Artifacts annotations in anesthesia blood pressure data by man and machine

Abstract

Physiologic data from anesthesia monitors are automatically captured. Yet erroneous data are stored in the process as well. While this is not interfering with clinical care, research can be affected. Researchers should find ways to remove artifacts. The aim of the present study was to compare different artifact annotation strategies, and to assess if a machine learning algorithm is able to accept or reject individual data points. Non-cardiac procedures requiring invasive blood pressure monitoring were eligible. Two trained research assistants observed procedures live for artifacts. The same procedures were also retrospectively annotated for artifacts by a different person. We compared the different ways of artifact identifications and modelled artifacts with three different learning algorithms (lasso restrictive logistic regression, neural network and support vector machine). In 88 surgical procedures including 5711 blood pressure data points, the live observed incidence of artifacts was 2.1% and the retrospective incidence was 2.2%. Comparing retrospective with live annotation revealed a sensitivity of 0.32 and specificity of 0.98. The performance of the learning algorithms which we applied ranged from poor (kappa 0.053) to moderate (kappa 0.651). Manual identification of artifacts yielded different incidences in different situations, which were not comparable. Artifact detection in physiologic data collected during anesthesia could be automated, but the performance of the learning algorithms in the present study remained moderate. Future research should focus on optimization and finding ways to apply them with minimal manual work. The present study underlines the importance of an explicit definition for artifacts in database research.

Effect of artifacts upon the pressure reactivity index

“…Pasma et al . 13 compared the algorithmic and manual annotation of artifacts in arterial blood pressure signals during anesthesia. There was a significant discrepancy among the direct observation of the artifacts and the retrospective analysis or the detection using machine learning algorithms.…”

“…Apart from the artifacts originating from disconnection of the monitors and from technical issues, there are several situations common in an intensive care unit that can distort the waveforms and values (patient positioning, coughing, etc.). Previous studies have investigated the pattern 8 and the origin 9 of the artifacts in the arterial blood pressure signal; some studies have also attempted to detect them 10 – 13 or even to eliminate them 9 , 14 , mainly using artificial intelligence tools and neural networks.…”

What is new in hemodynamic monitoring and management?

“…An interesting study on artifact detection in vital signs time series was conducted by Pasma et al [ 28 ]. In the study, two trained research assistants observed in a live setting the arterial blood pressure waveform for at least one hour in 88 surgical procedures and annotated artifacts.…”

“…The authors concluded that the overall performance remained moderate. It is not new that advanced algorithms have trouble in annotating artifacts, but the most important finding of the study is probably that “it is not only important to describe who annotated data, but also when and how data points were marked as artifacts, in order to make research reproducible” [ 28 ]. Automated or manual artifact rejection in the retrospective analysis of large datasets may be prone to a similar, unsolvable artifact issue, because information about the clinical context is not recorded in most datasets available for retrospective analysis.…”

Patient outcome after cerebrovascular interventions

“…Although we used an automated process, it is still time-consuming and presumable also still imperfect. 61 Second, anesthesia record-keeping systems still collect large amounts of free-text data, that is unstructured and often filled with spelling mistakes. Although filtering can also be partly automated, cleaning free-text data is very time-consuming.…”