sets include recordings of PPG and ECG using a commonly deployed patient monitor. We calculated the PTT signal offline. Main results We report a novel constant oscillatory pattern in the PTT signal and identify this pattern as a sawtooth artifact. We apply an approach based on the de-shape method to visualize, quantify and validate this sawtooth artifact. Significance The PPG and ECG signals not designed for the PTT evaluation may contain unwanted artifacts. The PTT signal should be calibrated before analysis to avoid erroneous interpretation of its physiological meaning. PLOS ONE | https://doi.org/10.1371/journal.pone.0221319 September 9, 2019 1 / 13 Sawtooth artifact in PTT Competing interests: The authors have declared that no competing interests exist. Introduction Calibration is one of the most important initial steps in any signal acquisition and experiment— the data collection equipment, or the quality of the data, needs to be calibrated before a meaningful data analysis can take place. By calibration, we mean the validity of the signal source and checking if the signal is correctly recorded for the specific purpose. While using clinical monitors as scientific instrument has been questioned [1], in our era of medical big data research, we rely on clinical monitors, such as patient vital signs monitors or Holter ECG, heavier than ever before to collect as much data as possible in the hospital research setting for the data analysis purposes [2,3]. While there has been a lot of discussion about the artifact issues in patient monitoring data [4–7], if and how the data is calibrated is often not discussed and, rather, implicitly assumed when collecting and analyzing data acquired from clinical monitors. Particularly, when multiple time series recorded from an off-the-shelf patient monitor are analyzed in the framework of sensor fusion [8], it is often implicitly assumed that on the device level the relationship between channels, such as synchronization, is not an issue. The calibration problem becomes more severe when we access the publicly available databases. Usually, less background information is available to the data analysts, which precludes a comprehensive judgement of the data quality. For example, while in the MIMIC III waveform database [3] the inter-waveform alignment problem is mentioned, there is no specific quantification of it but only a description. Without a specific quantification of the underlying problem, the information we can extract from the waveform is limited. For most online available databases, in general it is not consistently known which are suitable for which purposes, since the original clinical data acquisition device may not have been designed for the intended purpose of a secondary analysis and we do not have access to the device hardware or software details [9–11]. We hypothesize that this less discussed calibration issue for a secondary analysis will become an important source of artifacts in patient monitor data. To the best of our knowledge, this critical validation step in the work flow of any secondary (or even primary) analysis of data collected from clinical acquisition systems has not been reported, with relevance rising, particularly as more massively and passively collected databases become available. In this paper, we provide an evidence confirming our hypothesis. We analyzed two databases collected passively during patient care in a hospital environment. We identify a calibration problem and the artifact it produces. Specifically, we focus on the pulse transit time (PTT) signal derived from the electrocardiogram (ECG) and photoplethysmography (PPG). We demonstrate that an artifact in the PTT signal [12] referred to as sawtooth artifact can occur because the marketed patient monitor was not designed and calibrated for this specific purpose in the first place. Materials and methods Materials The data set used in the present manuscript comes from t (...truncated)